NEUROSCIENCE

Neuroscience, education and special educationUsha Goswami

IntroductionBoth educationists and neuroscientists are interested inlearning and how to optimise learning. Neuroscientistsinvestigate the processes by which the brain learns andremembers, from the molecular and cellular levels rightthrough to brain systems (for example, the system of neuralareas and pathways underpinning our ability to estimatenumerosities, calculate exact quantities and solve (or not)differential equations). Neuroscientists study learning at avariety of levels. Understanding cell signalling and synapticmechanisms (one brain cell connects to another via asynapse) is important for understanding learning, but so isexamination of the functions of specific brain structuressuch as the cerebellum. Brain cells (or neurons) transmitinformation via electrical signals, which pass from cell tocell via the synapses, tri gge ring the release ofneurotransmitters (chemical messengers). There are around

100 billion neurons in the brain, each with massiveconnections to other neurons. Patterns of neural activity arethought to correspond to particular mental states or mentalrepresentations. Learning essentially comprises changes inconnectivity: the release of neurotransmitters at the synapsecan be altered, or connections between neurons can bestrengthened or pruned. Successful teaching directly affectsbrain function by changing connectivity.

C l e a rly, successful learning is also dependent on thecurriculum and the teacher; the context provided by theclassroom and the family; and the context of the school andthe wider community. All of these factors also interact withthe ch a ra c t e ristics of individual brains. For ex a m p l e,children with high levels of the MAOA gene (monoamineoxidise A) who ex p e rience maltre atment and adve rs efamily environments seem to be protected from developingantisocial behaviours (Caspi, McClay, Moffitt, Mill, Martin,Craig, Taylor & Poulton, 2002). This protection may occurvia moderation of their neural response to stress. It is alsopossible to study the effects of various medications oncognitive function. Methylphenidate (Ritalin), a medicationfrequently prescribed for children with ADHD (AttentionDeficit Hyperactivity Disorder), has been shown to improvestimulus recognition in medicated children in terms ofimproving their attention to auditory and visual stimuli (asrevealed by neuro i m agi n g, see Seife rt, Sch e u e rp fl u g,Zillessen, Fa l l gater & Wa rn ke, 2003). Neuro i m agi n gtechniques thus offer the potential to study the effects ofdifferent medications, food additives and potential toxins oneducational performance.

Cognitive neuroscience also offers techniques for studyingthe effect of teaching on the brain. There are already studiesexploring the effects of different instructional programmesin literacy on brain function and eventually neurosciencemay be able to offer methods for the early identification ofspecial educational needs. Neuroscience is already able toassess the delivery of education for special needs in certainspecialised areas such as dyslexia. At the same time,however, it is worth noting that ‘neuromyths’ abound. Somepopular beliefs about wh at brain science can actuallydeliver to education are quite unrealistic. Although currentbrain science technologies offer exciting opportunities toe d u c ationists, they complement rather than rep l a c etraditional methods of educational enquiry.

A short introduction to brain developmentMany critical aspects of brain development are completeprior to birth (see Johnson, 1997, for an overview). The

The discipline of neuroscience draws from thefields of neurology, psychology, physiology andbiology, but is best understood in the wider worldas ‘brain science’. Of particular interest foreducation is the development of techniques for‘imaging’ the brain as it performs differe n tcognitive functions. Cognitive neuroimaging hasalready led to advances in understanding some ofthe basic functions involved in learning and raisedimplications for education and special education inparticular. For example, neuroimaging has enabledscientists to study the very complex processesunderpinning speech and language, thinking andreasoning, reading and mathematics. In this article,P rofessor Usha Goswami of the University ofCambridge Faculty of Education first reviews basicinformation on brain development. She provides abrief introduction to the tools used in neuroimagingthen considers recent findings from neurosciencethat seem relevant to educational questions.Professor Goswami uses this review to suggestparticular ways in which neuroscience researchcould inform special education. In its closingsections, this article provides authoritativeperspectives on some of the ‘neuromyths’ thatseem to have taken root in the popular imaginationand argues for increased dialogue, in the future,between the disciplines of neuroscience andeducation.

Key words: n e u roscience, brain, dyslexia, dyscalulia,learning disabilities.

© NASEN 2004 British Journal of Special Education • Volume 31 • Number 4 • 2004 175

development of the brain begins during the first weeks ofgestation, with the birth of the cells that compose the brain.These cells migrate to the different regions in the foetalbrain, the regions where they will be employed in themature brain, prior to birth. By seven months gestation,almost all of the neurons that will comprise the mature brainhave been formed. We already know that there are verys p e c i fic effects of mat e rnal drug addiction on bra i nd evelopment. For ex a m p l e, babies with foetal alcoholsyndrome are born with underdeveloped parietal lobes. Asthe parietal lobe is critical for numeracy, these babiesdevelop into children with specific problems in numberp rocessing and mat h e m atical cognition (Ko p e ra - Frye,Dehaene & Streissguth, 1996). Despite the plasticity of thedeveloping brain, some aspects of brain development arealready complete in the womb and therefore less amenableto the effects of later environment.

B rain development fo l l owing birth consists almostexclusively of the growth of fibre connections and synapsesbetween neurons: this process is called ‘synaptogenesis’.For vision and hearing (visual and auditory cortex), there isextensive early synaptogenesis. The density of connectionspeaks at around 150% of adult levels between four and 12months, and the connections are then extensively pruned.Synaptic density returns to adult levels between two andfour years in the visual cortex. For other areas such asp re f rontal cort ex (thought to underpin planning andreasoning), density increases more slowly and peaks afterthe first year. Reduction to adult levels of density takes atleast another 10–20 years, hence there is significant braindevelopment in the frontal areas even in adolescence. Brainmetabolism (glucose uptake, which is an approximate indexof synaptic functioning) is also above adult levels in theearly years. Glucose uptake peaks at about 150% of adultlevels somewhere around four to five years. By the age ofaround ten years, brain metabolism has reduced to adultlevels for most cortical regions. The general pattern is clear.Brain development consists of bursts of synaptogenesis,peaks of density, and then synapse rearrangement andstabilisation. This occurs at different times and differentrates for different brain regions. This means that there aredifferent sensitive periods for the development of differenttypes of knowledge.

In fact, brain volume quadruples between birth anda d u l t h o o d. This is because of the pro l i fe ration ofconnections. As the brain is highly plastic, significant newconnections fo rm all the time, even in adulthood, inresponse to new learning or to environmental events (suchas a stroke or motorbike crash). Similarly, sensitive periodsare not all-or-none. Windows of enhanced sensitivity do not‘close’; rather the ability of the brain to benefit fromspecialised input changes. For example, if early visual inputis lacking, the critical period for setting up a visual systemis not ‘missed’ (Fagiolini & Hensch, 2000). However, theeffects of early deprivation will vary depending on visualfunction. Functions that develop late (for example, depthp e rc eption) suffer more from early dep rivation thanfunctions that are relatively mature at birth (such as colour

perception, Maurer, Lewis & Brent, 1989). This means thatsome abilities have a lower likelihood of ach i eving full potentialthan others when the sensitive period is missed. However,particularly for more cognitive abilities, I would argue thatfocused intervention always has an effect on development.

Figure 1: The major subdivision of the cerebral cortex.The different lobes are specialised for different tasks.The frontal lobe is used for planning and reasoning andcontrols our ability to use speech and how we react tosituations emotionally. The temporal lobe is mainlyconcerned with memory, audition, language and objectre c ognition. The parietal lobe controls our sense of touchand is used for spatial processing and perception. Theoccipital lobe is specialised for vision. Structures such asthe hippocampus and the amygdala are internal to the bra i n ,situated beneath the cerebral cortex in the midbrain.

It is also important to realise that there are large individualdifferences between brains. There is striking variation in thesize of different brain structures and in the number ofneurons that different brains use to carry out identicalfunctions, even between ge n e t i c a l ly identical twins.Nevertheless, there is significant localisation of functionacross brains. A basic map of major brain subdivisions iss h own in Fi g u re 1. All adult brains show this basicstructure. However, it is thought that, early in development,a number of possible developmental paths and end states arepossible. The fact that development converges on the samebasic brain structure across cultures and gene pools isprobably to do with the constraints on development presentin the environment. Most children are exposed to verysimilar environments despite some cultural differences inrearing practices. Large differences in environment, such asbeing reared in darkness or without contact with otherhumans, are thankfully absent or ra re. When largeenvironmental differences exist, they have notable effectson cognitive function, some of which are beneficial. Forexample, neuroimaging studies show that blind adults arefaster at processing auditory info rm ation than sighted adultsand that congenitally deaf adults are faster at processingvisual information in the peripheral field than hearing adults(see, for example, Neville, Schmidt & Kutas, 1983; Röder,Rösler & Neville, 1999; Neville & Bavelier, 2000).

Even so, neurons themselves are interchangeable in theimmature system and so dramatic differences in rearing

176 British Journal of Special Education • Volume 31 • Number 4 • 2004 © NASEN 2004

environment can lead to different developmental outcomes.For example, deaf people’s brains use the area underpinnings p o ken language in hearing people to rep resent signlanguage (Neville, Bavelier, Corina, Rauschecker, Karmi,Lalwani, Braun, Clark, Jezzard & Turner, 1998). This isre m a rk abl e, as spoken language depends on auditoryanalysis whereas sign language depends on visual/spatialanalysis. Hence the same neurons have the potential toprocess either auditory or visual/spatial information. Visualbrain areas are recruited for braille reading (which requirestactile rather than visual analysis) in blind people (seeRöder & Neville, 2003). It has even been reported that ablind adult who suffered a stroke specific to the visual areasof her brain consequently lost her proficient braille readingability, despite the fact that her somatosensory perceptionabilities were unaffected (Jackson, 2000).

Another interesting perc eptual possibility is that allmodalities (vision, audition, touch) are initially mutuallylinked. Potential evidence for early multi-modal perceptionis that auditory stimulation during early infancy also evokesl a rge responses in visual areas of the brain, wh i l esomatosensory responses are enhanced by white noise(Neville, 1995). If ear ly mutual linkage of the senses is thenorm, a kind of simultaneous perception or ‘synaesthesia’could enable infants to extract schemas that are independentof particular modalities, schemas such as number, intensityand time (see Röder & Neville, 2003). For some, it may bethat early synaesthesia is never lost. Around one in 2,000adults is thought to be synaesthetic, for ex a m p l eexperiencing distinct colours (‘photisms’) for numbers andletters. These colours are always consistent and are evokedsimply by thinking about the number or letter: they are anintimate part of that person’s concept of numbers andletters. Synaesthesia seems to have beneficial effects onmemory (Smilek, Dixon, Cudahy & Merikle 2002). If thismutual linkage of usually distinct sensory systems is alsopresent in early childhood, it may explain why youngerchildren respond so well to teaching via multi-sensorymethods. Although there is as yet no re s e a rch withsynaesthetic children, synaesthesia is an area of growinginterest for cognitive neuroscientists.

A short primer on neuroimaging toolsA quick primer on methodology is worth introducing here,as current neuro i m aging techniques have import a n tlimitations. Neuroimaging is based on the assumption thatany cognitive task makes specific demands on the brain.These demands are met by changes in neural activity. Theb rain pumps more blood to meet demand. Cog n i t iveneuroimaging methods either measure local changes inblood flow directly (PET) or indirectly (fMRI) or measurethe extremely low-voltage electrical impulses associatedwith brain activity (EEG and ERP).

PET (positron emission tomography) tracks blood flow viaradioactive tracers. Brain areas with higher levels of bloodflow have larger amounts of the tracer. Due to the tracers,PET is unsuited to children. When blood flow to particularbrain areas increases, the distribution of water in the brain

tissue also changes. fMRI (functional magnetic resonanceimaging) depends on this property. It works by measuringthe magnetic resonance signal generated by the protons ofwater molecules in neural cells, generating a BOLD (bloodoxygenation level dependent) response. This techniquedepends on inserting participants into a large cylindricalmagnet. It is very noisy inside the magnet and participantsare given headphones to shield their ears and a panic button(the magnet is claustrophobic). Because of these factors, ithas been challenging to adapt fMRI for use with children( who also move a lot, impeding scanning accura cy ) .However, with the advent of specially adapted coils and lessclaustrophobic head scanners, fMRI studies of children aregrowing in number. Both PET and fMRI show where brainactivity is occurring (localisation of function). As imagesare often acquired over a few seconds (minimum possibleresolution 0.5 seconds), PET and fMRI cannot tell us aboutthe exact timing of mental events.

Fi g u re 2: A child we a ring a specially adapted headcap fo rmeasuring ERPs (event related potentials). I am gratefulto Professor Mark Johnson, Director of the Cognitiveand Brain Development Centre, Birk b e ck College, London,for this image.

A different neuroimaging technique that is highly sensitiveto timing is the event related potential (ERP). Sensitiveelectrodes are placed on the skin of the scalp in order tore c o rd brain activ i t y. When the spontaneous nat u ra lrhythms of the brain are recorded this is called EEG( e l e c t ro e n c ep h a l ograp hy). When particular events aredesigned by the ex p e rimenter to affect spontaneousrhythms, systematic deflections in electrical activity areevoked (hence the event related potential). ERP rhythms aretime-locked to specific events designed to study cognitivefunction and are widely measured in children. The usualtechnique is for the child to watch a video while wearing aheadcap (like a swimming cap) that holds the electrodes.An illustration is provided in Figure 2. Another advantageof this method is that the child does not have to attend to thespecific events in order for the brain to register them. Forlinguistic stimuli, the events can form a background noisewhile the child sits engrossed in a silent cartoon: the brainwill respond to the auditory events in the same way. ForERP studies, neuroscientists measure (1) the latency of thepotentials; (2) the amplitude (magnitude) and direction (positive

© NASEN 2004 British Journal of Special Education • Volume 31 • Number 4 • 2004 177

or negative) of the responses; and (3) the distribution of theactivity. Exact localisation (where the responses originate)is not yet possible with ERP, but millisecond differences intiming are measurable and remarkably consistent.

Different electrical potentials (characterised in countlessERP studies) are clearly visible when the potentials areplotted against time, and are called N100, P200, N400 andso on. These labels denote negative peak at 100 ms, positivepeak at 200 ms and so on. The amplitude and duration ofsingle ERP components such as the P200 increase until aget h ree to four ye a rs and then decrease until pubert y.Decreasing amplitude usually indexes development in theschool years. ERP latencies decrease within the first yearsof life and reach adult levels in late childhood. Therefore,decreasing latency usually indexes development in theschool years. If the brain can do something 50 millisecondsfaster in a ten-ye a r-old than in a seve n - ye a r- o l d, this has a bige ffect on cog n i t ive perfo rm a n c e. In fact, it has been s u gge s t e dt h at ‘g’, the ge n e ral factor proposed by Spearman (1927) tounderlie all individual differences in cognitive abilities, isactually speed of neural processing (Anderson, 2001).

Selected empirical studiesThe tools of cognitive neuroscience have the potential tooffer various exciting possibilities to education. For specialeducation, these include the early diagnosis of specialeducational needs; the monitoring and comparison of theeffects of different kinds of educational input on learning;and an increased understanding of individual differences inlearning and the best ways to suit input to learner. I will nowdescribe briefly some recent neuroscience studies in certainareas of cognitive development and try to give a flavour ofhow their methods could contribute to questions that aremore specific to special education.

LanguageOne striking aspect of a number of developmental disordersis the relative sparing of the language faculty. Even childrenwith very poor intellectual ability seem to acquire relativelynormal language, particularly in terms of vocabulary. Forreasons that are not yet well-understood, the brain systemsimportant for syntactic and grammatical processing aremore vulnerable than the brain systems responsible forsemantic and lexical functions. These two aspects oflanguage are also represented in different regions of thebrain. Studies of non-disabled adults show that grammaticalp rocessing relies more on frontal regions of the left hemisphere,wh e reas semantic processing and vo c abu l a ry learn i n gactivate posterior lateral regions of both hemispheres.

ERP studies show that when English is acquired late due toa u d i t o ry dep rivation or late immigration to an English-speakingcountry, syntactic abilities do not develop at the same rateor to the same extent (Neville, Coffey, Lawson, Fischer,Emmorey & Bellugi, 1997). Children acquiring English asa second language may thus need extra support withgrammatical aspects of their second language compared tosupport for basic vocabulary learning. Interestingly, brainimaging studies suggest that late learners do not rely on left

hemisphere systems for grammatical processing, but useboth hemispheres (Web e r- Fox & Nev i l l e, 1996). ERP studiesalso show that congenitally blind people show bilateralrepresentation of language functions (Röder et al., 2000).Blind people also process speech more efficiently (Hollins,1989); for example, they speed up cassette tapes of normalspeech, finding them too slow. Again, this shows that thed evelopment of certain sensory systems can be enhanced wh e nother systems are impaired or absent. The implications ofthis for learning disability have not yet been studied.

ReadingIn the field of specific reading disability or dyslexia,n e u ro i m aging studies of both ch i l d ren and adults arew i d e s p re a d. These studies suggest that, at least fo ra l p h abetic scripts, the major systems for reading arelateralised to the left hemisphere. Such studies typicallymeasure brain responses to single word reading, usingfMRI or ERPs. Rev i ews of such studies conclude thatm a ny neural regions are necessary for re a d i n g, asa l p h ab e t i c / o rt h ographic processing is associated withoccipital, temporal and parietal areas (see, for example,Pugh, Mencl, Je n n e r, Katz, Frost, Lee Shaywitz &Shaywitz, 2001). The occipital-temporal areas are mostactive when processing visual features, letter shapes andorthography. The inferior occipital-temporal area showse l e c t ro p hy s i o l ogical dissociations between wo rds (like‘ c at’) and non-wo rds (like ‘cet’) at around 180 ms,suggesting that these representations are not purely visualbut are linguistically stru c t u re d. Brain activation intemporo-occipital areas increases with reading skill (see,for example, Shaywitz, Shaywitz, Pugh, Mencl, Fulbright,Skudlarski, Constable, Marchione, Fletcher, Lyon & Gore,2002) and activation is decreased in ch i l d ren withdevelopmental dyslexia. There is also hyper-activation ofthis area in hy p e rl exic ch i l d ren (Tu rke l t a u b, Flowe rs ,Verbalis, Miranda, Gareau & Eden, 2004).

Studies of reading acquisition have emphasised thei m p o rtance of phonological awa reness (the ability torecognise and manipulate component sounds in words)across languages. Brain imaging shows that phonologicalprocessing appears to be focused on the temporo-parietaljunction. This ap p e a rs to be the main site support i n gletter-to-sound recoding and is also implicated in spellingd i s o rd e rs. Children with dy s l exia, who typically havep h o n o l ogical deficits, show reduced activation in thetemporo-parietal junction during tasks such as decidingwhether diffe rent letters rhyme (for ex a m p l e, P, T = ye s ,P, K = no). Ta rgeted reading re m e d i ation incre a s e sactivation in this area (see, for example, Simos, Fletcher,Bergman, Breier, Foorman, Castillo, Davis, Fitzgerald &Papanicolaou, 2002). This shows that special educationprogrammes can affect very specific areas of the brain.Finally, recordings of event-related magnetic fields (MEGrecordings) in children with dyslexia suggest that there isatypical organisation of the right hemisphere (Heim, Eulitz& Elbert, 2003). This is consistent with suggestions thatcompensation strategies adopted by the dyslexic brainrequire greater right hemisphere involvement in reading.

178 British Journal of Special Education • Volume 31 • Number 4 • 2004 © NASEN 2004

To date, it is probably fair to say that neuroimaging studieshave largely confirmed what was already known aboutreading and its development from behavioural studies.Imaging studies have essentially confirmed the centralimportance of the language system for reading and writing.Brain activation is reduced in phonological areas of thelanguage system in dyslexia and it is these areas thatincrease activation when targeted phonologically basedreading remediation packages are administered. However,neuroscience techniques also offer a way of distinguishingbetween different cognitive theories. This work has not yetbeen done. For example, it is still argued that dyslexia mayhave a visual basis (see, for example, Stein & Walsh, 1997)or may be due to a deficit in the cerebellum (see, forex a m p l e, Nicolson & Fawcett, 1999). Neuro i m agi n gtechniques could be used to measure activation in all threeof these brain systems in the same children. Neuroimagingcould also be used to measure the impact of trainingprogrammes devised in response to particular theories ofdyslexia (see, for example, the DDAT, an exercise-basedtreatment deriving from the cerebellar hypothesis, which isbased on motor exercises such as practice in catchingbeanbags while standing on a cushion on one leg, Reynolds,Nicolson & Hambly, 2003). If an exercise-based packageactually improves reading in children with dyslexia, thereshould be measurable effects in the neural systems forreading.

Neuroimaging techniques also offer a potential means fordistinguishing between deviance and delay when studyingdevelopmental disorders. Does the brain behave totallyd i ffe re n t ly in these disord e rs or are affected ch i l d re ndeveloping along the same trajectory as unaffected childrenbut more slowly? In studies of linguistic disorders such asspecific language impairment and dyslexia, ERP studiessuggest that the language system of the child is immaturerather than deviant (McArthur & Bishop, 2004; Thomson,Baldeweg & Goswami, 2004). In disorders such as autism,the brain may actually be different, lacking a ‘theory ofmind’ module (Frith & Happé, 1998), or it may be thatsocial cog n i t ive abilities are fo l l owing the usuald evelopmental tra j e c t o ry but are reliant on suchimpoverished input that this trajectory never looks normal(Baron-Cohen, 1997). The promise of neuroimaging is thatwe may actually be able to find out which of thesea l t e rn at ive possibilities is correct. This would haveenormous implications for intervention and remediation.

MathematicsFor mathematics, cognitive neuroscience has already gonebeyond existing behavioural-cognitive models. It has beenshown that there is more than one neural system for therepresentation of number. There is a phylogenetically old‘number sense’ system, found in animals and infants as wellas older humans, which seems to underpin knowledge aboutnumbers and their relations (Dehaene, Dehaene-Lambertz& Cohen, 1998). This system, located bilaterally in thei n t rap a rietal areas, is activated for example wh e ncomparing two numbers (‘is three larger or smaller thanfive?’). Mode of pre s e n t ation does not affect the locat i o n

of these parietal ERP components, as responses are thesame whether the comparisons involve Arabic numerals,sets of dots or number wo rds. Developmental ERPstudies have shown that young children activate exactly thesame parietal areas to perform number comparison tasks(Temple & Posner, 1998).

A different type of numerical knowledge is thought to bestored verbally, in the language system (Dehaene, Spelke,Pinel, Stanescu & Tsirkin, 1999). This system also storesknowledge about poetry and overlearned verbal sequences(such as the months of the year), and underpins countingand rote-acquired knowledge such as the multiplicationtables. The system seems to store ‘number facts’ rather thanto compute calculations. Many simple ari t h m e t i c a lproblems (for example, 3 + 4 or 3 x 4) are so overlearned,at least by adulthood, that they may be stored as linguisticknowledge. More complex calculation seems to involvev i s u o s p atial regions (Zago, Pesenti, Mellet, Crive l l o ,Mazoyer & Tzourio-Mazoyer, 2001). It is thought that thismay indicate the importance of visual mental imagery inmulti-digit operations (that is, a sophisticated form of anumber line, see Pesenti, Thioux, Seron & De Vo l d e r, 2000).Finally, a distinct parietal-premotor area is activated duringfinger counting and also calculation. This last observationm ay suggest that the neural areas activated during fi n ge r- c o u n t i n g(a developmental strategy for the acquisition of calculationskills) eventually come to partially underpin numericalmanipulation skills in adults.

Neuroimaging studies of young children doing mathematicsare still rare. One current growth area is dyscalculia.Dyscalculia is a specific difficulty in learning mathematics,despite good IQ and good performance in other curriculumareas. As there are distinct neural systems that contribute tomathematical cognition, not all children with dyscalculiamay be the same in neural terms. For example, somechildren with dyscalculia are also dyslexic (Landerl, Bevan& Butterwo rth, in press). If dy s l exia has a phonological basis,then it seems likely that the mathematical system affected inthese children with dyscalculia will be the verbal systemu n d e rpinning counting and calculation. Children withdy s l exia and mat h e m atical difficulties may not showc o m p a rable neural anomalies in the activation of thep a rietal and premotor number systems. Children withdyscalculia who do not have reading difficulties may showpatterns of impairment in these other neural systems fornumber. We do not know, but neuroimaging offers a way offinding out. Knowledge of the neural basis of the difficultiesexperienced by different children with dyscalculia couldthen inform individual remedial curricula.

General learning disabilityGeneral learning disability, as opposed to specific learningd i fficulties like dy s l exia and dyscalculia, encompassesmany different conditions. General learning disability isusually defined on the basis of low IQ (scores below 70).About 40% of individuals with IQ below 70 have a medicalbackground condition and this rises to 80% for individualswith IQ below 50 (Gillberg & Soderstrom, 2003). The most

© NASEN 2004 British Journal of Special Education • Volume 31 • Number 4 • 2004 179

common cause of learning difficulty is Down’s syndrome,which typically occurs because there is a third chromosome21 in some cells (‘trisomy’). Many children with Down’ss y n d rome go on to develop the clinical fe at u res ofAlzheimer’s disease as they get older. Other disordersl i n ked to ch romosomal ab n o rmality include Fragile Xs y n d ro m e, Angelman syndrome and Pra d e r- Wi l l isyndrome. Chromosomal abnormalities account for about50% of cases of severe learning difficulties (Rittey, 2003).Other prenatal causes include intrauterine infection (forexample, rubella) and intrauterine toxins (for example,alcohol), wh i ch can have deva s t ating effects on bra i nfo rm ation. Alcohol in particular seems to show a dose-re s p o n s erelationship (that is, the more alcohol, the more damage iscaused; Autti-Rämo, 2000). Birth asphyxia (lack of oxygen)and prematurity are other important causes.

Although brain imaging studies are rare within the area ofgeneral learning disabilities, there are some interestingneurophysiological hypotheses. For reasons that are notwell understood, epilepsy is much more common in peoplewith learning disabilities and affected individuals maysuffer continual clusters of small brain seizures on a regularbasis. This clearly impedes brain function. Other theoriesare linked to the disproportionate number of males whopresent with different medical conditions. For example, ithas been suggested that dy s reg u l ation of the norm a ldevelopmental trajectory of myelination may play a role insome disorders (Bartzokis, 2004). Myelin sheaths formaround the axons (nerve fibres) of brain cells and increaseneural transmission speed. Myelin also enables widelydistributed neural networks to fire at the same time, whichis necessary for higher- l evel skills like reasoning andmemory. There is extensive myelination through middle age(late fifties). When myelination is dysregulated, there isabnormal development of white matter and this has beenproposed to be characteristic of disorders such as Aspergersyndrome and non-verbal learning disorder (Ellis & Gunter,1999). Female hormones promote myelination, thereforeacting as a protective factor.

Developmental abnormalities in the amount/thickness ofmyelin would be expected to part i c u l a rly affect lat e - d eve l o p i n gbrain structures such as the frontal, temporal and parietallobes. Intere s t i n g ly, it has been proposed from bra i nimaging work that ‘g’, the general intelligence factor, maybe specific to areas of the frontal cortex (Duncan, Seitz,Kolodny, Bor, Herzog, Ahmed, Newell & Emslie, 2000). If‘developmental g’ is underpinned by frontal functions (seeA n d e rson, 2001), and ab n o rmalities in mye l i n at i o nparticularly affect frontal regions, then there is a theoreticalconnection between myelination and the development of‘g’. This is at present speculative. Nevertheless, if particularaspects of brain function can be related to general learningdisability, this would inform pharmacological intervention.Meanwhile, even if they do not raise IQ, educationali n t e rventions always improve the quality of life ofindividuals with learning disability. This is because of theirdocumented effects on behaviour and overall adjustment(see Gillberg & Soderstrom, 2003).

NeuromythsAn OECD report on understanding the brain coined theengaging term ‘neuromyths’ (OECD, 2002) to demonstratethe ease and rapidity with which scientific findings havealso translated into misinformation regarding education.Th e re are three myths given special attention in theOECD report, namely (1) the lay belief in hemisphericdifferences (‘left brain’ versus ‘right brain’ learning etc.);(2) the notion that the brain is only plastic for certain kindsof information during certain ‘critical periods’ and thattherefore education in these areas must occur during thecritical periods; and (3) the idea that the most effectiveeducational interventions need to be timed with periods ofsynaptogenesis.

The idea of ‘left brain’ ve rsus ‘right brain’ learn i n ghas virtually no credence in neuroscience. The idea appearsto stem from the fact that there is some hemispherics p e c i a l i s ation in terms of the localisation of diffe re n tskills. For example, many aspects of language processinga re left-lat e ralised (although not, as we have seen, inblind people or in those who emigrate in later childhood toa new linguistic community). Some aspects of fa c erecognition, in contrast, are right-lateralised. However, it isalso a fact that there are massive cro s s - h e m i s p h e reconnections in the normal brain. Both hemispheres workt ogether in eve ry cog n i t ive task so far ex p l o red withneuroimaging, including language and face recognitiontasks. So fa r, neuro i m aging data demonstrate that both‘left brain’ and ‘right brain’ are involved in all cognitivetasks.

Similarly, the conceptual notion of critical periods forl e a rning has been ove rextended from the actualneuroscience findings. In fact, this concept stems largelyfrom the neuroscience of the visual system rather than theneuroscience of cognition and learning. Although optimalperiods for certain types of learning clearly exist in visualdevelopment, they are sensitive periods rather than criticalones. The term ‘critical period’ implies that the opportunityto learn is lost forever if the biological window is missed. Infact, there seem to be almost no cognitive capacities thatcan be ‘lost’ at an early age. As discussed earlier, someaspects of complex processing suffer more than others fromdeprivation of early environmental input (for example,d epth perc eption in vision or grammar learning inlanguage), but nevertheless learning is still possible. It maybe better for the final performance levels achieved toeducate children in, for example, other languages during thesensitive period for language acquisition. Nevertheless, theexistence of a sensitive period does not mean that adultsare unable to acquire competent foreign language skillslater in life.

Fi n a l ly, the idea that the effe c t iveness of educat i o n a li n t e rventions depends on whether they coincide withperiods of synaptogenesis appears to be a misinterpretationof experimental work on learning in rats. This researchshowed that rodent brains form more connections if theyoung are reared in enriched and stimulating environments

180 British Journal of Special Education • Volume 31 • Number 4 • 2004 © NASEN 2004

(for example, Greenough, Black & Wallace, 1987; notehowever that these ‘enriched’ environments were withinlaboratory cages and did not come close to mimicking theintensely stimulating normative environment of the wildrat). Furt h e r, more connections fo rm in response toparticular environments throughout life. fMRI studies haveshown that skilled pianists (adults) have enlarged corticalrepresentations in auditory cortex, specific to piano tones( Pa n t ev, Oostenve l d, Engelien, Ross, Roberts & Hike, 1998).MEG studies show that skilled violinists have enlargedneural representations for their left fingers, those mostimportant for playing the violin (Elbert, Pantev, Wienbruch,Rockstroh & Taub, 1996). London taxi drivers who possess‘The Knowledge’ (detailed knowledge of the street map ofLondon) show enlarged hippocampus formations comparedto adults who do not drive taxis (Mag u i re, Gadian,Johnsrude, Good, Ashburner, Frackowiak & Frith, 2000;the hippocampus is a small brain area thought to bei nvo l ved in spatial rep re s e n t ation and nav i gat i o n ) .Hippocampal volume was found to be correlated with theamount of time spent as a taxi driver, just as the corticalrepresentation of piano tones was found to be correlatedwith amount of time spent in piano pra c t i c e. Th e s edemonstrations do not mean that greater synaptic densityp re d i c t s a gre ater capacity to learn. Rat h e r, theydemonstrate that the brain can always benefit from targetedinputs, even when these inputs are received exclusivelyd u ring adulthood (as for taxi drive rs learning ‘Th eKnowledge’).

Other neuromyths can also be identified. One is the ideathat a person can either have a ‘male brain’ or a ‘femalebrain’. The terms ‘male brain’ and ‘female brain’ werecoined to refer to differences in cognitive style rather thanbiological differences (Baron-Cohen, 2003) and applied toautism and autistic spectrum disord e rs. Baro n - C o h e na rgued that men we re better ‘systemisers’ (good atunderstanding mechanical systems) and women were better‘empathisers’ (good at communication and understandingo t h e rs). He there fo re suggested that autism could beconceptualised as an ‘extreme’ form of the male brain. Hedid not argue that male and female brains were radicallydifferent, and that females with autism had male brains.Rather, he was using the terms ‘male’ and ‘female’ brain asa psych o l ogical shorthand for (ove rl apping) cog n i t iveprofiles.

A final neuromyth is the idea that ‘implicit’ learning has thepotential to open new avenues educationally. Much humanlearning is ‘implicit’, in the sense that learning takes placein the brain despite lack of attention to and consciousawareness of what is being learnt (for example, Berns,Cohen & Mintun, 1997, but see Johnstone & Shanks, 2001).Almost all studies of implicit learning use perceptual tasksas their behavioural measures (that is, the participant getsbetter at responding appropriately to ‘random’ letter stringsin a computer task when the ‘random’ strings are actuallygenerated according to an underlying ‘grammar’ or rulesystem which can be learnt). There are no studies showingimplicit learning of the c og n i t ive skills underp i n n i n g

educational achievement. These skills most likely requireeffortful learning and direct teaching.

ConclusionsClearly, the potential is there for neuroscience to makeexciting contributions to educational research in generaland to special education in particular. Nevertheless, morebridges need to be built between basic neuroscience andresearch in education, and neuromyths need to be weededout. Bruer (1997) first made the point about buildingbridges between the disciplines (in an article provocativelysubtitled ‘a bridge too far?’) and suggested that cognitivep s y ch o l ogists are admirably placed to construct suchbridges. He also cautioned that, while neuroscience haslearnt a lot about neurons and synapses, it has not learntn e a rly enough to guide educational practice in anymeaningful way. In my view, this is too pessimistic.Cognitive developmental neuroscience has established anumber of findings relevant to education, as discussedab ove, and has also enabled the discove ry of neura l‘markers’ that can be used to assess development. Thesemarkers may prove very useful for investigating educationalquestions.

For example, consider the different ERP signatures oflanguage processing that have emerged over the last 20ye a rs of re s e a rch (Brown & Hago o rt, 1999). Diffe re n tERP parameters are robustly associated with semanticp rocessing (for ex a m p l e, N400), phonetic pro c e s s i n g( for ex a m p l e, mismat ch negat ivity or MMN) andsyntactic processing (for ex a m p l e, P600). Th edevelopment of these parameters can now be investigatedl o n gi t u d i n a l ly in ch i l d ren. Certain pat t e rns ofdevelopment may turn out to be indicative of certaindevelopmental disorders. For example, children at risk fordy s l exia may s h ow immat u re or atypical MMNs top h o n e t i c distinctions, such as /b/ versus /d/ (Csepe, 2003,for Hunga rian). Children with specific languagei m p a i rment may have ge n e ra l ly immat u re auditorysystems – systems resembling those of children three tofour ye a rs yo u n ger than them in terms of pro c e s s i n gbasic aspects of sound such as frequency (see McArthur &Bishop, 2004). The different ERP signatures may alsochange in response to t a rgeted educational progra m m e s .For ex a m p l e, the MMN to phonetic distinctions maybecome sharper (as indexed by faster latencies) inresponse to literacy tuition in phonics (see Csepe, 2003). Ifsuch findings were to be established across languages,education would have a neural tool for comparing theefficiency of different packages for remediating dyslexia.For example, one could measure whether the MMN tophonetic distinctions sharpened in response to differentc o m m e rc i a l ly marketed training regimes for dy s l ex i a .This is only one example of the potential for thec re at ive ap p l i c ation of neuroscience techniques toimportant issues in special education. Bearing in mind thelimitations of current technologies, it is time to think‘outside the box’ about how ava i l able neuro s c i e n c et e chniques can help to answer important educat i o n a lquestions.

© NASEN 2004 British Journal of Special Education • Volume 31 • Number 4 • 2004 181

ReferencesAnderson, M. (2001) ‘Annotation: conceptions of

intelligence’, Journal of Child Psychology andPsychiatry, 42, 287–298.

Autti-Rämo, I. (2000) ‘Twelve-year follow-up of childrenexposed to alcohol in utero’, Developmental Medicineand Child Neurology, 42, 406–411.

Baron-Cohen, S. (1997) ‘Is there a language of the eyes?’Visual Cognition, 4, 311–331.

B a ron-Cohen, S. (2003) The Essential Diffe rence: men, wo m e nand the ex t reme male bra i n . London: Penguin/Allen Lane.

Bartzokis, G. (2004) ‘Quadratic trajectories of brainmyelin content: unifying construct for neuropsychiatricdisorders’, Neurobiology of Aging, 25, 49–62.

Berns, G. S., Cohen, J. D. & Mintun, M. A. (1997) ‘Brainregions responsive to novelty in the absence ofawareness’, Science, 276, 1272–5.

B rown, C. M. & Hago o rt, P. (1999) (eds) The Neuro c og n i t i o nof Language. Oxford: Oxford University Press.

Bruer, J. T. (1997) ‘Education and the brain: a bridge toofar?’ Educational Researcher, 26 (8), 4–16.

Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J.,Craig, I. W., Taylor, A. & Poulton, R. (2002) ‘Role ofgenotype in the cycle of violence in maltreatedchildren’, Science, 297, 851–854.

C s ep e, V. (2003) ‘Au d i t o ry eve n t - re l ated potentials in study i n gdevelopmental dyslexia’, in V. Csepe (ed.) Dyslexia:different brain, different behaviour. New York: Kluwer.

Dehaene, S., Dehaene-Lambertz, G. & Cohen, L. (1998)‘Abstract representations of numbers in the animal andhuman brain’, Trends in Neuro s c i e n c e, 21 (8), 355–361.

D e h a e n e, S., Spelke, E., Pinel, P., Stanescu, R. & Tsirkin, S.(1999) ‘Sources of mathematical thinking: behaviouraland brain-imaging evidence’, Science, 284, 970–974.

Duncan, J., Seitz, R. J., Kolodny, J., Bor, D., Herzog, H.,A h m e d, A., Newell, F. N. & Emslie, H. (2000) ‘A neura lbasis for general intelligence’, Science, 289, 457–460.

E l b e rt, T., Pa n t ev, C., Wi e n b ru ch, C., Rock s t roh, B. & Ta u b, E.(1996) ‘Increased cortical representation of the fingersof the left hand in string playe rs’, S c i e n c e, 270, 305–307.

Ellis, H. D. & Gunter, H. L. (1999) ‘A s p e rger syndro m e :a simple matter of white matter?’ Trends in CognitiveSciences, 3, 192–200.

Fagiolini, M. & Hensch, R. K. (2000) ‘Inhibitorythreshold for critical-period activation in primaryvisual cortex’, Nature, 404, 183–186.

Frith, U. & Happé, F. (1998) ‘Why specific developmentaldisorders are not specific: on-line and developmentaleffects in autism and dyslexia’, DevelopmentalScience, 1, 267–272.

Gillberg, C. & Soderstrom, H. (2003) ‘Learningdisability’, The Lancet, 362, 811–821.

Greenough, W. T., Black, J. E. & Wallace, C. S. (1987)‘Experience and brain development’, ChildDevelopment, 58, 539–559.

Heim, S., Eulitz, C. & Elbert, T. (2003) ‘A l t e red hemispheri casymmetry of auditory P100m in dyslexia’, EuropeanJournal of Neuroscience, 17, 1715–22.

Hollins, M. (1989) Understanding Blindness. Hillsdale,NJ: Lawrence Erlbaum.

Jackson, S. (2000) ‘Seeing what you feel’, Trends inCognitive Sciences, 4, 257.

Johnson, M. H. (1997) Developmental CognitiveNeuroscience. Cambridge, MA: Blackwell.

Johnstone, T. & Shanks, D. R. (2001) ‘Abstractionist andprocessing accounts of implicit learning’, CognitivePsychology, 42, 61–112.

Kopera-Frye, K., Dehaene, S. & Streissguth, A. P. (1996)‘Impairments of number processing induced byprenatal alcohol exposure’, Neuropsychologia, 34,1187–1196.

Landerl, K., Bevan, A. & Butterworth, B. (in press)‘ D evelopmental dyscalculia and basic nu m e rical cap a c i t i e s :a study of eight- to nine-year-old students’, Cognition.

McArthur, G. M. & Bishop, D. V. M. (2004) ‘Whichpeople with specific language impairment haveauditory processing deficits?’ CognitiveNeuropsychology, 21, 79–94.

M ag u i re, E. A., Gadian, D. S., Jo h n s ru d e, I. S., Good, C. D. ,Ashburner, J., Frackowiak, R. S. & Frith, C. D. (2000)‘Navigation related structural change in thehippocampi of taxi drivers’, Proceedings of theNational Academy of Sciences of the United States ofAmerica, 97 (8), 4398–4403.

Maurer, D., Lewis, T. L. & Brent, H. (1989) ‘The effectsof dep rivation on human visual development: studiesin ch i l d ren tre ated with cat a racts’, in F. J. Morri s o n ,C. Lord. & D. P. Keating (eds) Applied DevelopmentalPsychology. San Diego, CA: Academic Press.

Neville, H. J. (1995) ‘Developmental specificity inneurocognitive development in humans’, in M. S.Gazzaniga (ed.) The Cognitive Neurosciences.Cambridge, MA: MIT Press.

Neville, H. J. & Bavelier, D. (2000) ‘Specificity andplasticity in neurocognitive development in humans’,in M. S. Gazzaniga (ed.) The Cognitive Neurosciences.Cambridge, MA: MIT Press.

Neville, H. J., Bavelier, D., Corina, D., Rauschecker, J.,Karni, A., Lalwani, A., Braun, A., Clark, V., Jezzard, P.& Tu rn e r, R. (1998) ‘Cereb ral orga n i s ation for languagein deaf and hearing subjects: biological constraints andeffects of experience’, Proceedings of the NationalAcademy of Sciences of the United States of America,95, 922–929.

Neville, H. J., Coffey, S. A., Lawson, D. S., Fischer, A.,Emmorey, K. & Bellugi, U. (1997) ‘Neural systemsmediating American Sign Language: effects of sensoryexperience and age of acquisition’, Brain & Language,57, 285–308.

Neville, H. J., Schmidt, A. & Kutas, M. (1983) ‘Alteredvisual-evoked potentials in congenitally deaf adults’,Brain Research, 266, 127–132.

Nicolson, R. I. & Fawcett, A. J. (1999) ‘Developmentaldyslexia: the role of the cerebellum’, Dyslexia: ani n t e rn ational journal of re s e a rch and pra c t i c e, 5, 155–177.

OECD (Organisation for Economic Co-operation andDevelopment) (2002) Understanding the Brain:towards a new learning science. [available online atwww.oecd.org] Paris: OECD.

182 British Journal of Special Education • Volume 31 • Number 4 • 2004 © NASEN 2004

Pa n t ev, C., Oostenve l d, R., Engelien, A., Ross, B., Roberts, L. E.& Hike, M. (1998) ‘Increased auditory corticalrepresentation in musicians’, Nature, 393, 811–814.

Pesenti, M., Thioux, M., Seron, X. & De Volder, A. (2000)‘ N e u ro a n atomical substrates of Arabic number pro c e s s i n g,nu m e rical comparison, and simple addition: a PET study ’ ,Journal of Cognitive Neuroscience, 12 (3), 461–479.

Pugh, K. R., Mencl, W. E., Je n n e r, A. R., Katz, L.,Frost, S. J., Lee, J. R., Shaywitz, S. E. & Shaywitz, B. A.(2001) ‘Neurobiological studies of reading and readingd i s ability’, Jo u rnal of Commu n i c ation Disord e rs, 34, 479–492.

Reynolds, D., Nicolson, R. I. & Hambly, H. (2003)‘Evaluation of an exercise-based treatment for childrenwith reading difficulties’, Dyslexia, 9, 48–71.

Rittey, C. D. (2003) ‘Learning difficulties: what theneurologist needs to know’, Journal of Neurology,Neurosurgery and Psychiatry, 74, 30–36.

Röder, G. & Neville, H. (2003) ‘Developmental functionalplasticity’, in J. Grafman & I. H. Robertson (eds)Handbook of Neuropsychology (second edition),Volume 9. Oxford: Elsevier Science.

Röder, G., Rösler, F. & Neville, H. J. (1999) ‘Effects ofinterstimulus interval on auditory event-relatedpotentials in congenitally blind and normally sightedhumans’, Neuroscience Letters, 264, 53–56.

R ö d e r, G., Rösler, F. & Nev i l l e, H. J. (2000) ‘Eve n t - re l at e dpotentials during language processing in congenitallyblind and sighted people’, N e u ro p s y ch o l ogi a , 38, 1482–1502.

Seifert, J., Scheuerpflug, P., Zillessen, K. E., Fallgater, A.& Warnke, A. (2003) ‘Electrophysiologicalinvestigation of the effectiveness of methylphenidate inchildren with and without ADHD’, Journal of NeuralTransmission, 110 (7), 821–829.

S h aywitz, B. A., Shaywitz, S. E., Pugh, K. R., Mencl, W. E.,Fulbright, R. K., Skudlarski, P., Constable, R. T.,M a rch i o n e, K. E., Fletch e r, J. M., Lyon, G. R. &Gore, J. C. (2002) ‘Disruption of posterior brainsystems for reading in children with developmentaldyslexia’, Biological Psychiatry, 52, 101–110.

Simos, P. G., Fletcher, J. M., Bergman, E., Breier, J. I.,Foorman, B. R., Castillo, E. M., Davis, R. N.,Fitzgerald, M. & Papanicolaou, A. C. (2002)‘Dyslexia-specific brain activation profile becomesnormal following successful remedial training’,Neurology, 58, 1203–1213.

Smilek, D., Dixon, M. J., Cudahy, C. & Merikle, P. M.(2002) ‘Synesthetic color experiences influencememory’, Psychological Science, 13 (6), 548–552.

Spearman, C. (1927) The Abilities of Man: their natureand measurement. New York: Macmillan.

Stein, J. & Walsh, V. (1997) ‘To see but not to read: themagnocellular theory of dyslexia’, Trends inNeuroscience, 20, 147–152.

Temple, E. & Posner, M. I. (1998) ‘Brain mechanisms ofquantity are similar in five-year-old children andadults’, Proceedings of the National Academy ofSciences of the United States of Ameri c a , 9 5 ,7836–41.

Thomson, J., Baldeweg, T. & Goswami, U. (2004)‘Amplitude envelope onsets and dyslexia: abehavioural and electrophysiological study’. Posterp resented at the annual meeting of the Society fo rthe Scientific Study of Reading, June 27–30,Amsterdam, NL.

Tu rke l t a u b, P. E., Flowe rs, D. L., Verbalis, A., Miranda, M.,Gareau, L. & Eden, G. F. (2004) ‘The neural basis ofhyperlexic reading: an fMRI case study’, Neuron, 41,11–25.

Weber-Fox, C. M. & Neville, H. J. (1996) ‘Maturationalconstraints on functional specialisation for languageprocessing: ERP and behavioural evidence in bilingualspeakers’, Journal of Cognitive Neuroscience, 8,231–256.

Zago, L., Pesenti, M., Mellet, E., Crivello, F., Mazoyer, B.& Tzourio-Mazoyer, N. (2001) ‘Neural correlates ofsimple and complex mental calculation’, NeuroImage,13, 314–27.

––––––––––––––––––––––––––––––––––––––––––––––––

Address for correspondenceProfessor Usha GoswamiUniversity of Cambridge Faculty of Education184 Hills RoadCambridgeCB2 2PQEmail: ucg10@cam.ac.uk

Manuscript submitted: May 2004Accepted for publication: July 2004

© NASEN 2004 British Journal of Special Education • Volume 31 • Number 4 • 2004 183

EDUCATION AND CHILDREN WITH DOWN

SYNDROME: NEUROSCIENCE, DEVELOPMENT,

AND INTERVENTION

Deborah J. Fidler1* and Lynn Nadel21Human Development and Family Studies, Colorado State University, Fort Collins, Colorado

2Department of Psychology, University of Arizona, Tucson, Arizona

Of the recent advances in education-related research in Downsyndrome, the characterization of the Down syndrome behavioral pheno-type has become a potentially critical tool for shaping education andintervention in this population. This article briefly reviews the literatureon brain–behavior connections in Down syndrome and identifies aspectsof the Down syndrome behavioral phenotype that are potentially relevantto educators. Potential challenges to etiologically informed educationalplanning are discussed. ' 2007 Wiley-Liss, Inc.MRDD Research Reviews 2007;13:262–271.

Key Words: Down syndrome; behavioral phenotypes; education; inter-

vention

As the most common chromosomal abnormality associ-ated with intellectual disability, Down syndrome hasbeen the focus of more behavioral and educational

research than most other genetic disorders. These efforts haveled to notable milestones in the advancement of education inDown syndrome, including cases in the literature of youngadults with Down syndrome attending university courses[Hamill, 2003; Casale-Giannola and Wilson Kamens, 2006]and achievements in the area of educating children and adoles-cents in inclusive environments [Buckley et al., 2006]. Thepast five years alone have brought innovations that includeteacher training interventions to shape attitudes toward inclu-sion in Down syndrome [Campbell et al., 2003], a refinedunderstanding of effective inclusive practice in Down syndrome[Wolpert, 2001], and new instructional approaches involvingcomputer technology [Lloyd et al., 2006; Ortega-Tudela andGomez-Ariza, 2006].

However, the innovation that has the potential to havethe greatest impact on educational practice in Down syndromeis the characterization of the Down syndrome ‘‘behavioralphenotype,’’ or the pattern of behavioral outcomes associatedwith this disorder throughout development. Research into thephenotypic outcomes associated with Down syndrome has ledto a better understanding of the learning profile associatedwith this disorder, and has offered new information regardingthe possible brain–behavior pathways leading to these out-comes. Over the past few decades, researchers have uncoveredcharacteristic patterns of functioning in the areas of cognition,

language development, social–emotional functioning, and per-sonality-motivation [see Dykens et al., 2000; Rondal andBuckley, 2001; Fidler, 2005]. Though many questions regard-ing development in this population remain unanswered,researchers have gained a clearer understanding of the devel-opmental trajectory associated with Down syndrome, and howthis chromosomal abnormality impacts development in adynamic and multisystemic way.

Amidst these advances in delineating the Down syn-drome behavioral phenotype, there remains a wide gapbetween these research findings and the development of inno-vative practice [Hodapp and Fidler, 1999; Fidler et al., 2007].While it has been argued that etiology-based informationcould be of importance for education in Down syndrome[Hodapp and Fidler, 1999; Freeman and Hodapp, 2000;Fidler, 2005], the use of these connections has not nearly metits full potential. It is true that relative to other disorders,greater research emphasis has been placed on identifying edu-cational strategies that might improve outcomes in Down syn-drome. Some of these recent approaches are informed by andspecifically target aspects of the Down syndrome behavioralphenotype [Laws et al., 1996; Kennedy and Flynn, 2003; Iarocciet al., 2006; van Bysterveldt et al., 2006], and others do not[Garcia and Conte, 2004; Park et al., 2005; Trent et al., 2005].Those existing syndrome-based recommendations may bepotentially quite useful, but lack empirical validation [Alton,1998; Fidler, 2005].

Though the gap between research findings and targetedpractice in Down syndrome remains wide, there is evidencethat educators and practitioners themselves recognize the im-portance of scientific progress in this area. Wolpert [2001]asked educators of children with Down syndrome in inclusivesettings to identify factors that might improve the outcomes inthe classroom. Among their answers was ‘‘. . . more informa-

*Correspondence to: Deborah J. Fidler, Human Development and Family Studies,Colorado State University, 502 W. Lake Street, Fort Collins, CO 80523-1570.E-mail: deborah.fidler@colostate.eduReceived 10 August 2007; Accepted 13 August 2007Published online in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/mrdd.20166

MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIESRESEARCH REVIEWS 13: 262 – 271 (2007)

' 2007Wiley -Liss, Inc.

tion on learning characteristics of chil-dren with Down syndrome’’ (p. 33).Similarly, parents seem to endorse etiol-ogy-specific modifications to theirchild’s educational planning [Fidler et al.,2003]. As part of an effort to addressthe gap between research and practicein this area, this article will first reviewthe current literature on the pathwayfrom brain to behavior in Down syn-drome. We then identify examples ofphenotypic outcomes that may haveeducational relevance, with particularattention to the development of cogni-tion and information processing inDown syndrome. Additional areas thatmay have educational relevance, includ-ing language development, social–emo-tional development, and personality-motivation, are briefly summarized. Weconclude with a brief discussion of thepotential challenges involved in shapingeducational instruction in etiologicallyrelevant ways. Although we start ourreview with discussion of neurobiology,we believe these kinds of data will bemost relevant to the development ofanimal models, which then can be usedto validate biological approaches to mit-igation of the syndrome’s impact oncognitive function. More important foreducational practice will be an under-standing of the phenotypic result ofthese neurobiological sequellae of tri-somy 21, and we will accordingly focuson such behavioral approaches below.

THE NEUROBIOLOGY OFDOWN SYNDROME

The brain of an individual withDown syndrome at or shortly beforebirth is in many respects indistinguish-able from the brain of a normal indivi-dual [Brooksbank et al., 1989; Wis-niewski and Schmidt-Sidor 1989; Fl�orezet al., 1990; Schmidt-Sidor et al., 1990;Bar-Peled et al., 1991; Pazos et al.,1994]. Normal values have been re-ported for brain and skull shape, brainweight, proportion of specific cerebrallobes, size of cerebellum and brainstem, and the emergence of most neu-rotransmitter systems. There is evidence,however, that some changes begin toemerge as early as 22 weeks gestationalage [e.g., Schmidt-Sidor et al., 1990;Golden and Hyman, 1994; Wisniewskiand Kida, 1994; Engidawork andLubec, 2003] and it is clear that by theage of 6 months a number of importantdifferences are already obvious. Some ofthese differences are expressed in termsof the proportion of individuals withDown syndrome who show abnormal

values, rather than in terms of a uni-form abnormality in all instances. Thisis important as it highlights the variabil-ity in this population sharing the geno-typic feature of trisomy 21.

While there is a postnatal delay inmyelination [Wisniewski, 1990], it isworth noting that in all cases myelina-tion is within normal range at birth,while in 75% of the cases it is withinnormal range throughout early develop-ment. Neuropathological differencesafter 3–5 months of age include a short-ening of the fronto-occipital length ofthe brain that appears to result from areduction in growth of the frontal lobes,a narrowing of the superior temporalgyrus (observed in about 35% of cases),a diminished size of the brain stem andcerebellum (observed in most cases),and a 20–50% reduction in the numberof cortical granular neurons [see Cromeet al., 1966; Benda, 1971; Blackwoodand Corsellis, 1976]. In sum, the overallpicture at birth or shortly thereaftershows that individuals with Down syn-drome tend to fall towards the bottomof the normal range (or outside it) onmost measures.

Investigations of neural function,as opposed to structure, in early infancysuggest some abnormalities: there is evi-dence of either delayed or aberrant au-ditory system development [Jiang et al.,1990], which might contribute to thewidespread hearing disorders observedin Down syndrome. Obviously, such adisorder, if organic, could be related tosubsequent difficulties seen in the learn-ing of language. Hill Karrer et al. [1998]have reported delayed development ofcerebral inhibition using visual event-related potentials (ERP) in a visual rec-ognition memory paradigm.

There have been few studies ofbrain function in adolescents and youngadults with Down syndrome and theexisting data are somewhat equivocal.Pinter et al. [2001] used high-resolutionMRI methods to analyze brain structurein 16 children (mean age 11.3 years)with Down syndrome. After correctingfor overall brain volume, hippocampal,but not amygdala, volume reductionswere seen in this group. This resultconfirms some earlier work using lowerresolution MRI methods [Jerniganet al., 1993]. Kates et al. [2002] lookedat a group of 12 children with Downsyndrome (all males, mean age 5.94years) and compared them with childrenwith fragile-X, developmental languagedelay, or typical development. The chil-dren with Down syndrome had smallerbrain volumes than any of the others,

with previously unreported reductionsin parietal cortex as well as the oft-reported reductions in the temporallobe. Pinter et al. [2001], on the otherhand, note the relative preservation ofparietal cortex.

Overall, study of neuropathologypoints to problems in certain regions ofthe cortex, including most prominentlythe temporal lobe and the hippocampalformation [Wisniewski et al., 1986], theprefrontal cortex, and the cerebellum.This conclusion meshes well with whathas been learned from the developmentof mouse models of the syndrome [e.g.,Kleschevnikov et al., 2004]. Thesemodels have generally demonstratedselective impairments in the anatomy,physiology, pharmacology, and behavior,which are associated with the hippo-campal formation, the prefrontal cortex,and the cerebellum.

These neurobiological findings inDown syndrome can inform educationin several ways. First, as only modestabnormalities are detectable at birth, therole of development is critical in build-ing into the pronounced profile ofstrengths and weaknesses observed inthis population. This offers educators animportant window of opportunity forintervention in the first few years of life,a point that will be revisited later. Sec-ond, the atypical development of spe-cific brain structures suggests that someareas of functioning may become moreimpaired than others throughout devel-opment. In addition, early in develop-ment, many of the differences observedare expressed in terms of the proportionof individuals with Down syndromewho show abnormal values, rather thanin terms of a uniform abnormality in allinstances. This is in line with the proba-bilistic approach to understanding out-comes in individuals with genetic dis-orders [Dykens, 1995], where childrenmay be at risk for certain neurobiologi-cal features, but not all children willshow those specific abnormalities. Thesethree themes are educationally relevantand will be explored in greater depth inthe following sections. Additional linksbetween brain and behavior will benoted throughout the next section whererelevant findings have been reported.

REVIEW OF EDUCATIONALLY-RELEVANT LITERATURE ONDEVELOPMENT IN DOWNSYNDROME

In this section, we explore thecurrent literature on development inDown syndrome and highlight some of

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 263

the main findings that are potentiallyrelevant to educators and intervention-ists. Before we examine the phenotypicprofile associated with Down syndrome,there are two important issues thatshould be noted. First, these findingsare part of a probabilistic approach tounderstanding the link between syn-drome and outcome in behavioral phe-notype research [Dykens, 1995]. Withinthis approach, it is understood that thereis a higher probability that childrenwith Down syndrome will show any ofthese specific outcomes relative to otherchildren who do not have Down syn-drome; however, not every child withDown syndrome will exhibit each ofthese phenotypic features [Dykens,1995]. Thus, educators might use thisinformation to adopt a ‘‘ready stance’’regarding areas of potential strength andweakness in children with Down syn-drome, but might expect that childrenwill likely vary in the degree to whichthey express any one of these predis-posed outcomes.

Second, researchers are currentlyexploring the question of whether thereis a subgroup of individuals with Downsyndrome who meet criteria for autismspectrum disorders. Some studies sug-gest that a small percentage of individu-als with Down syndrome have behav-ioral profiles consistent with autism[Collacott et al., 1992; Ghaziuddinet al., 1992; Kent et al., 1999; Paly andHurley, 2002; Hepburn et al., in press].These findings are not based on studiesthat have employed epidemiologicalmethods for estimating the prevalenceof comoribidity in this population, sothey should be taken with a note ofcaution. However, it is possible thatDown syndrome with children whohave comorbid autism may show a dif-ferent behavioral profile than theircounterparts with Down syndrome whodo not have autism, and may require adifferent educational approach [Hepburnet al., in press].

COGNITION IN DOWNSYNDROME

Early CognitionIn line with the neurobiological

findings described earlier, infants withDown syndrome show relatively normalabilities in learning and memory. Thisdoes not mean that either they, orindeed normally developing infants,have the full adult range of learning andmemory abilities at birth. In fact this isnot the case because some parts of the

brain mature postnatally and the formsof learning and memory dependent onthem are not available until some timeafter birth. The medial temporal lobe,and particularly, the hippocampus, aswell as parts of the prefrontal cortexand cerebellum, are included in this cat-egory. The fact that these late-develop-ing structures are apparently particularlyat risk in Down syndrome is probablyof considerable importance [see Nadel,1986].

In an early series of studies, Ohrand Fagen [1991, 1993] reported that3-month-old infants with Down syn-drome were entirely normal in learningabout the contingencies between theirown movements (leg kicking) and rein-forcement, including initial learning, ac-quisition speed, and retention. In a laterreport, Ohr and Fagen [1994] showedthat 9-month-old infants with Downsyndrome were impaired, as a group, inlearning about the contingency betweenarm movements and reinforcement.However, they noted that some infantswith Down syndrome were able tolearn. They concluded that there is arelative decline in conditionability ininfants with Down syndrome comparedto normally developing infants after 6months. Mangan [1992] tested controlinfants and infants with Down syn-drome on a variety of spatial tasks, oneof which, a place-learning task, wasdesigned especially to assess the state offunction of the hippocampal system.The pattern of results was consistentwith diffuse, but mild, neuropathologycombined with much more extensivepathology localized to the hippocampus.

Children with Down syndromehave typically been shown to acquirebasic object concept more slowly thannormal [see, e.g., Rast and Meltzoff,1995] but with extensive training theycan acquire it at more-or-less the sametime as normally developing infants[Wishart, 1993]. However, a differentkind of problem emerges in this tasksituation: instability of acquisition. Al-though the typical subject with Downsyndrome solved various levels of thetasks used to assess the object concept atages not very far from the norm, per-formance after acquisition could behighly variable and apparently beset bymotivational difficulties. These prob-lems, if representative of the learningstyle of children with Down syndrome,are extremely important in thinkingabout effective intervention. The resultsof Wishart’s studies using standard intel-ligence test batteries suggest that theyare indeed representative. Test–retest

reliability was very low because suc-cesses gained in one test might notappear upon retest, as soon as 2 weekslater. New skills show up, only to dis-appear shortly thereafter. One could spec-ulate that evidence of such ‘‘rapid for-getting’’ is consistent with damage inthe hippocampal formation but consid-erably more data are required beforethis conclusion can be accepted. Thesedevelopmental differences may have im-portant implications for educators. Forexample, if the learning process involvesmore observable regressions for childrenwith Down syndrome than for otherchildren, it might be important forteachers and interventionists to accountfor these differences with more frequentreviews of materials, and it may be im-portant to monitor the stability withwhich a child with Down syndrome hasacquired a skill.

Information ProcessingThe learning and memory prob-

lems that begin to emerge in lateinfancy become considerably more no-ticeable as the infant grows to child-hood and adolescence. These effectsvary as a function of the type of mem-ory being assessed. Explicit memoryinvolves things like facts and events thatsubjects consciously recollect, whereasimplicit memory can be demonstratedindirectly, without conscious recollec-tion. This distinction has been shownto be important in understanding or-ganic amnesia, since most amnesics areprofoundly impaired on explicit mem-ory tasks but can be relatively normalon implicit tasks. One common kind ofimplicit memory test looks at skills orprocedures, such as mirror-tracing;another common implicit memory testinvolves ‘‘priming,’’ where prior expo-sure to a word or picture can influencesubsequent performance on word-stemor partial-picture completion tasks eventhough the subjects might not recallhaving seen the relevant items before.

Carlesimo et al. [1997] examinedthe performance of children with Downsyndrome in the area of ‘‘implicit’’ (pro-cedural) and ‘‘explicit’’ (episodic) mem-ory paradigms, including word-stemcompletion, list learning, and proserecall. Robust priming effects were seenin the Down syndrome group, compa-rable to those observed in controls,indicating that implicit memory wasintact. However, deficits were observedin both explicit memory tasks. Perform-ance on these kinds of explicit memoryparadigms has been linked to functions

264 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

of the hippocampal system; hence, thedefects suggest differential impairmentin hippocampal function and therebyconverge with the data from study ofspatial cognition. This selective impair-ment of explicit, but not implicit,memory was also reported in Vicari[2001].

A great deal of research attentionhas been focused on deficits in theprocessing of verbal information inDown syndrome, and how these deficitscontribute to poor language and learn-ing outcomes [Byrne et al., 1995;Hesketh and Chapman, 1998; Laws,1998]. Most commonly, the verbalworking memory deficit in Down syn-drome is measured by performance onan auditory digit span task. Poor audi-tory digit span performance findings inDown syndrome were first reportedseveral decades ago [Bilvosky and Share,1965; Rempel, 1974].

More recently, a series of studiesby Hulme and Mackenzie [1992] de-scribed the development of poor verbalworking memory in children withDown syndrome in great detail. Theyfound that children in the Down syn-drome group scored lower than thetypical MA-matched controls on audi-tory digit span, and mental age andverbal working memory tasks weremore correlated in the typically devel-oping children than in the Down syn-drome group. Similar deficits in verbalworking memory are also observablewhen letters are substituted for num-bers in the paradigm [Varnhagen et al.,1987]. This deficit is also not subject tochanges when experimental design ismanipulated in order to reduce unre-lated demands on individuals with Downsyndrome [Marcell and Weeks, 1988;Marcell et al., 1988; Laws, 1998].

Deficits in verbal working mem-ory may relate to neuroanatomical char-acteristics associated with Down syn-drome, including a proportionatelysmaller planum temporale, which isreferred to as the auditory associationcortex [Frangou et al., 1997]. It is im-portant to note, however, that the rela-tionship between the planum temporalevolume deficit and cognitive-linguisticfunctioning in Down syndrome remainsunclear [Frangou et al., 1997].

Though there is a great deal ofevidence of verbal working memorydeficits in Down syndrome, it is impor-tant to note that the main evidence forthis dysfunction comes from studies thatmeasure the processing of acoustically-presented information only. Auditorydigit span, auditory sentence recall, and

other commonly used techniques audi-tory nonword repetition all involve theprocessing of acoustic, rather thanspeech-based visually presented infor-mation. Thus, one might argue that theverbal working memory difficulty inDown syndrome has only been demon-strated in the maintenance of auditorallypresented information, rather than theprocessing of all types of verbal infor-mation. The processing of speech-basedvisual information in Down syndromehas not yet been shown to be equallyimpaired, and importantly, may not be.

In contrast with their perform-ance on verbal processing tasks, individ-uals with Down syndrome show relativestrengths in visuospatial memory, andshow a profile of higher visuospatialprocessing than verbal processing [Sil-verstein et al., 1982; Thase et al., 1984;Pueschel et al., 1987; Wang and Bel-lugi, 1994; Jarrold et al., 1999; Kleinand Mervis, 1999]. The Corsi span, ablock tapping memory task, is most of-ten used as the test of visuospatial proc-essing. In terms of raw scores, mostindividuals without Down syndrome(both with developmental delay andtypically developing) tend to showhigher raw auditory digit span recallscores than Corsi span recall scores.However, individuals with Down syn-drome seem to perform equally well onthese tests, or even show an advantagefor Corsi span recall [Haxby, 1989;Azari et al., 1994; Wang and Bellugi,1994; Vicari et al., 1995; Jarrold andBaddeley, 1997], especially when partic-ipants were not required to includeorder in their responses. These resultshave also been demonstrated with tasksother than the Corsi and auditory digitspan [Pueschel et al., 1987; Hodappet al., 1992; Bower and Hayes, 1994;Klein and Mervis, 1999]. The visuospa-tial working memory advantage is alsodemonstrated when identical stimulusinformation is simply presented eithervisually versus auditorally [Varnhagenet al., 1987].

Neuroanatomical correlates forthis relative strength in Down syndromevisuospatial processing have been pos-ited. Pinter et al. [2001] reported whatthey call ‘‘striking preservation’’ of pari-etal and occipital cortical gray matter inan MRI study of 5–23-year olds withDown syndrome. Studies have shownthe importance of both parietal andoccipital lobe functioning for someaspects of visuospatial processing [Blackand Bernard, 1984; Jonides et al., 1993].

An information processing profilethat includes strengths in visual process-

ing and implicit memory, and deficits inverbal processing and explicit memorycould potentially inform educationalapproaches to working with childrenwith Down syndrome. Instruction thatis verbally based��and especially audito-rally mediated��might pose a greaterchallenge to children with Down syn-drome than instruction that is presentedwith visual supports. Minor and subtleteaching modifications can be made toaddress this issue, without any noticea-ble disruption to a larger classroomenvironment. In addition, instructionthat involves explicit memory, such aslogic problems, may be presented withgreater sensitivity given that this is anarea of challenge as well. Awareness ofthis area of deficit might allow an edu-cator to make informed decisionsregarding prompts, wait time, and sup-ports for a child who may have particu-lar difficulty with this type of informa-tion processing.

ReadingThe cognitive underpinnings of

reading are of particular relevance foreducation in Down syndrome. Despitethe impairments observed in other areasof cognition, many individuals withDown syndrome are able to show com-petence in some aspects of reading de-velopment. In particular, word identifi-cation appears to be an area of relativestrength within reading skills in thispopulation, while word attack and read-ing comprehension appear to be moreimpaired [Byrne et al., 1995; Cupplesand Iacono, 2000; Kay-Raining Birdet al., 2000]. Strengths in word identifi-cation have been linked to relativestrengths in visual processing [Fidleret al., 2005b], while deficits in wordattack skills have been linked to difficul-ties with verbal processing skills inDown syndrome [Boudreau, 2002].

There is some controversy regard-ing the implications of this profile forreading instruction in Down syndrome[see Hodapp and Freeman, 2003 for asummary], with some arguing that visu-ally-based approaches might be war-ranted [Buckley and Bird, 2002], andothers emphasizing the role of phono-logical processing in reading in thispopulation, despite impairments [Cup-ples and Iaocono, 2002]. Regardless ofthis debate, there is evidence that chil-dren with Down syndrome recruit boththeir visual and verbal processing skillswhen reading, particularly when identi-fying words [Cupples and Iacono, 2000;Fidler et al., 2005b].

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 265

BRAIN–BEHAVIORCONNECTIONS

In a series of recent studies, Pen-nington, Nadel and colleagues havetested several different groups of indi-viduals with Down syndrome on arange of tasks designed to directly assessthe function of specific brain systems.This ‘‘cognitive neuropsychological’’approach often uses tasks first developedin animal models, where the criticalunderlying brain circuits can be identi-fied and carefully studied in invasiveexperiments. The team started with afocus on three brain systems identifiedby the neuropathological data, much ofwhich was discussed above: the hippo-campal system, the prefrontal cortex,and the cerebellum. They developed aset of tasks that could, collectively, tellus something about how these brainsystems are faring. In the first set ofstudies, Pennington et al. [2003] foundevidence of specific hippocampal dys-function in our sample of 28 adolescents,using mental age matched controls.

Little evidence of prefrontal dys-function was observed in a battery ofnonverbal tasks. Subsequent pilot work,however, suggested that verbal tasksmight yield a different result, andindeed that is what is being observed(Moon et al., unpublished data). Usingverbal tasks to explore the prefrontalcortex, Moon et al. found in the young(aged 5–11) and old (aged 30–41)groups strong signs of dysfunction inboth the hippocampal and prefrontalsystems. Deficits were observed in arange of tasks although verbal mediationwas necessary to bring out the prefron-tal effect. Taken as a whole, these stud-ies show that particular problemsemerge in the memory domains servedby the hippocampal system and the pre-frontal system. The latter impairmentappears to be linked to the use of verbaltest materials. The impairment in hip-pocampal function could in principlereflect problems in any of the structuresof the hippocampal region; a recentstudy of two neuropsychological para-digms dependent on parahippocampaland perirhinal regions (delayed non-matching to sample and visual pairedcomparison), however, suggests thatthese areas are functioning appropriately,and that the impairment is more likelyto reflect improper development of thehippocampus itself [Dawson et al.,2001].

The prefrontal cortex, as notedalready, plays an important role in awide range of functions, including epi-sodic/explicit memory and working

memory. As noted earlier, explicitmemory is impaired in individuals withDown syndrome. There has beenextensive research on working memoryin this population, and clear deficitshave been observed in a number ofstudies [Varnhagen et al., 1987; Marcelland Weeks, 1988; Laws, 1998; Jarroldet al., 1999]. However, this impairmentseems to be limited to verbal informa-tion, as impairments are minimal invisuospatial domains. The deficitappears to be neither a motor nor artic-ulatory problem [Kanno and Ikeda,2002] and may relate to the so-calledphonological loop [Jarrold and Badde-ley, 2001; Laws, 2002].

OTHER AREAS OFEDUCATIONAL RELEVANCE

In addition to cognitive develop-ment and information processing, thereare other areas of development inDown syndrome that may be relevantto decisions made by educators. Briefdescriptions of findings in the area oflanguage development, social–emotionalfunctioning, and personality-motivationare presented in this section.

Language DevelopmentIndividuals with Down syndrome

generally show a profile of strongerreceptive language skills and weaker ex-pressive language skills. This profileseems to emerge in early childhood andbecome more pronounced as childrenprogress into middle childhood andbeyond [Miller, 1999]. Expressive lan-guage deficits are often manifested interms of morphosyntactic delays[Chapman et al., 1998; Eadie et al.,2002] and difficulties with speech intel-ligibility [Miller et al., 1999; Stoel-Gammon, 2003; Kumin, 2006]. Andwhile receptive language is strongerthan expressive language in most indi-viduals with Down syndrome, someareas such as receptive syntax, are morecompromised than others [Abbedutoet al., 2003]. In contrast, some areas ofpragmatic functioning seem to be anarea of relative strength in individualswith Down syndrome [Johnston andStansfield, 1997; Laws and Bishop,2004].

This language profile is potentiallyrelevant for educational planning in sev-eral ways. First, because many childrenwith Down syndrome have strongerreceptive than expressive language skills,they often understand much more lan-guage than they can produce. To anuninformed educator who is not aware

of the discrepancy between expressiveand receptive language, it might be nat-ural to address the child with input lan-guage and instruction at the level oftheir expressive language. But such anapproach would likely involve a markedunderestimation of the child’s academicand receptive abilities. When consider-ing how best to present class material, itmay be important for educators to iden-tify and target the receptive languagelevel of a child with Down syndromeso as to appropriately challenge themand engage them at their true level ofunderstanding [Roberts et al., 2007].

In addition, it may be useful foreducators to be aware that difficultieswith expressive language��and speechintelligibility in particular��may befrustrating for children with Down syn-drome in classroom contexts. Morpho-syntactic difficulties demand extra moti-vation from children with Down syn-drome to produce lengthier utterances[Miller and Leddy, 1999], and intelligi-bility problems may lead to situationswhere a child must repeat herself inorder to be understood. Thus, educa-tors may want to consider both thesocial and motivational consequences ofexpressive language difficulties. It maybe beneficial to identify ways to mini-mize the potential for negative experi-ences, while allowing the child withDown syndrome to benefit from theopportunity to build their speech, lan-guage, and communication skills.

Social–Emotional FunctioningThe majority of individuals with

Down syndrome show strengths in vari-ous aspects of social–emotional func-tioning, exhibiting behaviors that sug-gest evidence of intact social relatednessand some measure of social competencein early childhood [Fidler et al., 2006].In the first few years of life, markers ofprimary intersubjectivity, the earliestforms of dyadic social relating based onemotional displays and reciprocal signal-ing, are identified as emerging in adelayed, but competent manner in thispopulation [see Fidler, 2006 for areview]. In particular, young childrenwith Down syndrome show evidence ofprimary intersubjective development inthe form of increased eye gaze andvocalizations directed to other people[Gunn et al., 1982; Crown et al., 1992;Legerstee et al., 1992; Kasari et al.,1995], increased direction of positivefacial displays in the form of smiles[Kasari and Freeman, 1990; Fidler et al.,2005a].

266 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

In addition, there is also supportfor the notion that many aspects of sec-ondary intersubjectivity, the ability toengage with a social partner in a triadicfashion, also emerge with competence,albeit in a delayed fashion. Several stud-ies suggest that children with Downsyndrome show either MA-appropriateor even increased levels of joint atten-tion relative to other children withoutDown syndrome [Mundy et al., 1988;Kasari et al., 1995; Fidler et al., 2005c].Toddlers and preschoolers with Downsyndrome also show competent triadicrelating in the forms of play acts, turntaking, invitations, and object shows[Mundy et al., 1988; Sigman and Rus-kin, 1999].

In the context of these strengthsin the development of social relatednessand other social–emotional skills, thereis evidence that children with Downsyndrome may show difficulties whencognitive demands in social decisionmaking increase. While socializationskills remain a relative strength in mid-dle childhood [Dykens et al., 1994],and many children with Down syn-drome appear to be able to form recip-rocal friendships with peers [Freemanand Kasari, 2002], there is mixed evi-dence regarding whether individualswith Down syndrome show impair-ments in the ability perform more com-plex social cognition tasks [Baron-Cohen et al., 1985; Baron-Cohen, 1989;Yirmiya et al., 1996; Zelazo et al., 1996;Abbeduto et al., 2001]. Williams andWishart [in press] identify other factorsthat may contribute to the difficulty thatindividuals with Down syndrome haveon social–cognitive tasks, including exec-utive function or memory difficulties.Nevertheless, it may be that despitecompetence in the area of social related-ness, as the demands and complexitiesof social situations increase in middlechildhood and beyond, individuals withDown syndrome may show difficultieswith social adaptation and selectingappropriate social strategies, especially asthey enter adolescence and face changesin emotional functioning and mood[Dykens et al., 2002; Fidler et al., 2005a].

This social profile is relevant foreducational planning in several ways.First, these strengths may make it possi-ble for children with Down syndrometo learn through social techniques suchas modeling, peer collaboration, socialgroups [Lloveras and Fornells, 1998;Rogoff, 2003]. Though there is a sur-prising lack of empirical exploration ofthe efficacy of these techniques in chil-dren with Down syndrome, it could be

hypothesized that such techniqueswould capitalize on social motivationmay be a useful reinforcer for childrenwith Down syndrome. However, it isimportant to consider that the strongsocial motivation observed in this popu-lation may serve as a distractor as well[Wishart, 1996; Fidler, 2006; see discus-sion below]. There is also some evi-dence that, because of strengths in earlysocial relatedness in this population,affective cues put forth by a teacher orinterventionist can impact learning andmotivation in a particularly pronouncedway [Park et al., 2005]. It may also beimportant for educators to be mindfulof potential changes in mood and socialengagement as children with Downsyndrome transition into adolescence,and perhaps adopt modified strategies asthese behavioral changes become evi-dent [Dykens et al., 2002].

Personality-MotivationAnother aspect of the behavioral

profile in Down syndrome that may beeducationally relevant relates to motiva-tional orientation and task persistence[Gunn and Cuskelly, 1991]. In labora-tory settings, when presented with taskssuch as puzzles and other nonsocial/nonverbal tasks, children with Downsyndrome have been shown to abandonthe task sooner than other children atsimilar developmental levels, and toadopt strategies that divert attentionaway from the task [Landry and Cha-pieski, 1990; Pitcairn and Wishart,1994; Ruskin et al., 1994; Vlachou andFarrell, 2000; Kasari and Freeman,2001]. This, coupled with the strengthsin aspects of social initiation describedin the previous section, can lead to astyle that involves an over-reliance onsocial strategies, especially in contextsthat require instrumental thinking[Kasari and Freeman, 2001]. It may bethat the interaction between emergingdifficulties with instrumental thinkingand strengths in social relatedness leadto a personality-motivation orientationthat ultimately impacts the ability of achild with Down syndrome to learn ef-fectively [Wishart, 1996; Fidler, 2006].

A style that leads children toremove themselves from challenging sit-uations in favor of social interactionmay deprive children with Down syn-drome of important opportunities tochallenge themselves and gain new skillsthrough active engagement with theenvironment that surrounds them.Awareness of this profile may be impor-tant for educators when selecting tech-

niques for involving individuals withDown syndrome in classroom settings.It may be important for educators toidentify situations when the child withDown syndrome may be recruitingsocial strategies when engaging with thetask at hand is more appropriate [Fidler,2006]. Educators might also managebehavior using social consequences asreinforcement, not as a distracter duringtasks that might pose a cognitive chal-lenge [Fidler, 2006].

CHALLENGES IN LINKINGRESEARCH TO PRACTICE

Are Behavioral PhenotypesModifiable?

Though many aspects of the be-havioral phenotype in Down syndromeare potentially relevant for educators,there are several challenges that must beaddressed as researchers aim to translateresearch findings into educational prac-tice. The first challenge relates to thenotion of genes and modifiability. Thereis a danger in discussing the notion ofbehavioral outcomes associated withgenetic disorders in that genetic effectscan connote fixed, nonmodifiable path-ways. What we now know about themechanisms by which genes give rise tophenotypes, particularly behavioral phe-notypes, indicates that we need notworry about this danger. First, froma neurodevelopmental perspective, atevery step of the way, opportunities existto modulate the translational process. Inaddition, all learning and education isrooted in the notion that neurophysio-logical changes can be observed inresponse to environmental input, lead-ing the brain to undergo various typesof reorganization [Nelson, 2000]. Nel-son notes that it is commonly under-stood that, ‘‘. . . the success of earlychildhood intervention strategies rests toa great degree on the relative plasticityof the human brain (p. 222),’’ and thisapplies to children with and withoutgenetic disorders alike.

In addition, potential evidence ofthe modifiability of the Down syn-drome phenotypic profile has beenreported in a long-term study of Britishinclusion in this population. Buckleyet al. [2006] report that the practice ofincluding children with Down syn-drome in mainstream classrooms inEngland has had an impact on the phe-notypic profile in older children andadolescents with Down syndrome. Theynote that previous studies showed thatchildren with Down syndrome who

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 267

attended school in special education set-tings demonstrated a profile of strengthsin socialization and daily living skills,but deficits in adaptive communicationabilities [Dykens et al., 1994; Fidleret al., 2006]. However, in their sampleof children with Down syndrome whounderwent schooling in inclusive set-tings, they found that the marked defi-cits in adaptive communication werenot observable [Buckley et al., 2006].They argue that the social challengesassociated with being educated in aninclusive setting modified the phenotypicprofile and narrowed the gap betweenareas of strength and challenge. Theauthors of this study note that this war-rants replication, but if supported, therethis would be a critical means of modi-fying the profile of strengths and weak-nesses associated with Down syndrome.

When considering whether edu-cational approaches can modify pheno-typic profiles, it is also important tonote that the pattern of strengths andweaknesses associated with genetic dis-orders does not simply appear in a pro-nounced fashion in middle childhood.There is a developmental process thatleads to the more pronounced end statesof relative strengths and weaknesses inany genetic disorder. This is importantto recognize because there may beopportunities to target early emergingphenotypic characteristics in very youngchildren, before dissociations in profilebecome pronounced [see Karmiloff-Smith, 1997, 1998; Fidler, 2005, 2006].Fidler [2005] argued that for someaspects of the Down syndrome behav-ioral phenotype, it may be possible toidentify early developmental precursorsto later more pronounced outcomes. Ifthese early developmental precursorscan be identified and targeted withempirically validated intervention tech-niques, this too may be another meansfor altering the developmental pathwayand the phenotypic profile associatedwith Down syndrome.

What Constitutes EmpiricalSupport for Etiology-SpecificEducation/Intervention?

Another challenge in the attemptto bridge research and practice in thisarea relates to the empirical validationof techniques aimed at addressing phe-notype-specific dimensions. There is asmall, but growing literature thatdescribes the efficacy of educationaltechniques such as computer-basedlearning (Lloyd et al., 2006; Ortega-Tudela and Gomez-Ariza, 2006],instructional approaches to reading and

its component skills [Laws et al., 1996;Moni and Jobling, 2001; Cupples andIacono, 2002; Kennedy and Flynn,2003; van Bysterveldt et al., 2006],and math skills [Irwin, 1991; Nye andBuckley, 2006; Ortega-Tudela andGomez-Ariza, 2006] for children withDown syndrome. While more studies ofthis kind are warranted, only a few ofthese studies show educational benefitswhen using one specific technique overanother Cupples and Iacono, 2002].

The question remains whetherthis type of empirical validation is suffi-cient to warrant a syndrome-specificapproach to educational planning. Somemight argue that the efficacy of thesetechniques have little to do with thephenotypic profile associated withDown syndrome��rather they simplyshow that one technique is superior toanother regardless of the population towhich it is applied. By extension, itcould be argued that in order to justifya syndrome-specific set of recommenda-tions for educational practice, theremust be a set of techniques that workdifferentially across populations. That is,there must be techniques identified thatare effective for children with Downsyndrome, but not effective for childrenwho do not have Down syndrome orthe developmental profile associatedwith Down syndrome [see Fidler et al.,2007 for a discussion]. At present, thereare relatively few examples in the litera-ture that demonstrate such differentialeffects [Fey et al., 2006; Yoder andWarren, 2002]. Those that do exist sug-gest that the personality-motivationalorientation associated with Down syn-drome may be particularly important toconsider when selecting educational andintervention techniques [Yoder andWarren, 2002; see Fidler et al., 2007 fora discussion of this issue].

Is Syndrome-Specific EducationFeasible?

A third challenge to the idea oflinking phenotype research into practicerelates to issues of classroom manage-ment and the training of teachers.While future research may show thebenefits of etiology-specific instructionalapproaches, it could be argued that spe-cific techniques for different children inthe classroom would be too unwieldyand would require too great of a per-sonnel demand. It could also be arguedthat the training of teachers in etiology-specific instructional approaches wouldmake teacher education programs toolengthy of a process, requiring a masteryof approaches that target any number of

the many syndromes and behavioral dis-orders present in the student popula-tion.

While adopting an etiology-spe-cific approach in the classroom willundoubtedly place additional demandson educators, some potential approachesinvolve less of a diversion of resourcesfrom the larger classroom culture thanothers. Techniques to be developed inthe future can be imbedded in naturalis-tic ways, and might only involve subtleadjustments in teacher decision makingand presentation of material. For exam-ple, supplementing instruction withsupports that rely on a favored informa-tion processing modality might not bedetectable to the larger classroom, andin some instances could potentiallyenhance instruction for children in theclassroom without disabilities. In addi-tion, while some additional trainingmight be involved for teachers, thesecan come in the form of continuingeducational trainings, or useful informa-tional materials (websites, booklets) thatneed not burden a teacher in training.While the details of implementationwould need to be addressed in a real-world process, it is likely that theimplementation of some syndrome-spe-cific instructional approaches, if theyreceive empirical validation, might notnecessarily pose a prohibitively largechallenge to educators.

Future DirectionsDespite the many advances that

have been made in the study of brainand behavioral development in Downsyndrome, there is still a great deal ofprogress to be made both in the basicstudy of development in Down syn-drome and in the application of thesefindings to practice. In terms of thepotential contributions of the neurobio-logical approach, future work uncover-ing the neurobiological causes of thecognitive, language, and behavioralimpairments associated with Down syn-drome will ultimately lead to creationof an ever-more precise animal modelof Down syndrome. A more preciseanimal model of Down syndrome couldmake it possible to develop biologicalinterventions that might ultimatelyimpact development in this population.Thus, advances in this area will rely onthe close collaboration of behavioral sci-entists who are carefully delineating thenature of the Down syndrome behav-ioral phenotype, neurobiologists whoare able to map these phenotypic out-comes onto brain anatomy and brain

268 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

physiology in this population, and ani-mal model researchers who can use thisinformation to develop an ever moreprecise model of the disorder. This pro-cess will be aided by the neuropsycho-logical approach, which offers thepromise of identifying exactly thoseareas of disproportionate cognitive im-pairment that might guide the mappingfrom behavior to brain functioning.

Given the probabilistic nature ofphenotypic outcomes in genetic disor-ders, these approaches may also make itpossible to more deeply understand thenature of within syndrome variability inthe population of individuals withDown syndrome. As researchers collab-orate to uncover the pathway from geneto brain to behavior, it may be possibleto identify with greater precision thesources of within-syndrome variabilityin outcomes of interest, and it may bepossible to address the needs of childrenwho show variations around the pheno-typic profile that is associated with thelarger group of individuals with Downsyndrome. These advances offer thehope of even more targeted educationalplanning and the possibility of addressingthe variability in outcomes that is classi-cally associated with this population.

Another important future direc-tion for research in Down syndromerelates to the importance of detectingemerging phenotypes in early childhood[Fidler, 2005]. This type of researchtransforms the view of outcomes inDown syndrome from a static, cross-sectional approach into a dynamic, lon-gitudinal approach to studying this pop-ulation. In taking this more dynamicview, it may be possible to identify themore subtle developmental precursorsto more pronounced outcomes in laterchildhood and adolescence. These earlyprecursors may serve as potentially use-ful targets for early intervention in thispopulation. Rather than waiting tointervene once a split profile ofstrengths and weaknesses has becomepronounced, educators and interven-tionist may be able to target these earlyprecursors in their more subtle forms,which may set development on a moreoptimal pathway.

Finally, as the research communitysorts through the various controversiesrelated to syndrome-specific educationalapproaches, continued advances havestill been made in the education ofindividuals with Down syndrome.These efforts continue to challenge theoutmoded notions that children withgenetic disorders such as Down syn-drome have limited educational poten-

tial. Though the goal of bridgingresearch and practice in the study ofdevelopment in Down syndrome faceschallenges-especially the difficulty ofcollaboration among scientists acrossneighboring fields-it is a goal thatpromises greater returns than simplyeducating children with Down syn-drome according to their severity ofimpairment (mild, moderate, severe,profound intellectual disability) andignoring the complex profile associatedwith the disorder. It is likely that ourbest hope for improving outcomes ingenetic disorders such as Down syn-drome lies in our ability to use all ofthe scientific information that is avail-able, with developmentalists, educationscientists, and brain experts collaborat-ing to generate the most effective andinnovative practice approaches possible.n

REFERENCESAbbeduto L, Murphy MM, Cawthon SW, et al.

2003. Receptive language skills of adoles-cents and young adults with Down or fragileX syndrome. Am J Ment Retard 108:149–160.

Abbeduto L, Pavetto M, Kesin E, et al. 2001. Thelinguistic and cognitive profile of Down syn-drome: evidence from a comparison withfragile X syndrome. Down Syndrome: ResPract 7:9–15.

Alton S. 1998. Differentiation not discrimination:delivering the curriculum for children withDown’s syndrome in mainstream schools.Support Learn 13:167–173.

Azari NP, Horwitz B, Pettigrew KD, et al. 1994.Abnormal pattern of cerebral glucose meta-bolic rates involving language areas in youngadults with Down syndrome. Brain Lang46:1–20.

Baron-Cohen S. 1989. The autistic child’s theoryof mind: a case of specific developmentaldelay. J Child Psychol Psychiatry 30:285–297.

Baron-Cohen S, Leslie A, Frith U. 1985. Doesthe autistic child have a ‘‘theory of mind’’?Cognition 21:37–46.

Bar-Peled O, Israeli M, Ben-hur H, et al. 1991.Developmental patterns of muscarinic recep-tors in normal and Down’s syndrome fetalbrain��an autoradiographic study. NeurosciLett 133:154–158.

Benda CE. 1971.Mongolism. In: Minckler J, edi-tor. Pathology of the nervous system, Vol.2. New York: McGraw-Hill. p. 1867.

Bilvosky D, Share J. 1965. The Illinois Test ofPsycholinguistic ability and Down’s Syn-drome: an exploratory study. Am J MentDefic 70:78–82.

Black FW, Bernard BA. 1984. Constructionalapraxia as a function of lesion locus and sizein patients with focal brain damage. Cortex20:111–120.

Blackwood W, Corsellis JAN. 1976. Greenfield’sneuropathology. Chicago: Yearbook Medical.p 420–421.

Boudreau, D. 2002. Literacy skills in children andadolescents with Down syndrome. ReadWrit Interdiscipl J 15:497–525.

Bower A, Hayes A. 1994. Short-term memorydeficits and Down’s syndrome: a comparativestudy. Down Syndrome: Res Pract 2:47–50.

Brooksbank BWL, Walker D, Balazs R, et al.1989. Neuronal maturation in the foetalbrain in Down syndrome. Early Hum Dev18:237–246.

Buckley S, Bird G. 2002. Cognitive developmentand education: perspectives on Down syn-drome from a twenty-year research pro-gramme. In: Cuskelly M, Jobling A, BuckleyS, editors. Down syndrome across the life-span. London England: Whurr Publishers.p 66–80.

Buckley S, Bird G, Sacks B. 2006. Evidence thatwe can change the profile from a study ofinclusive education. Down Syndrome: ResPract 9:51–53.

Byrne A, Buckley S, MacDonald J, et al. 1995.Investigating the literacy, language, and mem-ory skills of children with Down’s syndrome.Down’s Syndrome: Res Pract 3:53– 58.

Campbell J, Gilmore L, Cuskelly M. 2003.Changing student teachers’ attitudes towardsdisability and inclusion. J Intellect Dev Disa-bil 28:369–379.

Carlesimo GA, Marotta L, Vicari S. 1997. Long-term memory in mental retardation: evi-dence for a specific impairment in subjectswith Down’s syndrome. Neuropsychologia35:71–79.

Casale-Giannola D, Wilson Kamens M. 2006.Inclusion at a university: experiences of ayoung woman with Down syndrome. MentRetard 44:344–352.

Chapman R, Seung H, Schwartz S, et al. 1998.Language skills of children and adolescentswith Down syndrome: II. Production deficits.J Speech Lang Hear Res 41:861–873.

Collacott RA, Cooper SA, McGrother A. 1992.Differential rates of psychiatric disorders inadults with Down’s syndrome compared toother mentally handicapped adults. Br J Psy-chiatry 161:671–674.

Crome L, Cowie V, Slater, E. 1966. A statisticalnote on cerebellar and brain-stem weight inMongolism. J Ment Defic 10:69–72.

Crown CL, Feldstein S, Jasnow MD, et al. 1992.Down’s syndrome and infant gaze: gazebehavior of Down’s syndrome and nonde-layed infants in interactions with their moth-ers. Eur J Child Adoles Psychiatry: ActaPaedopsychiatr 55:51–55.

Cupples L, Iacono T. 2000. Phonological aware-ness and oral reading skill in children withDown syndrome. J Speech Lang Hear Res43:595–608.

Cupples L, Iacono T. 2002. The efficacy of‘whole word’ versus ‘analytic’ readinginstruction for children with Down Syn-drome. Read Writ 15:549–574.

Dawson G, Osterling J, Rinaldi J, Carver L,McPartland J. 2001. Brief Report: Recogni-tion memory and stimulus-reward associa-tions: indirect support for the role of ventro-medial prefrontal dysfunction in autism. JAutism Dev Disord 31:337–341.

Dykens E, Shah B, Beck T, et al. 2002. Maladap-tive behavior in children and adolescentswith Down’s syndrome. J Intellect DisabilRes 46:484–492.

Dykens EM. 1995. Measuring behavioral pheno-types: provocations from the ‘‘New Genetics’’.Am J Ment Retard 99:522–532.

Dykens EM, Hodapp RM, Evans DW. 1994. Pro-files and development of adaptive behaviorin children with Down syndrome. Am JMent Retard 98:580–587.

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 269

Dykens EM, Hodapp RM, Finucane B. 2000.Genetics and mental retardation syndromes.Baltimore: Paul H. Brookes Publishing.

Eadie PA, Fey ME, Douglas JM, et al. 2002. Pro-files of grammatical morphology and sen-tence imitation in children with specific lan-guage impairment and Down syndrome. JSpeech Lang Hear Res 45:720–732.

Engidawork E, Lubec G. 2003. Molecular changesin fetal Down syndrome brain. J Neurochem84:895–904.

Fey ME, Warren SF, Brady N, et al. 2006. Earlyeffects of responsivity education/prelinguisticmilieu teaching for children with develop-mental delays and their parents. J SpeechLang Hear Res 49:526–547.

Fidler DJ. 2005. The emerging Down syndromebehavioral phenotype in early childhood:implications for practice. Infants YoungChild 18:86–103.

Fidler DJ. 2006.The emergence of a syndrome-specific personality-motivation profile inyoung children with Down syndrome. In:Rondal JA, Perera J, editors. Down syn-drome neurobehavioral specificity. Wiley.

Fidler DJ, Barrett KC, Most DE. 2005a. Agerelated differences in smiling and personalityin Down syndrome. J Dev Phys Disabil 17:263–280.

Fidler DJ, Hepburn S, Rogers S. 2006. Earlylearning and adaptive behavior in toddlerswith Down syndrome: evidence for anemerging behavioral phenotype? Down Syn-drome: Res Pract 10:37–44.

Fidler DJ, Lawson J, Hodapp RM. 2003. Whatdo parents want? an analysis of education-related comments made by parents of chil-dren with different genetic mental retarda-tion syndromes. J Intellect Dev Disabil 28:196–204.

Fidler DJ, Most DE, Guiberson MM. 2005b.Neuropsychological correlates of word iden-tification in Down syndrome. Res Dev Dis-abil 26:487–501.

Fidler DJ, Philofsky A, Hepburn S. 2007. Lan-guage phenotypes and intervention planning:bridging research and practice. Ment RetardDev Disabil Res Rev 13:47–57.

Fidler DJ, Philofsky A, Hepburn S, et al. 2005c.Nonverbal requesting and problem-solvingby toddlers with Down syndrome. Am JMent Retard 110:312–322.

Fl�orez J, Del Arco C, Gonzalez A, et al. 1990.Autoradiographic studies of neurotransmitterreceptors in the brain of newborn infantswith Down syndrome. Am J Med GenetSuppl 7:301–305.

Frangou S, Aylward E, Warren A, et al. 1997.Small planum temporale volume in Down’ssyndrome: a volumetric MRI study. Am JPsychiatry 154:1424–1429.

Freeman SFN, Hodapp RM. 2000. Educatingchildren with Down syndrome: linking be-havioral characteristics to promising inter-vention strategies. Down Syndrome Quart5:1–9.

Freeman SFN, Kasari C. 2002. Characteristics andqualities of the play dates of children withDown syndrome: emerging or true friend-ships. Am J Ment Retard 107:16–31.

Garcia SM, Conte EV. 2004. Improvement ofcognitive functions in children with Downsyndrome through the Bright Start Program.J Cogn Educ Psychol 4:65–86.

Gathercole SE, Adams AM. 1995. Phonologicalworking memory and speech production inpreschool children. J Speech Hear Res 38:403–414.

Ghaziuddin M, Tsai L, Ghaziuddin N. 1992. Au-tism in Down’s syndrome: presentation anddiagnosis. J Intellect Disabil Res 36:449–456.

Golden JA, Hyman BT. 1994. Development ofthe superior temporal neocortex is anoma-lous in trisomy 21. J Neuropathol Exp Neu-rol 53:513–520.

Gunn P, Cuskelly M. 1991. Down syndrome tem-perament: the stereotype at middle child-hood and adolescence. Int J Disabil DevEduc 38:59–70.

Gunn P, Berry P, Andrews RJ. 1982. Lookingbehavior of Down syndrome infants. Am JMent Defic 87:344–347.

Hamill L. 2003. Going to college: the experiencesof a young woman with Down syndrome.Ment Retard 41:340–353.

Haxby JV. 1989. Neuropsychological evaluation ofadults with Down’s syndrome: patterns ofselective impairment in non-demented oldadults. J Ment Defic Res 33:193–210.

Hepburn S, Fidler DJ, Rogers S. in press. Autismsymptoms in toddlers with Down syndrome.J Appl Res Intellect Disabil.

Hesketh LJ, Chapman RS. 1998. Verb use byindividuals with Down syndrome. Am JMent Retard 103:288–304.

Hill Karrer J, Karrer R, Bloom D, Chaney L,Davis R. 1998. Event-related brain poten-tials during an extended visual recognitionmemory task depict delayed development ofcerebral inhibitory processes among 6-month-old infants with Down syndrome. IntJ Psychophysiol 29:167–200.

Hodapp RM, Fidler DJ. 1999. Special educationand genetics: connections for the 21st cen-tury. J Spec Educ 33:130–137.

Hodapp RM, Freeman SFN. 2003. Advances ineducational strategies for children with Downsyndrome. Curr Opin Psychiatry 16:511–516.

Hodapp RM, Leckman JF, Dykens EM, et al.1992. K-ABC profiles in children with frag-ile X syndrome, Down syndrome, and non-specific mental retardation. Am J MentRetard 97:39–46.

Hulme C, Mackenzie S. 1992. Working memoryand severe learning difficulties. Hove: Law-rence Erlbaum.

Iarocci G, Virji-Babul N, Reebye P. 2006. TheLearn at Play Program (LAPP): mergingfamily, developmental research, early inter-vention, and policy goals for children withDown syndrome. J Policy Pract Intellec Dis-abil 3:11–21.

Irwin KC. 1991. Teaching children with Downsyndrome to add by counting-on. EducTreat Child 14:128–141.

Jarrold C, Baddeley AD. 1997. Short-term mem-ory for verbal and visuospatial informationin Down’s syndrome. Cogn Neuropschiatry2:101–122.

Jarrold C, Baddeley AD. 2001. Short-term mem-ory in Down syndrome: applying the work-ing memory model. Down Syndrome ResPract 7:17–23.

Jarrold C, Baddeley AD, Hewes AK. 1999. Ge-netically dissociated components of workingmemory: evidence from Down’s and Wil-liams syndrome. Neuropsychologia 37:637–651.

Jernigan TL, Bellugi U, Sowell E, Doherty S,Hesselink JR. 1993. Cerebral morphologicaldistinctions between Williams and Downsyndrome. Arch Neurol 50:186–191.

Jiang ZD, Wu YY, Liu XY. 1990. Early develop-ment of brainstem auditory evoked poten-tials in Down’s syndrome. Early Hum Dev23:41–51.

Johnston F, Stansfield J. 1997. Expressive prag-matics skills in pre-school children withand without Down’s syndrome: parentalperceptions. J Intellec Disabil Res 41:19–29.

Jonides J, Smith EE, Koeppe RA, et al. 1993.Spatial working memory in humans asrevealed by PET. Nature 363:623–625.

Kanno K, Ikeda Y. 2002. Word-length effect inverbal short term memory in individualswith Down’s syndrome. J Intellec DisabilRes 46:613–618.

Karmiloff-Smith A. 1997. Crucial differencesbetween developmental cognitive neuro-science and adult neuropsychology. DevNeuropsychol 13:513–524.

Karmiloff-Smith A. 1998. Development itself isthe key to understanding developmental dis-orders. Trends Cogn Sci 2:389–398.

Kasari C, Freeman SFN. 2001. Task-related socialbehavior in children with Down syndrome.Am J Ment Retard 106:253–264.

Kasari C, Freeman S, Mundy P, et al. 1995.Attention regulation by children with Downsyndrome: coordinated joint attention andsocial referencing looks. Am J Ment Retard100:128–136.

Kasari C, Mundy P, Yirmiya N, et al. 1990. Affectand attention in children with Down syn-drome. Am J Ment Retard 95:55–67.

Kates WR, Folley BS, Lanham DC, et al. 2002.Cerebral growth in fragile X syndrome:review and comparison with Down syn-drome. Microsc Res Tech 57:159–167.

Kay-Raining Bird E, Cleave PL, McConnell L.2000. Reading and phonological awarenessin children with Down syndrome. Am JSpeech-Lang Pathol 9:319–330.

Kennedy EJ, Flynn MC. 2003. Training, phono-logical awareness in children with Downsyndrome. Res Dev Disabil 24:44–57.

Kent L, Evans J, Paul M, et al. 1999. Comorbidityof autistic spectrum disorders in childrenwith Down syndrome. Dev Med ChildNeurol 41:153–158.

Klein BP, Mervis CB. 1999. Contrasting patternsof cognitive abilities of 9- and 10- year-oldswith Williams syndrome or Down syn-drome. Dev Neuropsychol 16:177–196.

Kleschevnikov AM, Belichenko PV, Villar AJ,et al. 2004. Hippocampal long-term poten-tiation suppressed by increased inhibitionin the Ts65Dn mouse, a genetic model ofDown syndrome. J Neurosci 24:8153–8160.

Kumin L. 2006. Speech intelligibility and childhoodverbal apraxia in children with Down syn-drome. Down Syndrome: Res Pract 10: 10–22.

Landry SH, Chapieski ML. 1990. Joint attentionand infant toy exploration: effects of Downsyndrome and prematurity. Child Dev 60:103–118.

Laws G. 1998. The use of nonword repetition asa test of phonological memory in childrenwith Down syndrome. J Child Psychol Psy-chiatry 39:1119–1130.

Laws G. 2002. Working memory in children andadolescents with Down syndrome: evidencefrom a colour memory experiment. J ChildPsychol Psychiatry 43:353–364.

Laws G, Bishop DVM. 2004. Verbal deficits inDown’s syndrome and specific languageimpairment: a comparison. Int J Lang Com-mun Disord 39:423–451.

Laws G, MacDonald J, Buckley S. 1996. Theeffects of a short training in the use of a re-hearsal strategy on memory for words andpictures in children with Down syndrome.Down Syndrome: Res Pract 4:70–78.

270 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

Legerstee M, Bowman TG, Fels S. 1992. Peopleand objects affect the quality of vocalizationsin infants with Down syndrome. Early DevParenting 1:149–156.

Lloveras RB, Fornells MG. 1998. The group: aninstrument of intervention for the globaldelve of the child with Down syndrome inthe process of social inclusion. Down Syn-drome: Res Pract 5:88–92.

Lloyd J, Moni KB, Jobling A. 2006. Breaking thehype cycle: using the computer effectivelywith learners with intellectual disabilities.Down Syndrome: Res Pract 9:68–74.

Mangan PA. 1992. Spatial memory abilities andabnormal development of the hippocampalformation in Down syndrome. Unpublisheddoctoral dissertation, University of Arizona,Tucson.

Marcell MM, Weeks SL. 1988. Short-term mem-ory difficulties and Down’s syndrome. JMent Defic Res 32:153–162.

Marcell MM, Harvey CF, Cothran LP. 1988. Anattempt to improve auditory short-termmemory in Down’s syndrome individualsthrough reducing distractions. Res Dev Dis-abil 9:405–417.

Miller JF. 1999. Profiles of language developmentin children with Down syndrome. In: MillerJ, Leddy M, Leavitt LA, editors. Improvingthe communication of people with Downsyndrome. Baltimore: Brookes Publishing. p11–40.

Miller JF, Leddy M. 1999. Verbal fluency, speechintelligibility, and communicative effective-ness. In: Miller JF, Leddy M, Leavitt LA,editors. Improving the communication ofpeople with Down syndrome. Baltimore:Brookes Publishing. p 81–91.

Miller J, Leddy M, Leavitt LA, editors. 1999.Improving the communication of peoplewith Down syndrome. Baltimore: BrookesPublishing.

Moni KB, Jobling A. 2001. Reading-related liter-acy learning of young adults with Downsyndrome: findings from a three year teach-ing and research program. Int J Disabil DevEduc 48:377–394.

Mundy P, Sigman M, Kasari C, et al. 1988. Non-verbal communication skills in Down syn-drome children. Child Dev 59:235–249.

Nadel L. 1986. Down syndrome in neurobiologi-cal perspective. In: Epstein CJ, editor. Theneurobiology of Down syndrome. NewYork: Raven Press. p 239–251.

Nelson C. 2000. The neurobiological bases ofearly intervention. In: Shokoff JP, Meisels J,editors. Handbook of early childhood inter-vention. New York: Cambridge UniversityPress. p 204–227.

Nye J, Buckley S. 2006. Evaluating the Numiconsystem of teaching number to children withDown syndrome. Down Syndrome: ResPract 11:97–104.

Ohr PS, Fagen JW. 1991. Conditioning and long-old infants with Down syndrome. Am JMent Retard 96:151–162.

Ohr PS, Fagen JW. 1993. Temperament, condi-tioning, old infants with Down syndrome. JAppl Dev Psychol 14:175–190.

Ohr PS, Fagen JW. 1994. Contingency learningDown syndrome. Am J Ment Retard 99:74–84.

Ortega-Tudela JM, Gomez-Ariza CJ. 2006. Com-puter-assisted teaching and mathematicallearning in Down syndrome children. JComput Assist Learn 22:298–307.

Paly RJ, Hurley AD. 2002. Down syndrome andautistic disorder. Ment Health Aspects DevDisabil 5:64–65.

Park S, Singer GH, Gibson M. 2005. The func-tional effect of teacher positive and neutralaffect on task performance of students withsignificant disabilities. J Positive Behav Interv7:237–246.

Pazos A, Del Olmo E, Diaz A, et al. 1994. Sero-tonergic (5-HhT1A and 5-HhT1D) andmuscarinic cholinergic receptors in DSbrains: an autoradiographic analysis. Posterpresented at International Down syndromeResearch Conference, Charleston, SouthCarolina, April 1994.

Pennington BF, Moon J, Edgin J, et al. 2003. Theneuropsychology of Down syndrome: evi-dence for hippocampal dysfunction. ChildDev 74:75–93.

Pinter JD, Eliez S, Schmitt E, et al. 2001. Neuro-anatomy of Down’s syndrome: a high-reso-lution MRI study. Am J Psychiatry 158:1659–1665.

Pitcairn TK, Wishart JG. 1994. Reactions ofyoung children with Down’s syndrome to animpossible task. Br J Dev Psychol12:485–489.

Pueschel SR, Gallagher PL, Zartler AS, et al.1987. Cognitive and learning profiles inchildren with Down syndrome. Res DevDisabil 8:21–37.

Rast M, Meltzoff AN. 1995. Memory and repre-sentation in young children with Downsyndrome: exploring deferred imitation andobject permanence. Dev Psychopathol 7:393–407.

Rempel ED. 1974. Psycholinguistic abilities ofDown’s syndrome children. In: Proceedingsof the Annual Meeting of the AmericanAssociation on Mental Deficiency, Toronto.

Roberts J, Price J, Malkin C. 2007. Language andcommunication development in Down syn-drome. Ment Retard Dev Disabil Res Rev13:26–35.

Rogoff B. 2003. The cultural nature of human de-velopment. Oxford: Oxford University Press.

Rondal J, Buckley S. 2001. Speech and languageintervention in Down syndrome. New York:Wiley.

Ruskin EM, Kasari C, Mundy P, et al. 1994.Attention to people and toys during socialand object mastery in children with Downsyndrome. Am J Ment Retard 99:103–111.

Schmidt-Sidor B, Wisniewski K, Shepard Th,et al. 1990. Brain growth in Down syn-drome subjects 15 to 22 weeks of gestationalage and birth to 60 months. Clin Neuropa-thol 9:181–190.

Sigman M, Ruskin E. 1999. Continuity andchange in the social competence of childrenwith autism, Down syndrome, and develop-mental delays. Monogr Soc Res Child Dev64:v–114.

Silverstein AB, Legutki G, Friedman SL, et al.1982. Performance of Down syndrome indi-viduals on the Stanford–Binet IntelligenceScale. Am J Ment Retard 86:548–551.

Stoel-Gammon C. 2003. Speech acquisition andapproaches to intervention. In: Rondal J,Buckley S, editors. Speech and languageintervention in Down syndrome. London:Whurr. p 49–62.

Thase ME, Tigner R, Smeltzer DJ, et al. 1984.Age-related neuropsychological deficits inDown’s syndrome. Biol Psychiatry 19:571–585.

Trent JA, Kaiser AP, Wolery M. 2005. The use ofresponsive interaction strategies by siblings.Top Early Child Spec Educ 25:107–118.

van Bysterveldt AK, Gillon GT, Moran C. 2006.Enhancing phonological awareness and letterknowledge in preschool children with Downsyndrome. Int J Disabil Dev Educ53:301–329.

Varnhagen CK, Das JP, Varnhagen S. 1987. Audi-tory and visual memory span: cognitiveprocessing by TMR individuals with Downsyndrome or other etiologies. Am J MentDefic 91:398–405.

Vicari S. 2001. Implicit versus explicit memoryfunction in children with Down and Wil-liams syndrome. Down Syndrome: Res Pract7:35–40.

Vicari S, Carlesimo A, Caltagirone C. 1995.Short-term memory in persons with intel-lectual disabilities and Down’s syndrome. JIntellect Disabil Res 39:532–537.

Vlachou M, Farrell P. 2000. Object mastery moti-vation in pre-school children with and with-out disabilities. Educ Psychol 20:167–176.

Wang PP, Bellugi U. 1994. Evidence from twogenetic syndromes for a dissociation betweenverbal and visuo-spatial short-term memory.J Clin Exp Neuropsychol 16:317–322.

Williams K, Wishart J. in press. Social cognitionin children with Down syndrome. MentRetard Dev Disabil Res Rev.

Wishart J. 1993. The development of learning dif-ficulties in children with Down syndrome. JIntellec Disabil Res 37:389–403.

Wishart JG. 1996. Avoidant learning styles andcognitive development in young children.In: Stratford B, Gunn P, editors. New ap-proaches to Down syndrome. London:Cassell. p 157–172.

Wisniewski K., Schmidt-Sidor B. 1989. Postnataldelay of myelin formation in brains fromDown syndrome infants and children. ClinNeuropathol 8:55–62.

Wisniewski K. 1990. Down syndrome childrenoften have brain with maturation delay, re-tardation of growth, and cortical dysgenesis.Am J Med Genet Suppl 7:274–281.

Wisniewski K, Kida E. 1994. Abnormal neuro-genesis and synaptogenesis in Down syn-drome. Dev Brain Dysfunct 7:289–301.

Wisniewski K, Laure-Kamionowska M, ConnellFF, et al. 1986. Neuronal density and synap-togenesis in the postnatal stage of brain mat-uration in Down syndrome. In: Epstein CJ,editor. The neurobiology of Down syn-drome. New York: Raven Press. p 29–44.

Wolpert G. 2001. What general educators have tosay about successfully including students withDown syndrome in their classes. J Res ChildEduc 16:28–38.

Yirmiya N, Solomonica-Levi D, Shulman C,et al. 1996. Theory of mind abilities in indi-vidual with autism, Down syndrome, andretardation of unknown etiology: The roleof age and intelligence. J Child Psychol Psy-chiatry 37:1003–1014.

Yoder P, Warren S. 2002. Effects of prelinguisticmilieu teaching and parent responsivity edu-cation on dyads involving children with in-tellectual disabilities. J Speech Lang HearRes 45:1158–1174.

Zelazo PD, Burack JA, Benedetto E, et al. 1996.Theory of mind and rule use in individualswith Down syndrome: a test of the unique-ness and specificity claims. J Child PsycholPsychiatry 37:479–484.

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 271

EDUCATION AND CHILDREN WITH DOWN

SYNDROME: NEUROSCIENCE, DEVELOPMENT,

AND INTERVENTION

Deborah J. Fidler1* and Lynn Nadel21Human Development and Family Studies, Colorado State University, Fort Collins, Colorado

2Department of Psychology, University of Arizona, Tucson, Arizona

Of the recent advances in education-related research in Downsyndrome, the characterization of the Down syndrome behavioral pheno-type has become a potentially critical tool for shaping education andintervention in this population. This article briefly reviews the literatureon brain–behavior connections in Down syndrome and identifies aspectsof the Down syndrome behavioral phenotype that are potentially relevantto educators. Potential challenges to etiologically informed educationalplanning are discussed. ' 2007 Wiley-Liss, Inc.MRDD Research Reviews 2007;13:262–271.

Key Words: Down syndrome; behavioral phenotypes; education; inter-

vention

As the most common chromosomal abnormality associ-ated with intellectual disability, Down syndrome hasbeen the focus of more behavioral and educational

research than most other genetic disorders. These efforts haveled to notable milestones in the advancement of education inDown syndrome, including cases in the literature of youngadults with Down syndrome attending university courses[Hamill, 2003; Casale-Giannola and Wilson Kamens, 2006]and achievements in the area of educating children and adoles-cents in inclusive environments [Buckley et al., 2006]. Thepast five years alone have brought innovations that includeteacher training interventions to shape attitudes toward inclu-sion in Down syndrome [Campbell et al., 2003], a refinedunderstanding of effective inclusive practice in Down syndrome[Wolpert, 2001], and new instructional approaches involvingcomputer technology [Lloyd et al., 2006; Ortega-Tudela andGomez-Ariza, 2006].

However, the innovation that has the potential to havethe greatest impact on educational practice in Down syndromeis the characterization of the Down syndrome ‘‘behavioralphenotype,’’ or the pattern of behavioral outcomes associatedwith this disorder throughout development. Research into thephenotypic outcomes associated with Down syndrome has ledto a better understanding of the learning profile associatedwith this disorder, and has offered new information regardingthe possible brain–behavior pathways leading to these out-comes. Over the past few decades, researchers have uncoveredcharacteristic patterns of functioning in the areas of cognition,

language development, social–emotional functioning, and per-sonality-motivation [see Dykens et al., 2000; Rondal andBuckley, 2001; Fidler, 2005]. Though many questions regard-ing development in this population remain unanswered,researchers have gained a clearer understanding of the devel-opmental trajectory associated with Down syndrome, and howthis chromosomal abnormality impacts development in adynamic and multisystemic way.

Amidst these advances in delineating the Down syn-drome behavioral phenotype, there remains a wide gapbetween these research findings and the development of inno-vative practice [Hodapp and Fidler, 1999; Fidler et al., 2007].While it has been argued that etiology-based informationcould be of importance for education in Down syndrome[Hodapp and Fidler, 1999; Freeman and Hodapp, 2000;Fidler, 2005], the use of these connections has not nearly metits full potential. It is true that relative to other disorders,greater research emphasis has been placed on identifying edu-cational strategies that might improve outcomes in Down syn-drome. Some of these recent approaches are informed by andspecifically target aspects of the Down syndrome behavioralphenotype [Laws et al., 1996; Kennedy and Flynn, 2003; Iarocciet al., 2006; van Bysterveldt et al., 2006], and others do not[Garcia and Conte, 2004; Park et al., 2005; Trent et al., 2005].Those existing syndrome-based recommendations may bepotentially quite useful, but lack empirical validation [Alton,1998; Fidler, 2005].

Though the gap between research findings and targetedpractice in Down syndrome remains wide, there is evidencethat educators and practitioners themselves recognize the im-portance of scientific progress in this area. Wolpert [2001]asked educators of children with Down syndrome in inclusivesettings to identify factors that might improve the outcomes inthe classroom. Among their answers was ‘‘. . . more informa-

*Correspondence to: Deborah J. Fidler, Human Development and Family Studies,Colorado State University, 502 W. Lake Street, Fort Collins, CO 80523-1570.E-mail: deborah.fidler@colostate.eduReceived 10 August 2007; Accepted 13 August 2007Published online in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/mrdd.20166

MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIESRESEARCH REVIEWS 13: 262 – 271 (2007)

' 2007Wiley -Liss, Inc.

tion on learning characteristics of chil-dren with Down syndrome’’ (p. 33).Similarly, parents seem to endorse etiol-ogy-specific modifications to theirchild’s educational planning [Fidler et al.,2003]. As part of an effort to addressthe gap between research and practicein this area, this article will first reviewthe current literature on the pathwayfrom brain to behavior in Down syn-drome. We then identify examples ofphenotypic outcomes that may haveeducational relevance, with particularattention to the development of cogni-tion and information processing inDown syndrome. Additional areas thatmay have educational relevance, includ-ing language development, social–emo-tional development, and personality-motivation, are briefly summarized. Weconclude with a brief discussion of thepotential challenges involved in shapingeducational instruction in etiologicallyrelevant ways. Although we start ourreview with discussion of neurobiology,we believe these kinds of data will bemost relevant to the development ofanimal models, which then can be usedto validate biological approaches to mit-igation of the syndrome’s impact oncognitive function. More important foreducational practice will be an under-standing of the phenotypic result ofthese neurobiological sequellae of tri-somy 21, and we will accordingly focuson such behavioral approaches below.

THE NEUROBIOLOGY OFDOWN SYNDROME

The brain of an individual withDown syndrome at or shortly beforebirth is in many respects indistinguish-able from the brain of a normal indivi-dual [Brooksbank et al., 1989; Wis-niewski and Schmidt-Sidor 1989; Fl�orezet al., 1990; Schmidt-Sidor et al., 1990;Bar-Peled et al., 1991; Pazos et al.,1994]. Normal values have been re-ported for brain and skull shape, brainweight, proportion of specific cerebrallobes, size of cerebellum and brainstem, and the emergence of most neu-rotransmitter systems. There is evidence,however, that some changes begin toemerge as early as 22 weeks gestationalage [e.g., Schmidt-Sidor et al., 1990;Golden and Hyman, 1994; Wisniewskiand Kida, 1994; Engidawork andLubec, 2003] and it is clear that by theage of 6 months a number of importantdifferences are already obvious. Some ofthese differences are expressed in termsof the proportion of individuals withDown syndrome who show abnormal

values, rather than in terms of a uni-form abnormality in all instances. Thisis important as it highlights the variabil-ity in this population sharing the geno-typic feature of trisomy 21.

While there is a postnatal delay inmyelination [Wisniewski, 1990], it isworth noting that in all cases myelina-tion is within normal range at birth,while in 75% of the cases it is withinnormal range throughout early develop-ment. Neuropathological differencesafter 3–5 months of age include a short-ening of the fronto-occipital length ofthe brain that appears to result from areduction in growth of the frontal lobes,a narrowing of the superior temporalgyrus (observed in about 35% of cases),a diminished size of the brain stem andcerebellum (observed in most cases),and a 20–50% reduction in the numberof cortical granular neurons [see Cromeet al., 1966; Benda, 1971; Blackwoodand Corsellis, 1976]. In sum, the overallpicture at birth or shortly thereaftershows that individuals with Down syn-drome tend to fall towards the bottomof the normal range (or outside it) onmost measures.

Investigations of neural function,as opposed to structure, in early infancysuggest some abnormalities: there is evi-dence of either delayed or aberrant au-ditory system development [Jiang et al.,1990], which might contribute to thewidespread hearing disorders observedin Down syndrome. Obviously, such adisorder, if organic, could be related tosubsequent difficulties seen in the learn-ing of language. Hill Karrer et al. [1998]have reported delayed development ofcerebral inhibition using visual event-related potentials (ERP) in a visual rec-ognition memory paradigm.

There have been few studies ofbrain function in adolescents and youngadults with Down syndrome and theexisting data are somewhat equivocal.Pinter et al. [2001] used high-resolutionMRI methods to analyze brain structurein 16 children (mean age 11.3 years)with Down syndrome. After correctingfor overall brain volume, hippocampal,but not amygdala, volume reductionswere seen in this group. This resultconfirms some earlier work using lowerresolution MRI methods [Jerniganet al., 1993]. Kates et al. [2002] lookedat a group of 12 children with Downsyndrome (all males, mean age 5.94years) and compared them with childrenwith fragile-X, developmental languagedelay, or typical development. The chil-dren with Down syndrome had smallerbrain volumes than any of the others,

with previously unreported reductionsin parietal cortex as well as the oft-reported reductions in the temporallobe. Pinter et al. [2001], on the otherhand, note the relative preservation ofparietal cortex.

Overall, study of neuropathologypoints to problems in certain regions ofthe cortex, including most prominentlythe temporal lobe and the hippocampalformation [Wisniewski et al., 1986], theprefrontal cortex, and the cerebellum.This conclusion meshes well with whathas been learned from the developmentof mouse models of the syndrome [e.g.,Kleschevnikov et al., 2004]. Thesemodels have generally demonstratedselective impairments in the anatomy,physiology, pharmacology, and behavior,which are associated with the hippo-campal formation, the prefrontal cortex,and the cerebellum.

These neurobiological findings inDown syndrome can inform educationin several ways. First, as only modestabnormalities are detectable at birth, therole of development is critical in build-ing into the pronounced profile ofstrengths and weaknesses observed inthis population. This offers educators animportant window of opportunity forintervention in the first few years of life,a point that will be revisited later. Sec-ond, the atypical development of spe-cific brain structures suggests that someareas of functioning may become moreimpaired than others throughout devel-opment. In addition, early in develop-ment, many of the differences observedare expressed in terms of the proportionof individuals with Down syndromewho show abnormal values, rather thanin terms of a uniform abnormality in allinstances. This is in line with the proba-bilistic approach to understanding out-comes in individuals with genetic dis-orders [Dykens, 1995], where childrenmay be at risk for certain neurobiologi-cal features, but not all children willshow those specific abnormalities. Thesethree themes are educationally relevantand will be explored in greater depth inthe following sections. Additional linksbetween brain and behavior will benoted throughout the next section whererelevant findings have been reported.

REVIEW OF EDUCATIONALLY-RELEVANT LITERATURE ONDEVELOPMENT IN DOWNSYNDROME

In this section, we explore thecurrent literature on development inDown syndrome and highlight some of

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 263

the main findings that are potentiallyrelevant to educators and intervention-ists. Before we examine the phenotypicprofile associated with Down syndrome,there are two important issues thatshould be noted. First, these findingsare part of a probabilistic approach tounderstanding the link between syn-drome and outcome in behavioral phe-notype research [Dykens, 1995]. Withinthis approach, it is understood that thereis a higher probability that childrenwith Down syndrome will show any ofthese specific outcomes relative to otherchildren who do not have Down syn-drome; however, not every child withDown syndrome will exhibit each ofthese phenotypic features [Dykens,1995]. Thus, educators might use thisinformation to adopt a ‘‘ready stance’’regarding areas of potential strength andweakness in children with Down syn-drome, but might expect that childrenwill likely vary in the degree to whichthey express any one of these predis-posed outcomes.

Second, researchers are currentlyexploring the question of whether thereis a subgroup of individuals with Downsyndrome who meet criteria for autismspectrum disorders. Some studies sug-gest that a small percentage of individu-als with Down syndrome have behav-ioral profiles consistent with autism[Collacott et al., 1992; Ghaziuddinet al., 1992; Kent et al., 1999; Paly andHurley, 2002; Hepburn et al., in press].These findings are not based on studiesthat have employed epidemiologicalmethods for estimating the prevalenceof comoribidity in this population, sothey should be taken with a note ofcaution. However, it is possible thatDown syndrome with children whohave comorbid autism may show a dif-ferent behavioral profile than theircounterparts with Down syndrome whodo not have autism, and may require adifferent educational approach [Hepburnet al., in press].

COGNITION IN DOWNSYNDROME

Early CognitionIn line with the neurobiological

findings described earlier, infants withDown syndrome show relatively normalabilities in learning and memory. Thisdoes not mean that either they, orindeed normally developing infants,have the full adult range of learning andmemory abilities at birth. In fact this isnot the case because some parts of the

brain mature postnatally and the formsof learning and memory dependent onthem are not available until some timeafter birth. The medial temporal lobe,and particularly, the hippocampus, aswell as parts of the prefrontal cortexand cerebellum, are included in this cat-egory. The fact that these late-develop-ing structures are apparently particularlyat risk in Down syndrome is probablyof considerable importance [see Nadel,1986].

In an early series of studies, Ohrand Fagen [1991, 1993] reported that3-month-old infants with Down syn-drome were entirely normal in learningabout the contingencies between theirown movements (leg kicking) and rein-forcement, including initial learning, ac-quisition speed, and retention. In a laterreport, Ohr and Fagen [1994] showedthat 9-month-old infants with Downsyndrome were impaired, as a group, inlearning about the contingency betweenarm movements and reinforcement.However, they noted that some infantswith Down syndrome were able tolearn. They concluded that there is arelative decline in conditionability ininfants with Down syndrome comparedto normally developing infants after 6months. Mangan [1992] tested controlinfants and infants with Down syn-drome on a variety of spatial tasks, oneof which, a place-learning task, wasdesigned especially to assess the state offunction of the hippocampal system.The pattern of results was consistentwith diffuse, but mild, neuropathologycombined with much more extensivepathology localized to the hippocampus.

Children with Down syndromehave typically been shown to acquirebasic object concept more slowly thannormal [see, e.g., Rast and Meltzoff,1995] but with extensive training theycan acquire it at more-or-less the sametime as normally developing infants[Wishart, 1993]. However, a differentkind of problem emerges in this tasksituation: instability of acquisition. Al-though the typical subject with Downsyndrome solved various levels of thetasks used to assess the object concept atages not very far from the norm, per-formance after acquisition could behighly variable and apparently beset bymotivational difficulties. These prob-lems, if representative of the learningstyle of children with Down syndrome,are extremely important in thinkingabout effective intervention. The resultsof Wishart’s studies using standard intel-ligence test batteries suggest that theyare indeed representative. Test–retest

reliability was very low because suc-cesses gained in one test might notappear upon retest, as soon as 2 weekslater. New skills show up, only to dis-appear shortly thereafter. One could spec-ulate that evidence of such ‘‘rapid for-getting’’ is consistent with damage inthe hippocampal formation but consid-erably more data are required beforethis conclusion can be accepted. Thesedevelopmental differences may have im-portant implications for educators. Forexample, if the learning process involvesmore observable regressions for childrenwith Down syndrome than for otherchildren, it might be important forteachers and interventionists to accountfor these differences with more frequentreviews of materials, and it may be im-portant to monitor the stability withwhich a child with Down syndrome hasacquired a skill.

Information ProcessingThe learning and memory prob-

lems that begin to emerge in lateinfancy become considerably more no-ticeable as the infant grows to child-hood and adolescence. These effectsvary as a function of the type of mem-ory being assessed. Explicit memoryinvolves things like facts and events thatsubjects consciously recollect, whereasimplicit memory can be demonstratedindirectly, without conscious recollec-tion. This distinction has been shownto be important in understanding or-ganic amnesia, since most amnesics areprofoundly impaired on explicit mem-ory tasks but can be relatively normalon implicit tasks. One common kind ofimplicit memory test looks at skills orprocedures, such as mirror-tracing;another common implicit memory testinvolves ‘‘priming,’’ where prior expo-sure to a word or picture can influencesubsequent performance on word-stemor partial-picture completion tasks eventhough the subjects might not recallhaving seen the relevant items before.

Carlesimo et al. [1997] examinedthe performance of children with Downsyndrome in the area of ‘‘implicit’’ (pro-cedural) and ‘‘explicit’’ (episodic) mem-ory paradigms, including word-stemcompletion, list learning, and proserecall. Robust priming effects were seenin the Down syndrome group, compa-rable to those observed in controls,indicating that implicit memory wasintact. However, deficits were observedin both explicit memory tasks. Perform-ance on these kinds of explicit memoryparadigms has been linked to functions

264 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

of the hippocampal system; hence, thedefects suggest differential impairmentin hippocampal function and therebyconverge with the data from study ofspatial cognition. This selective impair-ment of explicit, but not implicit,memory was also reported in Vicari[2001].

A great deal of research attentionhas been focused on deficits in theprocessing of verbal information inDown syndrome, and how these deficitscontribute to poor language and learn-ing outcomes [Byrne et al., 1995;Hesketh and Chapman, 1998; Laws,1998]. Most commonly, the verbalworking memory deficit in Down syn-drome is measured by performance onan auditory digit span task. Poor audi-tory digit span performance findings inDown syndrome were first reportedseveral decades ago [Bilvosky and Share,1965; Rempel, 1974].

More recently, a series of studiesby Hulme and Mackenzie [1992] de-scribed the development of poor verbalworking memory in children withDown syndrome in great detail. Theyfound that children in the Down syn-drome group scored lower than thetypical MA-matched controls on audi-tory digit span, and mental age andverbal working memory tasks weremore correlated in the typically devel-oping children than in the Down syn-drome group. Similar deficits in verbalworking memory are also observablewhen letters are substituted for num-bers in the paradigm [Varnhagen et al.,1987]. This deficit is also not subject tochanges when experimental design ismanipulated in order to reduce unre-lated demands on individuals with Downsyndrome [Marcell and Weeks, 1988;Marcell et al., 1988; Laws, 1998].

Deficits in verbal working mem-ory may relate to neuroanatomical char-acteristics associated with Down syn-drome, including a proportionatelysmaller planum temporale, which isreferred to as the auditory associationcortex [Frangou et al., 1997]. It is im-portant to note, however, that the rela-tionship between the planum temporalevolume deficit and cognitive-linguisticfunctioning in Down syndrome remainsunclear [Frangou et al., 1997].

Though there is a great deal ofevidence of verbal working memorydeficits in Down syndrome, it is impor-tant to note that the main evidence forthis dysfunction comes from studies thatmeasure the processing of acoustically-presented information only. Auditorydigit span, auditory sentence recall, and

other commonly used techniques audi-tory nonword repetition all involve theprocessing of acoustic, rather thanspeech-based visually presented infor-mation. Thus, one might argue that theverbal working memory difficulty inDown syndrome has only been demon-strated in the maintenance of auditorallypresented information, rather than theprocessing of all types of verbal infor-mation. The processing of speech-basedvisual information in Down syndromehas not yet been shown to be equallyimpaired, and importantly, may not be.

In contrast with their perform-ance on verbal processing tasks, individ-uals with Down syndrome show relativestrengths in visuospatial memory, andshow a profile of higher visuospatialprocessing than verbal processing [Sil-verstein et al., 1982; Thase et al., 1984;Pueschel et al., 1987; Wang and Bel-lugi, 1994; Jarrold et al., 1999; Kleinand Mervis, 1999]. The Corsi span, ablock tapping memory task, is most of-ten used as the test of visuospatial proc-essing. In terms of raw scores, mostindividuals without Down syndrome(both with developmental delay andtypically developing) tend to showhigher raw auditory digit span recallscores than Corsi span recall scores.However, individuals with Down syn-drome seem to perform equally well onthese tests, or even show an advantagefor Corsi span recall [Haxby, 1989;Azari et al., 1994; Wang and Bellugi,1994; Vicari et al., 1995; Jarrold andBaddeley, 1997], especially when partic-ipants were not required to includeorder in their responses. These resultshave also been demonstrated with tasksother than the Corsi and auditory digitspan [Pueschel et al., 1987; Hodappet al., 1992; Bower and Hayes, 1994;Klein and Mervis, 1999]. The visuospa-tial working memory advantage is alsodemonstrated when identical stimulusinformation is simply presented eithervisually versus auditorally [Varnhagenet al., 1987].

Neuroanatomical correlates forthis relative strength in Down syndromevisuospatial processing have been pos-ited. Pinter et al. [2001] reported whatthey call ‘‘striking preservation’’ of pari-etal and occipital cortical gray matter inan MRI study of 5–23-year olds withDown syndrome. Studies have shownthe importance of both parietal andoccipital lobe functioning for someaspects of visuospatial processing [Blackand Bernard, 1984; Jonides et al., 1993].

An information processing profilethat includes strengths in visual process-

ing and implicit memory, and deficits inverbal processing and explicit memorycould potentially inform educationalapproaches to working with childrenwith Down syndrome. Instruction thatis verbally based��and especially audito-rally mediated��might pose a greaterchallenge to children with Down syn-drome than instruction that is presentedwith visual supports. Minor and subtleteaching modifications can be made toaddress this issue, without any noticea-ble disruption to a larger classroomenvironment. In addition, instructionthat involves explicit memory, such aslogic problems, may be presented withgreater sensitivity given that this is anarea of challenge as well. Awareness ofthis area of deficit might allow an edu-cator to make informed decisionsregarding prompts, wait time, and sup-ports for a child who may have particu-lar difficulty with this type of informa-tion processing.

ReadingThe cognitive underpinnings of

reading are of particular relevance foreducation in Down syndrome. Despitethe impairments observed in other areasof cognition, many individuals withDown syndrome are able to show com-petence in some aspects of reading de-velopment. In particular, word identifi-cation appears to be an area of relativestrength within reading skills in thispopulation, while word attack and read-ing comprehension appear to be moreimpaired [Byrne et al., 1995; Cupplesand Iacono, 2000; Kay-Raining Birdet al., 2000]. Strengths in word identifi-cation have been linked to relativestrengths in visual processing [Fidleret al., 2005b], while deficits in wordattack skills have been linked to difficul-ties with verbal processing skills inDown syndrome [Boudreau, 2002].

There is some controversy regard-ing the implications of this profile forreading instruction in Down syndrome[see Hodapp and Freeman, 2003 for asummary], with some arguing that visu-ally-based approaches might be war-ranted [Buckley and Bird, 2002], andothers emphasizing the role of phono-logical processing in reading in thispopulation, despite impairments [Cup-ples and Iaocono, 2002]. Regardless ofthis debate, there is evidence that chil-dren with Down syndrome recruit boththeir visual and verbal processing skillswhen reading, particularly when identi-fying words [Cupples and Iacono, 2000;Fidler et al., 2005b].

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 265

BRAIN–BEHAVIORCONNECTIONS

In a series of recent studies, Pen-nington, Nadel and colleagues havetested several different groups of indi-viduals with Down syndrome on arange of tasks designed to directly assessthe function of specific brain systems.This ‘‘cognitive neuropsychological’’approach often uses tasks first developedin animal models, where the criticalunderlying brain circuits can be identi-fied and carefully studied in invasiveexperiments. The team started with afocus on three brain systems identifiedby the neuropathological data, much ofwhich was discussed above: the hippo-campal system, the prefrontal cortex,and the cerebellum. They developed aset of tasks that could, collectively, tellus something about how these brainsystems are faring. In the first set ofstudies, Pennington et al. [2003] foundevidence of specific hippocampal dys-function in our sample of 28 adolescents,using mental age matched controls.

Little evidence of prefrontal dys-function was observed in a battery ofnonverbal tasks. Subsequent pilot work,however, suggested that verbal tasksmight yield a different result, andindeed that is what is being observed(Moon et al., unpublished data). Usingverbal tasks to explore the prefrontalcortex, Moon et al. found in the young(aged 5–11) and old (aged 30–41)groups strong signs of dysfunction inboth the hippocampal and prefrontalsystems. Deficits were observed in arange of tasks although verbal mediationwas necessary to bring out the prefron-tal effect. Taken as a whole, these stud-ies show that particular problemsemerge in the memory domains servedby the hippocampal system and the pre-frontal system. The latter impairmentappears to be linked to the use of verbaltest materials. The impairment in hip-pocampal function could in principlereflect problems in any of the structuresof the hippocampal region; a recentstudy of two neuropsychological para-digms dependent on parahippocampaland perirhinal regions (delayed non-matching to sample and visual pairedcomparison), however, suggests thatthese areas are functioning appropriately,and that the impairment is more likelyto reflect improper development of thehippocampus itself [Dawson et al.,2001].

The prefrontal cortex, as notedalready, plays an important role in awide range of functions, including epi-sodic/explicit memory and working

memory. As noted earlier, explicitmemory is impaired in individuals withDown syndrome. There has beenextensive research on working memoryin this population, and clear deficitshave been observed in a number ofstudies [Varnhagen et al., 1987; Marcelland Weeks, 1988; Laws, 1998; Jarroldet al., 1999]. However, this impairmentseems to be limited to verbal informa-tion, as impairments are minimal invisuospatial domains. The deficitappears to be neither a motor nor artic-ulatory problem [Kanno and Ikeda,2002] and may relate to the so-calledphonological loop [Jarrold and Badde-ley, 2001; Laws, 2002].

OTHER AREAS OFEDUCATIONAL RELEVANCE

In addition to cognitive develop-ment and information processing, thereare other areas of development inDown syndrome that may be relevantto decisions made by educators. Briefdescriptions of findings in the area oflanguage development, social–emotionalfunctioning, and personality-motivationare presented in this section.

Language DevelopmentIndividuals with Down syndrome

generally show a profile of strongerreceptive language skills and weaker ex-pressive language skills. This profileseems to emerge in early childhood andbecome more pronounced as childrenprogress into middle childhood andbeyond [Miller, 1999]. Expressive lan-guage deficits are often manifested interms of morphosyntactic delays[Chapman et al., 1998; Eadie et al.,2002] and difficulties with speech intel-ligibility [Miller et al., 1999; Stoel-Gammon, 2003; Kumin, 2006]. Andwhile receptive language is strongerthan expressive language in most indi-viduals with Down syndrome, someareas such as receptive syntax, are morecompromised than others [Abbedutoet al., 2003]. In contrast, some areas ofpragmatic functioning seem to be anarea of relative strength in individualswith Down syndrome [Johnston andStansfield, 1997; Laws and Bishop,2004].

This language profile is potentiallyrelevant for educational planning in sev-eral ways. First, because many childrenwith Down syndrome have strongerreceptive than expressive language skills,they often understand much more lan-guage than they can produce. To anuninformed educator who is not aware

of the discrepancy between expressiveand receptive language, it might be nat-ural to address the child with input lan-guage and instruction at the level oftheir expressive language. But such anapproach would likely involve a markedunderestimation of the child’s academicand receptive abilities. When consider-ing how best to present class material, itmay be important for educators to iden-tify and target the receptive languagelevel of a child with Down syndromeso as to appropriately challenge themand engage them at their true level ofunderstanding [Roberts et al., 2007].

In addition, it may be useful foreducators to be aware that difficultieswith expressive language��and speechintelligibility in particular��may befrustrating for children with Down syn-drome in classroom contexts. Morpho-syntactic difficulties demand extra moti-vation from children with Down syn-drome to produce lengthier utterances[Miller and Leddy, 1999], and intelligi-bility problems may lead to situationswhere a child must repeat herself inorder to be understood. Thus, educa-tors may want to consider both thesocial and motivational consequences ofexpressive language difficulties. It maybe beneficial to identify ways to mini-mize the potential for negative experi-ences, while allowing the child withDown syndrome to benefit from theopportunity to build their speech, lan-guage, and communication skills.

Social–Emotional FunctioningThe majority of individuals with

Down syndrome show strengths in vari-ous aspects of social–emotional func-tioning, exhibiting behaviors that sug-gest evidence of intact social relatednessand some measure of social competencein early childhood [Fidler et al., 2006].In the first few years of life, markers ofprimary intersubjectivity, the earliestforms of dyadic social relating based onemotional displays and reciprocal signal-ing, are identified as emerging in adelayed, but competent manner in thispopulation [see Fidler, 2006 for areview]. In particular, young childrenwith Down syndrome show evidence ofprimary intersubjective development inthe form of increased eye gaze andvocalizations directed to other people[Gunn et al., 1982; Crown et al., 1992;Legerstee et al., 1992; Kasari et al.,1995], increased direction of positivefacial displays in the form of smiles[Kasari and Freeman, 1990; Fidler et al.,2005a].

266 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

In addition, there is also supportfor the notion that many aspects of sec-ondary intersubjectivity, the ability toengage with a social partner in a triadicfashion, also emerge with competence,albeit in a delayed fashion. Several stud-ies suggest that children with Downsyndrome show either MA-appropriateor even increased levels of joint atten-tion relative to other children withoutDown syndrome [Mundy et al., 1988;Kasari et al., 1995; Fidler et al., 2005c].Toddlers and preschoolers with Downsyndrome also show competent triadicrelating in the forms of play acts, turntaking, invitations, and object shows[Mundy et al., 1988; Sigman and Rus-kin, 1999].

In the context of these strengthsin the development of social relatednessand other social–emotional skills, thereis evidence that children with Downsyndrome may show difficulties whencognitive demands in social decisionmaking increase. While socializationskills remain a relative strength in mid-dle childhood [Dykens et al., 1994],and many children with Down syn-drome appear to be able to form recip-rocal friendships with peers [Freemanand Kasari, 2002], there is mixed evi-dence regarding whether individualswith Down syndrome show impair-ments in the ability perform more com-plex social cognition tasks [Baron-Cohen et al., 1985; Baron-Cohen, 1989;Yirmiya et al., 1996; Zelazo et al., 1996;Abbeduto et al., 2001]. Williams andWishart [in press] identify other factorsthat may contribute to the difficulty thatindividuals with Down syndrome haveon social–cognitive tasks, including exec-utive function or memory difficulties.Nevertheless, it may be that despitecompetence in the area of social related-ness, as the demands and complexitiesof social situations increase in middlechildhood and beyond, individuals withDown syndrome may show difficultieswith social adaptation and selectingappropriate social strategies, especially asthey enter adolescence and face changesin emotional functioning and mood[Dykens et al., 2002; Fidler et al., 2005a].

This social profile is relevant foreducational planning in several ways.First, these strengths may make it possi-ble for children with Down syndrometo learn through social techniques suchas modeling, peer collaboration, socialgroups [Lloveras and Fornells, 1998;Rogoff, 2003]. Though there is a sur-prising lack of empirical exploration ofthe efficacy of these techniques in chil-dren with Down syndrome, it could be

hypothesized that such techniqueswould capitalize on social motivationmay be a useful reinforcer for childrenwith Down syndrome. However, it isimportant to consider that the strongsocial motivation observed in this popu-lation may serve as a distractor as well[Wishart, 1996; Fidler, 2006; see discus-sion below]. There is also some evi-dence that, because of strengths in earlysocial relatedness in this population,affective cues put forth by a teacher orinterventionist can impact learning andmotivation in a particularly pronouncedway [Park et al., 2005]. It may also beimportant for educators to be mindfulof potential changes in mood and socialengagement as children with Downsyndrome transition into adolescence,and perhaps adopt modified strategies asthese behavioral changes become evi-dent [Dykens et al., 2002].

Personality-MotivationAnother aspect of the behavioral

profile in Down syndrome that may beeducationally relevant relates to motiva-tional orientation and task persistence[Gunn and Cuskelly, 1991]. In labora-tory settings, when presented with taskssuch as puzzles and other nonsocial/nonverbal tasks, children with Downsyndrome have been shown to abandonthe task sooner than other children atsimilar developmental levels, and toadopt strategies that divert attentionaway from the task [Landry and Cha-pieski, 1990; Pitcairn and Wishart,1994; Ruskin et al., 1994; Vlachou andFarrell, 2000; Kasari and Freeman,2001]. This, coupled with the strengthsin aspects of social initiation describedin the previous section, can lead to astyle that involves an over-reliance onsocial strategies, especially in contextsthat require instrumental thinking[Kasari and Freeman, 2001]. It may bethat the interaction between emergingdifficulties with instrumental thinkingand strengths in social relatedness leadto a personality-motivation orientationthat ultimately impacts the ability of achild with Down syndrome to learn ef-fectively [Wishart, 1996; Fidler, 2006].

A style that leads children toremove themselves from challenging sit-uations in favor of social interactionmay deprive children with Down syn-drome of important opportunities tochallenge themselves and gain new skillsthrough active engagement with theenvironment that surrounds them.Awareness of this profile may be impor-tant for educators when selecting tech-

niques for involving individuals withDown syndrome in classroom settings.It may be important for educators toidentify situations when the child withDown syndrome may be recruitingsocial strategies when engaging with thetask at hand is more appropriate [Fidler,2006]. Educators might also managebehavior using social consequences asreinforcement, not as a distracter duringtasks that might pose a cognitive chal-lenge [Fidler, 2006].

CHALLENGES IN LINKINGRESEARCH TO PRACTICE

Are Behavioral PhenotypesModifiable?

Though many aspects of the be-havioral phenotype in Down syndromeare potentially relevant for educators,there are several challenges that must beaddressed as researchers aim to translateresearch findings into educational prac-tice. The first challenge relates to thenotion of genes and modifiability. Thereis a danger in discussing the notion ofbehavioral outcomes associated withgenetic disorders in that genetic effectscan connote fixed, nonmodifiable path-ways. What we now know about themechanisms by which genes give rise tophenotypes, particularly behavioral phe-notypes, indicates that we need notworry about this danger. First, froma neurodevelopmental perspective, atevery step of the way, opportunities existto modulate the translational process. Inaddition, all learning and education isrooted in the notion that neurophysio-logical changes can be observed inresponse to environmental input, lead-ing the brain to undergo various typesof reorganization [Nelson, 2000]. Nel-son notes that it is commonly under-stood that, ‘‘. . . the success of earlychildhood intervention strategies rests toa great degree on the relative plasticityof the human brain (p. 222),’’ and thisapplies to children with and withoutgenetic disorders alike.

In addition, potential evidence ofthe modifiability of the Down syn-drome phenotypic profile has beenreported in a long-term study of Britishinclusion in this population. Buckleyet al. [2006] report that the practice ofincluding children with Down syn-drome in mainstream classrooms inEngland has had an impact on the phe-notypic profile in older children andadolescents with Down syndrome. Theynote that previous studies showed thatchildren with Down syndrome who

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 267

attended school in special education set-tings demonstrated a profile of strengthsin socialization and daily living skills,but deficits in adaptive communicationabilities [Dykens et al., 1994; Fidleret al., 2006]. However, in their sampleof children with Down syndrome whounderwent schooling in inclusive set-tings, they found that the marked defi-cits in adaptive communication werenot observable [Buckley et al., 2006].They argue that the social challengesassociated with being educated in aninclusive setting modified the phenotypicprofile and narrowed the gap betweenareas of strength and challenge. Theauthors of this study note that this war-rants replication, but if supported, therethis would be a critical means of modi-fying the profile of strengths and weak-nesses associated with Down syndrome.

When considering whether edu-cational approaches can modify pheno-typic profiles, it is also important tonote that the pattern of strengths andweaknesses associated with genetic dis-orders does not simply appear in a pro-nounced fashion in middle childhood.There is a developmental process thatleads to the more pronounced end statesof relative strengths and weaknesses inany genetic disorder. This is importantto recognize because there may beopportunities to target early emergingphenotypic characteristics in very youngchildren, before dissociations in profilebecome pronounced [see Karmiloff-Smith, 1997, 1998; Fidler, 2005, 2006].Fidler [2005] argued that for someaspects of the Down syndrome behav-ioral phenotype, it may be possible toidentify early developmental precursorsto later more pronounced outcomes. Ifthese early developmental precursorscan be identified and targeted withempirically validated intervention tech-niques, this too may be another meansfor altering the developmental pathwayand the phenotypic profile associatedwith Down syndrome.

What Constitutes EmpiricalSupport for Etiology-SpecificEducation/Intervention?

Another challenge in the attemptto bridge research and practice in thisarea relates to the empirical validationof techniques aimed at addressing phe-notype-specific dimensions. There is asmall, but growing literature thatdescribes the efficacy of educationaltechniques such as computer-basedlearning (Lloyd et al., 2006; Ortega-Tudela and Gomez-Ariza, 2006],instructional approaches to reading and

its component skills [Laws et al., 1996;Moni and Jobling, 2001; Cupples andIacono, 2002; Kennedy and Flynn,2003; van Bysterveldt et al., 2006],and math skills [Irwin, 1991; Nye andBuckley, 2006; Ortega-Tudela andGomez-Ariza, 2006] for children withDown syndrome. While more studies ofthis kind are warranted, only a few ofthese studies show educational benefitswhen using one specific technique overanother Cupples and Iacono, 2002].

The question remains whetherthis type of empirical validation is suffi-cient to warrant a syndrome-specificapproach to educational planning. Somemight argue that the efficacy of thesetechniques have little to do with thephenotypic profile associated withDown syndrome��rather they simplyshow that one technique is superior toanother regardless of the population towhich it is applied. By extension, itcould be argued that in order to justifya syndrome-specific set of recommenda-tions for educational practice, theremust be a set of techniques that workdifferentially across populations. That is,there must be techniques identified thatare effective for children with Downsyndrome, but not effective for childrenwho do not have Down syndrome orthe developmental profile associatedwith Down syndrome [see Fidler et al.,2007 for a discussion]. At present, thereare relatively few examples in the litera-ture that demonstrate such differentialeffects [Fey et al., 2006; Yoder andWarren, 2002]. Those that do exist sug-gest that the personality-motivationalorientation associated with Down syn-drome may be particularly important toconsider when selecting educational andintervention techniques [Yoder andWarren, 2002; see Fidler et al., 2007 fora discussion of this issue].

Is Syndrome-Specific EducationFeasible?

A third challenge to the idea oflinking phenotype research into practicerelates to issues of classroom manage-ment and the training of teachers.While future research may show thebenefits of etiology-specific instructionalapproaches, it could be argued that spe-cific techniques for different children inthe classroom would be too unwieldyand would require too great of a per-sonnel demand. It could also be arguedthat the training of teachers in etiology-specific instructional approaches wouldmake teacher education programs toolengthy of a process, requiring a masteryof approaches that target any number of

the many syndromes and behavioral dis-orders present in the student popula-tion.

While adopting an etiology-spe-cific approach in the classroom willundoubtedly place additional demandson educators, some potential approachesinvolve less of a diversion of resourcesfrom the larger classroom culture thanothers. Techniques to be developed inthe future can be imbedded in naturalis-tic ways, and might only involve subtleadjustments in teacher decision makingand presentation of material. For exam-ple, supplementing instruction withsupports that rely on a favored informa-tion processing modality might not bedetectable to the larger classroom, andin some instances could potentiallyenhance instruction for children in theclassroom without disabilities. In addi-tion, while some additional trainingmight be involved for teachers, thesecan come in the form of continuingeducational trainings, or useful informa-tional materials (websites, booklets) thatneed not burden a teacher in training.While the details of implementationwould need to be addressed in a real-world process, it is likely that theimplementation of some syndrome-spe-cific instructional approaches, if theyreceive empirical validation, might notnecessarily pose a prohibitively largechallenge to educators.

Future DirectionsDespite the many advances that

have been made in the study of brainand behavioral development in Downsyndrome, there is still a great deal ofprogress to be made both in the basicstudy of development in Down syn-drome and in the application of thesefindings to practice. In terms of thepotential contributions of the neurobio-logical approach, future work uncover-ing the neurobiological causes of thecognitive, language, and behavioralimpairments associated with Down syn-drome will ultimately lead to creationof an ever-more precise animal modelof Down syndrome. A more preciseanimal model of Down syndrome couldmake it possible to develop biologicalinterventions that might ultimatelyimpact development in this population.Thus, advances in this area will rely onthe close collaboration of behavioral sci-entists who are carefully delineating thenature of the Down syndrome behav-ioral phenotype, neurobiologists whoare able to map these phenotypic out-comes onto brain anatomy and brain

268 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

physiology in this population, and ani-mal model researchers who can use thisinformation to develop an ever moreprecise model of the disorder. This pro-cess will be aided by the neuropsycho-logical approach, which offers thepromise of identifying exactly thoseareas of disproportionate cognitive im-pairment that might guide the mappingfrom behavior to brain functioning.

Given the probabilistic nature ofphenotypic outcomes in genetic disor-ders, these approaches may also make itpossible to more deeply understand thenature of within syndrome variability inthe population of individuals withDown syndrome. As researchers collab-orate to uncover the pathway from geneto brain to behavior, it may be possibleto identify with greater precision thesources of within-syndrome variabilityin outcomes of interest, and it may bepossible to address the needs of childrenwho show variations around the pheno-typic profile that is associated with thelarger group of individuals with Downsyndrome. These advances offer thehope of even more targeted educationalplanning and the possibility of addressingthe variability in outcomes that is classi-cally associated with this population.

Another important future direc-tion for research in Down syndromerelates to the importance of detectingemerging phenotypes in early childhood[Fidler, 2005]. This type of researchtransforms the view of outcomes inDown syndrome from a static, cross-sectional approach into a dynamic, lon-gitudinal approach to studying this pop-ulation. In taking this more dynamicview, it may be possible to identify themore subtle developmental precursorsto more pronounced outcomes in laterchildhood and adolescence. These earlyprecursors may serve as potentially use-ful targets for early intervention in thispopulation. Rather than waiting tointervene once a split profile ofstrengths and weaknesses has becomepronounced, educators and interven-tionist may be able to target these earlyprecursors in their more subtle forms,which may set development on a moreoptimal pathway.

Finally, as the research communitysorts through the various controversiesrelated to syndrome-specific educationalapproaches, continued advances havestill been made in the education ofindividuals with Down syndrome.These efforts continue to challenge theoutmoded notions that children withgenetic disorders such as Down syn-drome have limited educational poten-

tial. Though the goal of bridgingresearch and practice in the study ofdevelopment in Down syndrome faceschallenges-especially the difficulty ofcollaboration among scientists acrossneighboring fields-it is a goal thatpromises greater returns than simplyeducating children with Down syn-drome according to their severity ofimpairment (mild, moderate, severe,profound intellectual disability) andignoring the complex profile associatedwith the disorder. It is likely that ourbest hope for improving outcomes ingenetic disorders such as Down syn-drome lies in our ability to use all ofthe scientific information that is avail-able, with developmentalists, educationscientists, and brain experts collaborat-ing to generate the most effective andinnovative practice approaches possible.n

REFERENCESAbbeduto L, Murphy MM, Cawthon SW, et al.

2003. Receptive language skills of adoles-cents and young adults with Down or fragileX syndrome. Am J Ment Retard 108:149–160.

Abbeduto L, Pavetto M, Kesin E, et al. 2001. Thelinguistic and cognitive profile of Down syn-drome: evidence from a comparison withfragile X syndrome. Down Syndrome: ResPract 7:9–15.

Alton S. 1998. Differentiation not discrimination:delivering the curriculum for children withDown’s syndrome in mainstream schools.Support Learn 13:167–173.

Azari NP, Horwitz B, Pettigrew KD, et al. 1994.Abnormal pattern of cerebral glucose meta-bolic rates involving language areas in youngadults with Down syndrome. Brain Lang46:1–20.

Baron-Cohen S. 1989. The autistic child’s theoryof mind: a case of specific developmentaldelay. J Child Psychol Psychiatry 30:285–297.

Baron-Cohen S, Leslie A, Frith U. 1985. Doesthe autistic child have a ‘‘theory of mind’’?Cognition 21:37–46.

Bar-Peled O, Israeli M, Ben-hur H, et al. 1991.Developmental patterns of muscarinic recep-tors in normal and Down’s syndrome fetalbrain��an autoradiographic study. NeurosciLett 133:154–158.

Benda CE. 1971.Mongolism. In: Minckler J, edi-tor. Pathology of the nervous system, Vol.2. New York: McGraw-Hill. p. 1867.

Bilvosky D, Share J. 1965. The Illinois Test ofPsycholinguistic ability and Down’s Syn-drome: an exploratory study. Am J MentDefic 70:78–82.

Black FW, Bernard BA. 1984. Constructionalapraxia as a function of lesion locus and sizein patients with focal brain damage. Cortex20:111–120.

Blackwood W, Corsellis JAN. 1976. Greenfield’sneuropathology. Chicago: Yearbook Medical.p 420–421.

Boudreau, D. 2002. Literacy skills in children andadolescents with Down syndrome. ReadWrit Interdiscipl J 15:497–525.

Bower A, Hayes A. 1994. Short-term memorydeficits and Down’s syndrome: a comparativestudy. Down Syndrome: Res Pract 2:47–50.

Brooksbank BWL, Walker D, Balazs R, et al.1989. Neuronal maturation in the foetalbrain in Down syndrome. Early Hum Dev18:237–246.

Buckley S, Bird G. 2002. Cognitive developmentand education: perspectives on Down syn-drome from a twenty-year research pro-gramme. In: Cuskelly M, Jobling A, BuckleyS, editors. Down syndrome across the life-span. London England: Whurr Publishers.p 66–80.

Buckley S, Bird G, Sacks B. 2006. Evidence thatwe can change the profile from a study ofinclusive education. Down Syndrome: ResPract 9:51–53.

Byrne A, Buckley S, MacDonald J, et al. 1995.Investigating the literacy, language, and mem-ory skills of children with Down’s syndrome.Down’s Syndrome: Res Pract 3:53– 58.

Campbell J, Gilmore L, Cuskelly M. 2003.Changing student teachers’ attitudes towardsdisability and inclusion. J Intellect Dev Disa-bil 28:369–379.

Carlesimo GA, Marotta L, Vicari S. 1997. Long-term memory in mental retardation: evi-dence for a specific impairment in subjectswith Down’s syndrome. Neuropsychologia35:71–79.

Casale-Giannola D, Wilson Kamens M. 2006.Inclusion at a university: experiences of ayoung woman with Down syndrome. MentRetard 44:344–352.

Chapman R, Seung H, Schwartz S, et al. 1998.Language skills of children and adolescentswith Down syndrome: II. Production deficits.J Speech Lang Hear Res 41:861–873.

Collacott RA, Cooper SA, McGrother A. 1992.Differential rates of psychiatric disorders inadults with Down’s syndrome compared toother mentally handicapped adults. Br J Psy-chiatry 161:671–674.

Crome L, Cowie V, Slater, E. 1966. A statisticalnote on cerebellar and brain-stem weight inMongolism. J Ment Defic 10:69–72.

Crown CL, Feldstein S, Jasnow MD, et al. 1992.Down’s syndrome and infant gaze: gazebehavior of Down’s syndrome and nonde-layed infants in interactions with their moth-ers. Eur J Child Adoles Psychiatry: ActaPaedopsychiatr 55:51–55.

Cupples L, Iacono T. 2000. Phonological aware-ness and oral reading skill in children withDown syndrome. J Speech Lang Hear Res43:595–608.

Cupples L, Iacono T. 2002. The efficacy of‘whole word’ versus ‘analytic’ readinginstruction for children with Down Syn-drome. Read Writ 15:549–574.

Dawson G, Osterling J, Rinaldi J, Carver L,McPartland J. 2001. Brief Report: Recogni-tion memory and stimulus-reward associa-tions: indirect support for the role of ventro-medial prefrontal dysfunction in autism. JAutism Dev Disord 31:337–341.

Dykens E, Shah B, Beck T, et al. 2002. Maladap-tive behavior in children and adolescentswith Down’s syndrome. J Intellect DisabilRes 46:484–492.

Dykens EM. 1995. Measuring behavioral pheno-types: provocations from the ‘‘New Genetics’’.Am J Ment Retard 99:522–532.

Dykens EM, Hodapp RM, Evans DW. 1994. Pro-files and development of adaptive behaviorin children with Down syndrome. Am JMent Retard 98:580–587.

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 269

Dykens EM, Hodapp RM, Finucane B. 2000.Genetics and mental retardation syndromes.Baltimore: Paul H. Brookes Publishing.

Eadie PA, Fey ME, Douglas JM, et al. 2002. Pro-files of grammatical morphology and sen-tence imitation in children with specific lan-guage impairment and Down syndrome. JSpeech Lang Hear Res 45:720–732.

Engidawork E, Lubec G. 2003. Molecular changesin fetal Down syndrome brain. J Neurochem84:895–904.

Fey ME, Warren SF, Brady N, et al. 2006. Earlyeffects of responsivity education/prelinguisticmilieu teaching for children with develop-mental delays and their parents. J SpeechLang Hear Res 49:526–547.

Fidler DJ. 2005. The emerging Down syndromebehavioral phenotype in early childhood:implications for practice. Infants YoungChild 18:86–103.

Fidler DJ. 2006.The emergence of a syndrome-specific personality-motivation profile inyoung children with Down syndrome. In:Rondal JA, Perera J, editors. Down syn-drome neurobehavioral specificity. Wiley.

Fidler DJ, Barrett KC, Most DE. 2005a. Agerelated differences in smiling and personalityin Down syndrome. J Dev Phys Disabil 17:263–280.

Fidler DJ, Hepburn S, Rogers S. 2006. Earlylearning and adaptive behavior in toddlerswith Down syndrome: evidence for anemerging behavioral phenotype? Down Syn-drome: Res Pract 10:37–44.

Fidler DJ, Lawson J, Hodapp RM. 2003. Whatdo parents want? an analysis of education-related comments made by parents of chil-dren with different genetic mental retarda-tion syndromes. J Intellect Dev Disabil 28:196–204.

Fidler DJ, Most DE, Guiberson MM. 2005b.Neuropsychological correlates of word iden-tification in Down syndrome. Res Dev Dis-abil 26:487–501.

Fidler DJ, Philofsky A, Hepburn S. 2007. Lan-guage phenotypes and intervention planning:bridging research and practice. Ment RetardDev Disabil Res Rev 13:47–57.

Fidler DJ, Philofsky A, Hepburn S, et al. 2005c.Nonverbal requesting and problem-solvingby toddlers with Down syndrome. Am JMent Retard 110:312–322.

Fl�orez J, Del Arco C, Gonzalez A, et al. 1990.Autoradiographic studies of neurotransmitterreceptors in the brain of newborn infantswith Down syndrome. Am J Med GenetSuppl 7:301–305.

Frangou S, Aylward E, Warren A, et al. 1997.Small planum temporale volume in Down’ssyndrome: a volumetric MRI study. Am JPsychiatry 154:1424–1429.

Freeman SFN, Hodapp RM. 2000. Educatingchildren with Down syndrome: linking be-havioral characteristics to promising inter-vention strategies. Down Syndrome Quart5:1–9.

Freeman SFN, Kasari C. 2002. Characteristics andqualities of the play dates of children withDown syndrome: emerging or true friend-ships. Am J Ment Retard 107:16–31.

Garcia SM, Conte EV. 2004. Improvement ofcognitive functions in children with Downsyndrome through the Bright Start Program.J Cogn Educ Psychol 4:65–86.

Gathercole SE, Adams AM. 1995. Phonologicalworking memory and speech production inpreschool children. J Speech Hear Res 38:403–414.

Ghaziuddin M, Tsai L, Ghaziuddin N. 1992. Au-tism in Down’s syndrome: presentation anddiagnosis. J Intellect Disabil Res 36:449–456.

Golden JA, Hyman BT. 1994. Development ofthe superior temporal neocortex is anoma-lous in trisomy 21. J Neuropathol Exp Neu-rol 53:513–520.

Gunn P, Cuskelly M. 1991. Down syndrome tem-perament: the stereotype at middle child-hood and adolescence. Int J Disabil DevEduc 38:59–70.

Gunn P, Berry P, Andrews RJ. 1982. Lookingbehavior of Down syndrome infants. Am JMent Defic 87:344–347.

Hamill L. 2003. Going to college: the experiencesof a young woman with Down syndrome.Ment Retard 41:340–353.

Haxby JV. 1989. Neuropsychological evaluation ofadults with Down’s syndrome: patterns ofselective impairment in non-demented oldadults. J Ment Defic Res 33:193–210.

Hepburn S, Fidler DJ, Rogers S. in press. Autismsymptoms in toddlers with Down syndrome.J Appl Res Intellect Disabil.

Hesketh LJ, Chapman RS. 1998. Verb use byindividuals with Down syndrome. Am JMent Retard 103:288–304.

Hill Karrer J, Karrer R, Bloom D, Chaney L,Davis R. 1998. Event-related brain poten-tials during an extended visual recognitionmemory task depict delayed development ofcerebral inhibitory processes among 6-month-old infants with Down syndrome. IntJ Psychophysiol 29:167–200.

Hodapp RM, Fidler DJ. 1999. Special educationand genetics: connections for the 21st cen-tury. J Spec Educ 33:130–137.

Hodapp RM, Freeman SFN. 2003. Advances ineducational strategies for children with Downsyndrome. Curr Opin Psychiatry 16:511–516.

Hodapp RM, Leckman JF, Dykens EM, et al.1992. K-ABC profiles in children with frag-ile X syndrome, Down syndrome, and non-specific mental retardation. Am J MentRetard 97:39–46.

Hulme C, Mackenzie S. 1992. Working memoryand severe learning difficulties. Hove: Law-rence Erlbaum.

Iarocci G, Virji-Babul N, Reebye P. 2006. TheLearn at Play Program (LAPP): mergingfamily, developmental research, early inter-vention, and policy goals for children withDown syndrome. J Policy Pract Intellec Dis-abil 3:11–21.

Irwin KC. 1991. Teaching children with Downsyndrome to add by counting-on. EducTreat Child 14:128–141.

Jarrold C, Baddeley AD. 1997. Short-term mem-ory for verbal and visuospatial informationin Down’s syndrome. Cogn Neuropschiatry2:101–122.

Jarrold C, Baddeley AD. 2001. Short-term mem-ory in Down syndrome: applying the work-ing memory model. Down Syndrome ResPract 7:17–23.

Jarrold C, Baddeley AD, Hewes AK. 1999. Ge-netically dissociated components of workingmemory: evidence from Down’s and Wil-liams syndrome. Neuropsychologia 37:637–651.

Jernigan TL, Bellugi U, Sowell E, Doherty S,Hesselink JR. 1993. Cerebral morphologicaldistinctions between Williams and Downsyndrome. Arch Neurol 50:186–191.

Jiang ZD, Wu YY, Liu XY. 1990. Early develop-ment of brainstem auditory evoked poten-tials in Down’s syndrome. Early Hum Dev23:41–51.

Johnston F, Stansfield J. 1997. Expressive prag-matics skills in pre-school children withand without Down’s syndrome: parentalperceptions. J Intellec Disabil Res 41:19–29.

Jonides J, Smith EE, Koeppe RA, et al. 1993.Spatial working memory in humans asrevealed by PET. Nature 363:623–625.

Kanno K, Ikeda Y. 2002. Word-length effect inverbal short term memory in individualswith Down’s syndrome. J Intellec DisabilRes 46:613–618.

Karmiloff-Smith A. 1997. Crucial differencesbetween developmental cognitive neuro-science and adult neuropsychology. DevNeuropsychol 13:513–524.

Karmiloff-Smith A. 1998. Development itself isthe key to understanding developmental dis-orders. Trends Cogn Sci 2:389–398.

Kasari C, Freeman SFN. 2001. Task-related socialbehavior in children with Down syndrome.Am J Ment Retard 106:253–264.

Kasari C, Freeman S, Mundy P, et al. 1995.Attention regulation by children with Downsyndrome: coordinated joint attention andsocial referencing looks. Am J Ment Retard100:128–136.

Kasari C, Mundy P, Yirmiya N, et al. 1990. Affectand attention in children with Down syn-drome. Am J Ment Retard 95:55–67.

Kates WR, Folley BS, Lanham DC, et al. 2002.Cerebral growth in fragile X syndrome:review and comparison with Down syn-drome. Microsc Res Tech 57:159–167.

Kay-Raining Bird E, Cleave PL, McConnell L.2000. Reading and phonological awarenessin children with Down syndrome. Am JSpeech-Lang Pathol 9:319–330.

Kennedy EJ, Flynn MC. 2003. Training, phono-logical awareness in children with Downsyndrome. Res Dev Disabil 24:44–57.

Kent L, Evans J, Paul M, et al. 1999. Comorbidityof autistic spectrum disorders in childrenwith Down syndrome. Dev Med ChildNeurol 41:153–158.

Klein BP, Mervis CB. 1999. Contrasting patternsof cognitive abilities of 9- and 10- year-oldswith Williams syndrome or Down syn-drome. Dev Neuropsychol 16:177–196.

Kleschevnikov AM, Belichenko PV, Villar AJ,et al. 2004. Hippocampal long-term poten-tiation suppressed by increased inhibitionin the Ts65Dn mouse, a genetic model ofDown syndrome. J Neurosci 24:8153–8160.

Kumin L. 2006. Speech intelligibility and childhoodverbal apraxia in children with Down syn-drome. Down Syndrome: Res Pract 10: 10–22.

Landry SH, Chapieski ML. 1990. Joint attentionand infant toy exploration: effects of Downsyndrome and prematurity. Child Dev 60:103–118.

Laws G. 1998. The use of nonword repetition asa test of phonological memory in childrenwith Down syndrome. J Child Psychol Psy-chiatry 39:1119–1130.

Laws G. 2002. Working memory in children andadolescents with Down syndrome: evidencefrom a colour memory experiment. J ChildPsychol Psychiatry 43:353–364.

Laws G, Bishop DVM. 2004. Verbal deficits inDown’s syndrome and specific languageimpairment: a comparison. Int J Lang Com-mun Disord 39:423–451.

Laws G, MacDonald J, Buckley S. 1996. Theeffects of a short training in the use of a re-hearsal strategy on memory for words andpictures in children with Down syndrome.Down Syndrome: Res Pract 4:70–78.

270 MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL

Legerstee M, Bowman TG, Fels S. 1992. Peopleand objects affect the quality of vocalizationsin infants with Down syndrome. Early DevParenting 1:149–156.

Lloveras RB, Fornells MG. 1998. The group: aninstrument of intervention for the globaldelve of the child with Down syndrome inthe process of social inclusion. Down Syn-drome: Res Pract 5:88–92.

Lloyd J, Moni KB, Jobling A. 2006. Breaking thehype cycle: using the computer effectivelywith learners with intellectual disabilities.Down Syndrome: Res Pract 9:68–74.

Mangan PA. 1992. Spatial memory abilities andabnormal development of the hippocampalformation in Down syndrome. Unpublisheddoctoral dissertation, University of Arizona,Tucson.

Marcell MM, Weeks SL. 1988. Short-term mem-ory difficulties and Down’s syndrome. JMent Defic Res 32:153–162.

Marcell MM, Harvey CF, Cothran LP. 1988. Anattempt to improve auditory short-termmemory in Down’s syndrome individualsthrough reducing distractions. Res Dev Dis-abil 9:405–417.

Miller JF. 1999. Profiles of language developmentin children with Down syndrome. In: MillerJ, Leddy M, Leavitt LA, editors. Improvingthe communication of people with Downsyndrome. Baltimore: Brookes Publishing. p11–40.

Miller JF, Leddy M. 1999. Verbal fluency, speechintelligibility, and communicative effective-ness. In: Miller JF, Leddy M, Leavitt LA,editors. Improving the communication ofpeople with Down syndrome. Baltimore:Brookes Publishing. p 81–91.

Miller J, Leddy M, Leavitt LA, editors. 1999.Improving the communication of peoplewith Down syndrome. Baltimore: BrookesPublishing.

Moni KB, Jobling A. 2001. Reading-related liter-acy learning of young adults with Downsyndrome: findings from a three year teach-ing and research program. Int J Disabil DevEduc 48:377–394.

Mundy P, Sigman M, Kasari C, et al. 1988. Non-verbal communication skills in Down syn-drome children. Child Dev 59:235–249.

Nadel L. 1986. Down syndrome in neurobiologi-cal perspective. In: Epstein CJ, editor. Theneurobiology of Down syndrome. NewYork: Raven Press. p 239–251.

Nelson C. 2000. The neurobiological bases ofearly intervention. In: Shokoff JP, Meisels J,editors. Handbook of early childhood inter-vention. New York: Cambridge UniversityPress. p 204–227.

Nye J, Buckley S. 2006. Evaluating the Numiconsystem of teaching number to children withDown syndrome. Down Syndrome: ResPract 11:97–104.

Ohr PS, Fagen JW. 1991. Conditioning and long-old infants with Down syndrome. Am JMent Retard 96:151–162.

Ohr PS, Fagen JW. 1993. Temperament, condi-tioning, old infants with Down syndrome. JAppl Dev Psychol 14:175–190.

Ohr PS, Fagen JW. 1994. Contingency learningDown syndrome. Am J Ment Retard 99:74–84.

Ortega-Tudela JM, Gomez-Ariza CJ. 2006. Com-puter-assisted teaching and mathematicallearning in Down syndrome children. JComput Assist Learn 22:298–307.

Paly RJ, Hurley AD. 2002. Down syndrome andautistic disorder. Ment Health Aspects DevDisabil 5:64–65.

Park S, Singer GH, Gibson M. 2005. The func-tional effect of teacher positive and neutralaffect on task performance of students withsignificant disabilities. J Positive Behav Interv7:237–246.

Pazos A, Del Olmo E, Diaz A, et al. 1994. Sero-tonergic (5-HhT1A and 5-HhT1D) andmuscarinic cholinergic receptors in DSbrains: an autoradiographic analysis. Posterpresented at International Down syndromeResearch Conference, Charleston, SouthCarolina, April 1994.

Pennington BF, Moon J, Edgin J, et al. 2003. Theneuropsychology of Down syndrome: evi-dence for hippocampal dysfunction. ChildDev 74:75–93.

Pinter JD, Eliez S, Schmitt E, et al. 2001. Neuro-anatomy of Down’s syndrome: a high-reso-lution MRI study. Am J Psychiatry 158:1659–1665.

Pitcairn TK, Wishart JG. 1994. Reactions ofyoung children with Down’s syndrome to animpossible task. Br J Dev Psychol12:485–489.

Pueschel SR, Gallagher PL, Zartler AS, et al.1987. Cognitive and learning profiles inchildren with Down syndrome. Res DevDisabil 8:21–37.

Rast M, Meltzoff AN. 1995. Memory and repre-sentation in young children with Downsyndrome: exploring deferred imitation andobject permanence. Dev Psychopathol 7:393–407.

Rempel ED. 1974. Psycholinguistic abilities ofDown’s syndrome children. In: Proceedingsof the Annual Meeting of the AmericanAssociation on Mental Deficiency, Toronto.

Roberts J, Price J, Malkin C. 2007. Language andcommunication development in Down syn-drome. Ment Retard Dev Disabil Res Rev13:26–35.

Rogoff B. 2003. The cultural nature of human de-velopment. Oxford: Oxford University Press.

Rondal J, Buckley S. 2001. Speech and languageintervention in Down syndrome. New York:Wiley.

Ruskin EM, Kasari C, Mundy P, et al. 1994.Attention to people and toys during socialand object mastery in children with Downsyndrome. Am J Ment Retard 99:103–111.

Schmidt-Sidor B, Wisniewski K, Shepard Th,et al. 1990. Brain growth in Down syn-drome subjects 15 to 22 weeks of gestationalage and birth to 60 months. Clin Neuropa-thol 9:181–190.

Sigman M, Ruskin E. 1999. Continuity andchange in the social competence of childrenwith autism, Down syndrome, and develop-mental delays. Monogr Soc Res Child Dev64:v–114.

Silverstein AB, Legutki G, Friedman SL, et al.1982. Performance of Down syndrome indi-viduals on the Stanford–Binet IntelligenceScale. Am J Ment Retard 86:548–551.

Stoel-Gammon C. 2003. Speech acquisition andapproaches to intervention. In: Rondal J,Buckley S, editors. Speech and languageintervention in Down syndrome. London:Whurr. p 49–62.

Thase ME, Tigner R, Smeltzer DJ, et al. 1984.Age-related neuropsychological deficits inDown’s syndrome. Biol Psychiatry 19:571–585.

Trent JA, Kaiser AP, Wolery M. 2005. The use ofresponsive interaction strategies by siblings.Top Early Child Spec Educ 25:107–118.

van Bysterveldt AK, Gillon GT, Moran C. 2006.Enhancing phonological awareness and letterknowledge in preschool children with Downsyndrome. Int J Disabil Dev Educ53:301–329.

Varnhagen CK, Das JP, Varnhagen S. 1987. Audi-tory and visual memory span: cognitiveprocessing by TMR individuals with Downsyndrome or other etiologies. Am J MentDefic 91:398–405.

Vicari S. 2001. Implicit versus explicit memoryfunction in children with Down and Wil-liams syndrome. Down Syndrome: Res Pract7:35–40.

Vicari S, Carlesimo A, Caltagirone C. 1995.Short-term memory in persons with intel-lectual disabilities and Down’s syndrome. JIntellect Disabil Res 39:532–537.

Vlachou M, Farrell P. 2000. Object mastery moti-vation in pre-school children with and with-out disabilities. Educ Psychol 20:167–176.

Wang PP, Bellugi U. 1994. Evidence from twogenetic syndromes for a dissociation betweenverbal and visuo-spatial short-term memory.J Clin Exp Neuropsychol 16:317–322.

Williams K, Wishart J. in press. Social cognitionin children with Down syndrome. MentRetard Dev Disabil Res Rev.

Wishart J. 1993. The development of learning dif-ficulties in children with Down syndrome. JIntellec Disabil Res 37:389–403.

Wishart JG. 1996. Avoidant learning styles andcognitive development in young children.In: Stratford B, Gunn P, editors. New ap-proaches to Down syndrome. London:Cassell. p 157–172.

Wisniewski K., Schmidt-Sidor B. 1989. Postnataldelay of myelin formation in brains fromDown syndrome infants and children. ClinNeuropathol 8:55–62.

Wisniewski K. 1990. Down syndrome childrenoften have brain with maturation delay, re-tardation of growth, and cortical dysgenesis.Am J Med Genet Suppl 7:274–281.

Wisniewski K, Kida E. 1994. Abnormal neuro-genesis and synaptogenesis in Down syn-drome. Dev Brain Dysfunct 7:289–301.

Wisniewski K, Laure-Kamionowska M, ConnellFF, et al. 1986. Neuronal density and synap-togenesis in the postnatal stage of brain mat-uration in Down syndrome. In: Epstein CJ,editor. The neurobiology of Down syn-drome. New York: Raven Press. p 29–44.

Wolpert G. 2001. What general educators have tosay about successfully including students withDown syndrome in their classes. J Res ChildEduc 16:28–38.

Yirmiya N, Solomonica-Levi D, Shulman C,et al. 1996. Theory of mind abilities in indi-vidual with autism, Down syndrome, andretardation of unknown etiology: The roleof age and intelligence. J Child Psychol Psy-chiatry 37:1003–1014.

Yoder P, Warren S. 2002. Effects of prelinguisticmilieu teaching and parent responsivity edu-cation on dyads involving children with in-tellectual disabilities. J Speech Lang HearRes 45:1158–1174.

Zelazo PD, Burack JA, Benedetto E, et al. 1996.Theory of mind and rule use in individualswith Down syndrome: a test of the unique-ness and specificity claims. J Child PsycholPsychiatry 37:479–484.

MRDD Research Reviews DOI 10.1002/mrdd � EDUCATION IN DOWN SYNDROME � FIDLER AND NADEL 271

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [UNAM]On: 27 May 2009Access details: Access Details: [subscription number 788841327]Publisher RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Educational ResearchPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713699076

Reading, dyslexia and the brainUsha Goswami a

a Centre for Neuroscience in Education, University of Cambridge, UK

Online Publication Date: 01 June 2008

To cite this Article Goswami, Usha(2008)'Reading, dyslexia and the brain',Educational Research,50:2,135 — 148

To link to this Article: DOI: 10.1080/00131880802082625

URL: http://dx.doi.org/10.1080/00131880802082625

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

Reading, dyslexia and the brain

Usha Goswami*

Centre for Neuroscience in Education, University of Cambridge, UK

(Received 17 August 2007; final version received 8 November 2007)

Background: Neuroimaging offers unique opportunities for understanding theacquisition of reading by children and for unravelling the mystery of developmentaldyslexia. Here, I provide a selective overview of recent neuroimaging studies, drawingout implications for education and the teaching of reading.Purpose: The different neuroimaging technologies available offer complementarytechniques for revealing the biological basis of reading and dyslexia. Functionalmagnetic resonance imaging (fMRI) is most suited to localisation of function, and henceto investigating the neural networks that underpin efficient (or inefficient) reading.Electroencephalography (EEG) is sensitive to millisecond differences in timing, hence itis suited to studying the time course of processing; for example, it can reveal whennetworks relevant to phonology versus semantics are activated. Magnetic sourceimaging (MSI) gives information about both location in the brain and the time course ofactivation. I illustrate how each technology is most suited to answering particularquestions about the core neural systems for reading, and how these systems interact, andwhat might go wrong in the dyslexic brain.Design and methods: Following a brief overview of behavioural studies of readingacquisition in different languages, selected neuroimaging studies of typical developmentare discussed and analysed. Those studies including the widest age ranges of childrenhave been selected. Neuroimaging studies of developmental dyslexia are then reviewed,focusing on (a) the neural networks recruited for reading, (b) the time course of neuralactivation and (c) the neural effects of remediation. Representative studies using thedifferent methodologies are selected. It is shown that studies converge in showing thatthe dyslexic brain is characterised by under-activation of the key neural networks forreading.Conclusions: Different neuroimaging methods can contribute different kinds of datarelevant to key questions in education. The most informative studies with respect tocausation will be longitudinal prospective studies, which are currently rare.

Keywords: reading; phonology; dyslexia; brain imaging

How universal are the neural demands made by learning to read in different languages?What are the core neural systems involved, and what goes wrong in the dyslexic brain?Current neuroimaging technologies are able to throw light on research questions such asthese, as will be illustrated below. In some instances, neuroimaging technologies cancontribute unique information that behavioural methodologies are simply unable toprovide. This includes information about the time processes in reading, and informationabout the parts of the brain that are affected by remedial packages for developmental

*Email: ucg10@cam.ac.uk

Educational Research

Vol. 50, No. 2, June 2008, 135–148

ISSN 0013-1881 print/ISSN 1469-5847 online

� 2008 NFER

DOI: 10.1080/00131880802082625

http://www.informaworld.com

Downloaded By: [UNAM] At: 00:52 27 May 2009

dyslexia. Some neuroimaging methodologies can gather data without requiringovert attention on the part of the child. These methodologies are particularly power-ful for contributing to our understanding of the biological basis of developmentaldyslexia.

The history of research on developmental dyslexia has been dominated by visualtheories of the disorder, ever since Hinshelwood (1896) described it as ‘congenital wordblindness’. Historically, theories of reading development also assumed that visualprocessing was core to individual differences in the acquisition of reading. In the 1970s,for example, there was much discussion of ‘Phonecian’ versus ‘Chinese’ readingacquisition strategies. It was assumed that children who were learning to readcharacter-based orthographies like Chinese required excellent visual memory skills inorder to distinguish between the visually complex characters that represented spokenwords. Hence visual memory or ‘logographic’ strategies were assumed core to readingacquisition of languages like Chinese and Japanese. Children who were learning to readlanguages like Greek or Italian, which were alphabetic and transparent (each lettercorresponding to one, and only one, sound) appeared to require code-breaking skills. Itwas assumed that once the brain had learned the symbol–sound code, reading shouldbe largely a process of phonological assembly. Many experiments were conductedwith children learning to read in English, to compare the contribution of ‘Chinese’versus ‘Phonecian’ acquisition strategies (e.g., Baron 1979). Dual-route models ofreading, originally developed using data from adults, were applied to childrenwho were learning to read (Stuart and Coltheart 1988). It was assumed that,developmentally, children could choose to learn to read by either Chinese or Phonecianstrategies.

These ideas about individual differences have not gone away (e.g., Stein and Walsh1997; Stuart 2006), but they are looking increasingly dated with the advent of brainimaging. Neuroimaging has also shed light on the processes underpinning the developmentof reading in deaf children, whom it was once assumed had no choice but to rely on visualmemorisation strategies (e.g., Conrad 1979). Essentially, neuroscience is showing thatdespite the apparently different demands on the brain made by learning to read English,Greek or Chinese, and the apparently different processing strategies used by children whoare deaf or who are dyslexic, reading across orthographies depends on the adequatefunctioning of the phonological system. Even for languages like Chinese, which wouldappear reliant on visual processing, it is oral language skills that underpin the acquisition ofreading.

As I show in this review, brain imaging studies demonstrate that reading beginsprimarily as a phonological process. In the early phases of learning to read, it is theneural structures for spoken language that are particularly active. As reading expertisedevelops, an area in the visual cortex originally named the ‘visual word form area’(VWFA) becomes increasingly active (Cohen and Dehaene 2004). This area is not alogographic system, even though it is very close to the visual areas that are active duringpicture naming. The VWFA is also active during nonsense word reading, and asnonsense words do not have word forms in the mental lexicon, the VWFA is thought tostore orthography–phonology connections at different grain sizes (Goswami and Ziegler2006a). Deaf readers rely on the same phonological system for reading as everyone else(MacSweeney et al. 2005). Children with developmental dyslexia show selective under-activation of key phonological areas of the brain, but targeted phonology-basedinterventions improve levels of activation in these areas, ‘normalising’ neural activity(Simos et al. 2002).

136 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

Learning to read: behavioural data

Many behavioural studies in developmental psychology show the critical role of‘phonological awareness’ in learning to read (for a recent review see Ziegler and Goswami2005). Phonological awareness is thought to develop via language acquisition. Between theages of 1 and 6 years, children acquire words at an exponential rate. For example, theaverage 1-year-old might have a productive vocabulary of around 50–100 words, but bythe age of 6 the average child’s receptive vocabulary will contain around 14,000 words(Dollaghan 1994). In order for the brain to represent each word as a distinct and uniquesequence of sounds, each entry in the ‘mental lexicon’ must incorporate phonologicalinformation along with information about meaning. For example, there must be implicitknowledge of the sounds that comprise a particular word, and the order in which theyoccur. Phonological awareness is essentially the child’s ability to reflect on this implicitknowledge, and to make judgements based on it. Hence phonological awareness istypically measured by a child’s ability to detect and manipulate component sounds inwords, for example, by deciding whether words rhyme, or by removing the initial soundfrom a spoken word.

The syllable is the primary processing unit across the world’s languages (Port 2006).In fact, there is an apparently language-universal sequence in the development ofphonological awareness, from syllable awareness, through ‘onset-rime’ awareness to‘phoneme’ awareness. Syllables (‘university’ has five syllables, ‘coffee’ has two syllables)can be segmented into sub-parts called onsets and rimes. The onset is the sound orsounds before the vowel, such as the ‘spr’ sound in ‘spring’ and the ‘st’ sound in ‘sting’.The rime is the vowel and any subsequent sounds in the syllable, such as the ‘ing’ soundin ‘spring’ and ‘string’. The phoneme is the smallest unit of sound that changesmeaning. ‘Spring’ and ‘string’ differ in meaning because the second sound is different ineach word (‘p’ versus ‘t’ respectively). In many of the world’s languages, onsets andrimes are the same as phonemes. This is because the dominant syllable structure acrossthe world’s languages is consonant–vowel (CV). Relatively few words in English are CVsyllables (5% of English monosyllables follow a CV structure: see De Cara andGoswami 2002). Examples of English words comprised of CV syllables are ‘go’, ‘do’and ‘yoyo’.

Behavioural studies across languages have shown that phonological sensitivity at allthree linguistic levels (syllable, onset-rime, phoneme) predicts the acquisition of reading(for a review see Ziegler and Goswami 2005). Furthermore, it has been demonstratedthat training phonological awareness has positive effects on reading acquisition acrosslanguages, particularly when it is combined with training about how letters or lettersequences correspond to sounds in that language (e.g., Bradley and Bryant 1983;Schneider, Roth, and Ennemoser 2000). Children with developmental dyslexia acrosslanguages appear to have specific problems in detecting and manipulating componentsounds in words (called a ‘phonological deficit’: see, e.g., Snowling 2000). For example,they find it difficult to count the number of syllables in different words, to recogniserhymes, to distinguish shared phonemes and to delete phonemes or substitute onephoneme for another (Korean: Kim and Davis 2004; German: Wimmer 1996; Greek:Porpodas 1999; Hebrew: Share and Levin 1999; for a comprehensive review see Zieglerand Goswami 2005). Dyslexic children are developing some awareness of phonology,but this is a slow and effortful process. Deaf children also develop phonological codes,for example, via lip reading (‘speech reading’) and vibrational cues. This is the caseeven if signing is their native language. Phonology is essentially the smallest contrastive

Educational Research 137

Downloaded By: [UNAM] At: 00:52 27 May 2009

units of a language that create new meanings. In signed languages, phonology dependson visual/manual elements, with handshapes, movements and locations combined toform signs (Sandler and Lillo-Martin 2006). For deaf children too, individualdifferences in phonological awareness are related to reading acquisition (e.g., Harrisand Beech 1998).

Reading and learning to read: neuroimaging data

To date, most neuroimaging studies of reading have been conducted with adults (see Priceand McCrory 2005 for a recent synthesis). This was partly due to the methodologiesavailable. The most popular methods for studying the brain during the act of readingdepended on imaging techniques like functional magnetic resonance imaging (fMRI) andpositron emission tomography (PET). The fMRI technique measures changes in bloodflow in the brain, which take approximately 6–8 seconds to reach a maximum value (i.e.maximum activity will be measurable 6–8 seconds after reading a particular word). fMRIworks by measuring the magnetic resonance signal generated by the protons of watermolecules in brain cells, generating a BOLD (blood oxygenation level dependent)response. The fMRI method is excellent for the localisation of function, but becausechanges in brain activity are summated over time, it cannot provide information about thesequence in which different neural networks become engaged during the act of reading. InPET, radioactive tracers are injected into the bloodstream and provide an index of brainmetabolism. Because of the use of radioactive tracers, PET is not suitable for studyingchildren.

More recently, the value of the electroencephalogram (EEG) methodology for studyingreading is being recognised. Neurons communicate on a millisecond scale, with the earlieststages of cognitive information processing beginning between 100 ms and 200 ms afterstimulus presentation. EEG methods can measure the low-voltage changes caused by theelectro-chemical activity of brain cells, thereby reflecting the direct electrical activity ofneurons at the time of stimulation (e.g., at the time of seeing a word). Initially, EEGmethods were less widely used in the neuroscience of reading, because it is difficult tolocalise function using EEG. However, developmentally, information about the timecourse of processing is very important. Data from EEG studies suggest that the brain hasdecided whether it is reading a real word or a nonsense word within 160–180 ms ofpresentation, for children and adults across languages (Csepe and Szucs 2003; Suaseng,Bergmann, and Wimmer 2004).

Adult studies of reading based on PET and fMRI have focused on a relatively smallrange of reading and reading-related tasks, and studies of children using fMRI havefollowed suit. Typical tasks include asking participants to read single words andthen comparing brain activation to a resting condition with the eyes closed; askingparticipants to pick out target visual features while reading print or ‘false font’ (false fontis made up of meaningless symbols matched to letters for visual features like the‘ascenders’ in the letters b, d, k); making phonological judgements while reading words ornonsense words (e.g., ‘do these items rhyme?’: leat, jete ) and making lexical decisions (e.g.,pressing a button when a word is presented, and a different button when a nonsense wordis presented). Adult experiments show a very consistent picture concerning the neuralnetworks that underpin skilled reading (e.g., Price et al. 2003; Rumsey et al. 1997; see forcomments on divergence Price and McCrory 2005). Word recognition in skilled readersappears to depend on a left-lateralised network of frontal, temporoparietal andoccipitotemporal regions, whatever language they are reading (see Figure 1). However,

138 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

there is some additional recruitment of visuo-spatial areas for languages with non-alphabetic orthographies (e.g., left middle frontal gyrus for Chinese: see meta-analysisby Tan et al. 2005). The frontal, temporoparietal and occipitotemporal regionsessentially comprise the language, auditory, cross-modal and visual areas of the brain.At a very simple level, semantic and memory processing is thought to occur intemporal and frontal areas, auditory and phonological processing in temporal areas,articulation in frontal areas, visual processing in occipital areas and cross-modalprocessing in parietal areas.

Although there are still relatively few neuroimaging studies of children reading, thestudies that have been done show a high degree of consistency in the neural networksrecruited by novice and expert readers. For example, work by Turkeltaub andcolleagues has used fMRI and the false font task to compare neural activation inEnglish-speaking participants aged from 7 to 22 years (Turkeltaub et al. 2003).Importantly, 7-year-olds can perform the ‘false font’ task as well as adults, hencechanges in reading-related neural activity are likely to reflect developmental differencesrather than differences in reading expertise. Turkeltaub et al. (2003) reported thatadults performing their task activated the usual left hemisphere sites, including left

Figure 1. A schematic depiction of some of the neural areas involved in reading (left hemispheredepiction) (from Price and McCrory 2005).

Educational Research 139

Downloaded By: [UNAM] At: 00:52 27 May 2009

posterior temporal and left inferior frontal cortex. They then restricted the analyses tochildren below 9 years of age. Now the main area engaged was left posterior superiortemporal cortex. This region is traditionally considered the focus of phonologicalactivity, and is thus thought to be active during grapheme–phoneme translation. Asreading developed, activity in left temporal and frontal areas increased, while activitypreviously observed in right posterior areas declined. This pattern was interpreted asshowing that reading-related activity in the brain becomes more left-lateralised withdevelopment.

In further analyses focusing just on the younger children, the researchersinvestigated the relationships between three core phonological skills and wordprocessing. The three core phonological skills are usually taken to be phonologicalawareness, phonological memory and rapid automatised naming (RAN). I will focuson the phonological awareness findings here. Turkeltaub et al. (2003) calculated partialcorrelations between activated brain regions and each of these three measures whilecontrolling for the effects of the other two measures. They reported that the threedifferent measures correlated with three distinct patterns of brain activity. Brainactivity during phonological awareness tasks appeared to depend on a network of areasin left posterior superior temporal cortex (phonology and grapheme–phonemetranslation) and inferior frontal gyrus (articulation). The level of the children’sphonological skills modulated the amount of activity in this network. As noted earlier,the left posterior temporal sulcus was the primary area recruited by young children atthe beginning of reading development. Therefore, neuroimaging data suggest thatphonological recoding to sound rather than logographic recognition is the key earlyreading strategy. Activity in the inferior frontal gyrus increased with reading ability.This area is also a key phonological area (Broca’s area), important for the motorproduction of speech. Left inferior frontal gyrus is also activated when deaf childrenperform phonological awareness tasks silently in fMRI studies (MacSweeney et al.2005).

An fMRI study of 119 typically developing readers aged from 7 years to 17 years byShaywitz and colleagues found a similar developmental pattern (Shaywitz et al. 2007).Instead of the false font task, this study used a rhyme decision task (e.g., ‘do these itemsrhyme?’: leat, kete ), and a visual line orientation task (e.g., ‘Do [\\V] and [\\V] match?’).Shaywitz and his colleagues reported that networks in both left and right superior andmiddle frontal regions were more active in younger readers, with activity declining asreading developed. In contrast, activity in the left anterior lateral occipitotemporal regionincreased. This region includes the putative visual word form area (VWFA). Hence bothTurkeltaub et al. (2003) and Shaywitz et al. (2007) found decreased right hemisphereinvolvement as reading developed, but found this for somewhat different neural networks.The difference in the behavioural tasks used (e.g., false font versus rhyme judgement) mayexplain some of these differences.

Overall, therefore, current neuroimaging data support a ‘single route’ model ofreading development, based on a process of developing orthographic–phonologicalconnections at different grain sizes (Ziegler and Goswami 2006). Reading is founded inphonology from the beginning (Goswami and Ziegler 2006b). The VWFA becomesmore active as reading develops, reflecting the development of an orthographic lexiconcontaining both whole words and fragments of familiar words such as orthographicrimes (Pugh 2006). The VWFA is not a logographic or visual lexicon, able to support‘Chinese’ processing or the ‘direct route’ from printed word to meaning postulated by‘dual-route’ theory. Neuroimaging studies of typically developing readers show that the

140 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

neural networks for spoken language play an important developmental role in readingfrom the outset.

Neuroimaging studies of dyslexia

The networks recruited for reading

Neuroimaging studies of adult readers with developmental dyslexia suggest that there isatypical activation in the three important neural sites for reading, namely the left posteriortemporal regions, the left inferior frontal regions and the left occipitotemporal regions(such as the VWFA). These data suggest both problems with the phonological aspects ofreading and with the efficient development of an orthographic lexicon (e.g., Brunswicket al. 1999). These fMRI and PET studies typically rely on tasks such as word andnonsense word reading (e.g., ‘valley’, ‘carrot’, ‘vassey’, ‘cassot’), and the ‘false font’ task.Again, the experimental picture is largely one of convergence across orthographies. Forexample, adult dyslexics in Italian, French and English all showed activation of a left-lateralised neural network based around posterior inferior temporal areas and middleoccipital gyrus (Paulesu et al. 2001). This was a cross-language comparison within onestudy. However, issues of experimental design become critical when comparing individualimaging studies across languages. When studying any kind of disability, it is crucial toequate participant groups for their overall ability in the actual tasks being used to acquirethe neuroimaging data. For example, it is impossible to interpret group differences in brainactivity if the dyslexics are worse at reading the nonsense words being used than thecontrol adults. In this case, differences in neural activation could simply reflect differentskill levels (i.e., behavioural differences in reading performance). Similarly, it is critical touse the same criteria for acquiring images of the brain in different studies if interpretationsabout cross-language differences are being drawn (e.g., Ziegler 2005). Otherwise, apparentlanguage-based differences could simply reflect differences in the significance thresholds orother experimental criteria used to acquire the images by different research groups.

Neuroimaging studies of children with developmental dyslexia report a very similarpattern to adult data (e.g., Shaywitz et al. 2002, 2007; Simos et al. 2000). For example,Shaywitz et al. (2002) studied 70 children with dyslexia aged on average 13 years, andcompared them to 74 11-year-old typically developing controls (although the controlswere not matched for reading level). Using fMRI, the children were scanned whileperforming a variety of reading-related tasks. These were letter identification (e.g., are tand V the same letter?); single letter rhyme (e.g., do V and C rhyme?); non-word rhyming(e.g., do leat and jete rhyme?); and reading for meaning (e.g., are corn and rice in the samesemantic category?). Brain activity in each condition was contrasted with activity in abaseline condition, the line orientation task (e.g., do [\\V] and [\\V] match?). Shaywitz et al.(2002) reported that the children with developmental dyslexia showed under-activation inthe core left temporoparietal networks, with older dyslexics showing over-activation inright inferior frontal gyrus. The children with developmental dyslexia also showedincreased activation in right temporoparietal networks. One drawback of the study,however, was that there were group differences in behavioural performance in some of thecomponent tasks. In the non-word rhyming measure, for example, the controls [79%] weresignificantly better than the children with dyslexia [59%]). This means that some of thedifferences found in brain activation could reflect differing levels of expertise rather thandifferences core to having developmental dyslexia. In a subsequent study of an expandedsample, Shaywitz et al. (2007) used in-magnet non-word reading ability as a covariate tocontrol for this problem. Shaywitz et al. compared 113 dyslexic children aged 7–18 years to

Educational Research 141

Downloaded By: [UNAM] At: 00:52 27 May 2009

the 119 typically developing readers discussed above in the non-word rhyme and visualline orientation tasks. Compared to the typically developing children, the dyslexic childrenshowed no age-related increase in the activity of the VWFA. Instead, activity in the leftinferior frontal gyrus (speech articulation) and the left posterior medial occipitotemporalsystem both increased, and reading did not become left-lateralized, with continued righthemisphere involvement.

There are also a few studies in the literature exploring the neural networks recruited forreading by dyslexic children in other languages. A study of 13 German dyslexic childrenaged 14–16 years was reported by Kronbichler et al. (2006). They used a sentenceverification task (e.g., ‘A flower needs water’ – TRUE), in an fMRI design, to try andreplicate natural reading. A false font task provided the control task. Consistent withstudies of English dyslexics, they found reduced activation of left occipitotemporalnetworks and increased activation of left inferior frontal areas. A study of eight Chinesechildren with developmental dyslexia reported by Siok et al. (2004) claimed biologicaldisunity, however. Their fMRI study used a homophone judgement task, in which thechildren had to decide whether two different Chinese characters made the same sound (anEnglish homophone is week – weak), and a character decision task, in which the childrensaw one Chinese character and had to decide whether it was a real word or not. The firsttask was intended to measure orthography–phonology connections, and the secondorthography–semantic relations. Siok et al. (2004) reported that the Chinese dyslexics didnot demonstrate the reduced activation in left temporoparietal regions that wouldtypically be found in developmental dyslexia in English during the homophone judgementtask. Instead, an area involved in visuo-spatial analysis showed reduced activity, the leftmiddle frontal gyrus. Siok et al. (2004) used this latter finding to argue that the biologicalmarker for developmental dyslexia in Chinese was reduced activation of left middle frontalgyrus. However, the design of this study does not yet permit this conclusion. A controlgroup matched for reading level is also required. Reduced activation in left middle frontalgyrus when making homophone judgements in Chinese might be expected for the level ofreading achieved by the children with dyslexia. If this were to be the case, then increasedinvolvement of networks for visuo-spatial analysis as reading develops would be part oftypical reading development in Chinese, rather than a unique biological marker fordevelopmental dyslexia.

Developmental differences in the time course of neural activation

While fMRI studies can provide important information about the neural networkssupporting reading in typically developing versus dyslexic readers, they do not provideinformation about the time course of neural processing. This is important, as in typicallydeveloping readers words are distinguished from non-words within around 180 ms,suggesting early contact with the VWFA and semantic sites. It seems likely that thisprocess would be delayed in developmental dyslexia. Similarly, it seems possible thatcognitive processes such as grapheme–phoneme conversion might take longer indevelopmental dyslexia.

A longitudinal study of 33 English-speaking children using magnetic source imaging(MSI) compared brain activation in a letter–sound task (the child sees a letter and has toprovide its sound) and a simple non-word reading task (e.g., ‘lan’) at the end ofkindergarten and again at the end of grade 1 (Simos et al. 2005). Magnetic source imagingdepends on a combination of magneto-encephalography (MEG) and MRI. The MEGmeasures the magnetic fields generated by the electrical activity in the brain rather than the

142 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

electrical activity itself (the latter is measured by EEG). These magnetic fields are tiny, theyare one billion times smaller than the magnetic field generated by the electricity in alightbulb. By combining this information with MRI scans, both the time course andspatial localisation of brain activity is possible. Of the 33 children studied, 16 were thoughtto be at high risk of developing dyslexia.

Simos et al. (2005) reported that the high-risk group were significantly slower to showneural activity in response to both letters and non-words in kindergarten in theoccipitotemporal region (320 ms compared to 210 ms for those not at risk). The high-riskgroup also showed atypical activation in the left inferior frontal gyrus when performingthe letter–sound task, with the onset of activity increasing from 603 ms in kindergarten to786 ms in grade 1. The typically developing readers did not show this processing timeincrease. Comparing the onset of activity of the three core neural networks for reading,Simos et al. (2005) reported that low-risk children showed early activity in the leftoccipitotemporal regions, followed by activity in temporoparietal regions, predominantlyin the left hemisphere, and then bilateral activity in inferior frontal regions. In contrast,high-risk children showed little differentiation in terms of the time course of activationbetween the occipitotemporal and temporoparietal regions. High-risk children who werenon-responsive to a phonological remediation package also being administered (n ¼ 3)were distinct in showing earlier onset of activity in inferior frontal gyrus compared to thetemporoparietal regions. Given the current dearth of time-course studies by other researchgroups in either English or in other languages, it is difficult to interpret these differences interms of the cognitive components of reading. Nevertheless, Simos et al. (2005) commentthat the increased inferior frontal activation probably reflects the role of compensatoryarticulatory processes. As noted earlier, deaf children also show increased inferior frontalactivation during phonological processing tasks. This may indicate that children withphonological difficulties rely more heavily on networks for articulation when phonologicalprocessing is required.

The neural effects of remediation

Although there are a variety of remediation packages for dyslexic children based ondifferent theories of developmental dyslexia, the most effective packages across languagesappear to be those offering intensive phonological intervention (e.g., Bradley and Bryant1983; Schneider, Roth, and Ennemoser 2000). Simos and his research group (2002) usedmagnetic source imaging to explore neural activation in eight children with develop-mental dyslexia who had received 80 hours of intensive training with such a package andwho had shown significant benefits from the remediation (Simos et al. 2002). MSI scanswere taken during a non-word rhyme matching task (e.g., ‘yoat’, ‘wote’) both before theintervention and following remediation. Simos et al. (2002) reported that prior to theintervention, the dyslexic children showed the expected hypoactivation of lefttemporoparietal regions. Following the intervention, all eight children showed adramatic increase in the activation of left temporoparietal regions, predominantly inthe left posterior superior temporal gyrus (the networks supporting grapheme–phonemerecoding in typically developing readers: see Turkeltaub et al. 2003). These activationprofiles were very similar to those of eight controls who also participated in the MSIstudy, but who did not require remediation. Nevertheless, even after remediation neuralactivity was delayed in the children with dyslexia relative to the controls. The peak inleft superior temporal gyrus activity occurred at 837 ms on average for the dyslexicchildren, and at 600 ms for the controls. The data were taken to show a normalisation

Educational Research 143

Downloaded By: [UNAM] At: 00:52 27 May 2009

of brain function with remediation. Nevertheless, Simos et al. (2002) commented thateven with intensive remediation, children with dyslexia are slow to achieve the readingfluency shown by non-dyslexic children.

Shaywitz and Shaywitz (2005) used retrospective examination of the large sample ofchildren with developmental dyslexia reported in Shaywitz et al. (2002) to compare thedifferent developmental trajectories for children at risk for reading difficulties. Shaywitzand Shaywitz (2005) distinguished three groups within this sample when they were youngadults. The first was a group of persistently poor readers (PPR), who had met criteria forpoor reading in both the 2nd/3rd and the 9th/10th grades. The second was a group ofaccuracy-improved poor readers (AIR), who had met criteria for poor reading in the 2nd/3rd grades but who did not meet criteria in the 9th/10th grades. The third was a controlgroup of non-impaired readers (C), who had never met criteria for poor reading (theparticipants had been studied since the age of 5 years). Shaywitz and Shaywitz (2005)reported that both the PPR and the AIR groups showed hypoactivation of the core lefthemisphere sites when required to manipulate phonology. For example, in a nonsenseword rhyming task, both groups of young adults still showed relative hypoactivity inneural networks in left superior temporal and occipitotemporal regions. However, thegroups were distinguished by their neural activity when reading real words. The AIRgroup still demonstrated under-activation in the usual left posterior areas for real wordreading, whereas the PPR group activated the left posterior regions to the same extent ascontrols (this was an unexpected finding).

Shaywitz and Shaywitz (2005) then carried out further analyses based on connectivity.Connectivity analyses examine the neural areas that are functionally connected to eachother during reading. The connectivity analyses suggested that reading achievementdepended on memory for the PPR group, and not on the normalised functioning of the leftposterior regions. The unimpaired controls demonstrated functional connectivity betweenleft hemisphere posterior and anterior reading systems, but the PPR group demonstratedfunctional connectivity between left hemisphere posterior regions and right prefrontalareas associated with working memory and memory retrieval. Shaywitz and Shaywitz(2005) speculated that the PPR group were reading primarily by memory. As the wordsused in the scanner were high-frequency, simple words, this is quite possible. However, thisdesign choice complicates the interpretation of the neural differences found, as the PPRgroup may not be able to use memory strategies to read less frequent or less simple words.For such stimuli, the PPR and AIR groups may show similar neural profiles. It may alsobe important that the PPR group had, in general, lower IQ scores than the AIR group.Prospective longitudinal studies comparing patterns of neural activation and connectivityin dyslexic children as high-frequency words become over-learned would clearly be veryvaluable.

Different technologies, different research questions: the promise of brain imaging for

understanding reading and developmental dyslexia

As will be clear from the foregoing review, most studies of reading development and ofdevelopmental dyslexia have relied on fMRI. These studies have provided excellent dataregarding the neural networks underpinning reading in typically developing and dyslexicreaders. They have also shown that the functional organisation of the networks forreading is similar in typical development and in dyslexia. Children with developmentaldyslexia do not recruit radically different neural networks when they are reading. Rather,they show hypoactivation of crucial parts of the network of areas involved in word

144 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

recognition, and an atypical pattern of continuing right hemisphere involvement.Although highly informative, these studies are essentially correlational studies. They cananswer research questions about the neural demands made by learning to read in differentlanguages, and they can answer research questions about the core neural systems involvedfor dyslexic and typically developing readers. They can also answer research questionsabout the patterns of connectivity between different neural networks. However, theycannot answer research questions about what ‘goes wrong’ in the dyslexic brain, althoughthey can help to rule out hypotheses (e.g., about the visual basis of developmental dyslexia;see Eden and Zeffiro 1998).

Neuroimaging methods that provide data on the time course of neural processing, suchas MEG (MSI) and EEG, can begin to answer causal questions. As might be expected, ithas been shown using MSI that neural activation is delayed in core components of thenetwork of areas recruited for reading by children at risk for dyslexia. However,behavioural studies showing that children with developmental dyslexia are slower to readwords aloud make the same point. When EEG or MSI techniques show that corecomponents of the reading network are activated in a different order in dyslexia comparedto typical reading, this is more informative with respect to causality. For example, Simosand his colleagues have shown atypically earlier onset of activity in inferior frontal gyrus(articulation) compared to the temporoparietal regions in three children at high risk fordyslexia who appear to be non-respondent to a phonological remediation package. Ifrobust with larger samples and diagnosed dyslexics, such findings could suggest that thereare different neuro-developmental routes to word recognition for dyslexic childrencompared to controls. Nevertheless, these different neuro-developmental routes are not thecause of dyslexia. Rather, they illustrate the response of a dyslexic brain to being trained tolearn to read.

In my view, the most informative studies with respect to causation in developmentaldyslexia are longitudinal prospective studies that use brain imaging to study basic sensoryprocessing in at-risk children, with a view to understanding the causes of the phonologicaldeficit. Here, the most promising studies to date are those investigating basic auditoryprocessing using methodologies sensitive to the time course of auditory processing at themillisecond level. For example, a large-scale Finnish study (the Jyvaskyla LongitudinalStudy of Dyslexia (JLD): see Lyytinen et al. 2004a) has followed babies at familial risk fordyslexia since birth. A large variety of behavioural and EEG measures has been taken asthe children have developed. EEG measures of auditory sensory processing (evokedresponse potentials to speech and non-speech cues) have been found to distinguish the at-risk babies from controls even during infancy (e.g., Lyytinen et al. 2005). For example,infants at risk for developmental dyslexia were less sensitive to the auditory cue ofduration at six months of age (Richardson et al. 2003). The infant participants had todiscriminate between two bisyllabic speech-like stimuli with a varying silent interval (e.g.‘ata’ versus ‘atta’). Duration discrimination was still impaired when the same childrenwere 6.5 years of age (Lyytinen et al. 2004b).

English children with developmental dyslexia are also impaired in this durationdiscrimination task (Richardson et al. 2004). In addition, English children are impaired indiscriminating the rise time of amplitude envelopes at onset, which is an importantauditory cue to the onset of syllables in the speech stream (Goswami et al. 2002;Richardson et al. 2004). Finnish adults with developmental dyslexia also show rise timeprocessing impairments, and individual differences in rise time sensitivity predicted up to35% of unique variance in phonological tasks like rhyme recognition (Hamalainen et al.2005). In the English studies, individual differences in rise time sensitivity predict unique

Educational Research 145

Downloaded By: [UNAM] At: 00:52 27 May 2009

variance in both phonological awareness measures (around 20%: Richardson et al. 2004)and in reading and spelling measures (around 25%: Goswami et al. 2002). We arecurrently collecting EEG data comparing rise time discrimination in English children withand without dyslexia. Data so far suggest that children with developmental dyslexia indeedshow atypical auditory processing of rise time stimuli, with N1 amplitude (an EEGmeasure of sound registration) failing to reduce as amplitude envelope rise times becomeextended (Thomson, Baldeweg, and Goswami 2005). This suggests that neural responsesin the dyslexic brain do not distinguish between different rise times, at least for theauditory processing comparisons used in our study (15 ms versus 90 ms rise times).

Conclusion

Different neuroimaging methodologies contribute complementary data regarding theneural networks underpinning reading acquisition and developmental dyslexia. WhilefMRI studies can identify the core neural systems involved in reading, EEG and MEGmethodologies are required to investigate the time course of activation of the differentnetworks that contribute to word recognition, and to investigate potential sensoryprecursors of the phonological deficit. With respect to key questions in education, eachneuroimaging method can contribute different kinds of data. For example, whenevaluating the claims made for different kinds of remediation package for developmentaldyslexia, fMRI will be useful in assessing whether interventions affect the core neuralnetworks for reading, or affect a different kind of network (e.g., motivational systems).When evaluating claims that the core cognitive difficulty in developmental dyslexia lies informing a high-quality phonological representation, methodologies that can explore thetime course of sensory processing such as EEG will be most useful. Neuroimagingmethods are of optimal use when they can provide experimental data that is not availablefrom behavioural investigations. For example, it is possible in principle to identify neuralmarkers of risk for developmental dyslexia that can be measured in pre-verbal infants andin older children without requiring their explicit attention (Szucs and Goswami 2007). It isthese areas of neuroscience that are likely to be of most potential benefit to educators.

References

Baron, J. 1979. Orthographic and word-specific mechanisms in children’s reading of words. ChildDevelopment 50: 60–72.

Bradley, L., and P.E. Bryant. 1983. Categorising sounds and learning to read: A causal connection.Nature 310: 419–21.

Brunswick, N., E. McCrory, C.J. Price, C.D. Frith, and U. Frith. 1999. Explicit and implicitprocessing of words and pseudowords by adult developmental dyslexics: A search for Wernicke’sWortschatz. Brain 122: 1901–17.

Cohen, L., and S. Dehaene. 2004. Specialization within the ventral stream: The case for the visualword form area. NeuroImage 22: 466–76.

Conrad, R. 1979. The deaf school child. London: Harper & Row.Csepe, V., and D. Szucs. 2003. Number word reading as a challenging task in dyslexia? An ERP

study. International Journal of Psychophysiology 51: 69–83.De Cara, B., and U. Goswami. 2002. Statistical analysis of similarity relations among spoken words:

Evidence for the special status of rimes in English. Behavioural Research Methods andInstrumentation 34, no. 3: 416–23.

Dollaghan, C.A. 1994. Children’s phonological neighbourhoods: Half empty or half full? Journal ofChild Language 21: 257–71.

Eden, G.F., and T.A. Zeffiro. 1998. Neural systems affected in developmental dyslexia revealed byfunctional neuroimaging. Neuron 21: 279–82.

146 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

Goswami, U., J. Thomson, U. Richardson, R. Stainthorp, D. Hughes, S. Rosen et al. 2002.Amplitude envelope onsets and developmental dyslexia: A new hypothesis. Proceedings of theNational Academy of Sciences of the United States of America 99: 10911–16.

Goswami, U., and J.C. Ziegler. 2006a. Fluency, phonology and morphology: A response to thecommentaries on becoming literate in different languages. Developmental Science 9, no. 5: 451–3.

———. 2006b. A developmental perspective on the neural code for written words. Trends inCognitive Sciences 10, no. 4: 142–3.

Hamalainen, J., P.H.T. Leppanen, M. Torppa, K. Muller, and H. Lyytinen. 2005. Detection ofsound rise time by adults with dyslexia. Brain and Language 94: 32–42.

Harris, M., and J.R. Beech. 1998. Implicit phonological awareness and early reading development inprelingually deaf children. Journal of Deaf Studies and Deaf Education 3, no. 3: 205–16.

Hinshelwood, J.A. 1896. A case of dyslexia: A peculiar form of word-blindness. Lancet 2: 1451.Kim, J., and C. Davis. 2004. Characteristics of poor readers of Korean Hangul: Auditory, visual and

phonological processing. Reading and Writing 17, no. 1–2: 153–85.Kronbichler, M., et al. 2006. Evidence for a dysfunction of left posterior reading areas in German

dyslexic readers. Neuropsychologia 44: 1822–32.Lyytinen, H., T. Ahonen, and T. Guttorm, et al. 2004a. Early development of children at familial

risk for dyslexia: Follow-up from birth to school age. Dyslexia 10: 146–78.Lyytinen, H., M. Aro, and K. Eklund, et al. 2004b. The development of children at familial risk for

dyslexia: Birth to school age. Annals of Dyslexia 5, no. 4: 185–220.Lyytinen, H., et al. 2005. Psychophysiology of developmental dyslexia: A review of findings

including studies of children at risk for dyslexia. Journal of Neurolinguistics 18: 167–95.MacSweeney, M., D. Waters, M. Brammer, B. Woll, and U. Goswami. 2005. ‘Phonological

processing of speech and sign in the deaf brain’. Poster presented at the Cognitive NeuroscienceSociety, March, in New York.

Paulesu, E., et al. 2001. Dyslexia: Cultural diversity and biological unity. Science 291, no. 5511: 2165–7.Porpodas, C.D. 1999. Patterns of phonological and memory processing in beginning readers and

spellers of Greek. Journal of Learning Disabilities 32: 406–16.Port, R. 2006. The graphical basis of phones and phonemes. In Second language speech learning: The

role of language experience in speech perception and production, ed. M. Munro, and O. Schwen-Bohm. Amsterdam: John Benjamins.

Price, C.J., M.-L. Gorno-Tempini, K.S. Graham, N. Biggio, A. Mechelli, K. Patterson, and U.Noppeney. 2003. Normal and pathological reading: Converging data from lesion and imagingstudies. NeuroImage 20, suppl. 1: S30–S41.

Price, C.J., and E. McCrory. 2005. Functional brain imaging studies of skilled reading anddevelopmental dyslexia. In The science of reading: A handbook, ed. M.J. Snowling andC. Hulme, 473–96. Oxford: Blackwell.

Pugh, K. 2006. A neurocognitive overview of reading acquisition and dyslexia across languages.Developmental Science 9: 448–50.

Richardson, U., P.H.T. Leppanen, M. Leiwo, and H. Lyytinen. 2003. Speech perception of infantswith high familial risk for dyslexia differ at the age of 6 months. Developmental Neuropsychology23: 385–97.

Richardson, U., J. Thomson, S.K. Scott, and U. Goswami. 2004. Auditory processing skills andphonological representation in dyslexic children. Dyslexia: An International Journal of Researchand Practice 10, no. 3: 215–33.

Rumsey, J.M., B. Horwitz, B.C. Donohue, K. Nace, J.M. Maisog, and P. Andreason. 1997.Phonological and orthographic components of word recognition: A PET rCBF study. Brain 120:739–59.

Sandler, W., and D. Lillo-Martin. 2006. Sign language and linguistic universals. Cambridge:Cambridge University Press.

Sauseng, P., J. Bergmann, and H. Wimmer. 2004. When does the brain register deviances fromstandard word spellings? An ERP study. Cognitive Brain Research 20: 529–32.

Schneider, W., E. Roth, and E. Ennemoser. 2000. Training phonological skills and letter knowledgein children at-risk for dyslexia: A comparison of three kindergarten intervention programs.Journal of Educational Psychology 92: 84–95.

Share, D., and I. Levin. 1999. Learning to read and write in Hebrew. In Learning to read and write: Across-linguistic perspective. In Cambridge Studies in Cognitive and Perceptual Development, ed.M. Harris and G. Hatano, 89–111. New York: Cambridge University Press.

Educational Research 147

Downloaded By: [UNAM] At: 00:52 27 May 2009

Shaywitz, B.A., et al. 2002. Disruption of posterior brain systems for reading in children withdevelopmental dyslexia. Biological Psychiatry 52, no. 2: 101–10.

———. 2007. Age-related changes in reading systems of dyslexic children. Annals of Neurology 61:363–70.

Shaywitz, S.E., and B.A. Shaywitz. 2005. Dyslexia (specific reading disability). Biological Psychiatry57: 1301–9.

Simos, P.G., J.I. Breier, J.M. Fletcher, E. Bergman, and A.C. Papanicolaou. 2000. Cerebralmechanisms involved in word reading in dyslexic children: A magnetic source imaging approach.Cerebral Cortex 10: 809–16.

Simos, P.G., et al. 2002. Dyslexia-specific brain activation profile becomes normal followingsuccessful remedial training. Neurology 58: 1203–13.

———. 2005. Early development of neurophysiological processes involved in normal reading andreading disability: A magnetic source imaging study. Neuropsychology 19, no. 6: 787–98.

Siok, W.T., C.A. Perfetti, Z. Jin, and L.H. Tan. 2004. Biological abnormality of impaired reading isconstrained by culture. Nature 43: 71–6.

Snowling, M.J. 2000. Dyslexia. Oxford: Blackwell.Stein, J., and V. Walsh. 1997. To see but not to read: The magnocellular theory of dyslexia. Trends in

Neuroscience 20: 147–52.Stuart, M. 2006. Teaching reading: Why start with systematic phonics teaching? Psychology of

Education Review 30, no. 2: 6–17.Stuart, M., and M. Coltheart. 1988. Does reading develop in a sequence of stages? Cognition 30: 139–

81.Szucs, D., and U. Goswami. 2007. Educational Neuroscience: Defining a new discipline for the study

of mental representations. Mind, Brain and Education 1, no. 3: 114–27.Tan, L.H., A.R. Laird, K. Li, and P.T. Fox. 2005. Neuroanatomical correlates of phonological

processing of Chinese characters and alphabetic words: A meta-analysis. Human Brain Mapping25, no. 1: 83–91.

Thomson, J.M., T. Baldeweg, and U. Goswami. 2005. ‘Developmental trajectories ofauditoryperception in dyslexia: an ERP study’. Poster presented at the 1st Course, International Schoolon Mind, Brain and Education, July, in Sicily, Italy.

Turkeltaub, P.E., L. Gareau, D.L. Flowers, T.A. Zeffiro, and G.F. Eden. 2003. Development ofneural mechanisms for reading. Nature Neuroscience 6, no. 6: 767–73.

Wimmer, H. 1996. The nonword reading deficit in developmental dyslexia: Evidence from childrenlearning to read German. Journal of Experimental Child Psychology 61: 80–90.

Ziegler, J.C. 2005. Do differences in brain activation challenge universal theories of dyslexia? Brainand Language 98: 341–3.

Ziegler, J.C., and U. Goswami. 2005. Reading acquisition, developmental dyslexia, and skilledreading across languages: a psycholinguistic grain size theory. Psychological Bulletin 131, no. 1:3–29.

———. 2006. Becoming literate in different languages: similar problems, different solutions.Developmental Science 9, no. 5: 429–53.

148 U. Goswami

Downloaded By: [UNAM] At: 00:52 27 May 2009

PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [UNAM]On: 27 May 2009Access details: Access Details: [subscription number 788841327]Publisher RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Educational ResearchPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713699076

Dyscalculia: neuroscience and educationLiane Kaufmann a

a Clinical Department of Pediatrics IV, Section Neuropediatrics, Innsbruck Medical University, Austria

Online Publication Date: 01 June 2008

To cite this Article Kaufmann, Liane(2008)'Dyscalculia: neuroscience and education',Educational Research,50:2,163 — 175

To link to this Article: DOI: 10.1080/00131880802082658

URL: http://dx.doi.org/10.1080/00131880802082658

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

Dyscalculia: neuroscience and education

Liane Kaufmann*

Clinical Department of Pediatrics IV, Section Neuropediatrics, Innsbruck Medical University, Austria

(Received 31 July 2007; final version received 8 November 2007)

Background: Developmental dyscalculia is a heterogeneous disorder with largelydissociable performance profiles. Though our current understanding of the neurofunc-tional foundations of (adult) numerical cognition has increased considerably during thepast two decades, there are still many unanswered questions regarding the develop-mental pathways of numerical cognition. Most studies on developmental dyscalculia arebased upon adult calculation models which may not provide an adequate theoreticalframework for understanding and investigating developing calculation systems.Furthermore, the applicability of neuroscience research to pedagogy has, so far, beenlimited.Purpose: After providing an overview of current conceptualisations of numericalcognition and developmental dyscalculia, the present paper (1) reviews recent researchfindings that are suggestive of a neurofunctional link between fingers (finger gnosis,finger-based counting and calculation) and number processing, and (2) takes the latterfindings as an example to discuss how neuroscience findings may impact on educationalunderstanding and classroom interventions.Sources of evidence: Finger-based number representations and finger-based calculationhave deep roots in human ontology and phylogeny. Recently, accumulating empiricalevidence supporting the hypothesis of a neurofunctional link between fingers andnumbers has emerged from both behavioural and brain imaging studies.Main argument: Preliminary but converging research supports the notion that fingergnosis and finger use seem to be related to calculation proficiency in elementary schoolchildren. Finger-based counting and calculation may facilitate the establishment ofmental number representations (possibly by fostering the mapping from concrete non-symbolic to abstract symbolic number magnitudes), which in turn seem to be thefoundations for successful arithmetic achievement.Conclusions: Based on the findings illustrated here, it is plausible to assume that fingeruse might be an important and complementary aid (to more traditional pedagogicalmethods) to establish mental number representations and/or to facilitate learning tocount and calculate. Clearly, future prospective studies are needed to investigatewhether the explicit use of fingers in early mathematics teaching might prove to bebeneficial for typically developing children and/or might support the mapping fromconcrete to abstract number representations in children with and without developmentaldyscalculia.

Keywords: dyscalculia; functional brain imaging; neuroscience; finger-based calculation;mental number representations

*Email: liane.kaufmann@i-med.ac.at

Educational Research

Vol. 50, No. 2, June 2008, 163–175

ISSN 0013-1881 print/ISSN 1469-5847 online

� 2008 NFER

DOI: 10.1080/00131880802082658

http://www.informaworld.com

Downloaded By: [UNAM] At: 00:55 27 May 2009

Introduction

Arithmetic learning disorders (developmental dyscalculia) denote circumscribed andoutstanding difficulties in the acquaintance of arithmetic skills. Importantly, dyscalculia isnot a unitary concept and the associated cognitive profiles might vary widely between andwithin individuals (for overviews see Kaufmann and Nuerk 2005; Wilson and Dehaene2007). With an estimated prevalence of 3% to 7%, developmental dyscalculia is about asfrequent as developmental dyslexia (APA 1994) and, similar to dyslexia, persists intoadulthood if untreated (Shalev and Gross-Tsur 2001). It is widely agreed that dyscalculia isa highly familial disorder (the risk for siblings of children suffering from dyscalculia is fiveto ten times higher than in the general population: Shalev et al. 2001). Thoughdevelopmental dyscalculia may present as a single-deficit disorder (e.g., core deficit of‘number sense’: Landerl, Bevan, and Butterworth 2004; for a review see Wilson andDehaene 2007), many affected children exhibit associated cognitive problems, both withinand outside the numerical domain (for respective overviews see Kaufmann and Nuerk2005; Shalev and Gross-Tsur 2001). The frequent occurrence of comorbidities coincideswith recent findings reporting a substantial genetic overlap between various developmentallearning disorders such as dyscalculia, dyslexia and attention-deficit hyperactivity disorder(Plomin, Kovas, and Haworth 2007). Nevertheless, dyscalculia research is much youngerthan dyslexia research and most of what is known is derived from adult studies involvingmature brain systems. Hence, our current understanding of the behavioural manifestationsand (neuro)cognitive foundations of developmental dyscalculia remains incomplete.

The main aim of this paper is to demonstrate the need to go beyond adultcalculation models when attempting to account for the peculiarities of developing brainsystems and developmental disorders such as developmental dyscalculia. Moreover, thepaper aims to illustrate how findings from brain imaging studies may inform educationalunderstanding and even classroom instructions/interventions. The example discussedhere concerns the association between fingers and numbers which has not beenconsidered explicitly in adult calculation models, but which nevertheless – as we shall see– plays a fundamental role in learning to count and calculate. Butterworth (1999) meritsreward for initiating the greatly renewed neuroscientific interest in the potentialimportance of finger use for the acquaintance of numerical skills. According toButterworth (1999), fingers are convenient and natural tokens to represent numbermagnitudes which are intuitively used by young children when learning the verbalcounting sequence and/or when executing first simple additions and subtractions (seealso Butterworth 2005). Moreover, even adults may use finger-based back-up strategies.Thus, it may be argued that finger-based number representations and finger-basedcalculation are deeply rooted in human ontology and phylogeny. Indeed, convergingevidence from brain imaging findings corroborate the latter notion of a neurofunctionallink between fingers and numbers (Andres, Seron, and Olivier 2007; Kaufmann et al.2008; Sato et al. 2007; Thompson et al. 2004). Before presenting the respective findingsand their potential educational implications in more detail, we present a brief overviewof current conceptualisations of typical and atypical developmental pathways ofnumerical cognition (as derived from neuropsychology and neuroscience).

Current conceptualisations of numerical cognition and developmental dyscalculia

Infants as young as 4–6 months are able to make number-based discriminations and evenexhibit additive expectation behaviour (up to set sizes of three objects: Wynn 1992, 1995).

164 L. Kaufmann

Downloaded By: [UNAM] At: 00:55 27 May 2009

Larger set size discriminations may be mastered, too, provided the ratio of the to-be-compared object sets is large enough (e.g., babies may discriminate 8 from 16, but not 8from 12 objects: Xu and Spelke 2000; and even 16 from 32 objects, but not 16 from 24: Xu,Spelke, and Goddard 2005). Interestingly, when continuous stimulus characteristics suchas contour length and total filled area are controlled, 6-month-old babies can successfullydiscriminate large, but not small, numerosities (Xu, Spelke, and Goddard 2005; see alsoXu 2003). Consequently, Xu (2003) and Xu, Spelke, and Goddard (2005) propose theexistence of two core systems for number magnitude representation in infants: onemediating small (exact) and the other supporting large (approximate) numericalmagnitudes (see also Feigenson, Dehaene, and Spelke 2004; for a similar distinction inmature brain systems see Dehaene and Cohen 1997).

In verbal individuals (i.e., children who have begun to master language) numericalconcepts emerge as soon as children use counting to refer to objects. Upon entering formaleducation, most typically developing children demonstrate a rudimentary understandingof number relations, are able to count up to 20 and may even master simple additions andsubtractions verbally when allowed to use their fingers or other reference objects. Thus,upon starting school (at age 6 in most European countries), most children already haveacquired some verbal counting and calculation skills, which are considered to be thefoundations for establishing more advanced school mathematics.

But what do we know about the development of these number processing andcalculation skills during infancy and preschool years? How do children develop quantityknowledge and number representations? Which neurocognitive processes and mechanismscome into play when children gradually acquire abstract (symbolic) number representa-tions during formal schooling? And which difficulties – within and outside the numericaldomain – accompany developmental dyscalculia? These latter questions are at the core ofcurrent research attempts aiming to delineate the developmental pathways of numericalcognition. However, dyscalculia research is relatively young and, in the absence of anempirically validated developmental calculation model, many neuropsychological studiestargeted at gaining a better understanding of the developmental pathways of typical and/or atypical numerical cognition rest on adult calculation models (e.g., Dehaene and Cohen1995; Dehaene et al. 2003; McCloskey, Caramazza, and Basili 1985).

According to popular adult calculation models, number processing and calculation ismulti-componential and the components constituting the calculation system are thoughtto be modularly organised (Dehaene and Cohen 1995; McCloskey, Caramazza, and Basili1985). The modularity assumption derives from adult cognitive neuropsychology whichconsiders double dissociations as evidence for a modular architecture of (neuro)cognitivesystems and/or mental representations (Shallice 1988). For example, a double dissociationis present if cognitive ability A is preserved while ability B is deficient in one individual,and in another individual the opposite pattern emerges (i.e., preserved ability B anddeficient ability A). Probably the most popular adult calculation model is the so-called‘triple code model’ of numerical cognition (Dehaene and Cohen 1995) postulating threemodularly organised but interrelated calculation components (i.e., analogue magnituderepresentation, auditory verbal word frame, visual Arabic number form), each of which isthought to be supported by distinct brain regions. A consistent finding in the adultliterature concerns the key role of the intraparietal sulcus (IPS) for number magnitudeprocessing (for an overview see Dehaene et al. 2003; Hubbard et al. 2005).

Up to now, numerous (mostly adult) studies have been published taking the Dehaenecalculation model (Dehaene and Cohen 1995, 1997) as a starting-point upon which to testtheir hypotheses (for respective reviews see Dehaene et al. 2003; Hubbard et al. 2005; see

Educational Research 165

Downloaded By: [UNAM] At: 00:55 27 May 2009

also for developmental issues Kaufmann and Nuerk 2005; Wilson and Dehaene 2007).However, because of crucial differences between developing and mature brain systems,adult models may not provide adequate theoretical frameworks for investigatingdevelopmental disorders (see Bishop 1997; Karmiloff-Smith 1992). For example, doubledissociations in developmental disorders need not necessarily reflect the presence ofmodularly organised neurofunctional networks (Karmiloff-Smith 1997; Pennington 2006).Indeed, double dissociations have been observed in non-modular cognitive architectures aswell (i.e., in the case of dyslexia the double dissociation between phonological and surfacedyslexia; Harm and Seidenberg 1999; Plaut 1995).

Behavioural studies suggest that number magnitude discrimination abilities may be aninnate capacity inherent to infants (Wynn 1992, 1995; Xu and Spelke 2000; Xu 2003) andeven non-human species (Brannon and Roitman 2003), yet the question arises whethernumber magnitude processing is supported by identical brain regions in infants and adults.Although present developmental findings are consistent with a neurofunctional linkbetween intraparietal regions and number magnitude processing, there is somecontroversy regarding the age-dependency of intraparietal (IPS) involvement in theformation of arithmetical skills. While some findings reveal similar activations inintraparietal regions extending across different ages during number magnitude processing(symbolic number processing: Temple and Posner 1998; non-symbolic number processing:Cantlon et al. 2006; Temple and Posner 1998), other studies suggest that the functionalspecialisation of the IPS for number magnitude processing increases with age (non-symbolic number processing: Ansari and Dhital 2006; symbolic number processing: Ansariet al. 2005; Kaufmann et al. 2005, 2006; Rivera et al. 2005). These conflicting results maypartly be explained by methodological differences between studies, making a directcomparison across studies (and paradigms) difficult. Alternatively, and as mentionedalready above, one may claim that adult models (resting on modularity assumptions) arenot apt to account for the complexity of developing brain systems.

Developmental dyscalculia: single- or multiple-deficit views?

Upon adopting Dehaene’s modularly organised adult calculation model (Dehaene andCohen 1995; Dehaene et al. 2003), some researchers propose that the neurocognitiveunderpinnings of developmental dyscalculia are best conceptualised as a (single) coredeficit of ‘number sense’ (e.g., Butterworth 2005; Landerl, Bevan, and Butterworth 2004).The core deficit hypothesis implies that children diagnosed with developmental dyscalculiadisplay specific difficulties to mentally represent and manipulate (non-symbolic) numbermagnitudes. Consequently, the core deficit of ‘number sense’ is thought to be related to amalfunctioning of intraparietal brain regions (i.e., the horizontal segment of theintraparietal sulcus (HIPS), which is supposed to mediate number magnitude processingaccording to Dehaene et al. 2003). In an excellent review, Wilson and Dehaene (2007)revisit this strong single-deficit view of developmental dyscalculia by arguing that the coredeficit of ‘number sense’ may be only one of several possibly underlying deficits. Accordingto Wilson and Dehaene (2007), other potential subtypes of dyscalculia – each of whichbeing supported by distinct brain regions – may rest on (1) deficient verbal symbolicrepresentations (manifesting themselves as arithmetic fact retrieval difficulties); (2)deficient executive functions (hampering fact retrieval as well as complex calculation);or (3) deficient spatial attention (leading to impaired quick recognition of smallnumerosities and possibly negatively affecting non-symbolic and symbolic numbermanipulations). Thus, Wilson and Dehaene (2007) propose that the behavioural

166 L. Kaufmann

Downloaded By: [UNAM] At: 00:55 27 May 2009

characteristics of developmental dyscalculia – and their neurocognitive underpinnings –might vary substantially between individuals, thus seriously questioning a strong single-deficit view.

Another argument challenging the single-deficit view of developmental dyscalculia isthe observation that many children exhibiting problems in learning to count and calculatealso have difficulties in other cognitive domains. Even within the arithmetical domain,children display quite distinguishable performance profiles (both at an intraindividual andinterindividual level of analysis: Dowker 2005). Consequently, various attempts to classifydevelopmental dyscalculia at a behavioural level have been undertaken (e.g., Geary 2000;Temple 1989, 1991; Von Aster 2000). As early as 1991, Temple reported a doubledissociation between arithmetic fact retrieval (e.g., number fact knowledge such as 3 6 5)and procedural knowledge (‘knowing how’ to solve a complex arithmetic problem) indevelopmental dyscalculia (however, see for a critical discussion of double dissociations indevelopmental disorders Pennington 2006). The distinction between arithmetic fact andprocedural knowledge was first acknowledged in the adult calculation model proposed byMcCloskey and colleagues (1985). Likewise, the theoretical foundations for Geary’s (2000)and Von Aster’s (2000) efforts to classify developmental dyscalculia were grossly based onthe Dehaene calculation model (Dehaene and Cohen 1995). A commonality of the latterclassification attempts is their effort to further differentiate developmental dyscalculiaaccording to specific performance profiles, which in turn imply the existence of distinctsingle cognitive deficits. And yet, according to Pennington (2006), any attempt to linkthese deficits to one – and only one – underlying neuroanatomical (and/or genetic)underpinning is likely to fail. Rather, developmental dyscalculia should be regarded as acomplex and dynamic developmental disorder (for similar views of dyslexia and attention-deficit hyperactivity disorder see Bishop 1997; Pennington 2006). Interestingly, andconsistent with the latter view, recent findings of quantitative genetic research report asubstantial genetic overlap between quite diverse cognitive (dis)abilities such as reading,language and arithmetic (Plomin and Kovas 2005; for a review see also Plomin et al. 2007).This genetic overlap may partly explain the repeatedly reported high incidence ofcomorbidity of developmental disorders such as dyslexia, dyscalculia and attentionaldisorders (for a review see Kaufmann and Nuerk 2005), and furthermore, may also partlyaccount for the considerable diversity of neurocognitive performance within onedevelopmental disorder (in our case, dyscalculia).

To summarise, although single-deficit models (e.g., core deficit of ‘number sense’:Butterworth 2005; Landerl, Bevan, and Butterworth et al. 2004; ‘number fact dyscalculia’:Temple 1991) are presently predominant in the neuroscientific literature – probablybecause they are simpler and hence more testable – multiple-deficit models ofdevelopmental dyscalculia seem to better fit our current understanding of the complexnature of developmental disorders.1 Thus, a change of paradigms from a modular andsingle-deficit view towards a dynamic, process-oriented and multiple-deficit view seems tobe essential for the development of empirically validated developmental calculation models,as well as for the production of mathematics curricula meeting children’s neurocognitivedevelopment (i.e., maturation-dependent readiness to grasp number-based concepts andskills) and the generation of tailored dyscalculia intervention programmes.

Neuroscience and education: the case of developmental dyscalculia

A frequent criticism of brain imaging studies involving learning is their restrictedapplicability to education and classroom interventions. Indeed, the great majority of

Educational Research 167

Downloaded By: [UNAM] At: 00:55 27 May 2009

neuroscientific research – including the realm of numerical cognition and/ordevelopmental dyscalculia – is targeted at basic research. As neuroscience is a ratheryoung discipline, which is further tightly connected to recent technological advances(and which underwent and still continues to undergo dramatic changes within veryshort periods), early respective studies are hardly comparable to more recent ones(regarding both methodological and practical issues). Further, it is important toacknowledge that significant activations reported in imaging studies reveal brainregions modulating a specific task which is not equivalent to regions being necessary toprocess the task at hand. In addition, and partly because of methodological andtechnical constraints, experimental paradigms generally focus on islets of skills ratherthan learning processes and mechanisms, the latter being a greater focus of interest foreducational researchers and classroom teachers. In general, imaging studies are only asgood as the behavioural paradigms they are implementing. Hence, the development ofadequate behavioural paradigms should be based on a sophisticated understanding ofthe interplay between neurocognitive (including genetic) and pedagogical factorsdetermining typical and atypical trajectories within particular cognitive domains (i.e., inour case, numerical cognition).

Recently, researchers of both disciplines (i.e., neuroscience and education) areslowly becoming aware of the urgent need to ameliorate communication and to fostercommon research efforts. The latter focus is reflected in continuously appearingscientific articles devoted to the topic of ‘neuroscience and education’ (Ansari and Coch2006; Fawcett and Nicolson 2007; Goswami 2004; Szucs 2005), as well as in the newlyfounded scientific journal Mind, Brain and Education, whose first issue was compiled in2007. Furthermore, Fawcett and Nicolson (2007) request the establishment of a newdiscipline of ‘pedagogical neuroscience’. These authors emphasise that diagnostic effortsbased on behavioural and/or cognitive symptoms are not sufficient to contribute to ourunderstanding of complex developmental disorders. Taking developmental dyslexia asan example, Fawcett and Nicolson (2007) stress the need to develop brain-basedtheories (by employing genetic and brain-based diagnostic methods), which eventuallymay advance not only our understanding of developmental disorders, but also lead totailored interventions. Hence, there is a clear need for research designs beingspecifically targeted at the educational implications of neuroscience research. In orderto accomplish the latter goal, educational experts must share their expertise inpedagogy, and neuroscience researchers must develop ecological paradigms that arecapable of investigating cognitive processes and learning mechanisms instead ofcircumscribed skills.

Below, I present a brain imaging study conducted at Innsbruck MedicalUniversity which was aimed at making a first step towards bridging the gap betweenneuroscience and education (Kaufmann et al. 2008). Although the main aim of ourstudy was to elucidate the link between non-symbolic numerical and spatialprocessing, here I will focus on the numerical task only and the potential implicationsfor educational sciences that arose from studying it. The numerical task requiredparticipants to make simple number comparisons. Stimuli were pictures of twohands, each hand showing a different finger pattern (e.g., the right hand raisingthree fingers, the left one two fingers). Thus, our experimental paradigm provokedfinger-based counting/number discriminations and these helped us address somequestions regarding the association between fingers and numbers. Before discussingthe results, I briefly present the relevant literature that led us to formulate our workinghypotheses.

168 L. Kaufmann

Downloaded By: [UNAM] At: 00:55 27 May 2009

Fingers and numbers

Empirical evidence for a link between fingers and numbers is derived from developmentalbehavioural studies (Fayol, Barrouillet, and Marinthe 1998; Gracia-Baffaluy and Noel2008; Landerl, Bevan, and Butterworth 2004; Noel 2005; Sato and Lalain 2008), patientstudies (Gerstmann syndrome: Gerstmann 1940; developmental Gerstmann syndrome:Benson and Geschwind 1970; Suresh and Sebastian 2000) and brain imaging studies(fMRI: Simon et al. 2002; Thompson et al. 2004; TMS: Andres, Seron, and Olivier 2007;Roux et al. 2003; Rusconi, Walsh, and Butterworth 2005; Sato et al. 2007).

Probably the earliest report of a neurofunctional association between fingers (i.e.,finger discrimination) and number processing was provided in 1940 by Gerstmann, whodescribed a patient with a right posterior parietal lesion accompanied by symptomscombining finger agnosia (difficulties to recognise and discriminate fingers), acalculia(outstanding calculation problems), right–left disorientation and agraphia (impairedwriting: see Benton 1997; for descriptions of developmental Gerstmann syndromes see,e.g., Benson and Geschwind 1970; Suresh and Sebastian 2000).2

Associations between fingers (finger gnosis) and calculation skills have also beenreported in developmental behavioural studies. For instance, in typically developingpreschool children, neuropsychological test scores (including finger recognition and fingerdiscrimination) were found to be a good predictor of calculation skills one year later(Fayol, Barrouillet, and Marinthe 1998). Furthermore, the findings of Noel (2005)revealed that finger gnosis seems to be a specific predictor for numerical abilities andfurther suggest that the link between finger gnosis and arithmetic is not restricted to tasksrelying on finger-based magnitude representations (but rather encompasses a wide rangeof number processing tasks). Noel (2005) argues that the latter findings are best explainedby the anatomical vicinity of brain regions supporting finger gnosis and those mediatingnumber magnitude processing and calculation.

Consistent with the latter suggestion, results of a functional magnetic resonanceimaging (fMRI) study (Simon et al. 2002) revealed neighbouring and partly overlappingactivations in posterior parietal brain regions for quite diverse abilities such as arithmeticand goal-directed hand movement (grasping/pointing), among others. In particular, brainregions supporting grasping (postcentral gyrus and anterior IPS) were found to borderthose mediating calculation (in and around the (H)IPS: see Simon et al. 2002, figure 2; fora review, see also Hubbard et al. 2005).

Further corroborating the notion of a neurofunctional link between finger use andnumber processing are the results of a repetitive transcranial magnetic resonance(rTMS) study revealing that both finger movements and number magnitude judgements(Arabic digits) are disrupted by left parietal stimulation in adults (i.e., angular gyrus:Rusconi, Walsh, and Butterworth 2005; see also Roux et al. 2003). Finally, two recentTMS studies assessing corticospinal excitability in hand muscles are suggestive of (1) aspecial role of right-hand muscles (left hemisphere) for small numerals (1–4, which wereinterpreted as reflecting culturally acquainted embodied finger counting strategies: Satoet al. 2007; for consistent behavioural findings see Sato and Lalain 2008); and (2) of a linkbetween hands (but not arms and/or legs) and enumeration (numbers and letters: Andres,Seron, and Olivier 2007). Interestingly, upon investigating 16- and 17-year-old adolescentswith and without a diagnosis of developmental dyscalculia (DD), Soltesz andcollaborators (2007) report that electrophysiological responses upon performing a simple,one-digit number comparison task were not comparable between the two groups (thoughboth groups displayed comparable behavioural performance on this task). In particular,

Educational Research 169

Downloaded By: [UNAM] At: 00:55 27 May 2009

and most interestingly, relative to their non-DD peers, individuals with DD displayed veryspecific neuropsychological performance profiles being characterised by preserved mentalrotation and body part knowledge (among others) but deficient performance on mentalfinger rotation and finger knowledge. Thus, the latter results provide the first evidencethat, in DD (or some groups of DD), deficient finger knowledge may be associated withatypical brain mechanisms for performing a basic numerical task.

The latter results are clearly exciting, but it has to be noted that all respective brainimaging studies were performed on adults and hence provide information about maturebrain systems only. Indeed, despite converging behavioural evidence for an associationbetween finger gnosis and numerical skills in children (e.g., Fayol, Barrouillet, andMarinthe 1998; Noel 2005; Sato et al. 2007), respective developmental brain imagingstudies are so far lacking. This gap in the research provided our motivation for conductinga developmental fMRI study that required 8-year-old children (and young adults) to makenumber magnitude judgements. More specifically, we presented stimuli that consisted ofpictures of two hands representing different numerosities (i.e., finger patterns).Participants were asked to indicate, by pressing a button, which hand displayed morefingers. Thus, by provoking finger-based comparison strategies, the experimentalparadigm required participants to make (non-symbolic) numerical classifications.

Besides number discriminations, participants were asked to make spatial and colourdiscriminations, too.3 As a thorough discussion of this research clearly goes beyond thescope of the present paper, results and educational implications presented here will focus onthe following questions: (1) do elementary school children and adults recruit identical brainregions upon solving a simple number comparison task (provoking finger-based numberrepresentations)?; and (2) is there a neurofunctional link between finger-based numberrepresentations and counting/number comparison, and if so, is there an age-related changein cerebral activation patterns related to finger-based magnitude extraction?

Results revealed highly interesting findings. Behaviourally, children and adultsperformed at ceiling upon making number classifications (99.3% correct). However,compared with adults, children were significantly slower (746 ms and 1017 msrespectively), although response latency patterns for different types of pairs were againcomparable between age groups. In particular, both age groups were significantly quickerto classify distant relative to adjacent number pairs (i.e., displaying shorter responselatencies upon comparing 1 versus 5 relative to 1 versus 2: children p 5 0.05; adultsp 5 0.001). The latter reaction time phenomenon has been coined the ‘distance effect’ andis thought to reflect the integrity of the mental number line (Dehaene 1991). Thus, thebehavioural data suggest that the task was performed flawlessly by both groups (asreflected by very high accuracy rates) and, moreover, children and adults alike processednumber magnitude all the way down to the semantic numerical level (as reflected by thepresence of the distance effect). However, a different picture emerged regarding brainactivation patterns. In particular, activation patterns in response to non-symbolic numberprocessing were clearly distinguishable between children and adults. Relative to adults,children recruited additional brain areas in lateral portions of anterior IPS, as well as inadjacent regions of the right post- and precentral gyrus upon making finger-based numbermagnitude classifications. Most interestingly, the latter regions were found to bedeactivated in adults. We interpret our findings as being suggestive of an age-dependentneurofunctional link between areas supporting finger use and non-symbolic numberprocessing (Kaufmann et al. 2008). Importantly, the latter findings imply that even in thecase of comparable behavioural performance between children and adults, brainactivation patterns need not be identical across age.

170 L. Kaufmann

Downloaded By: [UNAM] At: 00:55 27 May 2009

Potential educational implications of our findings

Our findings provide evidence for an age-dependent link between finger-use and numberprocessing that is not only interesting for neuroscience and numerical cognition research,but may also have significant implications for educational research and even classroomteaching. The demonstration of age-dependent activation differences when solving asimple number comparison task (which was performed at ceiling by both children andadults) emphasises the importance of going beyond behavioural (performance) issues. Ourfindings highlight the potential benefit of incorporating our steadily increasing under-standing of developing neurofunctional systems into efforts to design adequate and timelypedagogical curricula. In other words, although children may succeed in solving aparticular (numerical) task as well as adults in terms of performance, children may need toput more effort than adults into orchestrating the brain regions associated most closelywith the task. For instance, as illustrated above, a simple number comparison task requires8-year-old children – but less so adults – to recruit brain areas supporting finger use, thusrevealing age-dependent processing mechanisms at a neural level.

With respect to classroom teaching, the implications of the latter findings arestraightforward. As brain areas mediating finger use might be co-activated wheneverchildren need to access mental number representations, it does not seem advisable to forbidchildren using their fingers upon performing arithmetic problems. Rather, educators andteachers could take advantage of the fact that fingers may serve as concrete embodied tokensto represent number magnitude. Moreover, fingers mirror the base-10 number system, andmoreover, are readily available to be used as back-up strategies. Thus, it is plausible toexpect that the consistent use of fingers could positively affect the formation of mentalnumber representations (by facilitating the mapping from concrete non-symbolic quantityknowledge to abstract symbolic number processing) and thus also the acquisition ofcalculation skills. Indeed, preliminary evidence supporting the latter claim comes from arecent intervention study demonstrating that training finger gnosis significantly improvesarithmetic performance in 1st-graders (Gracia-Baffaluy and Noel 2008). Moreover, aprospective study of elementary school children revealed a predominance of split-fivecalculation errors (i.e., solutions deviating +5 from the correct result: Domahs, Krinzinger,and Willmes 2008). The latter authors interpreted their findings as reflecting ‘failure to keeptrack of ‘‘full hands’’ in counting or calculation’ (abstract: Domahs, Krinzinger, andWillmes 2008). Interestingly, with increasing age/schooling (i.e., grades 1 to 3) split-fiveerrors decreased, thus suggesting that children’s reliance on mental finger patterns (wholehand/five fingers) decreases with increasing schooling/calculation proficiency.

Finally, it is not far-fetched to argue that the explicit incorporation of finger-use innumeracy intervention programs could be beneficial for the establishment of mentalnumber representations in children suffering from developmental dyscalculia. However, inthe absence of respective empirical studies, the latter claim thus far remains speculative.

Last but not least, it is important to stress that finger knowledge is not the whole storyto becoming good at maths. Rather, several other skills like abstract thinking (e.g.,facilitating the mapping process between concrete and symbolic arithmetic), spatial skills(enabling the formation of a spatially oriented mental number line and, moreover, ourunderstanding of the base-10 system of the Arabic number system), working memory(enabling us to manipulate quantities when solving arithmetic tasks, to monitor multi-stepprocedures, etc.), language proficiency (underlying counting routines and arithmetic factretrieval, among others) are also considered to be important for becoming a proficientcalculator. Each of the latter domains is likely to mediate the acquisition and/orapplication of arithmetic skills and future – preferably longitudinal – research endeavours

Educational Research 171

Downloaded By: [UNAM] At: 00:55 27 May 2009

are needed to elucidate their impact on the development of numerical cognition (for areview see Kaufmann and Nuerk 2005; Wilson and Dehaene 2007).

To summarise, although the study described here nicely demonstrates the usefulness oftailoring research questions to pedagogical demands with respect to a core numerical skill,there is a clear need for future (prospective) research designs targeted at more complexlearning processes and mechanisms within the realm of numerical cognition. The mostsensible way to approach – and possibly achieve – the latter goal is to intensively fosterscientific communication and expertise between neuroscience and education.

Acknowledgement

The author was supported by the Austrian Science Foundation (grant number T286-B05).

Notes

1. A major disadvantage of ‘complex models’ is that they may render the verification and/orfalsification of working hypotheses difficult (because they typically encompass many dependentand/or unknown variables). Hence, researchers aiming to assess complex developmental disordersare especially required to (1) formulate very clear-cut hypotheses from the outset; (2) carefullydefine selection criteria for their study populations; and (3) employ paradigms that have beenfound previously to be adequate (and testable) for the research questions of interest. Reasons foradvocating ‘complex models’ are at least twofold: first, complex models readily acknowledgemodulating cognitive abilities (within and outside the numerical domain: Kaufmann and Nuerk2005; Wilson and Dehaene 2007); and second, complex models may lead to a better understandingof the link between mind (cognitive), brain (neurofunctional) and pedagogy (behavioural andeducational factors) mediating the acquaintance of number processing and calculation skills.

2. Though Benton (1997) seriously questioned the entity of the syndrome by stressing that asubstantial proportion of patients exhibit some, but not all four symptoms constituting the fullGerstmann syndrome, the Gerstmann syndrome has received increased interest recently.

3. Stimuli across the three tasks were identical (only instructions varying), thus enabling us tocontrol for domain-general perceptual and response-bound processing mechanisms. The strictcontrol of domain-general processing mechanisms is crucial in brain imaging research as the to-be-interpreted activation patterns should be attributable to task-relevant processing solely (or asfar as possible). The latter endeavour is achieved by a subtraction method whereby the cerebralactivation patterns obtained in response to a control task (which is preferably identical to theexperimental task in all but the variable of interest, in our case, number processing) aresubtracted from the activations obtained in response to the experimental task. According to thesubtraction logic, only the task-relevant – hence domain-specific – activations should remain. Inorder to achieve the best possible match between experimental and control tasks, we createdstimuli that could be used across all three task conditions. In particular, stimuli consisted of twosimultaneously displayed children’s hands with coloured thumbs. In half of the trials the palmsof the two hands showed in the same direction, while in the other half they did not. The spatialtask required participants to judge whether the palms of the two hands were showing in the samedirection or not. Likewise, in the colour condition, individuals were asked to state whether thecolours of the two thumbs were identical or not. The colour task served as a true control task.The spatial task was incorporated in the study because our main aim was to disentangle spatialand non-symbolic numerical processing (Walsh 2003; for a comprehensive review on theneurofunctional overlap between spatial and numerical processing see Hubbard et al. 2005).

References

American Psychiatric Association (APA). 1994. Diagnostic and statistical manual for mentaldisorders: DSM IV. Washington, DC: American Psychiatric Association.

Andres, M., X. Seron, and E. Olivier. 2007. Contribution of hand motor circuits to counting. Journalof Cognitive Neuroscience 19, no. 4: 563–76.

Ansari, D., and D. Coch. 2006. Bridges over troubled waters: Education and cognitive neuroscience.Trends in Cognitive Sciences 10, no. 4: 146–51.

172 L. Kaufmann

Downloaded By: [UNAM] At: 00:55 27 May 2009

Ansari, D., and B. Dhital. 2006. Age-related changes in the activation of the intraparietal sulcusduring non-symbolic magnitude processing: An event-related functional magnetic resonanceimaging study. Journal of Cognitive Neuroscience 1, no. 11: 1820–8.

Ansari, D., N. Garcia, E. Lucas, K. Hamon, and B. Dhital. 2005. Neural correlates of symbolicnumber processing in children and adults. NeuroReport 1, no. 16: 1769–73.

Benson, F., and N. Geschwind. 1970. Developmental Gerstmann syndrome. Neurology 20: 293–8.Benton, A.L. 1997. Reflections on the Gerstmann syndrome. Brain and Language 4: 45–62.Bishop, D.V. 1997. Cognitive neuropsychology and developmental disorders: Uncomfortable

bedfellows. Quarterly Journal of Experimental Psychology 5, no. 4: 899–923.Brannon, E.M., and J.D. Roitman. 2003. Nonverbal representation of time and number in animals

and human infants. In Functional and neural mechanisms of interval timing, ed. W.H. Meck, 143–82. New York: CRC Press.

Butterworth, B. 1999. The mathematical brain. London: Macmillan.———. 2005. The development of arithmetical abilities. Journal of Child Psychology and Psychiatry

46, no. 1: 3–18.Cantlon, J.F., E.M. Brannon, E.J. Carter, and K.A. Pelphrey. 2006. Functional imaging of

numerical processing in adults and 4-year-old children. PLoS Biology 4, no. 5: e125.Dehaene, S., and L. Cohen. 1995. Towards an anatomical and functional model of number

processing. Mathematical Cognition 1: 83–120.———. 1997. Cerebral pathways for calculation: Double dissociation between rote verbal and

quantitative knowledge of arithmetic. Cortex 33, no. 2: 219–50.Dehaene, S., G. Dehaene-Lambertz, and L. Cohen. 1998. Abstract representation of numbers in the

animal and human brain. Nature Neuroscience 21: 355–61.Dehaene, S., M. Piazza, P. Pinel, and L. Cohen. 2003. Three parietal circuits for number processing.

Cognitive Neuropsychology 20: 487–506.Domahs, F., H. Krinzinger, and K. Willmes. 2008. Mind the gap between both hands: Evidence for

internal finger-based number representations in children’s mental calculation. Cortex 44: 359–67.Dowker, A. 2005. Individual differences in arithmetic: Implications for psychology, neuroscience and

education. Hove: Psychology Press.Fawcett, A.J., and R.I. Nicolson. 2007. Dyslexia, learning, and pedagogical neuroscience.

Developmental Medicine and Child Neurology 49: 306–11.Fayol, M., P. Barrouillet, and C. Marinthe. 1998. Predicting arithmetical achievement from

neuropsychological performance: A longitudinal study. Cognition 68: B63–B70.Feigenson, L., S. Dehaene, and E. Spelke. 2004. Core systems of number. Trends in Cognitive

Sciences 8, no. 7: 307–14.Geary, D.C. 2000. From infancy to adulthood: The development of numerical abilities. European

Child and Adolescent Psychiatry 9, no. 2: 11–16.Gerstmann, J. 1940. Syndrome of finger agnosia, disorientation of right and left, agraphia and

dyscalculia. Archives of Neurology and Psychiatry 44: 298–408.Goswami, U. 2004. Neuroscience and education. British Journal of Educational Psychology 74: 1–14.Gracia-Baffaluy, M., and M.-P. Noel. 2008. Does finger training increase numerical performance?

Cortex 44: 368–75.Harm, M.W., and M.S. Seidenberg. 1999. Phonology, reading acquisition, and dyslexia: Insights

from connectionist models. Psychological Review 10, no. 3: 491–528.Hubbard, E.M., M. Piazza, P. Pinel, and S. Dehaene. 2005. Interactions between number and space

in parietal cortex. Nature Reviews Neuroscience 6: 435–48.Karmiloff-Smith, A. 1992. Beyond modularity: A developmental perspective in cognitive science.

Cambridge, MA: MIT Press/Bradford Books.———. 1997. Crucial differences between developmental cognitive neuroscience and adult

neuropsychology. Developmental Neuropsychology 13, no. 4: 513–24.Kaufmann, L., F. Koppelstaetter, M. Delazer, C. Siedentopf, P. Rhomberg, S. Golaszewski,

S. Felber, and A. Ischebeck. 2005. Neural correlates of distance and congruity effects in anumerical Stroop task: An event-related fMRI study. NeuroImage 25: 888–98.

Kaufmann, L., F. Koppelstaetter, C. Siedentopf, I. Haala, E. Haberlandt, L.-B. Zimmerhackl, S.Felber, and A. Ischebeck. 2006. Neural correlates of a number-size interference task in children.NeuroReport 1, no. 6: 587–91.

Kaufmann, L., and H.-C. Nuerk. 2005. Numerical development: Current issues and futureperspectives. Special issue: Brain and Number, Psychology Science 47, no. 1: 142–70.

Educational Research 173

Downloaded By: [UNAM] At: 00:55 27 May 2009

Kaufmann, L., S. Vogel, G. Wood, C. Kremser, M. Schocke, L.-B. Zimmerhackl, and J.W. Koten.2008. A developmental fMRI study of non-symbolic numerical and spatial processing. Cortex44: 376–85.

Landerl, K., A. Bevan, and B. Butterworth. 2004. Developmental dyscalculia and basic numericalcapacities: A study of 8–9-year-old students. Cognition 93, no. 2: 99–125.

McCloskey, M., A. Caramazza, and A. D. Basili. 1985. Cognitive mechanisms in number processingand calculation: Evidence from dyscalculia. Brain and Cognition 4: 171–96.

Noel, M.-P. 2005. Finger gnosia: A predictor of numerical abilities in children? ChildNeuropsychology 11: 413–30.

Pennington, B.F. 2006. From single- to multiple-deficit models of developmental disorders.Cognition 101, no. 2: 385–413.

Plaut, D. 1995. Double dissociations without modularity: Evidence from connectionist neuropsy-chology. Journal of Clinical and Experimental Neuropsycholgy 17: 291–321.

Plomin, R., and Y. Kovas. 2005. Generalist genes and learning disabilities. Psychological Bulletin131: 592–617.

Plomin, R., Y. Kovas, and C.M.A. Haworth. 2007. Generalist genes: Genetic links between brain,mind, and education. Mind, Brain, and Education 1, no. 1: 11–19.

Rivera, S.M., A.L. Reiss, M.A. Eckert, and V. Menon. 2005. Developmental changes in mentalarithmetic: Evidence for increased functional specialization in the left inferior parietal cortex.Cerebral Cortex 24, no. 1: 50–60.

Roux, F.E., S. Boetto, O. Sacko, F. Chollet, and M. Tremoulet. 2003. Writing, calculating, andfinger recognition in the region of the angular gyrus: A cortical stimulation study of Gerstmannsyndrome. Journal of Neurosurgery 99, no. 4: 716–27.

Rusconi, E., V. Walsh, and B. Butterworth. 2005. Dexterity with numbers: rTMS over leftangular gyrus disrupts finger gnosis and number processing.Neuropsychologia 43, no. 11: 1609–24.

Sato, M., L. Cattaneo, G. Rizzolatti, and V. Gallese. 2007. Numbers within our hands: Modulationof corticospinal excitability of hand muscles during numerical judgment. Journal of CognitiveNeuroscience 19, no. 4: 684–93.

Sato, M., and M. Lalain. 2008. On the relationship between handedness and hand-digit mapping infinger counting. Cortex 44: 393–99.

Shalev, R.S., and V. Gross-Tsur. 2001. Developmental dyscalculia. Pediatric Neurology 24, no. 5:337–42.

Shalev, R.S., O. Manor, B. Kerem, M. Ayali, N. Badichi, Y. Friedlander, and V. Gross-Tsur. 2001.Developmental dyscalculia is a familial learning disability. Journal of Learning Disabilities 34:59–65.

Shallice, T. 1988. From neuropsychology to mental structure. New York: Cambridge University Press.Simon, O., J.-F. Mangin, L. Cohen, D. Le Bihan, and S. Dehaene. 2002. Topographical layout of

hand, eye, calculation, and language-related areas in the human parietal lobe. Neuron 33:475–87.

Soltesz, F., D. Szucs, J. Dekany, A. Markus, and V. Csepe. 2007. A combined event-related potentialand neuropsychological investigation of developmental dyscalculia. Neuroscience Letters 417:181–6.

Suresh, P.A., and S. Sebastian. 2000. Developmental Gerstmann syndrome: A distinct clinical entityof learning disabilities. Pediatric Neurology 22: 267–78.

Szucs, D. 2005. Teachers can substantially inform cognitive psychological and cognitiveneuroscience research. Journal of the Professional Association of Teachers of Students withSpecific Learning Disabilities November: 4–7.

Temple, C.M. 1989. Digit dyslexia: A category-specific disorder in developmental dyscalculia.Cognitive Neuropsychology 6: 93–116.

———. 1991. Procedural dyscalculia and number fact dyscalculia: Double dissociation indevelopmental dyscalculia. Cognitive Neuropsychology 8, no. 2: 155–76.

Temple, E., and M. Posner. 1998. Brain mechanisms of quantity are similar in 5-year-olds andadults. Proceedings of the National Academy of Science USA 95: 7836–41.

Thompson, J.C., D.F. Abbott, K.J. Wheaton, A. Syngeniotis, and A. Puce. 2004. Digitrepresentation is more than just hand waving. Cognitive Brain Research 21: 412–17.

Von Aster, M. 2000. Developmental cognitive neuropsychology of number processing andcalculation: Varieties of developmental dyscalculia. European Child and Adolescent Psychiatry9, no. suppl. 2: 41–57.

174 L. Kaufmann

Downloaded By: [UNAM] At: 00:55 27 May 2009

Walsh, V. 2003. Cognitive neuroscience: Numerate neurons. Current Biology 13: 447–8.Wilson, A., and S. Dehaene. 2007. Number sense and developmental dyscalculia. In Human

behavior, learning, and the developing brain: Atypical development, ed. D. Coch, G. Dawson, andK. Fischer, 212–38. New York: Guilford Press.

Wynn, K. 1992. Addition and subtraction by human infants. Nature 358: 749–50.———. 1995. Origins of numerical knowledge. Mathematical Cognition 1: 35–60.Xu, F. 2003. Numerosity discrimination in infants: Evidence for two systems of representation.

Cognition 89, no. 1: B15–B25.Xu, F., and E.S. Spelke. 2000. Large number discrimination in 6-month-old infants. Cognition 74:

B1–B11.Xu, F., E.S. Spelke, and S. Goddard. 2005. Number sense in human infants. Developmental Science

8, no. 1: 88–101.

Educational Research 175

Downloaded By: [UNAM] At: 00:55 27 May 2009

Special Article

Relevance of Neuroscience to EffectiveEducation for Students With Readingand Other Learning Disabilities

Louisa Moats, EdD

ABSTRACT

New directions in educational assessment and instruction Eire supported by recent advances in the neurosciences. Among these are earlyidentification of potential learning problems through brief, efficient assessments of specific language skills that predict later readingoutcomes; early intervention that systematically targets critical linguistic processing skills; and the necessity of stimulating all functionsof a reading, writing, or computing brain. (J Child Neurol 2004;19:840-845).

Our understanding of language-based reading difficulties and theirremediation has grown considerably in the past three decades,although conceptual and practical challenges in the identification,classification, and treatment of such children have been numerous.''^Progress in understanding reading disabilities has resulted from acomprehensive. Federally-funded research program in basic andapplied sciences, including developmental psychology, cognitivepsychology, cognitive neuropsychology and functional neu-roimaging, pediatric neurology, genetics, and educational psy-chology.^ In pursuit of therapeutic models for reading disabilities,reseEu"chers have explored the mechanisms and processes of nor-mal reading development, discovering that the role of visual andtactile-kinesthetic (manual) pathways is secondary to the roles oflinguistic awareness and language proficiency.'' Even so, the func-tional and anatomic brain differences underlying reading disabil-ities are complex and involve multiple sensorimotor and cognitiveanomalies.^ To teach reading, then, is to teach language process-ing by ear, by voice, by eye, and by hand.

What we have learned about reading disabilities and effectivereading instruction is relevant to understanding a very large groupof children. About 4 in 5 of the 6% nationally who are classified aslearning disabled experience primary difficulty learning to read,write, and spell and can be diagnosed with language-related learn-ing disabilities." However, about 17% of all students exhibit the

Received April 26, 2004. Accepted for publication April 26, 2004.

From Sopris West Educational Services, Longmont, CO.

Address correspondence to Dr Louisa Moats, Sopris West EducationalServices, 411 Mother Lode Loop, Hailey, Idaho 83333. Tel: 208-788-0074;e-mail: louisam@sopriswest.com.

phonologic processing problems that put them at risk of reading fail-ure,^ and in high-poverty environments, many more are at risk. TheNational Assessment of Educational Progress reports that 38% ofall students nationally experience significant difficulty learning toread by fourth grade and are "below basic"; they do not read wellenough to participate in grade-level classroom instruction.

The vast majority of poor readers in our schools do not exhibitIQ achievement discrepancies on psychoeducational testing andthus would not meet the diagnostic criteria for learning disabili-ties and special education eligibility that have been in favor for thepast 30 years.^ Many of these children attend schools in which read-ing failure is the norm. Others do not demonstrate the comorbidbehavioral problems that lead to special education referrals. More-over, those who are classified do not have unique language or cog-nitive characteristics when they are compared with unclassifiedpoor readers,^ and children's response to instruction is predictedand accounted for by factors other than IQ, such as phonologic pro-cessing and speed of letter naming.'" Special education, as it is typ-ically managed and delivered today, leaves classified studentswithout significant gains or even specialized instruction" and usu-ally does little for the larger group of unclassified students who alsoneed research-based treatments.

Conceptions of reading disability, reading development, andreading instruction that are informed by the neurosciences couldmove us toward more effective treatment modalities. Such mod-els provide a rationale for early identification of all children at riskof reading failure, for the use of assessments designed to predictlong-term reading outcomes, and for instructional approaches andmethods that educate all relevant brain systems. Models consistentwith the findings of neuroscience can help educators reach almostall children who struggle to acquire basic academic skOls. Researchis far more extensive in early reading acquisition than in other areas

840

Neuroscience for Educating Students With Learning Disabilities I Moats 841

of educational performance, but other types of learning disabili-ties and academic problems can also be addressed by research-based interventions, including instruction in languagecomprehension, written expression, mathematics reasoning andcomputation, and adaptive behavior. This article offers a concep-tualization of informed treatments of the future, consistent withbrain science, that are possible with better teacher preparation,stronger administrative leadership, revised policies regarding stan-dards and the allocation of resources, and wiser use of technology.

TEACHING THE BRAIN TO READ

Preparing the Brain to ReadThe neural circuitry recruited for reading develops in interactionwith environmental stimulation. Studies of genetic influences onreading, for example, indicate that about half of the variance inachievement is attributable to biologic "wiring" or aptitudes andhalf is attributable to life experience.'^ The roots of literacy beginlong before entry into formal schooling with preschool develop-ment of language. Infants until the age of 6 to 8 months are capa-ble of processing all possible phonemes in all human languages,but before 10 months of age, they have become attuned to thephonologic systems and grammatic systems of their caregivers' lan-guage. Differentiation of phonemes m all language systems (eg, Chi-nese phonemes in English-exposed babies who have no experiencewith Chinese-speaking caregivers), which is possible for a 6-month-old child, is no longer observed in infants over age 10 months. ' Cir-cuitry for language is generated and fine-tuned, probably frombirth, to enable what should be the period of most rapid growthin vocabulary and syntax.

Normally developing children know the meanings of at least5000 base words before entering first grade." Children who residein low-verbal, less educated families, however, receive about onethird the hours of verbal stimulation, and the verbal interactionsto which they are exposed are more punitive, directive, and con-tent deprived than are those of more educated, middle class, or pro-fessional families.'^ Children who are at risk of reading failureoften demonstrate striking delays in vocabulary acquisition, inaddition to limited book exposure and other specific indicators ofrisk, such as a lack of familiarity with alphabet letters or the speechsounds they represent. To prepare preschool children for literacy,then, community-based early childhood programs could begin atabout 18 months when dendrites are proliferating in the left hemi-sphere and expressive language is emerging. The encouragementof parents and caregivers to use nurturing tones and articulatedphrases with toddlers while talking about shared experiences,even before the babies can articulate a response, is of criticalimportance because posterior cortical systems for language com-prehension are developing faster than anterior systems for verbalproduction.'"

Teachers and caretakers of toddlers can be taught conversa-tional skills and awareness of their use of language during mealtimes, play times, activities, and book reading." Teaching parentsand child care workers how to foster children's language devel-opment accelerates children's vocabulary acquisition and expres-sive language and increases their chances of reading success in firstgrade. Strategies such as the use of structured dialogue during

shared book reading, verbal extension of children's responses,modeling socially appropriate language, and narrating play and dailyactivities can significantly lower children's risk of reading dis-abilities. If children can be in the care of licensed preschool teach-ers who have learned to monitor and modulate their lcinguagepatterns to promote language development, the severity and impactof language processing difficulties can be substantially reduced.

Book distribution to families through pediatric clinics, libraryprograms, and other community contacts will continue to deservefunding from private and public sources. Children's familiaritywith alphabet letter names, exposure to the language and print con-ventions of books, and contact with worlds experienced onlythrough reading must be a goal of medical clinic, family literacy,and preschool programs.

Assessment to Locate Children at RiskAssessment procedures used by teachers from preschool onwardcan measure critical predictors of later reading achievement.'' Chil-dren at risk can be located before they have experienced emo-tionally damaging struggles with basic reading skills. Among thecritical indicators of later passage comprehension and reading fiu-ency that can be measured before children actually learn to readare knowledge of letter names, the ability to identify the phonemesor speech sounds in one-syllable words (chase = IchJIodlsf), the pro-duction of the speech sounds that graphemes (letters Eind letter com-binations that correspond to phonemes) represent, the ability tosound out simple nonsense syllables using phonic correspondences(fem, zis), and vocabulary knowledge. Moreover, such indicatorsare far more reliable and predictive when the measures are timedand the student's speed of response or fiuency is evaluated.'*

As reading is learned, the normally developing brain increas-ingly activates posterior areas involved in instant word recognition.'"Fluency, or speed of association of print with speech, is not onlya characteristic of proficient reading, but a lack of fiuency in com-ponent reading skills is one of the most enduring characteristicsof poor readers.^" In effect, the aim of early assessment is to deter-mine whether students are on course to fluent reading as a resultof typical classroom instruction or whether they need more inten-sive instruction that will normalize the brain's activation profile.^'Observations of the learning brain have verified the importance ofautomatizing the building blocks of reading, including speechsound identification and recall, letter recognition, alphabet recall,sound-symbol association, and instant word recognition.

In classrooms organized to prevent reading failure, childrenare screened three times per year from kindergarten onward andinterventions are begun immediately for those who fail to reach apredictive benchmark score on critical indicators of later readingsuccess. If children are "at risk," instruction can be designed to buildboth accuracy and fiuency in the weak skills, and the response toinstruction can be monitored with brief, focused assessments.Crucial to the efficacy of this endeavor, however, is the realizationthat reading is a multicomponent process subsumed by several func-tional brain networks, each recruited for a specific purpose: phono-logic processing, orthographic processing, morphologic andsemantic processing, and syntax and discourse processing. As thebrain learns to read, the component processors must be educatedto perform specific functions well so that smooth, automatic

842 Journal of Child Neurology I Volume 19, Number 10, October 2004

functioning of the reading brain is possible. Well-designed iessonswill include a number of components: explicit teaching about let-ters, speech sounds, phonics and spelling, vocabulary, and com-prehension, integrated into a coherent, systematic progression. ^

Educating the Phonologic ProcessorMost poor readers are slow and inaccurate at printed word recog-nition and are relatively weak on phonologic tasks, such as sayingthe individual phonemes in words, manipulating a phonemesequence by deleting and recombining sounds, or repeating a novelsequence of syllables.^^ Children will vary, however, in the extentto which they need remedial instruction in speech sound aware-ness. Although the majority of poor readers demonstrate a coredeficit in phonologic processing, some children are fast but inac-curate (phonologically impaired but not rate impaired), some areaccurate but slow (rate impaired), and some are inaccurate and slow(with a "double deficit").-'' All core classroom programs shouldinclude systematic and explicit instruction in speech sound aware-ness, ^ but those children who do not respond well and who requireremediation can vary in their need for explicit teaching of com-ponents such as speech sound awareness, fast word identification,vocabulary, or other reading comprehension skills. ^ Phonologic skillis necessary, but not sufficient, for proficient reading, and the pro-gram in use must allow the teaching emphasis to vary accordingto student need.

Phonologic awareness instruction should follow a develop-mental sequence of skill acquisition.^^ Methods that draw thechild's attention to the oral-motor formation of speech sounds areparticularly powerful. As the teacher models and instructs aboutthe mouth movements that characterize each speech sound, chil-dren with a "tin ear for language" can learn to distinguish similarspeech sounds such as fkj and /g/ and CEin then segment cind blendphonemes into whole words (/h/ + /ou/ + Is/). The teacher providessomething that technology cannot: modeling of the mouth formsand corrective feedback for the student who must focus on the fea-tures of the sounds. Once children learn to notice the internaldetails of the spoken word, they are more likely to be successfulat mapping print to speech.

To become fluent, the child must be helped to unitize perceptsof words and recognize them instantly, and this can be accomplishedeven with intermediate and older poor readers given instructionthat is sufficiently intense and skillful. * Intervention studies ver-ify that students with reading disabilities require about 30 minutesdaily in first grade and up to 2 hours daily in third grade andbeyond to normalize their functional brain systems for reading andthat conceptualization of the phoneme is the reference point forlearning an alphabetic orthography. Lessons begin with sounds, linksounds to symbols, and link words to meaning.

Neuroimaging studies show not only that children and adultswith reading disabilities are distinguished by underactivation of thetemporoparietal areas of the left hemisphere but also that youngerand older poor readers can activate right hemisphere networks thatare atypical of good readers.^" Thus, remediation for biologicallydyslexic students is more likely to result in compensatory adjust-ments in functional brain systems than to eradicate neurobiologicdifferences. In a well-designed treatment program, planning for stu-dents is long term. It includes direct remediation of phonologic sMUs

while simultaneously providing compensatory tools and strate-gies, such as proofreading assistance for poor spelling, time exten-sions for slow reading, books on tape, and academic studystrategies.

Educating the Orthographic ProcessorInformed instruction will enable students to recognize printedwords with fluency and accuracy, activating posterior corticalregions where orthographic processing takes place and where"sight word" images are stored. In the beginning stages of reading,children might know a good deal about letter sequences and printconventions through incidental exposure to them. Some childrendiscover sound-letter correspondence from just a few encounters;many others, however, are dependent on explicit instruction thatcalls their attention to letter forms and letter sequences. The goalof phonics and spelling instruction is to develop the child's explicitawareness phoneme-grapheme correspondence, which, in turn, sup-ports recognition and recall of whole words.™

Each individual lesson should juxtapose exercises aimed atstimulating the essential components of a unitized, functionalbrain system to support word recognition. Fast communicationamong multiple processing networks is more likely if each isaddressed separately and together. Partial approaches, for exam-ple, phonics instruction without direct cmd immediate applicationto reading and writing, have little justification within a neurosci-entific approach. Rather, a variety of skills are practiced briefly andsuccessively; children who are just learning to match sounds withsymbols might search a printed page for target words in a text; prac-tice naming upper and lower case letters with fluency; learn agrapheme for a speech sound and review known sound-symbolmatches; review previously learned sight words; practice blendingnew printed words using phonic correspondences; and read textwith learned words and patterns. Lesson design combines tar-geted work on component skills with the transfer and applicationof those skills in purposeful reading tasks.

Sound-symbol linkages, the meat of phonics instruction, arenot sufficient to educate a well-functioning orthographic processor.Proficient recognition of larger units is necessary for fluent decod-ing of longer words that are often found in literary and informationaltext. Lesson design at this level incorporates words from the Latinand Greek layers of English^' but focuses on syllabication and mor-phology because memory for these larger units facilitates pronun-ciation of novel words encountered in text. Morphemes are thesmallest meaningful parts of words and include inflectional endings(-S, -ed, -ing, -er, -est) and parts of compounds; Latin prefixes, suf-fixes, and roots; and Greek combining forms common in mathe-matical Emd scientific vocabulary. Morphemes are often spelledconsistently; thus, a direct link can be made between visual printpatterns and units of meaning (trans + port + ation; uni + eye + le).Reading lessons should aim to include comprehensive, sustainedinstruction about word structure, including phonologic, ortho-graphic, and morphologic correspondence units, throughout read-ing, writing, and vocabulary study from first grade to high school.

In summary, informed instruction about print will direct chil-dren's attention to the anatomy of words, both spoken and writ-ten. Mental cormections between subword units (sounds, syllables,and morphemes), whole words and their semantic networks.

Neuroscience for Educating Students With Learning Disabilities / Moals 843

syntax, and the structure of the text itself are established by com-bining skill practice with reading worthwhile material for well-defined purposes. The principles of redundancy within and acrosslessons and intensive practice to achieve automaticity reflect thebrain's need for repeated stimulation to establish communicativepathways. Because reading fluency enhances reading compre-hension and comprehension enhances reading fluency, lessonsare most effective when all components are addressed in ways thatpromote connections among multiple brain systems. Meaning-making, a complex enterprise, can then be facilitated by a purposefulteacher with the support of technology.

Path to ComprehensionEffective teaching of beginning reading engages attention, supportsmemory, and encourages self-regulation and self-correction. It isnot based on rote drill but rather encourages transfer of learningto reading for meaning through teacher guidance. The teacherleads the children into the insights necessary to comprehend; shedoes not assimie comprehension simply because the words arebeing read. Reading together, reading alternately, and readingsilently for specific purposes are the vehicles for active question-ing and discussion. The teacher's role is to preview the text withthe children, provide background knowledge, engage children inconnecting the topic to their own experience, and promote theactive use of reading comprehension strategies. Those strategiesinclude constructing mental images, predicting, summarizing, ques-tioning, and clarifying and are taught by the teacher's modeling or"thinking aloud" before, during, and after reading and by queriesduring shared reading. ^ Technology can enhance this process if textis presented electronically, and the reader can ask for clarificationand background information when words, phrases, or passages arenot understood.

A continual demand of comprehension is inference-making andthe mental construction of idea networks or mental models of theinformation presented. To this end, graphic organizers—visualmaps of conceptual relationships—can help children understandrelationships among ideas. The task of writing a summary orretelling a narrative requires children to form mental models of over-all text structure, but most children with language learning dis-abilities need considerable scaffolding or teacher support toapproximate this challenging task, along with specific steps tofollow. Skill is developed over time with extensive practice in rec-ognizing and formulating main ideas versus supporting details,ordering ideas in hierarchical networks, using graphic organizersand outlines, and recognizing various types of paragraphs. Throughcoaching and demonstration, teachers can learn to question atvarious levels, probing the meanings of words, phrases, sentences,and discourse that language-impaired students might misunderstandand modeling proven comprehension strategies as needed.

Technologic supports can assist teachers in matching chil-dren with text at the right difficulty level on topics of interest to thestudent. Computer programs now exist that parse any text intophrase-size units, so that the reader avoids word-by-word readingand can pick up the phrase contours in sentence structure. Otherinstructional packages provide videotaped background informationabout informational text in science or history to familiarize studentswith key concepts and terms before they tackle challenging passages.

TEACHING THE BRAIN TO WRITE

Language and writing difficulties can variously originate in grapho-motor, visual-spatial, attention, memory, and language formulationprocesses. Beginning writers must acquire automatic transcriptionskills to produce well-formulated compositions.^' Instruction thatapplies what we know about the brain will incorporate early train-ing in letter formation accuracy, phoneme-grapheme correspon-dences, spelling of high-frequency words, punctuation, andhandwriting fiuency so that attention and memory can be deployedin the service of language formulation, audience awareness, anddiscourse organization. Direct teaching of self-regulatory strategies**enhances writing length and quality in learning-disabled studentsin the intermediate and middle grades.

Parallel emphasis on direct teaching of component skillswithin a lesson that includes guided composition differs consid-erably from the currently dominant "writers' workshop" approachin regular classrooms, in which cumulative skill instruction is de-emphasized in favor of naturalistic, holistic, process-orientedwriting activities. It is true that writing instruction for all childrenshould be organized around real communicative goals estab-lished in social contexts so that skills are applied immediately inthe service of "real writing," but skills are a necessary foundationfor success. The skills part of a lesson, in which letter formation,letter naming, alphabet production, handwriting fluency, andspelling are emphasized, is brief (5 to 15 minutes, depending onstudents' ages). Frequent changes of activity and immediate, cor-rective feedback are necessary to engage short attention spansbecause the brain responds to novelty and to success. Afterworking on component skills in the first part of a lesson, youngchildren can be engaged in shared writing of a structured com-position, in which teachers model and discuss the processes ofplanning, translation of words into print, and review of what hasbeen written. Awareness of audience is fostered throughout writ-ing instruction through peer and teacher conferences and publi-cation of articles and books.

One reason for emphasizing component skills and habit for-mation in the beginning of the writing lesson is that practice ismore likely to generalize to composition that follows immediately.It is not productive, especially for children with learning dis-abilities, to give negative feedback on multiple errors of spelling,grammar, and sentence structure after a composition is com-pleted. Rather, good preparation for writing, supportive feed-back during writing, the use of reference materials such ashigh-frequency word lists, and grading that focuses on the con-tent and ideas are the best ways to avoid the negative emotionsand aversive reactions that many students develop when writingbecomes a punishing experience.

Other adaptations of instruction that help children with writ-ing disabilities include early instruction in keyboarding and wordprocessing, limiting handwriting instruction to either manuscriptor cursive, using voice-activated text generation devices on the com-puter, and planning with the help of interactive software thatprompts the writer to fonnulate key parts of the chosen genre. Atthe very least, a technology-based secretary can provide relieffrom the stress of transcription and rewriting and thus enhance moti-vation and erijoyment.

844 Journal of Child Neurology / Volume 19, Number 10, October 2004

TEACHING THE BRAIN TO COMPUTE

Learning disabilities in mathematics are more difficult to catego-rize and identify than disabilities in language, reading, and writing.Less research exists about what to teach, to whom, at what pointin development, and with what methods. The territory of mathe-matical reasoning differs from that of reading; it is more concep-tual than symbolic, although each academic domain encompassesconceptual and symbolic skills. To reason mathematically, thelearner develops both quantitative and logical mental models thatare distributed broadly in interconnected neural networks.

Many general principles important for addressing the atten-tion, memory, and linguistic processing problems of students withreading and writing difficulties apply to math instruction as well.'"Automaticity in lower-level skills frees up the mental "desktop" forhigher-level reasoning and problem solving. Thus, children need tolearn their math facts, graphomotor notation of numbers, and exe-cution of basic number operations through brief, frequent drills.Both mental arithmetic and paper-and-pencil production should bepracticed. The teacher must explain the concepts and proceduresusing manipulative aids and concrete representations when nec-essary and then must orchestrate the transfer and application ofskills to problem solving. Classroom discussion and collaborationduring problem solving is as important to mathematical learningas it is to reading and writing.

NEUROSCIENCES AND EDUCATION: WHAT WOULDBRAIN-BASED EDUCATION MEAN FOR STUDENTSWITH LEARNING DIFFERENCES?

If schools, classrooms, instructional programs, and teacher educationwere to be organized according to principles supported by neuro-science, assessments of global maturity, intelligence, and readi-ness would be replaced by assessments of proficiency in specificneurodevelopmental domains: gross motor; fine motor; reasoningand problem solving; visual-spatial and nonverbal cognition; languageprocessing at the levels of phonology, orthography, syntax, seman-tics, discourse, and pragmatics; attention; executive fimctions; andsocial-emotional development. Variations of development would notbe used to delay entry into school or to retain children who are devel-opmentally or academically delayed.'" Rather, children at risk of aca-demic failure would be identified early—at least as early as thebeginning of kindergarten, if not sooner—and given the benefit offocused intervention in small groups with similar instructionalneeds. Educators would not wait to see what "more time" wouldbring because organized instruction benefits children with devel-opmental delays more than maturation alone. Grade retentionwould go the way of other antiquated, unsupportable practices.Retention would be replaced by special instruction initiated for allchildren who scored below critical predictive benchmarks at thebeginning, middle, or end of an academic year.

Children in flexible, homogeneous intervention groups ofthree to six children would receive focused instruction eitherwithin the classroom or outside it. Focused intervention targets thespecific component skills of reading, writing, language, or math inwhich a student is delayed; uses validated instructional approaches;and gives ample practice so that students become accurate and

fluent in the application of component skills to academic endeav-ors. Additional practice occurs with the assistance of supportivetechnology, parent and paraprofessional tutoring, or peer-assistedlearning strategies. With progress monitoring assessments given ona frequent basis, teacher teams can decide who might need evenmore intensive, individual work. Interventions would be deliveredbefore expensive diagnostic testing and the development of an indi-vidual educational plan to maintain focus on the curriculum andthe child's response to instruction. When problems persist or theresponse to instruction continues to be poor, children would bereferred for a multidisciplinary evaluation to detennine the natureand extent of a handicapping condition.

The context for instruction would be schools organized toinvest resources preventively. The school schedule would permitteachers to work with whole groups, small groups, and individu-als for sufficient amounts of time to realize meaningful gains.Teachers would be expected by administrators to use validatedinstructional tools for specific purposes rather than to reinvent goodpractices on their own. Skillful implementation of proven practiceswould be the first expectation for teacher evaluation. Studentgrowth would be the primary consideration for all instructional deci-sion making rather than performance on or preparation for arbi-trary accountability measures that might have little relevance tostudents' instructional needs. Teachers would work with teams whoshared common understandings about neurodevelopmental vari-ations and the acquisition of reading, writing, language, and math-ematical competencies. Schools would be adequately staffed topermit reasonable class size, small-group and individual instruc-tion, and ancillary services as needed.

As this article has described, teaching is an enormously com-plex endeavor if it is done well. A skilled teacher who implementsbrain-compatible practices does many things, often simultane-ously and with groups of diverse individuals. The teacher seeksactively to develop all of the functional components of complexbehavioral and cognitive systems, separately and in combination.She or he plans instruction to include brief and varied routines tomaintain attention and motivation. The teacher plans for transferof skills to holistic application and ensures that sufficient practiceoccurs for long-term retention. The teacher often stands in forimmature executive functions in learners, providing organization,scaffolding, and feedback as needed. If we could substitute tech-nologic solutions for the human interactions required in education,society might be less concerned with the compelling issue ofteacher quality and teacher preparation.

Professional organizations such as the International DyslexiaAssociation, the American Federation of Teachers, and the Amer-ican Speech-Language-Hearing Association have called forchanges in the way teachers are taught to teach reading and theway speech-language therapists function to support literacydevelopment in schools. Teachers are not born with discipli-nary knowledge'^ any more than physicians are born with knowl-edge of anatomy, yet programs of teacher preparation do notnormally require teachers to learn the psychology of readingand language, the structure of language, the basics of brain-behavior relationships, or the implementation of research-basedpractices. When and if they do, children will be much morelikely to realize the benefits of neuroscientific research. To that

Neuroscience for Educating Students With Learning Disabilities / Moats 845

end, our most urgent challenge is how to attract capable indi-viduals into the profession of teaching, educate them for ademanding job, and support them while they are teaching. Per-haps neuroscientists can continue to inform the public not onlyabout our marvelous insights into the breiin at work but also aboutthe sustained human effort required to realize the benefits of thosediscoveries.

References1. Lyon GR, Fletcher JM, Shaywitz SE, et al: Rethinking learning dis-

abilities, in Finn CE, Rotherham RAJ, Hokanson C Jr (eds): Rethink-ing Special Education/or a New Century. Washington, DC, ThomasB. Fordham Foundation and Progressive Policy Institute, 2001,259-287.

2. Lyon GR, Moats LC: Critical conceptual and methodological consid-erations in reading intervention research. J Learn Disabil 1997;30:579-588.

3. Rayner K, Foorman BR, Perfetti CA, et al: How psychological scienceinforms the teaching of reading. Psyehol Sd 2001;2:31-74.

4. Beminger VW, Richards TL: Brain Literacy for Educators and Psy-chologists. Amsterdam, Academic Press, 2002.

5. Eden GF, ZefBro TA: Neural systems affected in developmental dyslexiarevealed by functional neuroimaging. Neuron 1998;21:279-282.

6. President's Commission on Excellence in Special Education: A NewEra: Revitalizing Special Education for Children and Their Fami-lies. Washington, DC: US Department of Education, Office of SpecialEducation and Rehabilitative Services, 2002, July 1. Available at:http://www.ed.gov/inits/commissions-boards.

7. Shaywitz SE: Overcoming Dyslexia. New York, Alfred Knopf, 2003.

8. Shaywitz SE, Escobar MD, Shaywitz BA, et al: Evidence that dyslexiamay represent the lower tail of a normal distribution of reading dis-ability. NEnglJMed 1992;326:145-150.

9. Hoskyn M, Swanson HL: Cognitive processing of low achievers andchildren with reading disabilities: A selective meta-analytic review ofthe published literature. School Psyehol Rev 2000;29:102-119.

10. Vellutino FR, Scanlon DM, Jaccard J: Toward distinguishing betweencognitive and experiential deficits as primary sources of difficulty inlearning to read: A two year follow-up of difficult to remediate and read-ily remediated poor readers, in Foorman B (ed): Preventing andRemediating Reading Difficulties: Bringing Science to Scale. Bal-timore, MD, York Press, 2003, 73-120.

11. Vaughn SR, Moody SW, Schumm JS: Broken promises: reading instruc-tion in the resource room. Except Child 1998;64:211-225.

12. Raskind W: Current understanding of the genetic basis of readingand spelling disability Learn Disabil Q 2001;24:141-157.

13. Kuhl P, Williams K, Lacerda F, et al: Linguistic experience alters pho-netic perception in infants by 6 months of age. Science 1992;255:606-608.

14. Biemiller A: Language and Reading Success. Brookline, MA, Brook-line Books, 1999.

15. Hart B, Risley T: Meaningful Differences in the Everyday Experienceof Young American Children. Baltimore, MD, Paul H. Brookes, 1995.

16. Eliot L: What's Going on in There?How the Brain and Mind Developin the Eirst Five Years of Life. New York, Bantam Books, 1999.

17. Lonigan CJ: Development and promotion of emergent literacy skillsin children at-risk of reading difficulties, in Foorman B (ed): Pre-venting and Remediating Reading Difficulties: Bringing Science toScale Baltimore, MD, York Press, 2003, 23-50.

18. Good RH, Simmons DC, Kameenui EJ: The importance and decision-making utility of a continuum of fluency-based indicators of founda-tional reading skUls for third-grade high-stakes outcomes. Sd Stud Read2001;5:257-288.

19. Simos PG, Breier JI, Fletcher JM, et al: Age-related changes in regionalbrain activation during phonological decoding and printed wordrecognition. Dev Neuropsyehol 2001; 19:191-210.

20. Torgesen JK, Rashotte CA, Alexander AW: Principles of fluency instruc-tion in reading: Relationships with established empirical outcomes,in Wolf MA (ed): Dyslexia, Fluency, and the Brain. Baltimore, MD,York Press, 2001, 333-355.

Simos PG, Fletcher JM, Foorman BR, et al: Dyslexia-specific brain acti-vation profile becomes normal following successful remedial training.

:1203-1213.

21.

22. Vellutino FR, Scanlon DM: The interactive strategies approach toreading intervention. Contemp Educ Psyehol 2002;27:573-635.

23. Torgesen J, Wagner R, Rashotte C: Longitudinal studies of phonolog-ical processing and reading. J Learn Disabil 1994;27:276-286.

24. Wolf ME, Bowers PG: The double-deficit hypothesis for the develop-mental dyslexias. J Educ Psyehol 1999;91:415^38.

25. National Reading Panel: Teaching Children to Read: An Evidence-Based Assessment of the Seientifie Research Literature on Readingand Its Implications For Reading Instruction. Washington, DC,National Institute of Child Health and Human Development, 2000.

26. Lovett MW, Steinbach KA, Frijters JC: Remediating the core deficitsof developmental reading disability: A double-deficit perspective.J Learn Disabil 2000;33:334-358.

27. Gillon GT: Phonological Awareness: From Research to Practice. NewYork, Guilford Press, 2004.

28. Torgesen JT, Alexander AW, Wagner RK, et al: Intensive remedialinstruction for children with severe reading disabilities: Immediate andlong-term outcomes from two instructional approaches. J Learn Dis-abil 2001;34:33-58.

29. Papanicolaou AC, Simos PG, Fletcher JM, et al: Early development andplasticity of neurophysiological processes involved in reading, inFoorman BF (ed): Preventing and Remediating Reading Difficulties:Bringing Science to Scale. Baltimore, MD, York Press, 2003, 3-21.

30. Ehri LC, Nunes SR, Stahl SA, Willows DM: Systematic phonics instruc-tion helps students learn to read: Evidence from the National Read-ing Panel's meta-analysis. Rev Educ Res 2001;71:393^47.

31. Henry M: Unlocking Literacy. Baltimore, MD, Paul H. Brookes, 2003.

32. Carlisle JF, Rice MSR: Improving Reading Comprehension: Research-Based Principles and Practices. Baltimore, MD, York Press, 2002.

33. Beminger V, Amtmann D: Preventing written expression disabilitiesthrough early and continuing assessment and intervention for hand-writing and/or spelling problems: Research into practice, in SwansonHL, Harris K, Graham S (eds): Handbook of Research on Learning Dis-abilities. New York, Guilford Press, 2004, in press.

34. Harris K, Graham S: Making the Writing Process Work: Strategies forComposition and Self-Regulation, 2nd ed. Newton, MA, BrooklineBooks, 1996.

35. Geary DC: Mathematics and learning disabilities. J Learn Disabil2004;37:4-15.

36. National Association of School Psychologists: Position Statementon Student Grade Retention and Social Promotion, Bethesda, MD,National Association of School Psychologists, 2003. http://www.naspon-line.org/information/posp aper_graderetent.html.

37. Moats LC, Foorman BF: Measuring teachers' content knowledge of lan-guage and reading. Ann Dyslexia 2003;53:23^5.