brain abnormal corticospinal function but normal axonal ... · normal corticospinal decussation and...

Brain (2001), 124, 2393–2406

Abnormal corticospinal function but normal axonalguidance in human L1CAM mutationsC. B. Dobson,1 F. Villagra,1 G. J. Clowry,1 M. Smith,1 S. Kenwrick,2 D. Donnai,3 S. Miller1 andJ. A. Eyre1

1Developmental Neuroscience, Department of Child Health, Correspondence to: Professor J. A. Eyre, Professor ofUniversity of Newcastle upon Tyne, 2Department of Paediatric Neuroscience, Department of Child Health,Medicine, University of Cambridge and 3Medical Genetics, The Royal Victoria Infirmary, Queen Victoria Road,St Mary’s Hospital, Manchester, UK Newcastle upon Tyne NE2 4LP, UK

E-mail: [email protected]

SummaryL1 cell adhesion molecule (L1CAM) gene mutations areassociated with X-linked ‘recessive’ neurologicalsyndromes characterized by spasticity of the legs. L1CAMknock-out mice show hypoplasia of the corticospinaltract and failure of corticospinal axonal decussation andprojection beyond the cervical spinal cord. The aim ofthis study was to determine if similar neuropathologyunderlies the spastic diplegia of males hemizygous forL1CAM mutations. Studies were performed on eightcarrier females and 10 hemizygous males. Transcranialmagnetic stimulation excited the corticospinal tract andresponses were recorded in biceps brachii and quadricepsfemoris. In contralateral biceps and quadriceps theresponses had high thresholds and delayed onset

Keywords: L1CAM; corticospinal; ipsilateral; man; TMS

Abbreviations: CAM � cell adhesion molecules; CMCD � central motor conduction delay; C-t � condition to test interval;EPSP � excitatory postsynaptic potential; GAP43 � growth associated protein 43 kilodaltons; IPSP � inhibitory post-synaptic potential; PMCD � peripheral motor conduction delay; TMCD � total motor conduction delay; TMS � transcranialmagnetic stimulation

IntroductionThe neural cell recognition molecule L1 cell adhesionmolecule (L1CAM), a member of the immunoglobulinsuperfamily of CAMs, is expressed on the surface of glial cellsand neurones in the developing nervous system (Lemmon andMcLoon, 1986; de la Rosa et al., 1990; Moscoso and Sanes,1995; Weiland et al., 1997). L1 protein is found at highlevels along major axonal pathways such as the corticospinaltract, corpus callosum and optic nerve (Hortsch, 2000). Ithas been implicated in a variety of processes important foraxonogenesis, including neuronal migration (Lindner et al.,1983), neurite growth and fasciculation (Stallcup and Beasley,1985; Lagenaur and Lemmon, 1987), myelination (Woodet al., 1990), axonal guidance (Castellani et al., 2000) andsynaptic plasticity (Luthl et al., 1994; Fields and Itoh, 1996).

© Oxford University Press 2001

compared with normal subjects. Ipsilateral responses inbiceps were smaller, with higher thresholds and delayedonsets relative to contralateral responses. Subthresholdcorticospinal conditioning of the stretch reflex of bicepsand quadriceps was abnormal in both hemizygous malesand carrier females suggesting there may also be areduced projection to inhibitory interneurones. Histo-logical examination of post-mortem material from a2-week-old male with an L1CAM mutation revealednormal corticospinal decussation and axonal projectionsto lumbar spinal segments. These data support a role forL1CAM in corticospinal tract development in hemizygousmales and ‘carrier’ females, but do not support a criticalrole for L1CAM in corticospinal axonal guidance.

The gene encoding L1CAM is located near the telomereof the long arm of the X chromosome (Xq28) (Djabali et al.,1990). In man, mutations in the L1CAM gene are responsiblefor a clinically variable, X-linked disorder, describedvariously as X-linked hydrocephalus, MASA (mentalretardation, adducted thumbs, spastic paraparesis, agenesisof the corpus callosum) syndrome, X-linked agenesis of thecorpus callosum or spastic paraplegia type 1 (Rosenthal et al.,1992; Jouet et al., 1994, 1995; Vits et al., 1994). The mostconsistent features in affected males are lower limb spasticity,mental retardation, hydrocephalus and flexion deformity ofthe thumbs (Kenwrick et al., 1996, 2000; Fransen et al.,1997), and it is now recognized that together these syndromesrepresent phenotypic variability of L1CAM mutations.

2394 C. B. Dobson et al.

Perhaps the most striking pathological observation ishypoplasia or apparent absence of two long axonal tracts,the corticospinal tract and corpus callosum (Chow et al.,1985; Yamasaki et al., 1995; Graf et al., 2000).

The L1CAM protein in mice is very similar in sequenceto human L1CAM protein and two transgenic mouse modelshave been generated that have a targeted null mutation inthe L1CAM gene (Cohen et al., 1997; Dahme et al., 1997).These mutations produce a phenotype similar to that observedin man. As in the human condition, the central nervoussystem of the mice was largely intact with major malformationlimited to hypoplasia of the corticospinal tract, corpuscallosum and cerebellar vermis and a degree of ventricularenlargement (Cohen et al., 1997; Dahme et al., 1997; Fransenet al., 1998; Demyanenko et al., 1999). The mutant micehad a reduced size of the medullary pyramids (Dahme et al.,1997) consistent with an overall reduction in the corticospinalprojection, and no corticospinal axons were detected in thespinal cord caudal to the cervical level (Cohen et al., 1997).Cohen and colleagues also found that a substantial proportionof corticospinal axons failed to cross the midline at themedullary decussation and instead passed ipsilaterally intothe dorsal column (Cohen et al., 1997). No abnormalities ofdecussation were found in the corpus callosum, optic chiasmor spinal commissural projection of L1CAM knock-out mice.This led to the suggestion that the ventral midline of thepyramidal decussation expresses a specific guidance cue,which functions as a ligand for L1CAM, facilitating crossingof corticospinal axons over the midline. Thus, L1CAM washypothesized to have a highly specific effect on the guidanceof axons of the corticospinal tract at the pyramidaldecussation.

The aim of this study was to investigate if there wasabnormality of corticospinal decussation in human subjectswith L1CAM mutations, and to determine whether failure ofprojection of the corticospinal tract to the lumbar enlargementunderlay the spastic paraparesis observed in hemizygousmales. Females with known L1CAM mutations were alsostudied to determine if L1CAM mutation was an X-linkedrecessive disorder or whether ‘carriers’ expressedabnormalities in corticospinal tract neurophysiology andfunction.

Material and methodsThe Joint Ethical Committee of Newcastle upon TyneUniversity and Health Authority gave ethical approval.Informed written consent was obtained from the subjects andtheir parents, where appropriate.

Neurophysiological studiesSubjectsEight mother–son pairs, two with two affected sons, wererecruited. All subjects had either class two or class three

mutations (Yamasaki et al., 1997). Class two mutationscomprise missense point mutations in the extracellular domainof L1CAM. Class three mutations include nonsense orframeshift mutations that produce a premature stop codon inthe extracellular domain. Mutant L1CAM molecules in Class3 do not remain integrated in the plasma membrane and thusL1CAM would not be expressed. The clinical details of thesubjects are summarized in Table 1.

Normative data for the threshold for evoking ipsilateraland contralateral responses following trancranial magneticstimulation (TMS) and their central motor conduction delays(CMCDs) to biceps brachii were obtained from previousstudies of normal subjects (Conway, 1995; Eyre et al., 2000,2001b). Normal values for cortical conditioning of the phasicstretch reflex in biceps were obtained for contralateralconditioning from the study by Eyre and colleagues (Eyreet al., 2000), and for ipsilateral conditioning from 25 normalsubjects, 11 males and 14 females, aged from 8 to 30 years.Normative data for the threshold for evoking responses andtheir CMCDs to quadriceps femoris were obtained from 100normal subjects: 65 children, 32 males and 33 females, agedfrom birth to 16 years; and 35 adults, 20 males and 15females, aged from 18 to 55 years. Normal values for corticalconditioning of the phasic stretch reflex in quadriceps wereobtained from 20 normal subjects, 10 males and 10 femalesaged from 12 to 30 years.

Surface electromyographyThe EMGs of left biceps and quadriceps were recorded usingminiaturized, skin-mounted differential amplifiers. A –3 dBbandpass of 5–1500 Hz was applied and the signals weresampled at 5000 Hz for computer analysis. For each musclethe EMG was also rectified and integrated (time constant1 s) and displayed as a horizontal line on an oscilloscope.The maximum voluntary contraction of each muscle wasrecorded as the greatest level of rectified, integrated EMGobtained by the subject in three attempts separated by 1-minintervals. During testing the subject contracted the muscle toraise the EMG display to meet a target line set at 10%maximum voluntary contraction.

Positioning of subjects and generation ofbackground muscle activitySubjects were seated with their head in the midline.Contraction of biceps was obtained with the arm adducted,the elbow held in 90° of flexion from full extension, and thehand supinated. Straps placed over the wrist resisted elbowflexion. Contraction of quadriceps was obtained with the legsadducted and the knee flexed in 90° from full extension.Straps placed over the ankle provided resistance to kneeextension.

Excitation of motor pathways projecting to bicepsCorticospinal tracts. TMS (MagStim Company Ltd,Whitland, UK), using a figure-of-eight coil with each circle

Corticospinal function in L1CAM mutation 2395

Table 1 Clinical details of subjects

having a diameter of 55 mm (SPC-ENG 8618; MagStimCompany Ltd), was used to excite corticospinal neurones(Winter et al., 1989; Edgley et al., 1990, 1997; Eyre et al.,2000, 2001a) projecting to left biceps α-motoneurones. Theleft (ipsilateral) and right (contralateral) motor cortices werestimulated. The induced current flow was maintained inposterior to anterior direction, approximately perpendicularto the central sulcus, by orientating the handle of thestimulating coil posteriorly and parallel to the parasagittalaxis. To ensure accurate and reproducible coil placement, themethod in a study by Metman and colleagues was used(Metman et al., 1993). The site with the lowest threshold forevoking responses in contracting left biceps was determinedfor the right (contralateral) and left (ipsilateral) motor cortex.To define the threshold for evoking responses, the intensityof TMS was increased until responses were evoked in thecontracting target muscle in 50% of trials. Twenty stimuli at1.2 times their respective thresholds were delivered andresponses recorded. The shortest onset latency of responses

following TMS of the right (contralateral) and left (ipsilateral)motor cortex was determined to estimate the total motorconduction delay (TMCD) from each cerebral hemisphere(Fig. 1A and B).

Peripheral motor nerves. Magnetic stimulation using a50 mm diameter circular coil (S/N ENG 014, MagStimCompany Ltd) placed over the spines of C5–8 vertebrae wasused to excite cervical spinal motor roots. The longest onsetlatency of at least 20 responses in biceps was determined toestimate the peripheral motor conduction delay (PMCD;Fig. 1D).

Excitation of motor pathways projecting toquadricepsCorticospinal tracts. To excite corticospinal neuronesprojecting to left quadriceps α-motoneurones, a ‘butterfly


Fig. 1 TMS. (A–F) Surface EMG of biceps or quadriceps. (A, B and D) Responses in biceps followingTMS of the (A) right (contralateral) cortex, (B) the left (ipsilateral) cortex and (D) over the cervicalspine. (C) Response in quadriceps following TMS over the vertex. (E and F) Phasic stretch reflexes in(E) biceps and (F) quadriceps. (G–J) CMCD and thresholds of responses following TMS of (G) thecortex, (H) biceps and (I and J) quadriceps. In G and H, filled symbols and continuous lines representdata from TMS of the right (contralateral) cortex, and open symbols and dashed lines from TMS of theleft (ipsilateral) cortex. Threshold is measured as the % of maximum TMS stimulator output. Eachsymbol represents a family (i.e. mutation) as indicated in Table 1. The lines represent themean � 2 SD for normal subjects.

coil’ (loop diameter 14 cm) was used and the centre of thecoil was placed over the vertex of the skull (Jalinous, 1991;Eyre and Miller, 1992; Terao and Ugawa, 1994). Twenty

stimuli at 1.2 � threshold were delivered and responsesrecorded. The shortest onset latency of responses wasdetermined to estimate the TMCD (Fig. 1C).


Peripheral motor nerves. A significant proportion of thepath of motor nerves projecting to quadriceps lies within thecauda equina, and hence magnetic excitation of these nervesat their point of emergence through the intervertebral foramina(L2–4) will provide an underestimate of the PMCD (Eyreand Miller, 1992). The PMCD to quadriceps was thereforeestimated from the onset of its homonymous phasic stretchreflex (Fig. 1F). The shortest onset latency of the quadricepsstretch reflex was determined (see below). One millisecondwas subtracted for trans-synaptic delays, and the peripheralconduction delay was scaled according to the relative meanconduction velocity for age in motor and Group Ia afferentnerves (Vecchierini-Blineau and Guiheneuc, 1979; Eyre andMiller, 1992).

Central motor conduction delaySubtraction of PMCD from TMCD estimated the CMCD forresponses in biceps and quadriceps (Fig. 1G and I).

Corticospinal conditioning of biceps andquadriceps phasic stretch reflexThe studies were only carried out in subjects aged 8 yearsand older, since we have previously shown that the patternof response in biceps does not achieve adult values until8 years (Conway, 1995). These studies could also only beundertaken in subjects who had demonstrated responses toTMS, since an estimate of CMCD was required.

Homonymous phasic stretch reflexA hand-held electromechanical tapper (model 201; Ling-Altec Ltd, Royston, UK) delivered small taps to the tendonof biceps or quadriceps. The tapper was held at right anglesto the tendon. A force transducer (type ELF-500 miniatureload cell; Entran, Watford, UK), was mounted in series withthe stylus of the tapper to encode the force profile of eachtap delivered to the muscle tendon. The static force withwhich the tapper pressed on the muscle tendon was kept at~2 N, indicated on a meter visible to the experimenter. Thepeak force achieved in each tap was fed to the computer foroff-line analysis. The peak force of the tap was increaseduntil a phasic stretch reflex was evoked in 50% of trials,defined as threshold (Fig. 1E and F). Using this techniquethe mean amplitude and onset latency of 20 phasic stretchreflexes in biceps evoked at 1.2 � threshold have beenshown to be consistent within a subject, both when recordedconsecutively within the same session and when recorded atdifferent experimental sessions (Miller et al., 2001). Inaddition the threshold is highly reproducible. Whenconsecutive estimates of threshold were made in 20 adults,the mean difference between repeat measurements madewithin the same experimental session was 0 � 0.22 N

(mean � standard error of the mean), and between differentexperimental sessions was 0 � 0.25 N (O’Sullivan, 1991).

Corticospinal conditioning of the phasic stretchreflex (Eyre et al., 2000)CMCD estimated the delay from TMS to the onset of thecorticospinal excitatory post-synaptic potential (EPSP) atbiceps or quadriceps α-motoneurones. The delay from thetap to arrival of the Group Ia EPSP was estimated bysubtracting PMCD from the shortest onset latency of thestretch reflex. TMS at 0.8 � threshold (conditioning stimulus,C), tap to biceps or quadriceps at 1.2 � threshold (teststimulus, t) and the two stimuli together at a defined conditionto test interval (C-t interval) were delivered in sequence.Trials, comprising 20 repeats of the sequence, were conductedover a range of C-t intervals such that the corticospinal EPSPwas estimated to arrive at biceps α-motoneurones from 2 msbefore (C-t interval –2 ms) to 8 ms after (C-t interval �8 ms)the estimated arrival of the Group Ia EPSP (Fig. 2). Thephasic stretch reflex in left biceps was conditioned bysubthreshold TMS of the left (ipsilateral) and right(contralateral) cortex, and that of left quadriceps wasconditioned by TMS over the vertex.

Functional studiesHand and leg dominance were assessed using the criteria ofPorac and Coren (Porac and Coren, 1977). Manual dexteritywas assessed using the nine-hole pegboard and the Purduepegboard (Sharpless, 1976; Mathiowetz and Weber, 1985).Normal control data were obtained from Mathiowetz andWeber (1985) for the nine-hole pegboard and from Tiffinand Asher for the Purdue pegboard (Tiffin and Asher, 1948).A nine-pad footboard was used to assess dexterity of leg andfoot movements (Dobson, 1999). The board comprised a3 � 3 array of 1.8 cm diameter pads, with 10 cm betweenthe centres of each pad. The subjects were timed whilst theycompleted depressing each of the pads, returning betweeneach to press a ‘home’ pad. The dominant leg was testedfirst and one practice run was allowed for each leg. The testwas repeated for each foot, first using the area over the firstmetatarsal (the ball of the foot) to press the pads, and thenthe first toe. Normal data were obtained from 100 subjects(45 males and 65 females) aged from 3 to 45 years.

Neuroanatomical studySubjectA male child hemizygous for a class 3 L1CAM mutation,who died aged 2 weeks. The genetics and clinical details ofthis child have been published (Jouet et al., 1995) (Case 9in Table 1).


Fig. 2 Time course and magnitude of subthreshold corticalconditioning of biceps stretch reflex. (A–C) EMG of biceps ina normal subject and male hemizygous for L1CAM mutation,(A) subthreshold TMS, (B) suprathreshold tap to biceps and(C) the two stimuli together at defined interstimulus (C-t)intervals. (D–F) The symbols represent the mean, and verticallines represent the 95% confidence limits of the amplitude ofthe conditioned stretch reflex expressed as a percentage of thetest stretch reflex (% test). The dashed line at 100% indicatesconditioned reflexes of the same size as the test. Values above100% indicate conditioned reflexes greater than the test(facilitation of the test reflex), values below 100% indicateconditioned responses less than the test (depression of the testreflex). Filled circles � males hemizygous for L1CAMmutation; open circles � female carriers; stars � normalsubjects. Time base C-t interval 0 � the interstimulus intervalof expected spatial convergence of a corticospinal and GroupIa volley at biceps α-motoneurones. Negative C-t intervals �Group Ia before corticospinal volley; positive C-t intervals �corticospinal before Group Ia volley. (D) Left (ipsilateral)cortical conditioning of biceps stretch reflex. (E) Right(contralateral) cortical conditioning of biceps stretchreflex. (F) Cortical conditioning of quadriceps stretchreflex.

MethodsFormalin-fixed and paraffin-embedded post-mortem samplesof the medulla and lumbar spinal cord were provided byDr M. V. Squier, Department of Neuropathology, RadcliffeInfirmary, Oxford, UK. Sections from each sample wereeither immunostained for neurofilaments, using a polyclonalantiserum to all neurofilament sub-units (Affiniti ResearchProducts, Manhead, Exeter, UK) diluted 1 : 1000 or, forgrowth associated protein 43 kilodaltons (GAP43), using amonoclonal antiserum clone GAP-7B10 (Sigma) diluted1 : 1000. Standard immunocytochemical procedures werefollowed. Appropriate biotinylated secondary antibodies andblocking sera, and streptavidin horseradish peroxidase wereobtained from Vector Laboratories (Newcastle upon Tyne,UK). Finally, sections were reacted with diaminobenzidineto visualize the immunostaining.

ResultsNeurophysiological studiesStudies were performed on all subjects with the exceptionof the son of case family 3 who refused to take part in theneurophysiological studies but completed the functional tests.

Responses in left biceps evoked by TMS of theright (contralateral) motor cortexL1CAM mutation hemizygous males. Responses wereevoked in all subjects following stimulation of the right(contralateral) motor cortex (Fig. 1A). The CMCDs wereeither abnormally prolonged or at the upper limit of thenormal range in all subjects aged 13 years or less, but forthe older four subjects CMCDs were within the normal range(Fig. 1G). Although the thresholds were within the normalrange, they were abnormally distributed towards high values(Fig. 1H), giving a mean standard deviate score of 0.95 with95% confidence limits of �0.22 to �1.66.

L1CAM mutation carrier females. Responses wereevoked in all subjects following TMS of the right(contralateral) motor cortex who had thresholds and CMCDswithin the normal adult range (Fig. 1G and H).

Responses evoked in left biceps by TMS of theleft (ipsilateral) motor cortexL1CAM mutation hemizygous males. Responsesfollowing TMS of the left (ipsilateral) motor cortex wereobtained in only four of the nine subjects tested. The CMCDswere within the normal range for two subjects and abnormallyprolonged in two (Fig. 1G). The thresholds of these responseslay within the normal range for age (Fig. 1H).


L1CAM mutation carrier females. Responses followingTMS of the ipsilateral (left) motor cortex were obtained inonly one subject with a CMCD and threshold within thenormal range (Fig. 1G and H).

Responses evoked in quadriceps by TMS over thevertex of the skullL1CAM mutation hemizygous males. Responses wereobtained in eight out of nine subjects tested. The CMCDsand thresholds for these subjects were either greater than orat the upper limit of the normal range for age (Fig. 1I andJ). Responses could not be obtained in Case 5 aged 10 years.

L1CAM mutation carrier females. Responses wereobtained in all subjects. While thresholds and CMCDs laypredominantly within the normal adult range, the CMCDsand thresholds were abnormally distributed towards the upperlimits of normal (standard deviate score: thresholds, mean�1.13, 95% confidence limits �0.25 to �1.99; CMCD,mean �0.67, 95% confidence limits –0.20 to �1.52 (Fig. 1Iand J).

Corticospinal conditioning of biceps andquadriceps stretch reflexes (Fig. 2)In normal subjects, hemizygous males and carrier femaleswith L1CAM mutations, subthreshold TMS of the ipsilateraland contralateral cortex for biceps, or the vertex forquadriceps, led initially to facilitation of the stretch reflexesin biceps and in quadriceps at C–t intervals appropriate forcoincidence of corticospinal and Group Ia afferent EPSPs atthe α-motoneuronal pools (Fig. 2).

Conditioning of left biceps stretch reflex bysubthreshold TMS of the right (contralateral) motorcortex. In normal subjects, the initial facilitation peakedwithin 1 ms of its onset. It was followed by significantdepression of the biceps stretch reflex that reached a maximumof 3 ms after the onset of facilitation (Fig. 2C and E). Forboth hemizygous males and carrier females of the L1CAMmutation, the peak of the facilitation occurred 1 ms laterthan for normal subjects and it was significantly greater thannormal for hemizygous males (P � 0.05). The mean durationof the facilitation was also prolonged (hemizygousmales �6 ms, carrier females 4 ms) and there was no periodof depression of the test reflex (Fig. 2C and E).

Conditioning of left biceps stretch reflex bysubthreshold TMS of the left (ipsilateral) motorcortex. The period of facilitation was brief in both normalsubjects and those with L1CAM mutations, so that theconditioned reflex amplitude had returned to baseline by2 ms for all subject groups (Fig. 2C). The peak facilitation

for males hemizygous for L1CAM mutations was significantlyless than for normal subjects (Fig. 2D; P � 0.05).

Conditioning of quadriceps stretch reflex bysubthreshold TMS of the vertex. In normal subjects,cortical facilitation of the quadriceps stretch reflex peakedwithin 1 ms of its onset and had a mean duration of 3 ms(Fig. 2F). For both hemizygous males and carrier females ofL1CAM mutation the peak of the facilitation occurred 1 mslater than for normal subjects and it was significantly greaterthan normal for hemizygous males (P � 0.05). The durationof the initial facilitation was prolonged (hemizygous males6 ms, carrier females 4 ms) and there was no period ofdepression of the test reflex.

Functional tests of dexterityTo allow comparison between the tests of dexterity, theindividual scores were converted to standard deviate scores,based on the mean and standard deviation score for normalsubjects (Fig. 3).

Nine-hole pegboard (Fig. 3A)All carrier females and hemizygous males were able tocomplete the test. Carrier females had normal test scores,while those for hemizygous males were significantly abnormalfor both their dominant (P � 0.01) and non dominant hands(P � 0.01).

Purdue pegboard (Fig. 3B)Males hemizygous for L1CAM mutations were unable tocomplete the Purdue pegboard tests. Carrier females were ableto complete the tests. Although their scores fell predominantlywithin the normal range, the scores were abnormallydistributed towards lower dexterity, such that the 95%confidence limits for mean standard deviate scores for eachcomponent of the test did not include zero.

The nine-pad footboard (Fig. 3C)Although the hemizygous males were able to complete thetest using the ball of their foot to depress the footboard pads,their scores were significantly abnormal (dominant footP � 0.001, non-dominant foot P � 0.01). They were unableto complete the test using their first toe to depress thefootboard pads. Female carriers completed both componentsof the nine-pad footboard test. While their individual scoreslay predominantly within the normal range, the scores wereskewed towards the lower limits of normality for dexterity,since the 95% confidence limits for their mean standarddeviate scores approached but did not include zero.


Fig. 3 Tests of dexterity: (A) nine-hole pegboard test; (B) Purdue pegboard test; and (C) nine-padfootboard test. The symbols represent the mean, and the vertical lines represent the 95% confidencelimits of the standard deviate scores (Z-scores). Filled circles � males hemizygous for L1CAMmutation; open circles � female carriers; dominant � dominant hand or foot; non-dominant: non-dominant hand or foot; both � score when both hands are used together; assembly � score forbuilding ‘assemblies’ of pins, washers and collars; NC � test not completed.

Neuroanatomical studyAnti-neurofilament immunocytochemistry gave clearlabelling of corticospinal fibres, which were visible in themedulla and could be seen crossing over the midline in bothdirections at the normal point of decussation (Fig. 4A andB). Far fewer corticospinal axons were GAP43 positive atthis level (Fig. 4C). In the lumbar spinal cord the lateralcorticospinal tract was identifiable and contained a mixtureof neurofilament-positive and GAP43-positive axons (Fig.4D and E).

DiscussionFunctional corticospinal projections to cervicaland lumbar spinal segments are present inmales hemizygous for the L1CAM mutationTMS has been shown in man and subhuman primates toexcite corticospinal tract axons, both directly and trans-synaptically at the level of the cortex (Edgley et al., 1990,1997; Burke et al., 2000). In hemizygous males, the responsesevoked in biceps and in quadriceps following TMS of themotor cortex establish that functional corticospinalprojections extending to the lumbar segments of the spinalcord can be present in subjects with L1CAM mutations. Thisconclusion is supported by histological evidence of clearlyvisible lateral corticospinal tracts in the lumbar spinal cordof Subject 9 (Fig. 4D and E). All but one of the reports ofabsent medullary pyramids in association with severeX-linked hydrocephalus have been made at post-mortem inchildren who died aged �3 months, or in foetuses who eitherdied or were terminated in utero (Chow et al., 1985; Yamasakiet al., 1995; Chow, 1996; Graf et al., 2000). Absence ofcorticospinal projections may therefore occur in cases with

the most extreme phenotypes associated with early death.Another possible explanation is marked delay in thedevelopment of the corticospinal projection, so that at thetime of death insufficient numbers of corticospinal axonshave projected to the midbrain to form clearly identifiablepyramids in severely affected foetuses and infants. Suchdevelopmental delay would be consistent with the slowfasciculation observed in neurones where there is a class 2or 3 mutation of the L1CAM gene (Stallcup and Beasley,1985; Lagenaur and Lemmon, 1987). The present study alsoprovides support for delayed corticospinal tract development,since the CMCDs of responses evoked in biceps andquadriceps were most likely to be abnormal in the youngestsubjects (Fig. 1G and I).

The majority of corticospinal axons decussate inhemizygous malesIn primates, including man, the majority (~90%) ofcorticospinal axons decussate in the medulla. The minority(~10%) descend into the ipsilateral spinal cord, mainly inthe dorsolateral column, but also in the ventromedial column(Nathan et al., 1990; Galea and Darian-Smith, 1994; Darian-Smith et al., 1996; Armand et al., 1997). Reflecting thispredominance of crossed projections, TMS of the motorcortex normally evokes responses in contralateral musclessignificantly more frequently than in ipsilateral muscles,and these responses have significantly lower thresholds andshorter CMCDs than responses evoked in ipsilateral muscles(Wassermann et al., 1994; Eyre et al., 2001b; Fig. 1). Cohenand colleagues have proposed that the ventral midline of thepyramidal decussation expresses a specific guidance cue,which functions as a ligand for L1CAM, permitting crossing


Fig. 4 Immunostaining of paraffin sections from the medulla and lumbar spinal cord of Subject 9 (seeTable 1). In the medulla (A) NF staining is prominent in axon tracts on the ventral surface of the brainand in fascicles of fibres projecting dorsolaterally and contralaterally from both sides of the brain(arrow). B shows a close up of this pyramidal decussation. Substantial numbers of fibres can be seencrossing over. Very few fibres are GAP43 positive, although the nearby neuropil is stronglyimmunopositive (C). D shows NF immunostaining in the lumbar spinal cord. All white matter tracts,along with motoneurone cell bodies (m) and fascicles of sensory axons in the dorsal horn are stronglystained. The lateral corticospinal tract (arrow) is clearly identifiable and also contains a proportion ofGAP43 positive axons (E, arrow). GAP43 immunoreactivity is also present in the superficial dorsalhorn and in the neuropil surrounding motoneurone cell bodies (m). Scale bars 500 µm in A; 200 µm inB–E.

of corticospinal axons over the midline (Cohen et al., 1997).If this hypothesis is correct, then males hemizygous for class2 or class 3 L1CAM mutations would be expected to have agreatly reduced decussation, i.e. the majority of fibres wouldnot cross the midline but would have an ipsilateral projection

to the spinal cord. In these circumstances, TMS of the cortexwould evoke responses more readily in muscles ipsilateralto the motor cortex than in contralateral muscles. In all ofthe hemizygous males tested we found evidence to thecontrary. Following TMS of the motor cortex, responses in


ipsilateral biceps occurred less frequently, were smaller(Fig. 1A and B) and had very much higher thresholds andlonger CMCDs than responses evoked in contralateral biceps(Fig. 1G and H). Indeed, the neurophysiological data of thepresent study imply a reduced rather than increased ipsilateralcorticospinal projection in the presence of L1CAM mutations.In normal subjects, ipsilateral responses in biceps can beobtained in all subjects under 12 years of age (Mulleret al., 1997; Eyre et al., 2001b), and in 65–100% of adults(Wassermann et al., 1994; Eyre et al., 2001b). In the presentstudy, however, ipsilateral responses were obtained in onlyfour out of nine (44%) males hemizygous for L1CAMmutations and in only one out of eight (13%) carrier females.While subthreshold TMS of the cortex led to facilitation ofthe phasic stretch reflex in ipsilateral biceps in hemizygousmales, the peak facilitation was significantly less than thatobserved in normal adult subjects, consistent with a reducedcorticospinal excitatory input to the ipsilateral motoneuronalpool. The possibility that the motoneuronal pool is lessresponsive to corticospinal input in subjects with L1CAMmutations is excluded by the significantly increasedfacilitation of the biceps stretch reflex observed followingsubthreshold TMS of the contralateral cortex.

The conclusion that the majority of corticospinal axonsdecussated is supported by the neuroanatomical findings inSubject 9. These revealed apparently normal decussation ofcorticospinal axons at the medullary pyramid with no evid-ence of the formation of substantial ipsilateral corticospinaltracts (Fig. 4A and B). Subject 9 had a nonsense mutationaffecting the Ig domain 6 and was thus unlikely to expressL1CAM at the cell membrane.

Evidence for abnormal corticospinal function inmales hemizygous for the L1CAM mutationThe responses elicited in contralateral biceps in hemizygousmales had CMCDs within the normal range, except for thoseaged 13 years or less, in whom CMCDs were abnormallyprolonged. The thresholds were abnormally skewed to lie inthe upper limits of normal. Responses evoked in quadricepshad CMCDs at or above the upper limit of normal andabnormally high thresholds. In one subject, responses couldbe evoked in biceps following stimulation of the contralateralcortex, but not following ipsilateral TMS, nor could responsesbe obtained in the quadriceps. These data indicate patho-physiology of the corticospinal system, which affects theprojection to the lumbar spinal segments more than that tothe cervical segments. In light of the hypoplasia of thecorticospinal tract, which has been demonstrated bothhistologically and radiologically in human subjects and inboth of the knock-out L1CAM mouse models, the data ofthe present study would be consistent with a reduced butfunctional corticomotoneuronal projection to both the cervicaland lumbar spinal segments (Chow et al., 1985; Yamasakiet al., 1995; Cohen et al., 1997; Dahme et al., 1997; Fransenet al., 1998; Graf et al., 2000).

TMS is most likely to excite axons within the cortex ratherthan cortical neurones (Di Lazzaro et al., 1998; Matthews,1999). There is increasing evidence that members of theL1CAM sub-family play a role in the clustering of voltage-gated sodium channels at the initial segment of the axon andat the nodes of Ranvier (Bennett and Lambert, 1999). Voltage-gated sodium channels in these domains are essential for theefficient initiation and propagation of action potentials, andfailure to appropriately locate these ion channels would alsolead to a higher threshold for activation of corticospinalaxons and slower saltatory conduction, thus increasingconduction delays.

In hemizygous males, cortical conditioning with TMS ledto both an increased peak and a prolonged duration offacilitation of both biceps and quadriceps phasic stretchreflexes. These findings at first appear incompatible with thehigh threshold and prolonged CMCDs of responses evokedby TMS alone, and previous morphological findings ofhypoplasic corticospinal tracts (Chow et al., 1985; Yamasakiet al., 1995; Graf et al., 2000) in subjects with L1CAMmutations. In the macaque monkey, cortically evokedcomposite EPSPs recorded in α-motoneurones have durationsof up to 15 ms, unless followed by an inhibitory post-synapticpotential (IPSP) mediated by corticospinal projections toGroup Ia inhibitory interneurones (Jankowska et al., 1975,1976). The very brief durations of cortical facilitation ofbiceps and quadriceps stretch reflexes in normal subjects andthe significant cortical depression of biceps stretch reflex,beginning within 1 ms of the onset of facilitation (Fig. 2F–I), are consistent with the arrival of an IPSP 1–2 ms afterthe onset of the EPSP (Jankowska et al., 1975, 1976; Eyreet al., 2000). To support this conclusion, subthreshold TMShas been demonstrated to evoke inhibition of bicepsα-motoneurones in man via monosynaptic corticospinalprojections to Group Ia inhibitory interneurones (Eyre et al.,2000). The increased peak and prolonged duration of thecortical facilitation of the phasic stretch reflex observedin hemizygous males may have arisen because a reducedcorticospinal projection to the Group Ia inhibitoryinterneurones resulted in a significantly smaller IPSPfollowing TMS. This conclusion is supported by the absenceof significant depression of biceps phasic stretch reflex inthe hemizygous males. However, since L1CAM is expressedby sensory nerves during development, it is also possiblethat there are abnormalities of muscle sensory afferentprojections, leading to changes in the rise time of the GroupIa EPSP at the α-motoneuronal pool (Honig and Rutishauser,1996). Finally, a further possibility is that of abnormaldevelopment of spinal inhibitory interneurones and theirprojections to α-motoneurones. To discriminate betweenthese possibilities, further studies will be required.

Impaired dexterity of both upper limb and lowerlimb function in hemizygous malesHemizygous males showed impaired manual dexterity aswell as abnormal dexterity of lower limb function (Fig. 3).


A degree of cognitive deficit will have contributed to theirpoor performance; however, skilled movements of the handand feet in primates have been shown to rely upon theintegrity of the corticospinal system (Kuypers, 1962, 1981;Lawrence and Hopkins, 1976; Bortoff and Strick, 1993). Theimpairment of dexterity in hemizygous males is similar tothat documented in the affected limbs of patients with spastichemiplegia who do not have a cognitive deficit (Tinnion,1999), and it is consistent with the demonstrated abnormalitiesof corticospinal neurophysiology.

Impaired corticospinal tract function in femalecarriers of L1CAM mutationsFemales with L1CAM mutations are described as beingasymptomatic and thus the syndromes associated withL1CAM mutations are thought to be X-linked recessivedisorders (Kenwrick et al., 1996; Fransen et al., 1997).However, the present neurophysiological studies revealabnormalities of corticospinal tract function, which are similarin pattern but less severe in nature to those observed inhemizygous males. Subtle abnormalities were alsodemonstrated in dexterity of both upper and lower limbfunction, which would be unlikely to be manifest in clinicalexamination, or to significantly impede the skill ofperformance of activities of daily living. In this context it isinteresting to note the much higher incidence of left-handedness in carrier females and hemizygous males thanoccurs in the normal population. A much higher incidenceof left-handedness has been shown to occur in children withclumsiness and mild cognitive impairment (Saigal et al.,1992). Our findings in carrier females are corroboratedby the occasional clinical reports of ‘expressing females’(Bianchine and Routtenberg, 1974; Halliday et al., 1986;Kaepernick et al., 1994), which may represent both pheno-typic variation and/or skewed X-chromosome inactivation inthese individuals.

There are discrepancies between theobservations in man and the knock-out mousemodelsThe present study demonstrates significant abnormality incorticospinal function in subjects with L1CAM mutations,as might be expected from the evidence for corticospinaltract hypoplasia observed both in man and in the knock-outmouse models (Chow et al., 1985; Yamasaki et al., 1995;Cohen et al., 1997; Dahme et al., 1997; Fransen et al., 1998;Graf et al., 2000) and the clinical signs of spasticity (Kenwricket al., 1996; Fransen et al., 1997). There are two observationsin this study, however, which are inconsistent with thosemade in the knock-out mouse by Cohen and colleagues(Cohen et al., 1997). First, corticospinal projections have beendemonstrated both neurophysiologically and histologically tothe lumbar spinal segments, and secondly, there is no

anatomical or neurophysiological evidence to supportsignificant failure of decussation of corticospinal axons.These discrepancies are unlikely to arise from differences inthe site of the mutation between the knock-out mouse andour subjects, since we have included in the study hemizygousmales with class 3 mutations who, like the knock-out mouseof Cohen and colleagues (1997), would not have expressedL1CAM protein (Cases 2, 3, 5 and 9). Indeed, Case 9, thesubject for the neuroanatomical study, had a nonsensemutation affecting Ig domain 6, and this is very similar tothe mutation used to create the knock-out mouse of Dahmeand colleagues (Dahme et al., 1997). The histologicalevidence in Case 9 for decussation of the majority ofcorticospinal axons at the medulla, combined with neuro-physiological evidence in the remaining cases that themajority of corticospinal axons projecting to bicepsmotoneurones decussated, makes it unlikely that there is acritical role for L1CAM proteins in corticospinal axonguidance at the pyramidal decussation. An alternativeexplanation may be that significant delay in the projectionof corticospinal axons, secondary to slow fasciculation, ledto the corticospinal axons in the knock-out mouse arrivingat the medulla too late for the optimal expression of guidancecues for decussation. Slow fasciculation might also explainthe failure of projection of corticospinal axons to the lumbarcord observed in the knock-out mouse. Since this projectionis the longest in the central nervous system, the axons mayarrive too late to be competitive in establishing synapses onmotoneurones and interneurones, and are thus predominantlyor even completely withdrawn (Stanfield et al., 1982; Martinand Lee, 1999; Eyre et al., 2001b). The findings in thepresent study that the neurophysiological abnormalities weremore marked in the corticospinal projection to the quadricepsα-motoneurones than to biceps α-motoneurones would beconsistent with this hypothesis. Axons projecting to bicepsmotoneurones would arrive relatively earlier and wouldtherefore have been more likely to be competitive in theestablishment of their synaptic projection. Finally, there isincreasing evidence that CAMs, including L1CAM, have apivotal role in activity-dependent establishment of synapses(Hoffman, 1998). The corticospinal projection to the lumbarsegments of the spinal cord must therefore be the mostvulnerable to abnormalities in L1CAM protein expression,since the axons are likely to arrive late and dysfunction incell adhesion will further compromise synapse formation.

AcknowledgementsWe thank all the subjects and physicians who gave permissionto study their patients. The study was supported by TheWellcome Trust.

ReferencesArmand J, Olivier E, Edgley SA, Lemon RN. Postnatal developmentof corticospinal projections from motor cortex to the cervicalenlargement in the macaque monkey. J Neurosci 1997; 17: 261–6.


Bennett V, Lambert S. Physiological roles of axonal ankyrins insurvival of premyelinated axons and localization of voltage-gatedsodium channels. [Review]. J Neurocytol 1999; 28: 303–18.

Bianchine L, Routtenberg A. The MASA syndrome: a new heritablemental retardation syndrome. Clin Genet 1974; 5: 298–306.

Bortoff GA, Strick PL. Corticospinal terminations in two new-worldprimates: further evidence that corticomotoneuronal connectionsprovide part of the neural substrate for manual dexterity. J Neurosci1993; 13: 5105–18.

Burke D, Bartley K, Woodforth I, Yakoubi A, Stephen JP. Theeffects of a volatile anaesthetic on the excitability of humancorticospinal axons. Brain 2000; 123: 992–1000.

Castellani V, Chedotal A, Schachner M, Faivre-Sarrailh C,Rougon G. Analysis of the L1-deficient mouse phenotype revealscross-talk between Sema3A and L1 signaling pathways in axonalguidance. Neuron 2000; 27: 237–49.

Chow CW. Anatomical pathology in the investigation of patientswith genetic diseases involving the central nervous system [MDthesis]. Melbourne: University of Melbourne; 1996.

Chow CW, Halliday JL, Anderson RM, Danks DM, Fortune DW.Congenital absence of pyramids and its significance in geneticdiseases. Acta Neuropathol (Berl) 1985; 65: 313–17.

Cohen NR, Taylor JS, Scott LB, Guillery RW, Soriano P, FurleyAJ. Errors in corticospinal axon guidance in mice lacking the neuralcell adhesion in molecule L1. Curr Biol 1998; 8: 26–33.

Conway E. The development of corticospinal projections in man[Ph D thesis]. Newcastle: University of Newcastle upon Tyne; 1995.

Dahme M, Bartsch U, Martini R, Anliker B, Schachner M, Mantei N.Disruption of the mouse L1 gene leads to malformations of thenervous system. Nat Genet 1997; 17: 346–9.

Darian-Smith I, Galea MP, Darian-Smith C. Manual dexterity: howdoes the cerebral cortex contribute? [Review]. Clin Exp PharmacolPhysiol 1996; 23: 948–56.

de la Rosa EJ, Kayyem JF, Roman JM, Stierhof YD, Dreyer WJ,Schwarz U. Topologically restricted appearance in the developingchick retinotectal system of Bravo, a neural surface protein:experimental modulation by environmental cues. J Cell Biol 1990;111: 3087–96.

Demyanenko GP, Tsai AY, Maness PF. Abnormalities in neuronalprocess extension, hippocampal development, and the ventricularsystem of L1 knockout mice. J Neurosci 1999; 19: 4907–20.

Di Lazzaro V, Restuccia D, Oliviero A, Profice P, Ferrara L,Insola A, et al. Effects of voluntary contraction on descendingvolleys evoked by transcranial stimulation in conscious humans.J Physiol (Lond) 1998; 508: 625–33.

Djabali M, Mattei MG, Nguyen C, Roux D, Demengeot J, Denizot F,et al. The gene encoding L1, a neural adhesion molecule of theimmunoglobulin family, is located on the X-chromosome in mouseand man. Genomics 1990; 7: 587–93.

Dobson C. Corticospinal function in subjects with mutations of theL1CAM gene [B Med Sci thesis]. Newcastle: University ofNewcastle upon Tyne; 1999.

Edgley SA, Eyre JA, Lemon RN, Miller S. Excitation of thecorticospinal tract by electromagnetic and electrical stimulation ofthe scalp in the macaque monkey. J Physiol (Lond) 1990; 425:301–20.

Edgley SA, Eyre JA, Lemon RN, Miller S. Comparison of activationof corticospinal neurons and spinal motorneurons by magnetic andelectrical transcranial stimulation in the lumbosacral cord of theanaesthetized monkey. Brain 1997; 120: 839–53.

Eyre JA, Miller S. Assessment of motor pathways. In: Eyre JA,editor. The neurophysiological examination of the newborn infant.Clin Dev Med, No. 120. London: MacKeith Press; 1992. p. 124–54.

Eyre JA, Miller S, Clowry GJ, Conway EA, Watts C. Functionalcorticospinal projections are established prenatally in the humanfoetus permitting involvement in the development of spinal motorcentres. Brain 2000; 123: 51–64.

Eyre JA, Miller S, Clowry GJ. Development of the corticospinaltract in man. In: Pascual-Leone A, Davey G, Wasserman EM,Rothwell J, editors. Handbook of transcranial magnetic stimulation.London: Arnold. In press 2001a.

Eyre JA, Taylor JP, Villagra F, Smith M, Miller S. Evidence ofactivity dependent withdrawal of corticospinal projections duringdevelopment in man. Neurology. In press 2001b.

Fields RD, Itoh K. Neural cell adhesion molecules in activity-dependent development and synaptic plasticity. [Review]. TrendsNeurosci 1996; 19: 473–80.

Fransen E, Van Camp G, Vits L, Willems PJ. L1-associated diseases:clinical geneticists divide, molecular geneticists unite. [Review].Hum Mol Genet 1997; 6: 1625–32.

Fransen E, D’Hooge R, Van Camp G, Verhoye M, Sijbers J,Reyniers E, et al. L1 knockout mice show dilated ventricles, vermishypoplasia and impaired exploration patterns. Hum Mol Genet1998; 7: 999–1009.

Galea MP, Darian-Smith I. Multiple corticospinal neuron populationsin the macaque monkey are specified by their unique cortical origins,spinal terminations, and connections. Cereb Cortex 1994; 4: 166–94.

Graf WD, Born DE, Shaw DW, Thomas JR, Holloway LW, MichaelisRC. Diffusion-weighted magnetic resonance imaging in boys withneural cell adhesion molecule L1 mutations and congenitalhydrocephalus. Ann Neurol 2000; 47: 113–17.

Halliday J, Chow CW, Wallace D, Danks DM. X linked hydro-cephalus: a survey of a 20 year period in Victoria, Australia. J MedGenet 1986; 23: 23–31.

Hoffman KB. The relationship between adhesion molecules andneuronal plasticity. [Review]. Cell Mol Neurobiol 1998; 18: 461–75.

Honig MG, Rutishauser US. Changes in the segmental pattern ofsensory neuron projections in the chick hindlimb under conditionsof altered cell adhesion molecule function. Dev Biol 1996; 175:325–37.

Hortsch M. Structural and functional evolution of the L1 family:Are four adhesion molecules better than one? [Review]. Mol CellNeurosci 2000; 15: 1–10.

Jalinous R. Technical and practical aspects of magnetic nervestimulation. [Review]. J Clin Neurophysiol 1991; 8: 10–25.


Jankowska E, Padel Y, Tanaka R. Projections of pyramidal tractcells to alpha-motoneurones innervating hind limb muscles in themonkey. J Physiol (Lond) 1975; 249: 637–67.

Jankowska E, Patel Y, Tanaka R. Disynaptic inhibition of spinalmotoneurones from the motor cortex in the monkey. J Physiol(Lond) 1976; 258: 467–87.

Jouet M, Rosenthal A, Armstrong G, MacFarlane J, Stevenson R,Paterson J, et al. X-linked spastic parplegia (SPG1), MASAsyndrome and X-linked hydrocephalus result from mutations in theL1 gene. Nat Genet 1994; 7: 402–7.

Jouet M, Moncla A, Paterson J, McKeown C, Fryer A, Carpenter N,et al. New domains of neural cell-adhesion molecule L1 implicatedin X-linked hydrocephalus and MASA syndrome. Am J Hum Genet1995; 56: 1304–14.

Kaepernick L, Legius E, Higgins J, Kapur S. Clinical aspects ofthe MASA syndrome in a large family; including expressing females.Clin Genet 1994; 45: 181–5.

Kenwrick S, Jouet M, Donnai D. X linked hydrocephalus andMASA syndrome. [Review]. J Med Genet 1996; 33: 59–65.

Kenwrick S, Watkins A, Angelis ED. Neural cell recognitionmolecule L1: relating biological complexity to human diseasemutations. [Review]. Hum Mol Genet 2000; 9: 879–86.

Kuypers HGJM. Corticospinal connections: postnatal developmentin the rhesus monkey. Science 1962; 138: 678–680.

Kuypers HGJM. Anatomy of the descending pathways. In: BrookhartJM, Mountcastle VB, Brooks VB, editors. Handbook of physiology;Sect. 1, Vol. 2, Pt. 1. Bethesda (MD): American PhysiologicalSociety; 1981. p. 597–666.

Lagenaur C, Lemmon V. An L1-like molecule, the 8D9 antigen, isa potent substrate for neurite extension. Proc Natl Acad Sci USA1987; 84: 7753–7.

Lawrence DG, Hopkins DA. The development of motor control in therhesus monkey: evidence concerning the role of corticomotoneuronalconnections. Brain 1976; 99: 235–54.

Lemmon V, McLoon SC. The appearance of an L1-like moleculein the chick primary visual pathway. J Neurosci 1986; 6: 2987–94.

Lindner J, Rathjen FG, Schachner M. L1 mono- and polyclonalantibodies modify cell migration in early postnatal mouse cere-bellum. Nature 1983; 305: 427–30.

Luthl A, Laurent JP, Figurov A, Muller D, Schachner M. Hippo-campal long-term potentiation and neural cell adhesion moleculeL1 and NCAM. Nature 1994; 372: 777–9.

Martin JH, Lee SJ. Activity-dependent competition betweendeveloping corticospinal terminations. Neuroreport 1999; 10:2277–82.

Mathiowetz V, Weber K. Adult norms for the nine-hole peg test offinger dexterity. Occup Ther J Res 1985; 5: 24–38.

Matthews PB. The effect of firing on the excitability of a modelmotoneurone and its implications for cortical stimulation. J Physiol(Lond) 1999; 518: 867–82.

Metman LV, Bellevich S, Jones SM, Barber MD, Streletz LJ.Topographic mapping of human motor cortex with transcranial

magnetic stimulation: homunculus revisited. Brain Topogr 1993; 6:13–19.

Miller S, Clarke J, Eyre JA, Kelly S, Lim E, McClelland VM, et al.Comparison of spinal myotatic reflexes in human adults investigatedwith cross-correlation and conventional signal averaging methods.Brain Res 2001; 899: 47–65.

Moscoso LM, Sanes JR. Expression of four immunoglobulinsuperfamily adhesion molecules (L1, Nr-CAM/Bravo, neurofascin/ABGP, and N-CAM) in the developing mouse spinal cord. J CompNeurol 1995; 352: 321–34.

Muller K, Kass-Iliyya F, Reitz M. Ontogeny of ipsilateralcorticospinal projections: a developmental study with transcranialmagnetic stimulation. Ann Neurol 1997; 42: 705–11.

Nathan PW, Smith MC, Deacon P. The corticospinal tracts in man:course and location of fibres at different segmental levels. Brain1990; 113: 303–24.

O’Sullivan M. The development of the phasic stretch reflex in manand its pathophysiology in central motor disorders [Ph.D. thesis].Newcastle: University of Newcastle upon Tyne; 1991.

Porac C, Coren S. Lateral preferences and human behavior. NewYork: Springer-Verlag; 1981.

Rosenthal A, Jouet M, Kenwrick S. Aberrant splicing of neuralcell adhesion molecule L1 mRNA in a family with X-linkedhydrocephalus. Nat Genet 1992; 2: 107–12.

Saigal S, Rosenbaum P, Szatmari P, Hoult L. Non-right handednessamong ELBW and term children at eight years in relation tocognitive function and school performance. Dev Med Child Neurol1992; 34: 425–33.

Sharpless J. The nine-hole peg test of finger-hand coordination forthe hemiplegic patient. In: Mossman PL, editor. A problem-orientedapproach to stroke rehabilitation. New York: Thomas; 1976.p. 428–31.

Stallcup W, Beasley L. Involvement of the nerve growth factor-inducible large external glycoprotein (NILE) in neurite fasciculationin primary cultures of rat brain. Proc Natl Acad Sci USA 1985; 82:1276–80.

Stanfield BB, O’Leary DD, Fricks C. Selective collateral eliminationin early postnatal development restricts cortical distribution of ratpyramidal tract neurones. Nature 1982; 298: 371–3.

Terao Y, Ugawa Y, Sakai K, Uesaka Y, Kohara N, Kanazawa I.Transcranial stimulation of the leg area of the motor cortex inhumans. Acta Neurol Scand 1994; 89: 378–83.

Tiffin J, Asher EJ. The Purdue Pegboard: norms and studies ofreliability and validity. J Appl Psychol 1948; 32: 234–47.

Tinnion R. A randomised double blind study of the effects ofbotulinum toxin A on heteronymous group 1a muscle reflexes ofbiceps brachii in spastic cerebral palsy [B Med Sci thesis] Newcastle:University of Newcastle upon Tyne; 1999.

Vecchierini-Blineau MF, Guiheneuc P. Electrophysiological studyof the peripheral nervous system in children. J Neurol NeurosurgPsychiatry 1979; 42: 753–9.


Vits L, Van Camp G, Coucke P, Fransen E, De Boulle K, Reyniers E,et al. MASA syndrome is due to mutations in the neural celladhesion gene L1CAM. Nat Genet 1994; 7: 408–13.

Wassermann EM, Pascual-Leone A, Hallett M. Cortical motorrepresentation of the ipsilateral hand and arm. Exp Brain Res 1994;100: 121–32.

Weiland UM, Ott H, Bastmeyer M, Schaden H, Giordano S,Stuermer CA. Expression of an L1-related cell adhesion moleculeon developing CNS fiber tracts in zebrafish and its functionalcontribution to axon fasciculation. Mol Cell Neurosci 1997; 9: 77–89.

Winter RM, Davies KE, Bell MV, Huson SM, Patterson MN. MASAsyndrome: further clinical delineation and chromosomal localisation.Hum Genet 1989; 82: 367–70.

Wood PM, Schachner M, Bunge RP. Inhibition of Schwann cellmyelination in vitro by antibody to the L1 adhesion molecule.J Neurosci 1990; 10: 3635–45.

Yamasaki M, Arita N, Hiraga S, Izumoto S, Morimoto K, Nakatani S,et al. A clinical and neuroradiological study of X-linked hydro-cephalus in Japan. J Neurosurg 1995; 83: 50–5.

Yamasaki M, Thompson P, Lemmon V. CRASH syndrome:mutations in L1CAM correlate with severity of the disease.Neuropediatrics 1997; 28: 175–8.

Received March 28, 2001. Revised June 4, 2001.Accepted July 10, 2001

brain abnormal corticospinal function but normal axonal ... · normal corticospinal decussation and...

Documents