Developmental cognitive neuroscience: progress and ?· Developmental cognitive neuroscience: progress…
Post on 19-Aug-2018
Developmental cognitiveneuroscience: progress and potentialYuko Munakata1, B.J. Casey2 and Adele Diamond3
1Department of Psychology, 345 UCB, University of Colorado, Boulder, Boulder, CO 80309-0345, USA2Sackler Institute for Developmental Psychobiology, Weill Medical College of Cornell University, NY 10021, USA3Center for Developmental Cognitive Neuroscience, University of Massachusetts Medical School, Shriver Center Campus, Waltham,MA 02452, USA
Developmental cognitive neuroscience is an evolvingfield that investigates the relations between neural andcognitive development. Lying at the intersection ofdiverse disciplines, work in this area promises to shedlight on classic developmental questions, mechanismssubserving developmental change, diagnosis and treat-ment of developmental disorders, and cognitive andneuroscientific topics traditionally considered outsidethe domain of development. Fundamental questionsinclude: What are the interrelations between develop-mental changes in the brain (e.g. in connectivity, chem-istry, morphology) and developmental changes inchildrens behavior and cognitive abilities (e.g. rep-resentational complexity, ability to sustain selectiveattention, speed of processing)? Why, and how, is learn-ing enhanced during certain periods in development?How is our knowledge organized, and how does thischange with development? We discuss preliminaryinvestigations of such questions and directions forfuture work.
Developmental cognitive neuroscience is an evolving fieldthat investigates the relations between brain developmentand cognitive development. It has the potential to informunderstanding of the mechanisms that subserve percep-tion, attention, memory, language and other cognitiveprocesses at different points in the life cycle, andmechanisms subserving developmental changes in thoseprocesses. This field draws upon many others, such asdisciplines within psychology, neuroscience, cognitivescience, genetics and social science (see Box 1 for therelation between developmental cognitive neuroscienceand developmental neuroscience). Work in this area usesmethods from all of the related disciplines, includingbehavioral studies, neuroimaging, molecular genetics,computational modeling, single-cell recording and neuro-chemical assays. An emphasis is placed on complementarymethods evaluating multiple aspects or levels ofdevelopmental processes (from molecular to systemslevels), in typical and atypical development, in humansand other species.
Developmental cognitive neuroscience research canspeak to classic developmental issues, such as nature
versus nurture and continuity versus discontinuity indevelopment. For example, everyone now agrees that theanswer to the naturenurture debate includes some ofeach (see Box 2 for changes in terminology with advancesin the study of development). The real question is howgenetic and environmental factors interact during thecourse of development to shape the brain, mind andbehavior. Developmental cognitive neuroscience researchcan help to address this question. For example, suchresearch has demonstrated how specific genes can affectthe pruning and maturation of synapses, which in turnaffects the ability to learn from experience , and howexperience can affect which genes are turned on, when,and how they are expressed (e.g. ). Computationalmodels have demonstrated how small variations in initialprocessing, which might be genetically governed, can leadthrough experience to large differences in cognitive out-comes [3,4].
By investigating both typical and atypical development,developmental cognitive neuroscience research can informa variety of practical applications, such as earlierdiagnosis and more effective treatment of developmentaldisorders. For example, such research has informed the
Box 1. Developmental neuroscience
Whereas developmental cognitive neuroscience is squarely at theintersection of brain development and cognitive development, thefield of developmental neuroscience pertains to the physicaldevelopment of the brain, the cellular and molecular events thatunderlie nervous system development . The two fields also differin that developmental neuroscientists have traditionally focusedprimarily on embryonic and very early postnatal development.
Many developmental neuroscientists study mechanisms thatdirect the molding of the anatomical and functional organization ofbrain, such as axonal guidance, synapse formation and thedevelopment of neuronal connectivity. Intensive analyses areconducted, for example, of the cellular and molecular mechanisms(including signal transduction pathways) that stimulate, inhibit,regulate or perturb axonal and dendritic outgrowth, neuronal andglial migration, and establishment of ordered neural maps.
A central problem in developmental neuroscience is to identify thesignaling molecules (such as neurotrophic factors and cytokines)that control and arrest cell differentiation and migration. Anothercentral problem is the expression and function of regulatory genes(which control such aspects of brain development as neuronalmorphology and excitability, or the expression of neurotransmittersand their receptors) and regionalization of gene transcription.
Corresponding author: Yuko Munakata (firstname.lastname@example.org).
Review TRENDS in Cognitive Sciences Vol.8 No.3 March 2004
www.sciencedirect.com 1364-6613/$ - see front matter q 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2004.01.005
treatment for the genetic disorder, phenylketonuria(PKU). Individuals with PKU cannot convert phenyl-alanine to tyrosine, the precursor to the neurotransmitterdopamine. Treatment involves restricting dietary intake ofphenylalanine, thus allowing ingested tyrosine moreopportunity to compete with phenylalanine for transportinto the brain. Developmental cognitive neuroscienceresearch showed that the recommended diet had notrestricted phenylalanine intake enough and resulted indeficits in executive control functions, but that astricter diet could prevent and reverse those deficits[5,6]. Further, if the gross elevations in blood phenyl-alanine levels are not reduced before Postnatal Day 11,they impair the visual system, which develops rapidlyafter birth, and deficits in sensitivity to low contrastare still evident many years later , similar to thelong-lasting deficits seen after neonatal cataractscorrected in the first month .
As another example, schizophrenia is a devastatingneuropsychiatric disorder with an onset usually in youngadulthood. Animal research has indicated that neonatalperturbation of the hippocampus disrupts normal pre-frontal cortex development and its regulation bydopaminergic activity [9,10]. Behavioral effects are notseen immediately, but are evident years later in heigh-tened reactivity to stress and poorer executive control,which are characteristic of schizophrenia. Similar pertur-bation of the hippocampus in adulthood does not producesuch effects. Not only might this provide insight intoschizophrenia etiology, it also provides insight into thecomplex interactions that are important for normaldevelopment. This is but one example of how studying
atypical development can inform understanding of typicaldevelopment.
Developmental cognitive neuroscience work can alsoinform issues that have traditionally been consideredoutside the domain of development. For example, to whatextent is adult cognition subserved by domain-specificsystems that are intrinsically specialized for particularkinds of inputs (e.g. a face processing system ), asopposed to domain-general systems that become shapedfor particular kinds of inputs through learning ? Bothpossibilities are compatible with newborn preferences tolook towards face-like patterns, which probably reflectsubcortical processes rather than the cortical specializ-ations debated in adults . Developmental cognitiveneuroscience research could inform these debates  byevaluating whether domain-specific systems appear earlyin development, before much learning has occurred(supporting the intrinsically specialized view), or onlylater with the development of expertise (supporting thedomain-general view).
In the selective review that follows, we discuss severalspecific questions of investigation in the field and assesspreliminary answers (for additional discussion and cover-age of a broad range of topics in developmental cognitiveneuroscience, see ).
Neural changes during learningHow does brain organization and function change duringthe process of learning, and how does this compare withchanges observed across development? The increasingavailability of non-invasive tools, such as functionalmagnetic resonance imaging (fMRI, [21,22]), provides uswith the opportunity to ask such brain- and behavior-related questions in the developing human that was notpossible only a decade ago. With fMRI, we can safely trackchanges in cortical activation following extensive learningin the same individual, and we can compare such changeswith those observed in younger versus older children.
One of the first studies to track cortical changes over anextensive period of time with fMRI  showed rapidlearning effects in primary motor areas. Changes wereshown during motor sequence learning within a singlesession and increased over weeks of training. Corticalactivity became less diffuse and increased over time. Thisexample of initial diffuse cortical activity early in learning,followed by an increase in focal activity, parallels resultsfrom developmental fMRI studies. These studies showdiffuse activity in children relative to adolescents andadults, with adolescents showing the greatest focalactivity during performance of behavioral tasks, evenwhen performance across groups is equated .
Differences in brain activity between age groups are notdue to experience alone, as even without normal stimu-lation, changes in neuronal connections and synapticpruning occur with brain maturation . Rather, thesefindings highlight a possible approach for investigatingmaturational and experiential contributions. For example,to determine whether the immature brain after extendedpractice engages in the same neural processes as themature one, we could compare brain activity in the maturesystem with brain activity in the immature system both
Box 2. Terminology in the study of development
Early terminology focused on hard-and-fast contrasts in develop-ment (such as nature vs. nurture, genes vs. environment, andmaturation vs. learning) and hard-and-fast time windows (such ascritical periods):Learning: changes in response to experience with the environment(the nurture side of naturenurture, when this experience comesfrom a caretaker)Maturation: changes driven by genetic processes according tospecific timetables (the nature side of naturenurture)Critical periods: time-limited windows when specific experiencesmust occur to drive typical or maximal development. Learning isineffective outside these time windows.However, it became evident that such hard-and-fast distinctions aretoo simplistic, and this is acknowledged in later terminology (such asexperience-expectant and experience-dependent processes, whichincorporate genetic and environmental factors):Experience-expectant processes : processes that utilize environ-mental information that is highly reliable for all members of thespecies (e.g. for humans, hearing a language)Experience-dependent processes : processes that utilize environ-mental information that can vary across individuals (e.g. for humans,the particular language that is heard)Sensitive periods: time-limited windows when specific experienceshave their largest effects. Learning can still be effective outside thosetime windows.In addition, more neutral terminology refers to general aspects ofchange:Plasticity: the capacity for modificationDevelopment: processes of change across the lifespan
Review TRENDS in Cognitive Sciences Vol.8 No.3 March 2004 123
before and following extended experience. This use of fMRIto trace learning-related changes in cortical areas iscurrently being used to investigate the impact of beha-vioral and cognitive interventions on developmentaldisorders like dyslexia and obsessive-compulsive disorder.
Neuroanatomical changes also occur with learning anddevelopment, notably changes in the strength and numberof neuronal connections and the myelination of fibers.During early development, the neural connections in thebrain undergo dramatic organization, generating moreneuronal processes and connections than will ultimatelysurvive (e.g. [31,32]). Learning plays a key role in anactivity-mediated competition process through whichsome of these synapses are eliminated or pruned, andothers are stabilized and strengthened [1,33,34].
Developmental studies have challenged some acceptednotions about neural organization and learning. It haslong been known that monocular deprivation causeschanges in ocular dominance columns in primary visualcortex (V1), but it had been assumed that the effects ofvisual experience were passed along from the eye to thethalamus (the lateral geniculate) and from there to V1.However, recent work has shown that physiologicalchanges occur more rapidly in V1 than in the thalamus. Moreover, protein synthesis in V1 is necessary forrapid plasticity; anatomical changes in thalamocorticalafferents are not . This work suggests that corticalcircuitry is probably the substrate of the rapid plasticity inresponse to visual experience (or the lack thereof),whereas thalamocortical changes might then makethose changes hard to reverse.
Learning across developmentWhy is learning sometimes enhanced during certainperiods in development? That is, why do there appear tobe sensitive periods in development, during whichlearning is most effective? For example, the ability tolearn the grammar of a language declines with age [37,38].Sensitive periods for other kinds of learning fall within thefirst few years or months of life, for example, for thephonemes of ones mother tongue  and for certainaspects of face processing .
Computational models, specifically neural networksimulations, have been used to investigate potentialmechanisms for such sensitive periods (e.g. [41,42]).These models allow researchers complete control oversimulated learning systems and their environments, tohelp identify factors contributing to enhanced learningduring particular points in development.
One set of simulations  investigated the possibilitythat sensitive periods in language learning arise from theadvantages of starting small (see also ). Specifically,less-developed working memory abilities might facilitatelanguage learning, by restricting attention to keyelements of language input, highlighting the grammaticalstructure of the language. Simulations testing this ideademonstrated that limitations in working memo...