Asymmetric bilateral effect of the supplementary motor area proper in the human motor system

Clinical Neurophysiology 123 (2012) 324–334

Contents lists available at ScienceDirect

Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph

Asymmetric bilateral effect of the supplementary motor area proper in the human motor system Takayuki Kikuchi a,c, Riki Matsumoto b, Nobuhiro Mikuni a,⇑, Yohei Yokoyama a,c, Atsuhito Matsumoto a,c, Akio Ikeda b, Hidenao Fukuyama c, Susumu Miyamoto a, Nobuo Hashimoto d a

Department of Neurosurgery, Kyoto University Graduate School of Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 6068507, Japan Department of Neurology, Kyoto University Graduate School of Medicine, Kyoto, Japan Human Brain Research Center, Kyoto University Graduate School of Medicine, Kyoto, Japan d National Cerebral and Cardiovascular Center, Suita, Japan b c

a r t i c l e

i n f o

Article history: Accepted 11 June 2011 Available online 27 July 2011 Keywords: Human motor areas Functional connectivity Corticospinal tract Single-pulse electric stimulation Functional mapping

h i g h l i g h t s  Single-pulse electric cortical stimulation of human supplementary motor area proper (SMA) elicited motor-evoked potentials (MEPs) and following silent periods (SPs) in extremities bilaterally, but with a longer latency than with primary motor area (MI) stimulation.  The responses and the latencies of MEPs in both contra- and ipsilateral upper extremities to stimulation of the SMA indicate the existence of direct descending volleys from the SMA.  The latency difference of MEPs elicited by SMA and MI stimulation, together with the simplicity of single-pulse electric stimulation, may be useful in differentiating SMA and MI, especially intra-operatively.

a b s t r a c t Objective: This study aimed to clarify the function of human supplementary motor area proper (SMA) by the single-pulse electric stimulation method and its clinical usefulness for SMA mapping. Methods: We studied five patients with epilepsy or brain tumour who underwent invasive functional mapping with subdural electrodes. Single-pulse electric stimulation of primary motor area (MI) and SMA was carried out through pairs of subdural electrodes, and motor-evoked potentials (MEPs) were recorded from surface electromyogram on both sides and also cortico-cortical-evoked potentials (CCEPs) from electrocorticogram. Results: SMA stimulation elicited: (1) MEPs and following silent periods (SPs) in the contralateral upper and lower extremities, (2) SPs with or without minimal MEPs in the ipsilateral upper extremity and (3) CCEPs in the somatotopically corresponding region of the ipsilateral MI. Compared with MI stimulation, SMA stimulation required higher stimulus intensities (mean 14.2 mA (SMA) vs. 8.5 mA (MI)) to elicit MEPs and showed significantly longer onset latencies in upper extremity (range: 4–10 ms). Conclusions: The results demonstrated an asymmetric bilateral effect of human SMA upon the corticospinal pathway. Significance: Single-pulse electric cortical stimulation would be clinically useful for distinguishing SMA from MI. The asymmetric bilateral effect of SMA might be conveyed through the direct descending pathway. Ó 2011 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

1. Introduction The supplementary motor area proper (SMA) has an important role in controlling motor output. This area has preparatory (Wiesendanger, 1986; Deecke and Kornhuber, 1978) and executive ⇑ Corresponding author. Address: Department of Neurosurgery, Sapporo Medical University School of Medicine, S1W17, Chuo-ku, Sapporo 0608556, Japan. Tel.: +81 11 611 2111; fax: +81 11 614 1662. E-mail address: [email protected] (N. Mikuni).

(Humberstone et al., 1997) functions for motor output and coordinates bilateral movements (Brinkman, 1984). Located anterior to the foot primary motor area (MI) anatomically, the SMA has been shown to have association fibres with other motor areas, transcallosal connections with the contralateral SMA and projection fibres descending directly to the spinal cord in animals and humans (Tokuno and Nambu, 2000; Rouiller et al., 1994; He et al., 1995). Although SMA per se contains higher motor-control function for memory-guided, sequential, bimanual movements, connections from SMA to other areas also have been assumed to play important

1388-2457/$36.00 Ó 2011 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2011.06.011

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

roles in its function. The importance of association fibres in motor control has been implicated in the reports of a functional interaction between SMA and MI by human studies applying either transcranial magnetic stimulation (TMS) or direct electric stimulation to SMA (Hamada et al., 2009; Okabe et al., 2003; Matsunaga et al., 2005; Matsumoto et al., 2007). Transcallosal connections between bilateral SMA were suggested to have some roles in bimanual coordination in macaques (Rouiller et al., 1994) and humans (Grefkes et al., 2008). By directly evaluating excitatory (EPSP) and inhibitory (IPSP) postsynaptic potentials (Maier et al., 2002) or measuring facilitatory and inhibitory effects on an electromyogram (EMG) (Boudrias et al., 2006), intracortical microstimulation studies in macaques have shown that projection fibres from SMA convey both excitatory and inhibitory effects to the spinal motoneuron. Only a few attempts have been made using similar direct electric cortical stimulation in humans. A recent study using stereo-electroencephalogram (EEG) electrodes revealed that single-pulse electric stimulation of SMA induced a silent period (SP) without a preceding motor-evoked potential (MEP) in the surface EMG placed on the contralateral deltoid muscle (Rubboli et al., 2006). By contrast, another study using a short train of high-frequency (500 Hz) stimulation showed that SMA can elicit MEPs in several muscles on the contralateral side (Usui et al., 2008). Train stimulation, however, can cause temporal facilitation and activate a broader cortical network. Whether SMA has excitatory, inhibitory or both roles in human motor control is still unresolved; thus, further investigation is needed to fully understand the excitatory and inhibitory effects of SMA in humans. The localisation of SMA is clinically important. Surgical resection of SMA causes global akinesia more prominently on the side contralateral to the resection and impairment of alternating movements of the upper extremities. These symptoms, referred to as SMA syndrome, recover almost completely in several weeks (Laplane et al., 1977; Zentner et al., 1996). It is important to differentiate the SMA syndrome from the irreversible injury of the corticospinal tract from MI, especially during the surgery for lesions in either the dorsal or medial frontal areas. The gold-standard technique for SMA mapping has been high-frequency electric cortical stimulation (HFECS). However, this method cannot be used during surgery which requires the head to be firmly fixed because HFECS can induce massive body movements and potentially cause neck and spinal cord injury. Thus, development of a simpler and safer method for SMA mapping is needed. In this study, we sought to determine the role of SMA in motor output using single-pulse electric cortical stimulation; this technique is considered to activate a relatively limited neuronal network compared with either HFECS or train stimulation. By applying this method and evaluating EMG activity under contraction as well as resting conditions, we could delineate excitatory and inhibitory effects on motor output with higher sensitivity compared with previous electric-stimulation studies. Furthermore, we investigated the clinical relevance of single-pulse electric stimulation as a method for the functional mapping of SMA. Tracts that convey excitatory and inhibitory effects from SMA to limb muscles were also discussed.

2. Methods 2.1. Subjects We enrolled five patients with partial epilepsy or brain tumours in the frontoparietal areas, who underwent invasive functional mapping as a part of their pre-surgical evaluations. All of the patients were right-handed, as assessed by the Edinburgh Handedness Inventory (Oldfield, 1971). Functional mapping was

325

performed using subdural electrodes which were implanted either chronically (patient 1) or intra-operatively (patients 2–5). In the intra-operative condition, patients were awakened before the functional mapping. The subdural electrodes were made of platinum and had a round contact (3 mm in diameter) with an interelectrode distance of 1 cm (ADTECH, Racine, WI, USA or Unique Medical Co., Ltd., Tokyo, Japan). The electrodes were placed on the lateral frontoparietal convexity (herein termed ‘grid A’) and the medial cortex (grid B) to cover the lateral and medial cortical areas that are related to motor control. We used the results of a three-dimensional magnetic resonance imaging (3D-MRI) taken after implantation to confirm the location of the chronic electrode implants (Matsumoto et al., 2004). For the intra-operative implants, we visually inspected the location of the electrodes in reference to the cerebral sulci in the lateral convexity. An intraoperative image-guided frameless localisation system (VectorVision; BrainLab, Heimstetten, Germany) was employed to confirm the location of electrodes on the medial interhemispheric surface (Muacevic et al., 2000). A pair of electrodes for recording the surface EMG was attached on the skin over each of the following muscles in the bilateral extremities: the deltoid (DLT), extensor carpi radialis (ECR), abductor pollicis brevis (APB) and tibialis anterior (TA). We note that the EMG of DLT was not recorded in patient 3, and the EMG was recorded from first dorsal interosseous (FDI) muscle instead of the APB in patient 5. The demographics of the subjects are summarised in Fig. 1. All of the subjects had no motor deficits in their preoperative neurological examinations. Before the implantation or operation, informed consent was obtained according to the Clinical Research Protocols (No. 79 for the chronic implants and No. C-175 for intra-operative implants). These protocols were approved by the Ethics Committee of Kyoto University Graduate School of Medicine. All of the patients had right-sided lesions, and all the electrodes were placed on the right cerebral hemispheres. During the operation, the cortical area including SMA was removed in patients 2 and 5. In patient 3, the cortical area including SMA was preserved but the subcortical area near SMA was removed. Both SMA and the surrounding subcortical area were preserved in patients 1 and 4. After the operation, three patients (patients 2–4) demonstrated transient hemiparesis or coordination disturbance, which was consistent with SMA syndrome, and recovered completely within 1–2 weeks. 2.2. Conventional functional cortical mapping We first recorded cortical somatosensory-evoked potentials (SEPs). For the functional mapping of the lateral cortex, we electrically stimulated the median nerve at a rate of 4 Hz and recorded cortical responses with subdural electrodes (see Supplementary Data, patients 1–5). In all cases, the location of the central sulcus was determined based on the visual inspection of cortical sulci and the phase reversal of the cortical response of SEPs (N20 and P20) (Lüders et al., 1986). Tibial-nerve SEPs were also recorded to identify the primary foot sensorimotor areas (foot SM) (Allison et al., 1996). To identify the location of motor cortices more precisely, we employed 5-train electric-pulse stimulation at a frequency of 500 Hz and recorded the motor-evoked potential (MEP) from surface EMG (Kombos et al., 2001). In patient 1 with chronic implantation, we applied a 50-Hz HFECS instead of 5-train stimulation to all the electrodes and analysed the clinical motor manifestation (e.g., clonic or tonic contractions) to create a precise functional map including the higher motor areas (Lüders et al., 1988). Regarding functional mapping of the medial cortex, we defined the foot MI as the area with a typical MI response, that is, MEP or muscle twitch elicited in the lower extremity contralateral to the side of stimulation. We defined SMA as the area which, upon

326

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

Patient

1

2

3

4

5

Age(year)/sex

27/Male

41/Female

28/Male

32/Male

62/Male L

Diagnosis

Parietal lobe epilepsy (FCD type 2A)

Anaplastic astrocytoma

Frontal lobe epilepsy (FCD type 2B)

Diffuse astrocytoma

Glioblastoma

Recording condition

Chronic

Intraoperative

Intraoperative

Intraoperative

Intraoperative

Antiepileptic drugs (mg/day)

VPA 800 PHT 225 CBZ 800

VPA 800 PHT 600

ZNS 300 PB 105 CLB 5

PHT 300

VPA 800 ZNS 200

L

R

Preoperative MRI

Conventional functional mapping (viewed from top and medial)

Lesion

A01

B08

B05

B08 B16 CentralRsulcus

R

Electrode

A05

A05

A05

A20

A20 A01

B01

B01

R

B16 A05

A01 A05

B09 A01

A16

Electrode pair for single pulse stimulation

Maximum response of tibial SEP

B09

B14 B01

A16 B01 B09

R

B06

Positive motor (left hand)

Positive motor (left foot)

A01

R

A20

B01 B09

A16

Positive motor (right foot)

A electrodes in the lateral frontoparietal cortex, B electrodes in the medial frontoparietal cortex

Fig. 1. The demographics of the patients are listed with the results of the conventional functional mapping. Representative MRI slices are also shown in neurological convention. In the functional brain maps, only the relevant hemisphere is shown. The convexity surface was viewed from the top, and the interhemispheric surface is viewed from the medial side. ‘‘A’’ electrodes are placed on the lateral frontoparietal cortex and ‘‘B’’ electrodes on the medial cortex. Positive motor indicates tonic or clonic contraction of limb muscles by HFECS or MEP in the 5-train stimulation. Electrodes enclosed by a rectangle indicate a pair where the single pulse electric stimulation is applied in this study. Results of the conventional functional mapping (HFECS or 5-train stimulation) are shown as a symbol in each electrode (see legend). FCD, focal cortical dysplasia; VPA, valproate; PHT, phenytoin; CBZ, carbamazepine; ZNS, zonisamide; PB, phenobarbiturate; CLB, clobazam.

stimulation, elicited (1) a positive motor response of the contralateral or bilateral upper extremities, (2) a positive motor response of the bilateral lower extremities or (3) mixture of (1) and (2). Tonic contraction was regarded as a typical SMA response in HFECS (Lim et al., 1994; Ikeda et al., 1999). In intra-operative settings, we employed only response distribution criteria (the former criteria) because tonic contraction was difficult to identify in 5-train stimulation. The results of the conventional functional mapping are

summarised in Fig. 1 and Table 1. MI was determined in the lateral convexity in all patients (one to three pairs of electrodes per patient) and in the medial cortex in patient 3 (B01–B09: foot MI). SMA was determined in all but patient 4 in the frontal medial cortex (one to three pairs of electrodes per patient). In patient 3, B02–B10 was determined as SMA because the 5-train stimulation elicited not only MEP in the left TA but also a negative motor response, that is, a short lapse of the posture in the left arm. In

Table 1 Summary of conventional functional mapping. Patient

Pair of electrodes

HFECS

1

A01–A02 A03–A04 A05–A10 B06–B14 B07–B15 B08–B16

Contralateral Contralateral Contralateral Contralateral Contralateral Contralateral

MEP (5-train stimulation)

Functional localisation

– – – – – –

Lower extremity MI Upper extremity MI Upper extremity MI SMA proper SMA proper SMA proper

2

A03–A04 B01–B02 B03–B04

– – –

Contralateral DLT, ECR Contralateral DLT No response

Upper extremity MI SMA proper

3

A03–A04 B01–B09 B02–B10 B03–B11

– – – –

Contralateral ECR Contralateral TA Contralateral TA, arm lapse No response

Upper extremity MI Lower extremity MI SMA proper

4

A02–A03 B01–B02 B03–B04 B05–B06

– – – –

Contralateral ECR, APB No response No response No response

Upper extremity MI

5

A06–A07 A11–A12 B04–B12 B05–B13 B06–B14

– – – – –

Contralateral Contralateral Contralateral Contralateral Contralateral

Upper extremity MI Upper extremity MI SMA proper SMA proper SMA proper

ankle flexion elbow flexion wrist extension shoulder and bilateral foot abduction shoulder abduction wrist extension

ECR, DLT, DLT, DLT, DLT,

FDI ECR ECR, TA ECR ECR, FDI

327

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

patient 5, although the tibial SEP response (P40) was recorded at B12, the 5-train stimulation of one electrode pair (B04–B12) elicited MEPs in the left DLT, ECR and TA, with larger amplitudes in the upper extremity. While a partial overlap with foot SM cannot be excluded, taken together with our stimulation findings, this pair was regarded as SMA for the following single-pulse-stimulation studies. The tibial SEP finding would be consistent to the result of 5-train stimulation because an early tibial SEP response could be also recorded from SMA (Terada et al., 2009). 2.3. Single-pulse electric cortical stimulation After the completion of the conventional cortical mapping, a repetitive single-pulse electric stimulation was applied to pairs of adjacent electrodes using previously published methodologies (Ikeda et al., 2000; Matsumoto et al., 2007). Briefly, electric square-wave currents with a pulse width of 0.3 ms were delivered with alternating polarity at a frequency of 1 Hz to minimise stimulus artefacts and the polarisation of platinum electrodes. All of the patients were asked to maintain their monitored muscles in a moderately contracted condition. We note that in patient 5, the contraction condition could be performed only in ipsilateral ECR because the patient’s arousal level was insufficient to keep adequate muscle contraction in the contralateral extremities, which were placed beneath the body in the lateral decubitus position. The stimulus intensity was gradually increased until one of the following conditions occurred: (1) MEPs were elicited in at least one targeted muscle or (2) the intensity reached the maximum of 15– 20 mA. We set the stimulus intensity to <40 lC cm 2 per phase to avoid brain-tissue injury (Agnew and McCreery, 1987; Gordon et al., 1990). Recordings were also performed at the same stimulus intensities in the resting condition, in which the patients were asked to relax all of the monitored muscles. A total of 22 pairs of electrodes (three to six pairs per patient) were investigated (Table 2). Anatomically, single-pulse stimuli were delivered in the medial cortex at 14 stimulus sites (two to three per patient) and in the lateral cortex at eight sites (one to

three per patient). Functionally, stimulus sites were located at SMA (eight sites, 0 to three per patient), MI (nine sites, one to three per patient) and the non-functional cortex (five sites, 0 to three per patient). Surface EMG and electrocorticogram (ECoG) were sampled at a frequency of 5000 Hz for the simultaneous recording of MEPs and cortico-cortical-evoked potentials (CCEPs). We note that, in patient 5, the signals were sampled at a frequency of 2000 Hz due to a technical issue. The band-pass filter was set to 0.5–1500 Hz for patients 1–4 and to 0.5–600 Hz for patient 5 in both ECoG and surface EMG. A 32-channel intra-operative monitoring device (Neuromaster MEE-1000, Nihon-Koden, Tokyo, Japan) was used for the electrical stimulation and recording of surface EMG and ECoG. Two stimulus sets were applied to each pair of electrodes. In each stimulus set, at least 30 pulses with alternating polarity were applied to the cortex. 2.4. Analysis Before averaging the EMG data, the raw EMG data were filtered with a low-frequency filter set to 5 Hz and then rectified in an offline manner. Averaging of the EMG data was performed with a time window of 210 ms ( 10 ms to +200 ms), time locked to the single-pulse stimulus. Two stimulus sets were averaged separately to confirm the reproducibility. When MEPs were identified in the same muscle by the stimulation of both lateral and medial cortices in an averaged waveform, the analysis was conducted on individual responses from the raw EMG data (i.e., in each single response). This analysis allowed the characterisation and comparison of the following features between the stimulation of MI and SMA: positive rate of MEP and SP and onset latency and amplitude of MEP. For the analysis of foot responses, these comparisons were done only when foot areas were identified both in the lateral and medial cortices. In the individual response analyses, the first 30 responses of each stimulus set were analysed. Onset latencies were determined with millisecond precision because the sampling of the signals was performed at a frequency of either 2000 or 5000 Hz. If

Table 2 MEP of contralateral muscles induced by single pulse stimulation of MI and SMA proper. Patient

Pair of electrodes

Functional localisation

Intensity (mA)

1

A01–A02 A03–A04 A05–A10 B06–B14 B07–B15 B08–B16

Lower extremity MI Upper extremity MI Upper extremity MI SMA proper SMA proper SMA proper

4 2 6 15 8 12

A03–A04 B01–B02 B03–B04

Upper extremity MI SMA proper No function

A03–A04 B01–B09 B02–B10 B03–B11

4

5

2

3

Deltoid

ECR

+b +

+

+a +

+ +a

7 20 20

+

+ +

Upper extremity MI Lower extremity MI SMA proper No function

15 15 15 15

nr nr nrc nr

A02–A03 B01–B02 B03–B04 B05–B06

Upper extremity MI No function No function No function

15 15 15 15

+

A06–A07 A11–A12 B04–B12 B05–B13 B06–B14

Upper extremity MI Upper extremity MI SMA proper SMA proper SMA proper

20 20 15 15 15

+

nr, No recording. a Largest response among medial cortex stimulation. b Largest response among lateral cortex stimulation. c Lapse induced.

+a + +

APB

FDI

TA

nr nr nr nr nr nr

nr nr nr nr nr nr

+

nr nr nr

nr nr nr

+

+ nr nr nr

nr nr nr nr

+ +

+ +

nr nr nr nr + +b + +a

nr nr nr nr nr

+ +b +

328

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

MEP was elicited by more than two stimulus sites in the same muscle, the stimulus site which elicited the largest MEP amplitude in each surface was selected for the subsequent analysis of individual MEP responses because onset latency was difficult to identify in the contraction condition when the MEP was small and because the site was supposed to have the most reproducible response. The largest MEP had the earliest onset latency in all but two situations (DLT of patient 1 and TA of patient 2 by medial cortex stimulation). In these situations, the largest MEPs had the onset latencies later by 1 ms compared to the earliest onset latencies (see Supplementary Table S1). For the analysis of ipsilateral responses, a high-frequency filter set to 60 Hz was also applied to the EMG data before rectification and averaging to better delineate the onset of SPs. Averaged EMG waveforms were used to determine the onset latencies of MEP and SP of the ipsilateral muscles because the onset of the timelocked ipsilateral EMG activity was not apparent by visual inspection on an individual basis. Averaging of the ECoG data was also performed to obtain CCEPs with a time window of 60 ms ( 10 ms to +50 ms), time locked to single-pulse stimuli. For the analysis of CCEPs, peak latencies of the earliest positive and negative components were measured and compared on averaged waveforms because the onset of those components were difficult to identify on an individual basis due to stimulus artefact in some records. For statistical analysis, a Mann–Whitney U test was used for group comparisons, and a Wilcoxon signed-rank test was used for related samples. A Spearman’s correlation test was used for the correlation analysis. A p-value of <0.05 was considered significant. All statistical analyses were done with Statistical Package for Social Sciences (SPSS) for Windows software version 11.5.1. 3. Results 3.1. Localisation of SMA by single-pulse stimulation In the medial cortex, the stimulation at eight electrode pairs defined as SMA by conventional mapping showed EMG responses and/or clinical motor responses, while no responses were elicited by the stimulation of the non-functional cortex. The distribution of the responses with single-pulse stimulation corresponded with the conventional mapping results regarding the combination of limbs, except one pair in patient 1 (B06–B14: contralateral lower extremity (single-pulse stimulation) vs. bilateral lower and contralateral upper extremities (HFECS)). 3.2. Evoked responses of the contralateral muscles upon single-pulse stimulation of SMA and MI In four patients (patients 1, 2, 3 and 5), a brief contraction or short lapse of the posture in the contralateral limbs was elicited by single-pulse stimulation of the medial cortex. The results are summarised in Table 2, and representative responses are shown in Fig. 2. In patient 1, stimulation at SMA elicited both MEPs and SPs in the contralateral DLT (B07–B15 and B08–B16), ECR (B07–B15 and B08–B16) and TA (B06–B14 and B08–B16). Stimulation at MI elicited MEPs and SPs in the ECR (A05–A10), DLT (A03–A04) and TA (A01–A02). In patient 2, stimulation at SMA (B01–B02) elicited MEPs and SPs in the contralateral ECR and TA, while stimulation at the non-functional cortex (B03–B04) produced SP while no reproducible MEP was elicited. As for stimulation at MI, MEPs and SPs in the contralateral DLT and ECR were elicited by stimulation at A03–A04. In patient 3, stimulation at SMA (B02–B10) and MI (B01–B09) elicited MEP and SP in the contralateral TA in the ab-

sence of detectable EMG changes in the contralateral forearm and hand (ECR and APB). Although EMG activity of the DLT was not recorded in this particular patient, a clinically short lapse of the posture in the contralateral arm was reproducibly induced by the stimulation at the SMA (B02-B10). Stimulation at MI (A03–A04) elicited MEP and SP in the contralateral APB. In patient 4, stimulation at the non-functional cortex (B01–B02, B03–B04 and B05– B06) did not elicit any MEPs. Stimulation at MI elicited MEP at the contralateral DLT. In patient 5, stimulation at SMA elicited MEPs in the contralateral DLT (B04–B12, B05–B13 and B06–B14), ECR (B04–B12 and B05–B13) and TA (B04–B12), while stimulation at MI showed MEPs in the contralateral DLT (A06–A07) and ECR (A06–A07 and A11–A12). In the resting condition, the number of electrode pairs that elicited MEP was small compared with that in contraction condition (six of eight pairs in the lateral cortex stimulation, four of 14 pairs in the medial cortex stimulation). When comparing the eliciting rate of MEP between MI and SMA in the resting condition, stimulation of eight of nine pairs elicited MEP in MI, while that of five of eight pairs elicited MEP in SMA. Comparison of onset latencies between resting and contraction condition was available in five pairs on MI and two pairs on SMA using averaged waveforms. The onset latencies determined from averaged waveforms in the resting condition were longer by 0–2 ms and 2–4 ms in MI and SMA stimulation, respectively (Fig. 2), compared with those in the contraction condition (see Supplementary Table S2).

4. Comparison of MEP and SP characteristics between the stimulation of SMA and MI MEPs were identified in averaged EMGs of the same muscle by the stimulation of both SMA and MI in three patients (patients 1, 2 and 5), and SPs were identified in two patients (patients 1 and 2). In this session, comparison of MEP and SP indices were performed on an individual response basis. The result of the stimulation of B06–B14 in patient 1 was not used for these analyses because of a discrepancy between the results of HFECS and single-pulse stimulation.

4.1. Positive rate of MEP and SP in single-pulse stimulation The positive rate of individual responses was analysed at the selected muscles which showed MEPs upon single-pulse stimulation of both SMA and MI (Table 3). In patient 1, 25 (83.3%), 22 (73.3%) and nine (30.0%) of stimulus pulses elicited MEPs in the DLT, ECR and TA, respectively, in the stimulation of SMA, while nine (30%), 30 (100%) and 30 (100%) pulses elicited MEPs in MI stimulation. In patient 2, MEPs were elicited in the ECR by seven (23.3%) pulses of SMA stimulation and 30 (100%) pulses of MI stimulation. In patient 5, MEPs were elicited in the DLT and ECR by 29 (96.7%) and 24 (80.0%) pulses of SMA stimulation, while 30 (100%) and 29 (96.7%) pulses elicited MEPs in MI stimulation. SPs were elicited in 80– 100% of the stimulus pulses with SMA stimulation compared with 100% with MI stimulation in patients 1 and 2, with the exception of the deltoid muscle (53.3%) in A03–A04 (MI) stimulation in patient 1. Upon SMA stimulation, SPs were elicited more frequently than MEPs (80–100% compared with 30–83.3% in patient 1, 90% compared with 23.3% in patient 2). While SPs always followed MEPs in MI stimulation except for A03–A04 in Patient 1 (MEP only: three; SP after MEP: six; SP without preceding MEP: 10), SPs by SMA stimulation were sometimes induced without preceding MEPs (In patient 1, DLT (B07–B15 stimulation): two; ECR (B08– B16): eight; TA (B08–B16): 15; in patient 2, ECR (B01–B02): 20). Regarding the stimulus intensity to elicit MEP and/or SP, there

329

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

Stimulation pair / intensityAveraged MEP waveform LDLT

Medial surface

B06-B14 (SMA)

15 mA

B07-B15 (SMA)

8 mA

B08-B16 (SMA)

12 mA

LECR

LTA

Patient 1

Lateral surface

A03-A04 (forearm MI) 2 mA

(MI)

A05-A10 6 mA (forearm MI)

4 mA 100 µV

0

100 ms

0

0 LDLT

B01-B02 (SMA)

A01-A02 (foot MI)

LECR

LTA

20 mA

Medial surface B03-B04 20 mA (no function)

Patient 2

Lateral surface

A03-A04 (arm MI)

60 µV

7 mA 0

0 LDLT (not available)

Patient 3

Medial surface

B01-B09 (foot MI)

15 mA

B02-B10 (SMA)

* Arm lapse (+) 15 mA

100 ms

0 LECR

LTA

B03-B11 15 mA (no function)

200 µV 0

0 LDLT

Medial surface

Patient 5

B04-B12 (SMA)

15 mA

B05-B13 (SMA)

15 mA

B06-B14 (SMA)

15 mA

A06-A07 (hand MI)

20 mA

A11-A12 (arm MI)

15 mA

100 ms

0 LECR

LTA

Lateral surface 100 µV 100 ms 0

0

0

MEP waveform (resting condition)

MEP waveform (contraction condition)

Fig. 2. Representative contralateral EMG responses. The broken line represents MEP onset latencies by the stimulation of MI, and the arrowhead represents MEP onset latencies by the stimulation of SMA. Note that in patient 5, the EMG was recorded in the resting condition only.

was a tendency for a higher intensity requirement for SMA stimulation than MI (mean 14.2 mA (SMA) vs. 8.5 mA (MI), p = 0.067).

Mean onset latencies were longer by 4–10 ms with SMA stimulation compared with MI stimulation (see Table 5).

4.2. Onset latency of MEP in single-pulse stimulation

4.3. MEP amplitude

The onset latency of MEP elicited by SMA stimulation was significantly longer compared with that elicited by MI stimulation, except for the TA in patient 1 (Table 4). In patient 1, the mean onset latency of the 30 individual MEP responses by SMA stimulation was 22 ms in DLT, 24 ms in ECR, 38 ms in TA, compared with 11, 16 and 31 ms by MI stimulation, respectively. In patient 2, the mean onset latency by SMA stimulation was 19 ms in ECR compared with 13 ms by MI stimulation. In patient 5, the mean onset latency by SMA stimulation was 18 ms in DLT and 21 ms in ECR compared with 14 and 17 ms by MI stimulation, respectively.

Shown in Table 4, the amplitude of MEP elicited by MI stimulation was significantly larger compared with the amplitude of MEP of the corresponding muscles by SMA stimulation, except for the DLT in patients 1 and 5. 4.4. Responses of the ipsilateral muscles Upon SMA stimulation, SPs were elicited in the ipsilateral ECR in patients 1, 2 and 5 (Fig. 3). The decrement of EMG activity in the averaged waveform began at 27 and 33 ms by B07–B15 and

330

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

Table 3 Induction rate and stimulus intensity of contralateral muscles.

a

Patient

Muscle

MEP positive rate (%)/ stimulus intensity (mA)

SP induction rate (%)/ stimulus intensity (mA)

1

Contralateral DLT Contralateral ECR Contralateral TA

SMA

MI

SMA

MI

83.3/8.0 73.3/12.0 30.0/15.0

30.0/2.0 100/6.0 100/4.0

90.0/8.0 100/12.0 80.0/15.0

53.3/2.0 100/6.0 100/4.0

2 5a

Contralateral ECR

23.3/20.0

100/7.0

90.0/20.0

100/7.0

Contralateral DLT Contralateral ECR

96.7/15.0 80.0/15.0

100/15.0 96.7/15.0

– –

– –

Comparison in resting condition.

B08–B16 stimulation, respectively, in patient 1; at 37 ms by B01– B02 stimulation in patient 2; and at 35 ms by B05–B13 and B06– B14 stimulation in patient 5. In patient 5, a small MEP preceded SP with an onset latency of 28 ms. With MI stimulation in patient 1 (A05–A10), SP was elicited in the ipsilateral ECR. Although the waveform did not have clear onset in one subaverage (light grey line in Fig. 3), the decrement began at 26 ms in the other subaverage (dark grey line in Fig. 3). 4.5. CCEPs at MI in response to SMA stimulation In patients 1, 2 and 5, localised short-latency negative potentials (N1CCEP) were evoked in the lateral convexity by the stimulation of SMA (Table 5). The peak latencies of these components were 9–12 ms (average 10.6 ms). N1CCEP components were observed at all of the corresponding lateral electrodes which were defined as MI of the same muscle. Of note, the maximum response always located at the electrode over MI of the same muscle (Fig. 4). Although earlier positive components (P1CCEP) were observed at latencies of 4–7 ms (average 5.7 ms) in all patients, they were widely distributed and not localised. The comparison of N1CCEP and MEP latencies was also performed for each muscle to investigate a possible indirect pathway from SMA to the spinal cord via MI for producing MEP. Peak latencies of N1CCEP at MI were significantly longer compared with the difference of the MEP-onset latencies elicited by SMA and MI stimulations (average 10.6 vs. 6.3 ms, p = 0.027). P1CCEP components were also compared with the MEP-onset differences; they were not significantly different (p = 0.671). No clear tendency was observed among patients; P1CCEP of patient 1 was shorter compared with the MEP-onset difference (average 5.7 vs. 8.0 ms), while P1CCEP of patient 2 or 5 was longer than the MEP-onset difference (average 6.0 vs. 4.7 ms). 5. Discussion We found that stimulation of the unilateral SMA caused (1) MEPs and SPs to be recorded in the contralateral upper- and low-

er-extremity muscles; (2) SP with or without minimal preceding MEPs to be elicited at the ipsilateral upper-extremity muscle; and (3) these MEPs to have significantly longer latencies and higher stimulus intensities compared with MI stimulation. 5.1. Single-pulse stimulation of the SMA in humans In macaques, intracortical microstimulation of SMA (F3) has been extensively studied. For example, stimulation of SMA elicited a D wave and an I wave in the spinal cord in addition to EPSP and IPSP in the spinal motoneuron (Maier et al., 2002). Another report demonstrated that both facilitatory and inhibitory effects were elicited in the upper-extremity muscles by SMA stimulation (Boudrias et al., 2006). In our human study, single-pulse stimulation of SMA elicited both MEPs and SPs, and SPs were more frequently elicited compared with MEPs. Only a few studies have investigated the effects of single-pulse stimulation of the medial frontal cortex upon final motor output in humans. Several authors of the present study previously investigated single-pulse stimulation of the pre-SMA, and it did not induce either MEP or SP using a procedure quite similar to that detailed in our study; however, the caudal part of SMA, the area posterior to the vertical anterior commissural (VAC) line, was not investigated (Ikeda et al., 2000). A study using stereo-EEG electrodes for medial frontal cortex stimulation demonstrated that only SP was elicited by SMA stimulation (Rubboli et al., 2006). Some of the stimulus sites in the medial cortex described in this study were located anterior to the VAC line (Talairach coordinates: y = 22 to 8 mm), but even stimulation of a site located posterior to the VAC line failed to demonstrate MEP. The discrepancies with our results presumably depend on the modality and intensities of the stimulus. Rubboli et al. applied a single-pulse stimulation of 0.2–3 mA with a 3-ms pulse width using a depth electrode with a 2-mm length and 0.8-mm diameter. Current densities were calculated as 179.1 lC cm–2, whereas the maximum current density of our study was calculated as 84.9 lC cm–2. Although a stronger current density was used in this study, they used depth electrodes which deliver electric currents in a more restricted fashion compared with subdural electrodes

Table 4 Comparison of MEP between MI and SMA proper stimulation. Patient

1

a b

Muscle

Contralateral DLT Contralateral ECR Contralateral TA

MEP amplitude (lV (SD))

MEP onset latency (ms (SD)) SMA

MI

p value

SMA

MI

p value

22 (3.8) 24 (5.7) 38 (11.4)

11 (1.3) 16 (1.0) 31 (3.5)

<0.001 <0.001 0.109

414 (113)b 498 (135) 381 (93)

278 (49) 1198 (339) 941 (255)

<0.001 <0.001 <0.001

2

Contralateral ECR

19 (2.8)

13 (1.3)

<0.001

154 (63)

465 (166)

<0.001

5a

Contralateral DLT Contralateral ECR

18 (1.5) 21 (2.4)

14 (1.7) 17 (1.9)

<0.001 <0.001

206 (147)b 191 (195)

21 (11) 239 (93)

<0.001 0.038

Comparison in resting condition. Larger MEP by SMA stimulation.

331

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334 Table 5 Comparison of CCEP and MEP latencies. Patient

Muscle

Stimulation

Peak latency of CCEP component

Latency difference of MEP

SMA to MI (N1, ms)

SMA to MI (P1, ms)

SMA stimulation–MI stimulation (ms)

B07–B15 B08–B16 B08–B16

12 12 9

6 7 4

10 8 6

1

Contralateral DLT Contralateral ECR Contralateral TA

2

Contralateral ECR

B01–B02

11

6

5

5

Contralateral DLT Contralateral ECR

B04–B12 B05–B13

9 11

6 6

4 5

N1, P1: first negative and positive component of CCEP on MI by the stimulation of SMA.

(Yeomans, 1990). The absence of MEP at the higher current density in this study could be ascribed to the use of a different type of stimulating electrode. Another study used train stimulation to SMA with subdural grids and reproducibly elicited MEP, although SP was not investigated (Usui et al., 2008). In this study, the authors used a unidirectional 5-train stimulation of a 0.3-ms pulse width at a rate of 500 Hz and stimulus intensity of 1–10.4 mA (max. 1103 lC cm–2), indicating that a stronger current was needed for excitation of the direct corticospinal pathway in SMA. Indeed, there was a tendency for SMA to require a higher stimulus intensity than MI for eliciting MEP in our study. Compared with conventional cortical stimulation, single-pulse stimulation of SMA induced the same response regarding the com-

bination of limbs except for two electrode pairs (B06–B14 in patient 1 and B03–B04 in patient 2). This corroboration indicates the clinical validity of single-pulse stimulation for functional mapping of SMA. At one site of SMA (B06–B14) in patient 1, single-pulse stimulation induced MEP only in the contralateral TA, while HFECS induced bilateral leg tonic contraction and contralateral shoulder abduction. The absence of upper and ipsilateral lower-extremity responses by single-pulse stimulation could be ascribed to a more restricted activation compared with HFECS, indicating the possibility of detailed functional mapping with single-pulse stimulation. The similarity of latencies of MEP by the stimulation of B06–B14 and A01–A02 (Fig. 2) might indicate that this medial cortical area was the foot MI, not SMA. The response in the shoulder and ipsilateral

MI A05-A10 LECR(contralateral)

Patient 1

SMA proper B06-B14

B07-B15

B08-B16

RECR (ipsilateral)

0

26 ms

0

0

27 ms

0

33 ms

50 µV 100 ms

B01-B02

B03-B04

LECR (contralateral)

Patient 2

RECR (ipsilateral)

20 µV 100 ms 0

A11-A12

B04-B12

37 ms

0

B05-B13

B06-B14

LECR (contralateral, resting condition)

Patient 5

RECR (ipsilateral)

0

0

0

35 ms

0

35 ms 20 µV 100 ms

Fig. 3. Comparison of ipsilateral EMG responses to contralateral responses. The broken line represents the onset of SP in the ipsilateral (right) ECR in response to the stimulation of SMA.

332

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

A

C

B

A05 11.0 ms

Ipsilateral

Contralateral

A01 DLT 5.8 ms B05

STIMULATION

ECR A05 19 ms

A20

B01 A20

TA A01 A16

A16

200 µV 25 ms

20 µV -

25 ms

Fig. 4. Cortico-cortical evoked potentials (CCEPs) by the stimulation of SMA from a representative case (patient 2). (A) Averaged ECoG waveform (CCEP) recorded from the lateral cortex; (B) averaged MEP waveform simultaneously obtained with CCEP; (C) electrode configuration. Note single pulse stimulation of B01–B02 elicited MEP at the contralateral (left) ECR and maximum CCEP (N1) at the MI of the somatotopically corresponding region (A3). The peak and onset latencies are shown at the maximum N1 response (A3) and MEP (LECR), respectively.

foot by HFECS might be induced by the propagated electric current from the foot MI to the SMA. At B03–B04 in patient 2, single-pulse stimulation induced SP without reproducible MEP, while 5-train stimulation evoked no response. Although we judged that this area has no function, this negative response in single-pulse stimulation might indicate that this area was a part of SMA. As 5-train stimulation was applied only at the resting condition, we might fail to detect a negative motor response of SMA. In patient 3, the single-pulse stimulation of SMA (B02–B10) induced MEP in the TA with a latency similar to MI (B01–B09), while a short lapse of the posture in the contralateral arm was observed (Fig. 2). It is possible that the propagation of electric current similar to B06–B14 in patient 1 occurred in this area, but this is unlikely because single pulse stimulation elicited a response of the contralateral arm similar to that of conventional mapping (5-train stimulation). The unexpected similarity in latencies could be caused by the simultaneous coverage of the foot MI and the SMA with B02– B10 electrodes, indicating a close proximity of the foot MI and the SMA. To discriminate these two foot motor areas, single-pulse stimulation with smaller spacing of electrodes, such as with a 5mm inter-electrode distance, is warranted. 5.2. Characteristics of MEP parameters in the contralateral muscles In our single-pulse stimulation study, onset latencies of MEPs were significantly longer with SMA stimulation compared with MI stimulation. The difference in onset latencies between MI and SMA stimulation was 6–10 ms in patient 1, 5 ms in patient 2 and 4–5 ms in patient 5. In macaques, direct descending volleys from the SMA have a slower conduction velocity compared with that from MI. In an intracortical microstimulation study with epicortical recording from the cervical and thoracic cord, the velocity was reported to be 61 m s–1 from SMA compared with 79 m s–1 from MI (Maier et al., 2002). In the same study, a collision technique was also used between MI and SMA and showed no clear involvement of MI in the spinal response induced by the stimulation of SMA. Another study using intracortical microstimulation also demonstrated a similar difference between SMA and MI in that the stimulation of SMA elicited the facilitation of EMG of the wrist muscle at a latency of 15.3 ms compared with 9.3 ms by the stimulation of MI (Boudrias et al., 2006). The longer MEP latency by SMA stimulation indicates two possible routes for the excitation of spinal motoneurons: direct slower descending volley and indirect corticospinal pathway through MI.

We attempted to solve this issue by performing a CCEP study in the same patients. CCEP has been used to study interareal connections in vivo in patients undergoing invasive evaluations (Matsumoto et al., 2004, 2007). Compared with diffusion tractography, this technique physiologically traces the cortico-cortical connections and provides spatial as well as temporal information on connectivity. In our study, stimulation of SMA elicited maximum CCEP response at MI of the same somatotopy as we previously reported (Matsumoto et al., 2007). Although the generator mechanisms are not fully known, CCEP seems to be generated by orthodromic activation of cortico-cortical projection neurons (see Matsumoto et al., 2004, 2007 for a detailed discussion). We cannot guarantee that the first negative peak represents the very first volley into MI; nevertheless, CCEP was the only available technique to study interareal connections in the intra-operative setting and, thus, we employed the negative peak (N1CCEP) of physiological significance (i.e., restricted distribution at and around MI of the same somatotopic representation). The N1CCEP latency at MI by SMA stimulation (10.6 ms, average of six stimulations in three patients) seems to be a reasonable measure in humans, considering the latency of induced spikes in MI pyramidal neurons (4.3 ms) in response to intracortical microstimulation of SMA in macaques (Tokuno and Nambu, 2000) and the difference in brain size between the two species. Compared with the N1CCEP latency at MI, the latency difference (6.3 ms) of the MEP onsets between SMA and MI stimulation was shorter by 4.3 ms on average, suggesting the existence of direct slower descending volleys from the SMA to the spinal cord. Despite its broad distribution, P1CCEP latency was comparable to the latency difference of MEP induced by SMA and MI (Table 5). Although this finding might indicate the involvement of MI in generation of MEP by SMA stimulation (i.e., indirect pathway through MI), no consistent tendency was found between patients (shorter P1 latency compared with MEP difference in patient 1, in contrast to longer P1 latency in patients 2 and 5). We compared the eliciting rates of MEP and SP between SMA and MI stimulation in two patients (patients 1 and 2) and found that the rates were lower for SMA stimulation than for MI stimulation (MEP: 30.0–73.3% vs. 100%; SP: 80.0–100% vs. 100%), with the exception of the recording from the DLT in patient 1. The same findings were observed in macaques: intracortical microstimulation of SMA elicited EPSPs in 36/75 (48.0%) spinal motoneurons and that of MI in 74/84 (88.1%) (Maier et al., 2002). In that study, a detailed tracer investigation demonstrated denser corticospinal projections from MI than SMA to the motor nuclei of the hand

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

muscles in the spinal cord, providing an anatomical substrate of the lower rate of SMA effects. Our finding in the stimulation of A03–A04 in patient 1, i.e. lower eliciting rates of MEP and SP at the contralateral DLT by MI stimulation compared to the SMA stimulation, seems to be due to suboptimal stimulation for the DLT region of MI because of (1) lower MEP amplitude at DLT by MI stimulation compared with SMA stimulation (B07–B15) and (2) HFECS findings at A03–A04 that showed left-elbow flexion without apparent shoulder movement. Nevertheless, the short onset latency of MEP (11 ms) supports the view that the response was induced by MI directly through the corticospinal tract. For SMA stimulation, the eliciting rate of SP was higher than that of MEP at the same intensity, although the comparison was limited in two patients. This is consistent with the TMS study of SMA showing that SP has a lower threshold than MEP and can occur independently of MEP (Hallett, 1995). 5.3. Response of ipsilateral muscles We also assume that the EMG response was mediated by a direct descending pathway from SMA because of the ipsilateral response and its latency. For example, in three patients (patients 1, 2 and 5), SPs with or without minimal preceding MEP were elicited on the upper-extremity muscles ipsilateral to the side of SMA stimulation. The beginning of the EMG decrement or the peak latency of minimal preceding MEPs ranged from 33 to 37 ms. In humans, electric stimulation of SM has been reported to elicit SP without preceding MEPs in the contralateral upper-extremity muscle with a latency of 44.5 ms in the APB (Ikeda et al., 2000) and 40–48 ms in the DLT (Rubboli et al., 2006). The latencies of ipsilateral SPs were comparable to or even shorter than that of SM. Because interhemispheric conduction takes at least 5–10 ms (Terada et al., 2008), it seems unlikely that the impulse travelled through the contralateral SM to reach the ipsilateral muscle. It would seem more reasonable to conclude that the impulse was conveyed directly from the SMA to the spinal cord through the corticospinal tract. 5.4. Coverage of cerebral cortex and muscles In the current study, we could not thoroughly cover the medial and lateral cortices and the distal and proximal muscles because of the limitation of number of electrodes available for simultaneous ECoG and EMG recording and of time allowed for experimental research in the clinical practice. There remains the possibility that we failed to completely locate the MI and SMA. In fact, we could not identify SMA and foot MI in patients 3 and 4. In other patients, in whom the SMA and MI was successfully identified, there was a possibility that we only investigate a limited portion of these functional cortices. In addition, we applied only 5-train stimulation intra-operatively, which is certainly inferior to HFECS for functional mapping, especially for higher motor areas. Further study is needed to clarify the precise location and functional representation of SMA and MI by using wider electrode placement and monitoring of proximal muscles, such as neck, truncal and hip muscles. 5.5. Effects of antiepileptic drugs All of the five patients received antiepileptic drug medication during the study. Zonisamide has been reported to reduce MEP amplitude, carbamazepine to lengthen SP and valproate to reduce intracortical excitability, according to TMS studies of MI (Ziemann et al., 1996; Joo et al., 2008; Cantello et al., 2006). Zonisamide was administered to patients 3 and 5, carbamazepine to patient 1 and valproate to patients 1, 2 and 5. It could not be ruled out that these drugs affected the results of this study. In addition, the motor

333

threshold might be elevated by these medications. Nevertheless, the comparison of MEP/SP parameters within the patient (e.g., paired associations in Table 4) remains significant under this condition.

6. Clinical implications The present study clearly showed differences in MEP latencies and in the distribution of responses between SMA and MI stimulation. The latency difference might be useful in differentiating MI from SMA. Usui et al. (2008) demonstrated a similar conclusion by recording MEPs in resting condition applying a train of electric pulses to the SMA and MI. The distribution of EMG responses was also different. By SMA stimulation, responses of the ipsilateral muscles, in addition to the contralateral muscles, were induced, while no response was elicited by the stimulation of MI. In addition, stimulation of SMA elicited CCEPs at the somatotopically corresponding region of MI, corroborating the result of our previous study (Matsumoto et al., 2007). Clinically, the differences between SMA and MI in MEP latency, the distribution of EMG responses, as well as connectivity findings could be useful for the functional mapping of SMA. Practically, single-pulse electric stimulation did not require additional electrode placement, and each online average from a given stimulus site only took a minute (two sets of 30 responses). This time requirement is reasonably short to perform in the intra-operative condition, where the conventional HFECS may carry the risk of injury. The smaller electrode spacing, however, would be needed to discriminate between the foot MI and SMA, as even single-pulse stimulation, the weakest stimulation among the mapping techniques, did show the same MEP latencies at the TA with the two adjacent stimulus sites. In this study, patient 4 demonstrated transient hemiparesis after surgery despite the fact that the SMA and the surrounding subcortical area were preserved. On the contrary, patient 5 had no post-surgical deficit after the removal of SMA. The reason for these inconsistencies might be the brain oedema after surgery and the extent of its removal. In patient 4, we removed the subcortical area very near to functional fibres around the SMA. Indeed, white-matter electrical stimulation of the removal cavity performed as a clinical investigation elicited MEP on DLT and ECR (data not shown). We thought the postoperative symptom of this patient was due to the oedema of this subcortical area. Conversely, in patient 5, only limited part of the SMA (B06 and B14) was removed. The patient was supposed not to demonstrate any postoperative symptom because the main part of the SMA (B04–B12 and B05–B13) was preserved. Our findings might also implicate the involvement of a direct descending pathway from the SMA to the spinal cord in the pathophysiology of SMA syndrome. Patient 3 demonstrated impairments of alternating movements of the upper extremities more prominently on the side contralateral to the resection after the operation. His symptoms were compatible with the SMA syndrome, without removing the medial cortex. Although it is possible that removing the dorsal premotor area itself caused the SMA syndrome, this seems unlikely because cortical stimulation of that area induced no motor response. It is more likely that the direct descending pathway from the SMA was injured in the deep white matter and caused the SMA syndrome. A collision study between MI and SMA, as well as white-matter stimulation, is needed to clarify this alternative mechanism of the SMA syndrome. During the intra-operative awake setting, once we encountered acute symptoms compatible with the SMA syndrome, it is clinically important to differentiate injuries of the MI itself or the corticospinal pathway from the MI from those of the SMA or the corticospinal pathway from the SMA. Whether the SMA is removed or not, the SMA

334

T. Kikuchi et al. / Clinical Neurophysiology 123 (2012) 324–334

syndrome is more likely if MEP is preserved upon stimulation of the MI of the corresponding extremity. In the present study, organic lesions were located close to the SMA in some patients (patients 2 and 5). This proximity might have prolonged the latency of MEP; however, the latencies of MEPs in these patients were comparable to those of patient 1 who had a lesion in the parietal lobe. In the epileptic patients, it is also possible that a plastic change was introduced in their motor system due to the presence of an epileptic focus near SMA. Further investigation is warranted to determine the significance of SMA function and establish the clinical usefulness of our methods. Acknowledgements This work was supported by Grants-in-Aid for Scientific Research (C) 20591713, (C) 20591022 and Grant-in-Aid for Challenging Exploratory Research 22659258 from the Japan Society for the Promotion of Science (JSPS) and The Research Grant (19A-6) for Nervous and Mental Disorders from the Ministry of Health, Labour and Welfare. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.clinph.2011.06.011. References Agnew WF, McCreery DB. Considerations for safety in the use of extracranial stimulation for motor evoked potentials. Neurosurgery 1987;20:143–7. Allison T, McCarthy G, Luby M, Puce A, Spencer DD. Localization of functional regions of human mesial cortex by somatosensory evoked potential recording and by cortical stimulation. Electroencephalogr Clin Neurophysiol 1996;100: 126–40. Boudrias MH, Belhaj-Saif A, Park MC, Cheney PD. Contrasting properties of motor output from the supplementary motor area and primary motor cortex in rhesus macaques. Cereb Cortex 2006;16:632–8. Brinkman C. Supplementary motor area of the monkey’s cerebral cortex: short- and long-term deficits after unilateral ablation and the effects of subsequent callosal section. J Neurosci 1984;4:918–29. Cantello R, Civardi C, Varrasi C, Vicentini R, Cecchin M, Boccagni C, et al. Excitability of the human epileptic cortex after chronic valproate: a reappraisal. Brain Res 2006;1099:160–6. Deecke L, Kornhuber HH. An electrical sign of participation of the mesial ‘supplementary’ motor cortex in human voluntary finger movement. Brain Res 1978;159:473–6. Gordon B, Lesser RP, Rance NE, Hart Jr J, Webber R, Uematsu S, et al. Parameters for direct cortical electrical stimulation in the human: histopathologic confirmation. Electroencephalogr Clin Neurophysiol 1990;75:371–7. Grefkes C, Eickhoff SB, Nowak DA, Dafotakis M, Fink GR. Dynamic intra- and interhemispheric interactions during unilateral and bilateral hand movements assessed with fMRI and DCM. Neuroimage 2008;41:1382–94. Hallett M. Transcranial magnetic stimulation. Negative effects. Adv Neurol 1995;67:107–13. Hamada M, Hanajima R, Terao Y, Okabe S, Nakatani-Enomoto S, Furubayashi T, et al. Primary motor cortical metaplasticity induced by priming over the supplementary motor area. J Physiol 2009;587:4845–62. He SQ, Dum RP, Strick PL. Topographic organization of corticospinal projections from the frontal lobe: motor areas on the medial surface of the hemisphere. J Neurosci 1995;15:3284–306. Humberstone M, Sawle GV, Clare S, Hykin J, Coxon R, Bowtell R, et al. Functional magnetic resonance imaging of single motor events reveals human presupplementary motor area. Ann Neurol 1997;42:632–7. Ikeda A, Ohara S, Matsumoto R, Kunieda T, Nagamine T, Miyamoto S, et al. Role of primary sensorimotor cortices in generating inhibitory motor response in humans. Brain 2000;123(Pt. 8):1710–21.

Ikeda A, Yazawa S, Kunieda T, Ohara S, Terada K, Mikuni N, et al. Cognitive motor control in human pre-supplementary motor area studied by subdural recording of discrimination/selection-related potentials. Brain 1999;122(Pt. 5):915–31. Joo EY, Kim SH, Seo DW, Hong SB. Zonisamide decreases cortical excitability in patients with idiopathic generalized epilepsy. Clin Neurophysiol 2008;119: 1385–92. Kombos T, Suess O, Ciklatekerlio O, Brock M. Monitoring of intraoperative motor evoked potentials to increase the safety of surgery in and around the motor cortex. J Neurosurg 2001;95:608–14. Laplane D, Talairach J, Meininger V, Bancaud J, Orgogozo JM. Clinical consequences of corticectomies involving the supplementary motor area in man. J Neurol Sci 1977;34:301–14. Lim SH, Dinner DS, Pillay PK, Lüders H, Morris HH, Klem G, et al. Functional anatomy of the human supplementary sensorimotor area: results of extraoperative electrical stimulation. Electroencephalogr Clin Neurophysiol 1994;91:179–93. Lüders H, Dinner DS, Lesser RP, Morris HH. Evoked potentials in cortical localization. J Clin Neurophysiol 1986;3:75–84. Lüders H, Lesser RP, Dinner DS, Morris HH, Wyllie E, Godoy J. Localization of cortical function: new information from extraoperative monitoring of patients with epilepsy. Epilepsia 1988;29(Suppl. 2):S56–65. Maier MA, Armand J, Kirkwood PA, Yang HW, Davis JN, Lemon RN. Differences in the corticospinal projection from primary motor cortex and supplementary motor area to macaque upper limb motoneurons: an anatomical and electrophysiological study. Cereb Cortex 2002;12:281–96. Matsumoto R, Nair DR, LaPresto E, Bingaman W, Shibasaki H, Lüders HO. Functional connectivity in human cortical motor system: a cortico-cortical evoked potential study. Brain 2007;130:181–97. Matsumoto R, Nair DR, LaPresto E, Najm I, Bingaman W, Shibasaki H, et al. Functional connectivity in the human language system: a cortico-cortical evoked potential study. Brain 2004;127:2316–30. Matsunaga K, Maruyama A, Fujiwara T, Nakanishi R, Tsuji S, Rothwell JC. Increased corticospinal excitability after 5 Hz rTMS over the human supplementary motor area. J Physiol 2005;562:295–306. Muacevic A, Uhl E, Steiger HJ, Reulen HJ. Accuracy and clinical applicability of a passive marker based frameless neuronavigation system. J Clin Neurosci 2000;7:414–8. Okabe S, Hanajima R, Ohnishi T, Nishikawa M, Imabayashi E, Takano H, et al. Functional connectivity revealed by single-photon emission computed tomography (SPECT) during repetitive transcranial magnetic stimulation (rTMS) of the motor cortex. Clin Neurophysiol 2003;114:450–7. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97–113. Rouiller EM, Babalian A, Kazennikov O, Moret V, Yu XH, Wiesendanger M. Transcallosal connections of the distal forelimb representations of the primary and supplementary motor cortical areas in macaque monkeys. Exp Brain Res 1994;102:227–43. Rubboli G, Mai R, Meletti S, Francione S, Cardinale F, Tassi L, et al. Negative myoclonus induced by cortical electrical stimulation in epileptic patients. Brain 2006;129:65–81. Terada K, Usui N, Umeoka S, Baba K, Mihara T, Matsuda K, et al. Interhemispheric connection of motor areas in humans. J Clin Neurophysiol 2008;25:351–6. Terada K, Umeoka S, Baba K, Sakura Y, Usui N, Matsuda K, et al. Generators of tibial nerve somatosensory evoked potential: recorded from the mesial surface of the human brain using subdural electrodes. J Clin Neurophysiol 2009;26:13–6. Tokuno H, Nambu A. Organization of nonprimary motor cortical inputs on pyramidal and nonpyramidal tract neurons of primary motor cortex: an electrophysiological study in the macaque monkey. Cereb Cortex 2000;10:58–68. Usui N, Terada K, Baba K, Matsuda K, Tottori T, Umeoka S, et al. Extraoperative functional mapping of motor areas in epileptic patients by high-frequency cortical stimulation. J Neurosurg 2008;109:605–14. Wiesendanger M. Recent developments in studies of the supplementary motor area of primates. Rev Physiol Biochem Pharmacol 1986;103:1–59. Yeomans JS. Principles of brain stimulation. New York: Oxford University Press; 1990. Zentner J, Hufnagel A, Pechstein U, Wolf HK, Schramm J. Functional results after resective procedures involving the supplementary motor area. J Neurosurg 1996;85:542–9. Ziemann U, Lonnecker S, Steinhoff BJ, Paulus W. Effects of antiepileptic drugs on motor cortex excitability in humans: a transcranial magnetic stimulation study. Ann Neurol 1996;40:367–78.