DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY REVIEW Motor mapping in cerebral palsy GEORGE F WITTENBERG MD PHD Baltimore VA Medical Center Geriatric Research, Education and Clinical Center, Baltimore, MD, USA; Department of Neurology, University of Maryland School of Medicine, Baltimore, MD, USA. Correspondence to George F Wittenberg at Baltimore VA Medical Center Geriatric Research, Education and Clinical Center, 10 N Green St. (BT ⁄ 18 ⁄ GR) 21201 Baltimore, MD, USA. E-mail: GWittenb@GRECC.UMaryland.edu LIST OF ABBREVIATIONS COG FDI fMRI TMS Center of gravity First dorsal interosseus Functional magnetic resonance imaging Transcranial magnetic stimulation ACKNOWLEDGEMENTS The work described in this paper was carried out at Wake Forest University School of Medicine, Winston-Salem, NC with the collaboration of: Lumy Sawaki, Normann Cabrera, Kathleen Kolaski, Beth Smith, Michael O'Shea, and Andrew Koman. Funding was provided by US Public Health Service Grant No R21 HD049019 to GFW. CONFLICTS OF INTEREST The author declares no conflicts of interest. The measurement of motor deficits in individuals with cerebral palsy (CP) has been largely based on clinical criteria. Yet functional imaging and non-invasive stimulation methods provide a means to measure directly abnormalities of the motor system. The size and location of muscles and movement representations can be determined with transcranial magnetic stimulation (TMS) and functional magnetics resonance imaging. Thus the homunculus can be individually mapped in children with CP. Because size of representation within the homunculus relates to quality of motor control, measurement of the distance between body parts provides a metric that may be useful in classifying deficits. Bilateral motor control in one hemisphere, while normal in neonates, persists variably in CP, providing another physiological metric. In this study, we used TMS to measure hand and ankle representations in a convenience sample of children with spastic CP. Overlapping thumb and ankle maps were found in children with both hemiplegia and diplegia, and these maps may be from either side of the body. While more participants are required to make conclusions about disability and compression ⁄ bilaterality of the homunculus, it appears as if TMS-derived metrics relate to motor abnormalities. These abnormal motor maps also are a therapeutic target, as stimulation methods are being developed as adjuncts to physical means of rehabilitation. Diagnostic and interventional strategies in children with cerebral palsy (CP) are largely empirical. The effects of growth and development, limited validated outcome instruments, and the absence of techniques with which to assess central nervous system organization and function have confounded analysis of treatment paradigms. The severity and types of symptoms associated with CP are subject to wide individual variations, ranging from severe generalized motor disabilities to mild spasticity limited to the lower limbs. Although there are known underlying pathologies for many forms of CP, classification and severity measures have generally been based solely on clinical grounds. Classification has been particularly unreliable, as mixed features may exist, leaving some patients with unclassifiable or overlapping disorders. This has been highlighted in a provocative paper calling for abandonment of the commonly used category of ‘diplegia.’1 134 Relationship between area of motor cortex and motor ability The amount of motor cortex dedicated to a body part is related to the complexity of movement possible with that body part. This is reflected in the familiar motor homunculus described in neuroanatomy, with a larger area of cortex related to hand movement than to elbow movement, for example. The area devoted to a body part is influenced by experience, throughout development and into adult life, for both sensory2 and motor3 cortex. Recent work has even demonstrated rapid expansion of the non-primary sensory cortex with practice of a motor skill.4 This suggests, but does not prove, that a larger area of cortex devoted to primary sensory or primary motor, or sensory association function is associated with enhanced ability. Imaging the motor map Standard anatomical imaging, with computed tomography and magnetic resonance imaging, provides the methodol- ª 2009 The Author Journal compilation ª 2009 Mac Keith Press Developmental Medicine & Child Neurology 2009, 51 (Suppl. 4): 134–139 DOI: 10.1111/j.1469-8749.2009.03426.x ogy to determine, in patients with motor developmental disorders, lesion type and location, including the presence of gross disorders in neuronal migration and gyrus formation. There have been studies relating the type of anatomical lesion to motor disability5–7 and quantitative correlation between anatomical lesions and motor ability.8,9 However, anatomical imaging does not provide information regarding the adaptive response to such macro- or micro-anatomical abnormalities. Such information can be provided through functional imaging10,11 or non-invasive neurophysiology methods.12–14 Such studies have suggested several mechanisms of recovery: (1) sprouting of ipsilateral corticospinal connections from the healthy hemisphere or recrossing into the spinal cord; (2) stabilization of the bilateral fetal corticospinal projections; (3) enhancement of normally non-functional ipsilateral corticospinal pathways; (4) reinforcement of bilaterally distributed corticoreticulospinal connections; and (5) shifting or spreading of the cortical motor representation within the damaged cortex over adjacent (presumably intact) areas, resulting in a reshaping of cortex somatotopy.15 All but the last of these are mechanisms that reflect recruitment of pathways outside of the normal adult pattern of dominance of the contralateral projection to lower motor neurons from the primary motor cortex. The mechanisms of cortical reorganization and the positive or detrimental role of specific patterns of reorganization are poorly understood and probably differ in the adult and the immature brain. Transcranial magnetic stimulation One of the principal methods used to study the human motor system is transcranial magnetic stimulation (TMS), a non-invasive technique that activates the primary motor cortex. TMS is performed by generating a large, transient electrical current through an insulated coil placed over the scalp. The rapidly changing magnetic field that results induces small electrical currents in the brain.16 When the coil is situated so as to induce sufficient current in the motor cortex, motor-evoked potentials (MEPs) appear in one or more muscles and may be recorded with surface electrodes. These MEPs indicate activation of a population of upper and lower motor neurons with sufficient synchrony to produce a measurable compound action potential in a muscle. Safe use and standardization of TMS techniques have been established.17,18 Mapping Introduction of the double-circular coil, also known as a figure-of-eight coil, has allowed detailed mapping of muscle representations in the primary motor cortex.19 Even without stereotactic methodology, the relationships between muscle representations can be measured with sufficient accuracy.20 (Interpretation of the size of the motor representation is more complicated,21 but larger motor maps have been associated with functional recovery in stroke patients after rehabilitation.22) TMS safety Although TMS is not used in people with implanted metallic devices, it is thought to be safe in patients with hydrocephalus and ventricular shunts. Although high-rate, repetitive TMS has the potential to induce seizures,23 TMS rates of 0.2Hz or less are safe in patients with epilepsy, and even higher rates may have a protective effect in the case of intractable seizures.24 No seizures have been reported in patients with CP during or after TMS, even in children with epileptic foci on electroencephalogram or with known epilepsy.25–27 TMS in children with and without CP Electrophysiological maturation of the corticospinal tract is complete by age 13.28 Before that, the threshold for evoking a motor potential in extremity muscles is higher than it is in adults. The reason for this higher threshold is not entirely clear, although both smaller brain size and immature myelination are theoretical considerations. TMS thresholds increase over the first 3 months of life29 and then decrease monotonically after age 8 years. Corticospinal conduction velocity reaches adult values by age 11 and correlates with the acquisition of fine-motor skills. Central motor conduction time, corrected for increasing height, decreases until age 10.30 Such normative data in healthy children14,31 allow comparative TMS studies in patients with CP. TMS has also been used to measure the effectiveness of connections within the brain; for example, the phenomenon of transcallosal inhibition is absent in bilateral patients with spastic CP.32 In this study, patterns of upper-extremity central motor reorganization, using TMS mapping principally, correlated with functional motor performance and outcomes in patients with CP. METHOD Participants The 10 participants analyzed so far in this study had a diagnoses of CP with motor impairment, without chorea, athetosis, or ballismus. Children with postnatal onset of motor impairment were excluded. The two ICD-9 diagnoses were: 343.0 (CP, diplegic) and 343.1 (CP, hemiplegic), but two of the participants with diplegia were noted to be asymmetrically affected. Participants were 9 to 16 years of age, with younger (<7y) patients excluded mainly because of the higher motor thresholds that occur in younger Motor Mapping in CP George F Wittenberg 135 children. Patients on anticonvulsants were excluded for the same reason. Patients had to be able to cooperate with verbal instruction and be competent to give assent to participate. Because one of the long-term goals of this study was to establish prognostic criteria for response to treatment, we excluded children who have had tendon transposition, who have implanted baclofen pumps, or who have had recent (within 3mo) botulinum toxin injections. Children who required antispasticity agents were similarly excluded. Clinical scales Motor measures were the Gross Motor Function Measure (GMFM)-66,33 a reduced version of the original GMFM,34 and the Melbourne Assessment of Unilateral Upper Limb Function.35,36 The reason for two primary outcome measures is the lack of a clinical scale that accurately measures both fine motor and gross motor function. Experimental procedure TMS protocol The motor cortex was stimulated using a MagStim 200 Magnetic Stimulator (MagStim Ltd., Wales, UK) with a double-circular coil to perform focal hemispheric stimulation. MEPs were recorded bilaterally by surface electrodes fixed over both first dorsal interosseus (FDI; or abductor pollicis brevis in pilot studies) and both tibialis anterior (TA) muscles. Responses were amplified by a battery-powered surface EMG amplifier (James Long Co., Canada Lake, NY, USA), using an amplification of 1000· (bandpass 30–3kHz), and fed into a personal computer through a multifunctional I ⁄ O board and LabView acquisition ⁄ analysis software (National Instruments, Austin, TX, USA). A 100-msec period after stimulus was examined. Amplitudes were measured peak to peak. Motor threshold was determined for each participant using modified International Federation of Clinical Neurophysiology criteria.17 Ipsilateral thresholds were determined as well; these are generally higher than the contralateral thresholds and were measured only after the contralateral hot spots had been found. For mapping, stimulation intensity was set at 120% of motor threshold. The participant wore a closely fitting electroencephalography cap with a grid of anteroposteriorly and mediolaterally oriented 1-cm squares. During the mapping, the TMS coil was positioned in a parasagittal line with the handle pointing posteriorly, with its center in contact with, and tangential to, the scalp at each stimulated grid position. Ten consecutive stimuli were delivered at each position. The rate of stimulation did not exceed 0.2Hz. Stimulation was continued on every other line of the grid, thus mapping at 2-cm intervals on the scalp, until a border position without 136 Developmental Medicine & Child Neurology 2009, 51 (Suppl. 4): 134–139 a response of at least 50lV peak-to-peak in five, or <50% of 10, successive stimulations was encountered. The average response of every series of 10 was calculated after each mapping session. Map metrics The TMS-derived map can be viewed as a matrix of values, with the position in the matrix representing the location on the scalp grid, and the values, the MEP size. The simplest means to describe the location and spread of such a map is by the center of gravity (COG) and the number of active positions.22 The COG is a particularly accurate measure, because it is a weighted average of all points with a response and is thus less sensitive to error in any one measurement. Map COG is given by: COGðxÞ ¼ i (MEP size)i  x( coordinate)I = (map volume) similarly for y (and z, if performing measurement in 3D) The main outcome measures were the distance between ispilateral and contralateral FDI COG in each hemisphere (if both were present) and between the COG of the FDI and TA within the same hemisphere. RESULTS Pilot mapping both upper- and lower-extremity representations in children with CP revealed a variety of abnormalities. Figure 1 is an example, from an 11-year-old with a clinical diagnosis of hemiplegic CP, of the affected side map of abductor pollicus brevis (thumb) and tibialis anterior (ankle dorsiflexor). Note the overlapping thumb and ankle maps, with the thumb being slightly more lateral at higher stimulation strengths. This was the key finding that initiated the current study. It had been reported once before.27 Figure 2 shows the clinical characteristics and TMSderived map COG of participants in this study. They fell into three clinical categories: diplegic, diplegic with left body side more affected, or hemiplegic with right side more affected. Age and brain side affected are shown. If TMS elicited motor maps, then COG of those maps is shown. The affected brain side is highlighted. Although the number of participants in whom we could measure map distance and GMFM has been limited so far, Figure 3 shows that the expected correlation holds. Because two participants had crossed hand ⁄ ankle maps (hand and ankle of opposite sides mapped within the same hemisphere), their data were plotted separately, and both these participants had diplegia. The ipsilateral maps were present in participants with hemiplegia. For reference, the 14 14 12 RAPB –9 –7 –5 –3 12 RTA 10 10 8 8 6 6 4 4 2 2 0 –1 –9 –7 –5 –3 0 –1 Figure 1: Right thumb motor map (left) and right ankle motor map (right). The bubble graphs show relative motor-evoked potentials size evoked at each location, with standard definitions of x (left ⁄ right) and y (anterior ⁄ posterior) for the scalp surface. The vertex is the origin (0,0). RAPB, right abductor pollicus brevis; RTA, right tibialis anterior. ID ICD-9 8 diplegia 11 diplegia 12 diplegia 14 diplegia 10 L diplegia 13 L diplegia 1 R Hemi 2 R Hemi 4 R Hemi 7 R Hemi Age Aff. (yrs.) Br. 12 B 12 B 15 B 9B 10 R 9R 11 L 12 L 16 L 11 L Left L FDI x – – –5.92 – – – – – – – Brain L FDI y – – –3.32 – – – – – – – COG R FDI x –6.15 –3.16 –5.80 –5.76 –1.97 –7.06 – – –3.75 –5.07 cm R FDI y 3.50 2.63 3.08 –1.46 1.72 2.71 – – 4.32 1.94 L TA x – – – –5.39 – – – – – –3.21 L TA y – – – –1.42 – – – – – 1.83 R TA x – – – – – – – – – –3.58 R TA y – – – – – – – – – 1.30 Right L FDI x 5.48 3.81 – 3.79 3.33 3.23 4.24 7.53 3.61 3.52 Brain COG cm L FDI R FDI R FDI L TA y x y x 3.46 – – – 2.88 – – – – – – – –2.25 – – – 2.47 – – – 1.21 – – – 2.39 3.72 2.86 4.34 4.14 6.95 4.17 – 1.89 3.59 1.70 – – – 1.74 – L TA y – – – – – – 0.97 – – – R TA x – – – – 2.67 – – – – – Ispil. R TA Map(s)? y – N – N – Y – Y 2.44 TA – N – Y – Y – Y – TA Figure 2: Clinical characteristics and transcranial magnetics stimulation-derived map center of gravity (COG). The study ID, clinical classification, age, affected brain side (B-bilateral), and location of muscle COG on the scalp coordinate system are shown. The presence of ipsilateral maps is indicated in the column at the far right (Y – present for FDI (SD) TA, TA – present for TA only). FDI, First dorsal interossers; TA, tibialis anterior. GMFM 1.05 0.95 0.85 Ipsilateral Crossed 0.75 0 1 TA-FDI (cm) 2 Figure 3: Correlation between gross motor function measure and hand-ankle distance. normal distance between hand and ankle representations is at least 4cm. DISCUSSION The following three conclusions can be made on the basis of the results of this study to date: (1) The affected brain may have a map of the affected, but not the unaffected, hand, while the unaffected brain may have a map of both hands. (2) Ipsilateral hand or ankle maps may occur in patients with CP defined as either hemiplegic or diplegic so that the clinical subtypes do not precisely divide the TMS abnormalities. (3) Bilateral maps in one hemisphere are not geometrically distinct, suggesting that the upper motor neuron has a bilateral corticospinal projection. TMS can demonstrate a localized motor representation in a manner that is complementary to functional magnetic resonance imaging (fMRI; also performed on these participants, but results not shown). While TMS may reveal a bilateral motor representation, this shared drive of the affected side is often not apparent on motor task–related fMRI, suggesting that there are motor representations that are untapped in voluntary behavior. The lateralization of the ankle map probably arises from a remapping of the body representation onto undamaged motor cortex. The hand map appears to move laterally to compensate, but there is insufficient space to prevent overlap of these usually very distinctly represented body parts. The distance Motor Mapping in CP George F Wittenberg 137 that does exist between the COG provides an opportunity for training better differential control, with TMS as an adjunctive measure in that process. The study described here may provide a unique method to: (1) assess treatment efficacy; (2) analyze the effects of intervention on long-term outcome; and (3) define the relationship between pathophysiology and disability in CP. The results have implications for treatment of older children and adults with CP as these abnormalities are likely to persist and are more extreme examples of abnormalities seen in stroke survivors. On the basis of the results of this study, we plan to design further studies to test the usefulness of motor map abnormalities as predictive for outcome of treatment strategies (e.g. occupational therapy, electrical stimulation, neuromuscular blockade, and tendon transfer). In another disorder, epilepsy, non-invasive neurophysiology (electroencephalography) has led to better selection of anticonvulsants and to better selection of candidates for surgical procedures. Similarly, in patients with CP, understanding the adaptive and maladaptive changes in motor maps should allow the design of specific pharmacological, rehabilitative, and surgical procedures. stimulation parameters. Electroencephalogr 10. Thickbroom GW, Byrnes ML, Archer SA, Clin Neurophysiol 1997; 105: 415–21. Nagarajan L, Mastaglia FL. Differences in 1. Colver AF, Sethumadhavan T. The term 19. Wassermann EM, Pascual-Leone A, Vallssensory and motor cortical organization diplegia should be abandoned. Arch Dis REFERENCES Child 2003; 88: 286–90. 2. Merzenich MM, Recanzone G, Jenkins following brain injury early in life. Ann Sole J, et al. Topography of the inhibitory Neurol 2001; 49: 320–7. and excitatory responses to transcranial WM, Allard TT, Nudo RJ. Cortical rep- 11. Cioni G, Montanaro D, Tosetti M, Canapicchi R, Ghelarducci B. Reorganisation resentational plasticity. In: Rakic P, Singer magnetic stimulation in hand muscle. Electroencephalogr Clin Neurophysiol 1993; 89: 424–33. W, editors. Neurobiology of Neocortex. of the sensorimotor cortex after early focal Chichester: Wiley, 1988: 41–67. brain lesion: a functional MRI study in 20. Ziemann U, Wittenberg GF, Cohen LG. Stimulation-induced within-representation monozygotic twins. Neuroreport 2001; 12: 3. Plautz EJ, Milliken GW, Nudo RJ. Effects of repetitive motor training on movement 1335–40. and across-representation plasticity in human motor cortex. J Neurosci 2002; 22: representations in adult squirrel monkeys: 12. Farmer SF, Harrison LM, Ingram DA, 5563–71. Stephens JA. Plasticity of central motor role of use versus learning. Neurobiol Learn pathways in children with hemiplegic 21. Thickbroom GW, Sammut R, Mastaglia Mem 2000; 74: 27–55. 4. Dranganski B, Gaser C, Busch V, et al. Neuroplasticity: changes in grey matter cerebral palsy. Neurology 1991; 41: 1505– FL. Magnetic stimulation mapping of 10. motor cortex: factors contributing to map induced by training. Nature 2004; 427: 13. Kalil K. Regeneration of pyramidal tract axons. Adv Neurol 1988; 47: 67–85. 311–12. area. Electroencephalogr Clin Neurophysiol 1998; 109: 79–84. 5. De Vries LS, Connell JA, Dubowitz LM, 14. Koh TH, Eyre JA. Maturation of corti- 22. Liepert J, Bauder H, Wolfgang HR, et al. Treatment-induced cortical reorganization cospinal tracts assessed by electromagnetic et al. Neurological, electrophysiological and MRI abnormalities in infants with stimulation of the motor cortex. Arch Dis Child 1988; 63: 1347–52. after stroke in humans. Stroke 2000; 31: 1210–16. 15. Rossini RM, Caltagirone C, Castriota- 23. Wassermann EM, Cohen LG, Flitman SS, Scanderbeg A, et al. Hand motor cortical Chen R, Hallett M. Seizures in healthy 6. Cioni G, Sales B, Paolicelli PB, et al. MRI area reorganization in stroke: a study with people with repeated ‘‘safe’’ trains of transand clinical characteristics of children with fMRI, MEG and TCS maps. Neuroreport cranial magnetic stimuli. Lancet 1996; hemiplegic cerebral palsy. Neuropediatrics 1998; 9: 2141–6. 347: 825–6. 1999; 30: 249–55. extensive cystic leukomalacia. Neuropediatrics 1987; 18: 61–6. 7. Cioni G, Bartalena L, Biagioni E, Boldrini 16. Barker AT, Jalinous R, Freeston IL. 24. Tergau F, Naumann U, Paulus W, SteinNon-invasive magnetic stimulation of hoff BJ. Low-frequency repetitive transA, Canapicchi R. Neuroimaging and funchuman motor cortex. Lancet 1985; 11: cranial magnetic stimulation improves tional outcome of neonatal leukomalcia. 1106–7. intractable epilepsy. Lancet 1999; 353: Behav Brain Res 1992; 49: 7–19. 2209. 8. Staudt M, Pavlova M, Bohm S, Grodd W, 17. Rossini PM, Barker AT, Berardelli A, et al. Non-invasive electrical and magnetic stim- 25. Maegaki Y, Yamamoto T, Takeshita K. Krageloh-Mann I. Pyramidal tract damage ulation of the brain, spinal cord and roots: Plasticity of central motor and sensory correlates with motor dysfunction in bilatbasic principles and procedures for routine pathways in a case of unilateral extensive eral periventricular leukomalacia (PVL). Neuropediatrics 2003; 34: 182–8. 9. Melhem ER, Hoon AH Jr, Ferrucci JT, et al. Periventricular leukomalacia: rela- clinical application. Report of an IFCN cortical dysplasia: investigation of mag- committee. Electroencephalogr Clin Neuro- netic resonance imaging, transcranial mag- physiol 1994; 9: 79–92. netic tionship between lateral ventricular vol- 18. Chen R, Gerloff C, Classen J, et al. Safety of different inter-train intervals for ume on brain MR images and severity of cognitive and motor impairment. Radiology 2000; 209: 199–204. stimulation, and short-latency somatosensory evoked potentials. Neurol- ogy 1995; 45: 2255–61. repetitive transcranial magnetic stimulation 26. Maegaki Y, Maeoka Y, Ishil S, et al. Mechand recommendations for safe ranges of anisms of central motor reorganization in 138 Developmental Medicine & Child Neurology 2009, 51 (Suppl. 4): 134–139 pediatric hemiplegic patients. Neuropediat- 30. Muller K, Homberg V, Lenard HG. Magrics 1997; 28: 168–74. netic stimulation of motor cortex and nerve 27. Maegaki Y, Maeoka Y, Ishil S, et al. roots in children. Maturation of cortico- ing the assumptions of the Rasch model to create and interval-level measure. Arch Phys Med Rehabil 2003; 84: 697–705. Central motor reorganization in cerebral motoneuronal projections. Electroencepha- 34. Rosenbaum PL, Walter SD, Hanna SE, palsy patients with bilateral cerebral logr Clin Neurophysiol 1991; 81: 63–70. et al. Prognosis for gross motor function in lesions. Pediatr Res 1999; 45:(4 Pt 1) 559– 31. Masur H, Althoff S, Kurlemann G, Strater cerebral palsy: creation of motor develop- 67. ment curves. JAMA 2002; 288: 1357–63. 28. Nezu A, Kimura S, Uehara S, et al. Magnetic stimulation of motor cortex children: maturity of corticospinal pathway and R, Oberwittler C. Inhibitory period and late muscular responses after transcranial 35. Randall M, Carlin JB, Chondros P, Redmagnetic stimulation in healthy children. Brain Dev 1995; 17: 149–52. dihough D. Reliability of the Melbourne assessment of unilateral upper limb func- problem of clinical application. Brain Dev 32. Heinen F, Kirschner J, Fietzek U, et al. tion. Dev Med Child Neurol 2001; 43: 1997; 19: 176–80. 761–7. 29. Eyre JA, Taylor JP, Villagra F, Smith M, Miller S. Evidence of activity-dependent Absence of transcallosal inhibition in ado- lescents with diplegic cerebral plasy. Mus- 36. Bourke-Taylor H. Melbourne assessment cle Nerve 1999; 22: 255–7. of unilateral upper limb function: construct withdrawal of corticospinal projections 33. Avery LM, Russell DJ, Raina PS, Walter validity and correlation with the Pediatric during human development. Neurology SD, Rosenbaum PL. Rasch analysis of the Evaluation of Disability Inventory. Dev 2001; 57: 1543–54. Gross Motor Function Measure: validat- Med Child Neurol 2003; 45: 92–6. Motor Mapping in CP George F Wittenberg 139