Plasticity of central motor and sensory pathways in a case of unilateral extensive cortical dysplasia

Early brain damage may present advantages over acquired brain damage in both humans ^[1-3] and animals ^[4,5] because central motor or sensory reorganization occurs, as demonstrated by transcranial magnetic stimulation ^[6-9] or magnetoencephalography (MEG). ^[10] We report a patient with unilateral extensive cerebral cortical dysplasia. She had mild hemiparesis and no sensory deficit. Her neurologic signs were milder than expected based on MRI. We sought to determine the presence and site of compensation for central motor and sensory representation of the affected extremity.

Case report.

A 13-year-old girl was referred to our hospital for evaluation of motor function. She was the second child of healthy, nonconsanguineous parents, and was delivered at term with no complications. Motor and mental development were delayed since early childhood. Insufficient spontaneous movement of left hand and leg was noticed since infancy. She could walk unaided at the age of 2 years. At 12 years, her overall intellectual performance was below the normal range according to the Japanese Wechsler Intelligence Scale for Children-Revised test (verbal IQ equals 47, performance IQ equals 48, total IQ equals 42). On examination, she had normal motor function in the right upper and lower extremities. There was a mild left hemiparesis and she was slightly disabled in fine motor tasks with her paretic hand. Mild mirror movement was observed when the child attempted to open or close each hand. She was able to walk unassisted but had a spastic hemiplegic gait. She had no sensory disturbance of her sense of touch, pain, temperature, joint, or vibration on either side. There were no clinical episodes of seizure. She was not taking any drugs. EEG examination revealed a right temporal spike.

Magnetic resonance imaging.

Brain MRI was performed on a 0.5-tesla unit (Yokogawa Medical System) with 10mm section thickness. Various spin-echo images including T_1-weighted (TR/TE 400/15), T_2-weighted (TR/TE 2,000/100), and proton-density-weighted (TR/TE 2,000/40) were obtained in axial and coronal sections. There was an extensively thickened cortex and effacement of sulci indicative of pachygyria or polymicrogyria in the right frontal, parietal, and temporal lobes, including the area where left-hand representation would be expected Figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Proton-density-weighted (left) and T_1-weighted (right) MRI. There is an extensively thickened cortex and effacement of sulci indicative of cortical dysplasia in the right frontal, parietal, and temporal lobes.

Magnetic brain stimulation.

Our patient and 10 normal subjects, consisting of five children (10 to 15 years old) and five adults (27 to 72 years old), were studied. Their informed consent was obtained. Transcranial focal magnetic brain stimulation (TMS) was carried out using an SMN-1100 magnetic stimulator (Nihon Kohden) with a figure-eight-shaped magnetic coil, producing a maximum magnetic field of 1.0 tesla. The center of the coil was moved over different scalp positions 1.0 cm apart. C_z, C₃, and C₄ were defined according to the International 10-20 System. Subjects sat on a comfortable chair in a quiet room and were instructed to keep their arms relaxed. Muscle relaxation was assessed by EMG monitoring during the experiment. Motor evoked potentials (MEPs) were recorded from the abductor pollicis brevis muscle (APB) using surface electrodes positioned over the muscle belly and the tendon. MEPs in both hands were recorded simultaneously, with maximal stimulator intensity. Filters were set from 100 Hz to 5 kHz. Four stimuli were delivered at each position. The peak-to-peak amplitudes of four MEPs at each scalp position were averaged. Onset latencies were measured at each stimulation.

In normal subjects, the highest MEP amplitudes were elicited from scalp positions between 3 cm medial and 0.5 cm lateral to C₃ or C₄, and between 0 cm and 3 cm anterior to C₃ or C₄, which indicated the motor cortex with the highest concentration of corticospinal neurons projecting to APBs. ^[11] MEPs were not seen in the ipsilateral APBs by TMS in normal subjects. In our patient, MEPs following TMS to the left hemisphere were recorded in both the contralateral and ipsilateral APBs. But MEPs of both APBs were never elicited by TMS to the affected right hemisphere. A map of MEP amplitude in each APB showed the same scalp distribution and the highest peak lying on the same scalp site, namely 2.5 cm medial and 2 cm posterior to C₃, which was different from the scalp sites obtained in normal subjects Figure 2. There were no significant differences in MEP amplitudes and latencies of either APB at each TMS. Minimal MEP latency of left APB (20.2 msec) and that of right APB (20.6 msec) were elicited at the same scalp site, where the highest MEP amplitudes were obtained. There was a strong positive correlation between left and right MEP amplitudes elicited at each stimulation (r equals 0.836, N equals 136 stimuli).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Transcranial focal magnetic brain stimulation (TMS). Motor evoked potentials (MEPs) were recorded from both left and right abductor pollicis brevis muscles (APB) simultaneously. At TMS to the unaffected left hemisphere, bilateral APB responses were obtained, with similar latency and amplitude. Neither MEP was elicited by TMS to the affected hemisphere. (A and B) Superimposed MEPs of the right (A) and left (B) APBs by four trials of TMS at each scalp position 1 cm apart over the left hemisphere. (C and D) MEP maps of right (C) and left (D) APBs show similar distributions.

Short-latency somatosensory evoked potentials.

The patient under study and five normal subjects (one child and four adults) reclined in the supine position on a bed in a quiet room, and were asked to relax and keep their eyes closed. The median nerve was stimulated at the wrist. Square-wave pulses of 0.2-msec duration were adjusted to obtain slight thumb twitches at a rate of 5 Hz. The surface electrodes were placed on the scalp contralaterally and ipsilaterally to the stimulus site, according to the International 10-20 System, including F₃, F₄, C₃, C₄, T₃, T₄, P₃, and P₄, and were 2.0 cm apart along the sagittal line crossing C_3-C4 Figure 3. Linked electrodes applied to both earlobes were used as reference. Filters were set from 20 Hz to 3 kHz and two series of 500 responses were averaged separately. In normal subjects, N20s were only seen in the contralateral hemisphere, and N20 amplitudes were higher at P_3-P4 than at C_3-C4 or T_3-T4. P20 and N24 were recorded frontally in both hemispheres. P25, which indicates the central sulcus, ^[12] was located at C_3-C4 or 2 cm posterior to C_3-C4. *Figures 3 and 4* show somatosensory evoked potentials (SEPs) in our patient. Upon right-side stimulation, N20 was elicited with peak amplitude at P₃, and P20 and N24 were recorded frontally at both sides. P25 was found in C₃. N20 was not seen in the ipsilateral hemisphere. There was no abnormality in SEP by right-hand stimulation. In paretic left-hand stimulation, however, N20 was seen only in the ipsilateral hemisphere, but not the contralateral one. N20 amplitudes obtained at F₃ or C₃ were higher than at P₃ or T₃ Figure 4. N20-P30 amplitude was highest at 2 cm anterior to C₃ Figure 3. P20 and N24 were recorded in the contralateral hemisphere (C₄, T₄, and P₄), and P25 was not obtained at any site Figure 4.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Short-latency somatosensory evoked potentials recorded with scalp electrodes along the sagittal line. (A) Location of scalp electrodes 2 cm apart along the sagittal line crossing C₃. (B) On right-hand stimulation, P20 was recorded frontally, and N20 parietally. Because P25 was recognized at C₃, the central sulcus must cross near C₃. (C) On stimulation of the affected left hand, N20 was observed diffusely in the ipsilateral hemisphere. N20-P30 amplitude was highest at 2 cm anterior to C₃. P20 and P25 were not observed.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Short-latency somatosensory evoked potentials recorded from the scalp by electrical right (A) and left (B) wrist stimulation. (A) At right-hand stimulation, N20 (peak latency, 17.4 msec) was seen at contralateral P₃, and P20 (peak latency, 17.4 msec) was recorded at F₃, F₄, and C₄. N24 (peak latency, 20.6 msec) was recorded at F₃ and F₄. (B) At paretic left-hand stimulation, N20 (peak latency, 17.9 msec) was elicited only in the ipsilateral hemisphere, and P20 (peak latency, 17.9 msec) and N24 (peak latency, 21.8 msec) were recorded in the contralateral hemisphere (C₄, T₄, and P₄). N18 (peak latency, 14.9 msec) was observed diffusely in both hemispheres following stimulation at each side.

Discussion.

The APB responses following TMS of the undamaged hemisphere had similar latencies and amplitudes on the two sides. Bates ^[13] reported that patients with congenital, or early acquired, hemiplegia who underwent hemispherectomy had ipsilateral as well as contralateral leg responses when the cortical surfaces of the unaffected hemisphere were stimulated electrically. Carr et al ^[6] showed that bilateral responses to magnetic stimulation were elicited with similar latencies and amplitudes in patients with congenital hemiplegic cerebral palsy, revealing intense mirror movements. There are many reports of the formation of an ipsilateral motor pathway, ^[14-27] and the following mechanisms are proposed: (1) that transient fetal connections persist without the normal regression; (2) that normally occurring ipsilateral projections establish more extensive and enhanced connections; (3) that ipsilateral projections originate from an abnormal branching of the undamaged corticospinal axons; and (4) that other aberrant ipsilateral projections occur. In neonatal rats, many of the neurons in the occipital cortex send into the pyramidal tract a collateral that is later eliminated. ^[17,18] The maps of MEP amplitude in our patient, which showed the peak at the same scalp site in both APBs, indicated that both APB responses originated from the same motor cortex. The first hypothesis (persistence of transient fetal connections) is unlikely because of the same distribution of the MEP maps in both APBs. Wassermann et al ^[19] reported that the MEP maps of the ipsilateral hand muscle following TMS were different from those of the contralateral one in normal subjects, and that onset latencies of the ipsilateral MEP were delayed in comparison with those of the contralateral one. Similar data were observed in rats by microstimulation study. ^[20] Hence, the second hypothesis (normally occurring ipsilateral projections, which may include ipsilateral corticospinal tracts or corticoreticulospinal tracts) can be ruled out. The MEP response of the paretic side in our patient probably originates from an axon sprouting from the corticospinal axon of the undamaged left motor cortex. ^[6,7,21-24] We cannot exclude other aberrant ipsilateral pathways after unilateral cortical lesion in the prenatal or neonatal period. ^[25,26] Reinoso and Castro ^[25] demonstrated intermingled distribution of ipsilateral and contralateral corticospinal neurons after unilateral neonatal cortical lesion in rats. If this mechanism was present in our patient, the similar distribution of both MEP maps can be explained. While MEP amplitude varies considerably at each stimulation even at the same scalp site, there was a strong positive correlation between right and left MEP amplitudes at each TMS. These data may imply a connection between right and left corticospinal axons. Hence we conclude that the ipsilateral potentials observed in our patient must originate from the undamaged motor cortex by axonal sprouting rather than other mechanisms.

Axonal sprouting is maximal after early lesion and declines with age in animals. ^[22] Merline and Kalil ^[27] showed that the intact cortex responded to contralateral pyramidotomy by sprouting into denervated areas of the spinal cord, provided that pyramidotomy took place before the corticospinal axons innervated the spinal target. The corticospinal tracts for the upper extremity are completed by 17 weeks' gestation at the brainstem or spinal cord in the human fetus. ^[28] Pachygyria dates near or after the 13th week of gestation, and polymicrogyria dates near the 20th to 24th weeks. ^[29] Hence we conclude that collateral sprouting had occurred early in gestation in our patient, and innervated the motor neuron pools for the paretic extremities.

N20 was recorded in the ipsilateral hemisphere following paretic left-hand stimulation and must be a near-field potential because of N20 latency similar to that of contralateral hand stimulation. The generator of far-field N18 potential is the thalamus, thalamocortical radiations, ^[30] or the brainstem. ^[31,32] In our patient, N20-N18 interpeak latency was 3.0 msec at left-hand stimulation and 2.5 msec at right-hand stimulation. This interpeak latency indicates that the source of ipsilateral N20 following affected-hand stimulation is the cortical generator rather than the thalamus or thalamocortical radiation. This N20 spread diffusely in the ipsilateral hemisphere. The source of N20 for a paretic hand differs from that for a nonparetic hand because of the difference in N20 or P20 distribution at each hand stimulation and the absence of P25 at paretic hand stimulation. Since N20-P30 amplitude (peak-to-peak) at paretic hand stimulation was highest at 2 cm anterior to C₃, we suspect N20 or P30 originates from the cortical generator anterior to the central sulcus. We cannot completely deny the existence of a primary sensory cortex in the right hemisphere. There are two apparent potentials after N18 (peak latency of the first positive wave was 17.9 msec and that of the negative wave was 21.8 msec) in the right hemisphere at left-hand stimulation. We suspect that the first positive wave may correspond to P20 and the second negative wave may correspond to N24, which are usually elicited frontally at both sides. ^[33] Since the affected cortex has a considerably altered anatomy, the waveform of the contralateral SEP may be altered. We think this unlikely, however, because such changes in morphology of SEP were not observed in patients with lissencephaly. ^[34] Moreover, an absence of cortical components (N20 and other late potentials after N20) was observed in these patients. We cannot deny, of course, the possibility of changed contralateral SEP, because there are only a few reports in the neurophysiology literature concerning unilateral cortical dysplasia or other types of cortical dysplasia. ^[34,35]

The source of N20 of SEP is the primary sensory cortex. ^[12,36-38] Lewine et al ^[10] showed that MEG following electrical median nerve stimulation indicated that the primary sensory cortex was in the contralateral inferior temporal gyrus in a patient with a neonatal infarct involving the middle cerebral artery, who had mild impairment of primary sensation. Jenkins and Merzenich ^[39] found a functional reorganization of cortical representations of the skin surfaces in the cortical zones surrounding focal cortical lesions in adult monkeys. In the SEP study of adult patients with an acquired unilateral cerebral lesion, contralateral N1 responses (peak latency, 19 msec) following affected-side stimulation were reduced or absent when the patients had sensory loss, and normal contralateral N1 responses were obtained when they had no sensory loss. However, ipsilateral N1 responses could not be elicited in either patient group. ^[40] Since our patient had congenital and extensive hemispheric brain lesion without any sensory impairment, the ipsilateral hemisphere might compensate for the sense of the paretic hand. Anand et al ^[41] described the somatosensory pathway as being completed at 20 to 24 weeks' gestation. We suspect that the reorganization of the primary sensory cortex occurs in the ipsilateral hemisphere if brain damage involves almost the whole hemisphere and if it occurs at early gestation, probably up to 20 to 24 weeks.

Our patient showed two different patterns of cortical plasticity; one was axonal sprouting arising from the motor cortex of the unaffected hemisphere, and the other was the development of a new ipsilateral projection of the somatosensory pathway. We can't explain why two different mechanisms of plasticity occurred when the brain was damaged at the same gestational age. The functional distribution of the primary motor and sensory cortex shows a significant difference from a functional map of the normal brain. Penfield and Boldrey ^[42] demonstrated that the cortical points responsive to finger movement were commonly located in the precentral gyri along the central sulci, within 1 cm from the sulci in most subjects. Maps of MEP following TMS of normal subjects in our study, and those of Wassermann et al, ^[11] indicated that the scalp positions where the maximal MEP amplitudes of APB were evoked were located 0 to 4 cm anterior to the C_z-external ear canal line. Levy et al ^[43] reported that maximal MEPs of APB following TMS were usually found over the precentral gyrus when the scalp positions were corrected by MR cortex images. In our SEP study of normal subjects, P25s were elicited in, or in the vicinity of, C_3-C4, 0 to 2 cm posterior to the scalp position where maximal MEPs were obtained. These data imply an abnormal functional distribution of the motor cortex in our patient, which was located approximately 2 cm behind the central sulcus. In addition, normal cortical representation points of the finger sensation are located along the central sulci, mostly behind the sulci. ^[42] In our patient, the cortical representation point of the paretic hand sensation might be located anterior to the central sulcus in the ipsilateral hemisphere. Thus, rearrangement of the functional map of the brain occurs if an extensive hemispheric brain lesion develops in early gestation.