Neural activity of supplementary and primary motor areas in monkeys and its relation to bimanual and unimanual movement sequences

Pergamon

PII:

Neuroscience Vol. 89, No. 3, pp. 661–674, 1999 Copyright  1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/99 $19.00+0.00 S0306-4522(98)00348-0

NEURAL ACTIVITY OF SUPPLEMENTARY AND PRIMARY MOTOR AREAS IN MONKEYS AND ITS RELATION TO BIMANUAL AND UNIMANUAL MOVEMENT SEQUENCES O. KAZENNIKOV,* B. HYLAND,† M. CORBOZ,‡ A. BABALIAN,§ E. M. ROUILLER§ and M. WIESENDANGER0¶ *Institute of Problems in Information Transmission, Russian Academy of Sciences, Ermolova Street, Moscow, Russia †Department of Physiology, Medical School, University of Otago, Dunedin, New Zealand ‡Service de Neurologie, CHU, CH-1211 Gene`ve, Switzerland §Institute of Physiology, **University of Fribourg, CH-1700 Fribourg, Switzerland 0Motor Systems Laboratory, Department of Neurology, Inselspital BHH-M130, CH-3010 Bern, Switzerland Abstract––A chronic single-unit study of motor cortical activity was undertaken in two monkeys trained to perform a bimanually coordinated task. The hypothesis was tested that the supplementary motor area plays a specific role in coordinating the two hands for common goal-oriented actions. With this objective, a special search was made for neurons that might exhibit properties exclusively related to bimanual task performance. Monkeys learned to reach for and to pull open a spring-loaded drawer with one hand, while the other hand reached out to grasp food from the drawer recess. The two hands were precisely coordinated for achievement of this goal. Monkeys also performed, in separate blocks of trials, only the pulling or grasping movements, using the same hands as in the bimanual task. Task-related activity of 348 neurons from the supplementary motor area and 341 neurons from the primary motor area, each examined in the bimanual and in both unimanual tasks, was recorded in the two hemispheres. Most neurons from the supplementary motor area were recorded within its caudal microexcitable portion. Contrary to expectation, the proportion of neurons with activity patterns related exclusively to the bimanual task was small, but somewhat higher in the supplementary motor area (5%) than in the primary motor cortex (2%). Another group of neurons that were equally modulated during the bimanual as well as to both unimanual task components might also contribute in controlling bimanual actions. Such ‘‘task-dependent’’ rather than ‘‘effector-dependent’’ activity patterns were more common in neurons of the supplementary motor area (19%) than of the primary motor cortex (5%). Bilateral receptive fields were also more numerous among the supplementary motor area neurons. However, a large majority of neurons from primary and supplementary motor areas had activity profiles clearly related only to contralateral hand movements (65% in the primary motor and 51% in the supplementary motor area). A similar group of neurons showed an additional slight modulation with ipsilateral movements; they were equally common in the two areas (14% and 16%, respectively) and their significance for bimanual coordination is questionable. Summed activity profiles of all neurons recorded in the primary and supplementary motor areas of the same hemisphere were compared. The modulations of the three histograms, corresponding to the two unimanual and the bimanual tasks, were similar for the two motor areas, i.e. prominent with bimanual and contralateral movements and weak with ipsilateral movements. It is concluded that the supplementary motor area is likely to contribute to bimanual coordination, perhaps more than the primary motor cortex, but that it is not a defining function for the former cortical area. Instead, it is suggested that the supplementary motor area is part of a callosally interconnected and distributed network of frontal and parietal cortical areas that together orchestrate bimanual coordination.  1999 IBRO. Published by Elsevier Science Ltd. Key words: bimanual coordination, supplementary motor area, motor cortex, single units, motor skill, monkey.

Many motor skills of primates require cooperation of both hands in which the movements of the two limbs need to be precisely coordinated in space and time. ¶To whom correspondence should be addressed. **Institution where the work was carried out. Abbreviations: BIM task, bimanual task; ICMS, intracortical microstimulation; M1, primary motor cortex; SMA, supplementary motor area.

The brain mechanisms involved in this important coordination are still poorly understood. The clinical literature indicates that damage to some cortical areas may interfere with proper execution of bimanual skills in a manner that cannot be accounted for completely by deficits in performance of the limb contralateral to the damaged hemisphere. This is the case with lesions of the parietal association cortex (e.g., Ref. 26) and mesial frontal cortex, either with

661

662

O. Kazennikov et al.

or without involvement of the corpus callosum (see Ref. 68 for review). Although callosal interconnections in the two hemispheres would seem an obvious substrate for such coordination, callosal transection alone has rather slight and only transient effects on practiced bimanual everyday tasks.50 Another possible anatomical substrate for bimanual coordination is that an area or areas in each hemisphere contribute to coordination of the two upper limbs by virtue of bilateral projections, so that each side of the brain can address both sides of the body. In the human brain, the dominant hemisphere may perhaps exert a hierarchically superior role in bimanual coordination.41 Penfield and Jasper49 and Travis63 found that only bilateral, but not unilateral, lesions of the supplementary motor area (SMA) produced significant deficits (clumsy movements, forced grasping, spasticity). They therefore suggested that the SMA is a bilaterally organized system. However, there is no mention about any deficits in bimanual coordination as a consequence of SMA lesions in these reports. Similarly, early brain imaging studies in humans indicated that the SMA is bilaterally activated during unilateral hand and finger movements (e.g., Ref. 53). In their pioneering study on single-unit activity of the SMA, Brinkman and Porter4 found that neurons were typically activated with movements of either limb. However, a bilateral functional relationship of a cortical area does not in itself necessarily indicate a role in coordination, especially of the kind where different movements of the two limbs must be coordinated for goal achievement. More direct evidence for the role of the SMA in bimanual coordination comes from lesion experiments by Brinkman,3 who found a bimanual dyscoordination, somewhat akin to mirror movements, which occurred in two monkeys subjected to large unilateral mesial–frontal lesions. Further evidence for the bimanual hypothesis of the SMA was provided by Tanji et al.,61 who discovered neuronal activity in the SMA specifically related to simple bimanual key presses finger flexion/extension movements. However, the interesting results emerging from the above lesion and unit recording experiments do not settle the question about the specific role of the SMA in coordinating bimanual movements. Firstly, in Brinkman’s lesion study, the abnormal coordination of the two hands appeared only during the second postoperative week, after the monkey’s immediate postoperative disturbances of clumsiness and abnormal hand postures had already disappeared. Possibly, the late occurrence of mirror movements was an adaptive (though not successful) strategy of the monkey to retrieve the food from the holes in the foodboard, rather than a primary post-lesion deficit.36 Secondly, the bimanual key presses studied by Tanji et al. are relatively simple mirror movements of the two hands, without need for precise temporal

coordination across different muscle groups in the two limbs that typically occurs during goal-directed bimanual actions. In view of the above-mentioned uncertainties, it was deemed important to further clarify the possible role of the SMA in natural, coordinated actions of the two hands as required for everyday tasks of primates, including humans. In this perspective, the present study was undertaken to examine activity patterns of SMA neurons in monkeys performing a complex bimanual task, which we had shown to be performed normally with exquisite temporal coordination between the two hands.37 We asked three questions. (1) Is the SMA involved in bimanual coordination? (2) If so, are there numerical differences between SMA and primary motor cortex (M1) neurons potentially involved in bimanual coordination? (3) If the SMA is involved, how important is its role in comparison with other functions of the SMA?37 Although there is considerable evidence that the SMA is less lateralized than the M1, it remains to be seen whether the SMA is a key structure in coordinating bimanual goal-directed movements. Some of the results have been presented at a conference.67 The same behavioural task was also used in a parallel study in monkeys subjected to mesial–frontal lesions.36 EXPERIMENTAL PROCEDURES

The experiments were carried out in accordance with the European Communities Council Directive. The Fribourg State Commission for Animal Experimentation approved the experimental protocol and made on-site visits. All efforts were made to minimize animal suffering. Behavioural paradigm and data acquisition Two monkeys (referred to as Monkey P and Monkey F, Macaca fascicularis, 2.5 and 3 kg) were born in the animal house of the Physiology Institute in Fribourg. As they reached adulthood, they were operantly conditioned for the present study. Before learning the task, the monkeys were trained to accept handling by the experimenters and to be carried from their animal quarters to the experimental room without struggling. The task was a bimanual goal-oriented synergy, as reported recently.37 Briefly, the animals, sitting in a primate chair, reached out with the left arm to grasp the knob of a baited drawer and pulled it open. As the left hand moved towards the drawer, the right hand followed the leading left hand with only a short delay and moved at a speed that was timed accurately with the left hand opening the drawer. The instance the right hand entered the opened drawer for taking out the food morsel with the precision grip is referred to as ‘‘goal achievement’’ (see inset of Fig. 1). During the grasp, the left hand had to keep the spring-loaded drawer open until the right hand withdrew. Note that, in this task, fine coordination of the two hands is not imposed; the monkeys could, in principle, have held open the drawer for an extended period, and entered the drawer with the other at any time to obtain the food. However, they did not perform in this way. Instead, we have shown that, faced with such a task, monkeys naturally and spontaneously employ an extremely fine-tuned and invariant temporal coordination between the two hands as they approach the goal.37 On this basis, the task seems ideal for investigations into central mechanisms of bimanual coordination.

Role of SMA neurons in bimanual coordination To initiate a trial, the animals had to place the two hands on to separate starting platforms in front of two closed windows in a vertical panel separating the monkey from the drawer. Sensors on the two platforms detected that the two hands were in the correct starting position. After a waiting period of 1 s, the screens of both windows above the platforms were opened, thus providing access to the baited drawer. At the same time, two light emission diodes, positioned above the windows, were illuminated. Thus, the monkeys initiated each trial at their own pace by placing their hands on the two platforms. The bimanual synergy was characterized by a typical sequence of discrete left- and right-hand events that were detected by sensors for assessing the time structure of the synergy. The grasping hand lagged behind the leading pulling hand by 200–250 ms. The two limb movements were quite asymmetric; the leading hand rapidly reached out for opening the drawer while the right hand moved more slowly for grasping the food as the drawer opened. The monkeys usually promptly initiated the next trial by replacing the two hands on to the starting platforms while eating the food. The whole movement sequence was performed in about 1.5–2.5 s. Left- and righthand event markers (rectangular pulses of different amplitudes for each event) were fed into two separate channels of a laboratory computer. These signals served as triggers for aligning neural activity in dot rasters and peri-event time histograms. In addition, an analogue signal of the drawer displacement was recorded and digitized. In order to evaluate the specificity of any involvement of neural activity in the bimanual task (BIM task), monkeys were also trained to perform separately either the unimanual pulling component or the unimanual grasping component of the full task. Both monkeys spontaneously and consistently used the left hand for pulling the drawer and the right hand for picking up the food. We therefore refer to these as the LEFT task and RIGHT task. In the LEFT task, only the left window opened and the left arm had to pull the drawer, which was then automatically fixed when it reached the open position. The left arm could then release the drawer and pick up the food with the left hand. When the left hand withdrew from the food well, the fixation was released and the drawer automatically closed. Thus, in the LEFT task, only the reach-and-pull movement corresponded to the left-hand movement of the BIM task, while the later picking of food with the left hand was additional, but sufficiently delayed so as to not interfere with histogram analyses centred on reach onset or time of handle touch. During the entire unimanual LEFT task, the right hand had to remain in the resting position on the platform. In the RIGHT task, only the right window opened. The right hand reached through the window and picked up the food as the drawer opened automatically by a computer-controlled motor. During this time, the left hand had to remain on the platform. Thus, the right hand mimicked, in all respects, the movement of this hand during the BIM task (except for the requirement for temporal coordination with the other limb). The most difficult part of training was to teach animals to keep the nonperforming hand quiet on the starting platform in the unimanual tasks. Otherwise, insight into the pull-and-grasp task was rapid. The BIM, LEFT and RIGHT tasks, consisting of blocks of 20–30 trials each, were performed in varying order. We used highly palatable biscuit morsels as food rewards, and the animals were highly motivated, easily performing 300–400 trials per session. Surgery Stable performance of the three different tasks was attained after three to four months of daily training. The monkeys were then deeply anaesthetized with pentobarbital (30 mg/kg) and a rectangular recording chamber (40 mm20 mm) was chronically implanted under sterile

663

conditions. This allowed transdural microelectrode penetrations in the cortical arm areas of the M1 and SMA of both hemispheres (Fig. 2; also see Ref. 8). Recordings Following recovery from surgery, daily recording sessions (1–2 h) were held, usually with one penetration aimed at either the M1 or SMA. Tungsten microelectrodes had typical impedances of 2–5 MÙ. At the majority of recording sites, it was possible to clearly discriminate two action potentials. The activity was fed into two separate amplitude discriminators and the output pulses were entered into two spike channels of the computer. We checked the quality of spike discrimination by making waveform comparisons on a digital storage oscilloscope with pretrigger viewing and fast sweep speed. The sampling interval of the computer was set at 2 ms. Responsiveness of neurons to gentle touch stimuli or passive joint movements was examined whenever possible. After completion of the recording in a given track, conventional intracortical microstimulation (ICMS) was applied at sites of recorded unit activity (also see Ref. 8). Responding muscle groups and threshold intensities were noted. Intensity was generally limited to 30 µA, but in SMA tracks, the intensity was occasionally increased up to 100 µA. Units were recorded in blocks of 20–30 trials while the monkey performed the BIM, RIGHT and LEFT tasks. In order to check consistency, one or more of the tasks were sometimes repeated. Data analysis Analogue data from the drawer movement, left- and right-hand event times, and spike data were recorded online by the same computer controlling the task. The analysis time started 500 ms before both hands were placed on the two starting platforms (=zero time). When the monkey left the platform with one or both hands before the imposed waiting time of 1 s, the trial was aborted and a reset occurred. The analysis time for the correct bimanual or unimanual movement sequence lasted for 3 s. Dot rasters, peri-event histograms of neuron activity (single units and populations of units as detailed in the Results) and averages of drawer position were constructed off-line and aligned to various behavioural events, as indicated in the scheme of Fig. 1. The activity of all neurons was examined in histograms centred on the events ‘‘window(s) open(s)’’ and ‘‘movement onset’’ of the left leading hand. The event ‘‘movement onset right’’ was used routinely for the RIGHT task, and also when recordings were made in the left hemisphere. Additional events for alignment of neuronal discharges were also used in an effort to identify the best association of the activity changes with various events or phases of the task: ‘‘passing through window’’ (left or right), ‘‘touch knob’’, ‘‘onset of drawer displacement’’ (left hand), ‘‘drawer fully opened’’ (left hand), ‘‘drawer released’’ (left hand), ‘‘pick in’’ and ‘‘pick out’’ (right index finger enters and leaves food well). Dot rasters were primarily used to decide whether an early peak or trough of activity was related to ‘‘window opening’’ or to ‘‘movement onset’’ (left or right). To this end, trials were ordered with increasing intervals between window opening and movement onset to reveal shifts in the raster, depending on whether activity was related to the visuo-acoustic signal ‘‘window opens’’ or to ‘‘movement onset’’ (also see Fig. 6). Anatomical and functional attributes of each unit were entered into a spreadsheet (Excel). The recorded data from a given neuron were accepted into the database when the ratio of peak task activity to spontaneous activity was two or more, for at least one of the three tasks; the average ratio was 11.2. Numerical differences (ICMS thresholds, response latencies) were subjected

664

O. Kazennikov et al.

Fig. 1. Experimental paradigm. Upper panel: temporal sequence of left-hand events, with drawer displacement (above) and right-hand events (below) during performance of bimanual synergy. Goalrelated events are marked by asterisks. Lower panel: two frames of the bimanual task (redrawn from serial photographs). to t-tests. The significance of differences in the proportions of neurons distributed in the two areas was analysed using the Chi-square test.

movement onset=500120 and 52090 ms; goal reaching=820160 and 950180 ms.

Histology After completion of recording sessions, fine steel pins were implanted at known stereotaxic coordinates; two were near the central sulcus of the right- and left-hand areas, and two near the caudal and rostral borders of the recording sites in the SMA. The monkeys were then deeply anaesthetized with pentobarbital (50 mg/kg) and perfused through the heart with 300 ml of a physiological saline solution, followed by 4000 ml of a 4% paraformaldehyde fixative solution. The brain was removed and photographed with landmarks on the cortical surface. It was immersed in 10% sucrose for one day and then in 30% sucrose for three days. The solution was buffered at a pH of 7.4. The two hemispheres were cut on a freezing microtome in the coronal plane at 60 µm thickness. Nissl-stained sections were dehydrated and coverslipped. The position of the electrode tracks was checked and reconstructed from serial sections, viewed at 100 magnification, as described previously.55

Overview of neuronal properties and intracortical microstimulation effects. In total, 689 neurons (341 in the M1 and 348 in the SMA) that showed activity changes related to at least one of the three tasks have been included in this study. The distribution of these neurons in the two hemispheres of Monkey P and Monkey F is displayed in Fig. 2, together with the electrode penetrations relative to the cortical surface. Single units were isolated at depths ranging from 0.5 to 7.5 mm; the largest density of recorded M1 and SMA neurons was between 2 and 3 mm. In addition to histological localization, the two populations of M1 and SMA neurons recorded had properties that conformed to well-established findings reported for these areas, as did the responses to ICMS applied in the two areas, confirming that the sample was representative of SMA and M1 neurons. The average discharge rate, measured during the delay period, was similar for the two populations (5.6 spikes/s in the M1 and 6.2 spikes/s in the SMA). Long-lead activity, as exemplified in Fig. 3, i.e. preparatory activity that evolved during the waiting period while the hands were resting on the platforms and before opening of the access windows, was encountered in both areas. The incidence was higher in the SMA than in the M1 (19% vs 9%, respectively), as reported previously in other tasks.54,60 In movement-related neurons, i.e.

RESULTS

General Behavioural task performance. Both monkeys executed the three tasks (BIM, LEFT, RIGHT) reliably and in a similar manner to that described before.37 The temporal evolution of the bimanual synergy was also comparable in Monkey P and Monkey F; the main values (meanS.D. relative to window opening) were, respectively: initiation (left hand)=24040 and 30040 ms; right-hand

Role of SMA neurons in bimanual coordination

665

Fig. 2. Top view of the explored cortical areas with entry points of microelectrode. In Monkey P, 66 penetrations were in the M1 and 68 in the SMA; for Monkey F, the corresponding numbers were 41 and 60. Positioning was made stereotaxically with reference to the two rectangular frames representing the borders of the recording chamber in the two monkeys. The outlines of the central sulcus (CS) and arcuate sulcus (AS) are drawn approximately on the basis of several visible reference points measured during implantation of the chamber. Arrows indicate the anterior border of the microexcitable caudal SMA; this was slightly caudal to the genu (G) of the arcuate sulcus, indicating that most of the recordings were from ‘‘SMA proper’’. The number (n) of recorded neurons is indicated for each cortical area.

those with activity changes occurring after the opening of the windows but before movement onset, the lead time for the bimanual synergy was similar among M1 and SMA populations. Even if only units best related to movement onset of the left leading hand were selected, no significant inter-areal differences in lead time were obtained, as expected on the basis of previous work.8 Activity peaks (or more rarely troughs) could be related to any of the discrete behavioural events which marked epochs of the movement sequence (movement initiation, reaching, pulling drawer, picking food). Most neurons exhibited activity changes in relation to one or two events of the bimanual synergy (70% of neurons, in equal proportions for the M1 and SMA). A minority (30% for both areas) had distinct activity changes related to three events, or had more tonic changes of activity over the entire sequence. Finally, consistent with previous reports,28,69 responsiveness to gentle cutaneous stimuli and to passive joint movements was significantly higher among M1 than SMA neurons (96% vs 57%, respectively).

In order to be guided in the outlay of somatotopic representations, ICMS was effected at 498 sites in the two monkeys (284 sites in the M1 and 214 in the SMA). As expected,8 mean thresholds were significantly lower in the M1 (2411 µA for proximal and 2014 µA for distal effects) than in the SMA (4518 µA for proximal and 5221 µA for distal effects). However, in the latter area, the proportion of proximal effects was higher than in the M1. The arrows in Fig. 2 indicate the border between the caudal microexcitable forelimb representation of the SMA (usually less than 50 µA) and an adjoining rostral subdivision in which ICMS effects were rarely observed, but if present most often involved saccadic eye movements. Associations performance

of

neuronal

activity

during

task

Since the aim of this study was to identify neural activity specifically related to the requirements of bimanual coordination, we compared activity in the

666

O. Kazennikov et al.

alone (i.e. LEFT or RIGHT task, depending on the hemisphere recorded from). In addition, there was no detectable activity during performance of the limb ipsilateral to the hemisphere. This activity pattern, typically encountered in neurons of the motor cortex, is thus completely lateralized. Activity during bimanual performance probably relates only to the contralateral limb. This was the most common type of neuron in both cortical areas and was more represented in the M1 than in the SMA. Complex contralateral relation. These neurons were similar to those of subclass (a), in that they only showed activity during bimanual and contralateral movement, but they were either more, or less, activated in the BIM task than with the unimanual contralateral movement. The differences were moderate and the pattern of discharge was comparable in the two tasks, as illustrated in the example of Fig. 4. Movement parameters were not always the same in the contralateral and the BIM tasks, and this may easily explain the difference in the amount of activation. These neurons were almost equally distributed in the M1 and in the SMA.

Fig. 3. Long-lead SMA and M1 neurons with ramp increases of activity while the monkey waits for the windows to open, giving access to the drawer (event used for alignment). (A) Activity of single SMA neuron (18 trials). (B, C) Activity of clusters of neurons were cumulated (eight M1 and five SMA neurons, respectively). Response amplitude of each bin was normalized for the probability of discharges.

BIM task with that seen in the LEFT and RIGHT tasks. This should allow us to dissociate activity simply relating to unimanual components of the task from that which might have been specially involved in coordinating the two hands in space and time, i.e. independently of the effectors. Neural activity patterns examined across the three tasks led to the following two broad categories: (A) lateralized neurons with activity changes exclusively related to the contralateral and the BIM task, but not or only spuriously to the ipsilateral task, and (B) bimanual neurons with exclusive relations with the BIM task or with all three tasks. The first category of neurons was by far the largest in both the M1 and SMA. The numbers and details of all neuronal subclasses are listed in Table 1, which also contains an additional category of neurons (C: unclassified neurons) not fitting into the two main categories of interest. The subclasses are briefly characterized below. Lateralized neurons Simple contralateral relation. In these, activity during the BIM task was indistinguishable from that associated with movement of the contralateral limb

Partial bilateral relation. The activity pattern of these neurons was again similar to that in subclass (a), except that a much smaller involvement was also seen during movements of the ipsilateral limb. This subclass of neurons was similarly represented in the M1 and in the SMA. Bimanual task-related neurons Complete bilateral relation. In these neurons, similar levels of activity modulations were seen during BIM, LEFT and RIGHT tasks. The neurons were therefore activated at specific phases of task performance whether one or both limbs were used, and independent of which limb was used in the unimanual cases. Such activity patterns are illustrated in Fig. 5 for an M1 and an SMA neuron. Of particular interest in this figure is the SMA neuron, which was equally activated during the goal-related epoch of picking up the food morsel from the opened drawer, independent of whether the right or the left hand was grasping (Fig. 5B). This ‘‘grasp neuron’’ was thus task dependent rather than effector dependent. The majority of neurons in this class had activity peaks related to movement initiation of all three tasks, as shown in the example of an M1 neuron in Fig. 5A. Exclusively related to bimanual vs unimanual performance. This category is small, but most relevant to the present study. It consists of cells that were either only activated during the BIM task, and not during LEFT or RIGHT tasks, or not activated during the BIM task, despite being active during performance of either one or both of the unimanual components. Two examples of BIM-exclusive SMA

Role of SMA neurons in bimanual coordination

667

Table 1. Neuron categories A–C with subclasses a–f and their distribution in the primary motor cortex and supplementary motor area M1 Relation of activity changes

SMA

n

%

n

%

(A) Lateralized neurons Subclass (a) BIMYcontralateral Subclass (b) BIM > or < contralateral Subclass (c) BIM, contralateralnipsilateral Subtotal

120 103 48 271

35 30 14 79

83 95 56 234

24 27 16 67

(B) BIM task-related neurons Subclass (d) BIMYcontralateralYipsilateral Subclass (e) BIM exclusive Subtotal

10 5 15

3 2 5

49 17 66

14 5 19

(C) Unclassified neurons Subclass (f)

55

16

48

14

341

100

348

100

Grand total

Fig. 4. Activity patterns of lateralized neurons recorded from the M1 (30 trials) and from the SMA (20 trials). Both neurons had similar activity patterns during execution of the contralateral task and the BIM task, but the number of spikes was in these cases smaller during performance of the unimanual contralateral task [=subclass (b)]. Activity was aligned to event ‘‘window(s) open(s)’’. Note the right shift in the dot display when trials were ordered from the shortest (bottom) to the longest (top) movement onset times. This indicates that the activity of both neurons was related to movement onset and not to the signal of windows opening.

neurons with excitatory activity bursts are illustrated in Fig. 6. Overall, the areal distribution among M1 and SMA neurons listed in these two neuron categories did not differ significantly (Chi-square test). Unclassified neurons. These neurons, also equally distributed in both areas, had activation patterns that

did not fit into any of the above subclasses. The majority exhibited weak tonic changes in the discharge rate over the entire task performance; others had sharp bursts at a short latency (<100 ms) after window opening (providing access to the drawer). These bursts may thus have been responses to the auditory/visual stimulus generated by the opening of the window.

668

O. Kazennikov et al.

Fig. 5. Examples of two effector-independent neurons of category B [subclass (d)]. The activity of the M1-R neuron (A; 15 trials) was well related to initiation of the movement in all three tasks, i.e. independently of the arm initiating the movement. The activity of the SMA-R neuron (B; 23 trials) was closely related to the goal-related behavioural act of grasping food from the drawer, again independently of the involved hand (‘‘grasp neuron’’).

Fig. 6. Examples of two effector-independent SMA neurons of subclass (e) with rare patterns of activity changes exclusively related to the BIM task, but not (or only minimally) to the unimanual RIGHT and LEFT tasks): (A) 25 trials; (B) 20 trials.

Laterality of somatosensory receptive fields and of intracortical microstimulation responses Ipsilateral or bilateral receptive fields were found in 14 of 196 M1 neurons tested (=7%) and in 23 of 78

SMA neurons tested (=29%). The difference between the two motor areas was significant (P<0.001). Forelimb ICMS effects were contralateral at all cortical sites tested, for both the M1 (n=252) and SMA (n=85).

Role of SMA neurons in bimanual coordination

669

Fig. 7. Population histograms comparing summed activity of all M1 and SMA neurons of the left hemisphere recorded while Monkey F executed the three tasks. Alignment is with event ‘‘window(s) open(s)’’. The bin amplitudes have been calculated as spikes per trial and number of neurons (n=80 for the M1 and n=95 for the SMA, as indicated in Fig. 2).

Task relationship and laterality in population averages Lateralization is a general property of a neuron population as compared to the more specific functional attributes of each single neuron. Are there areal differences in lateralization between SMA neurons and M1 neurons? To answer the above question, we analysed grand averages of pooled data of all M1 and SMA neurons, computed separately for the three tasks and for the right and left hemispheres. In Fig. 7, the neuron populations were taken from the two areas of the left hemisphere while Monkey F performed the three tasks. The activity profiles were most modulated in the BIM and RIGHT tasks (reaching and picking food). Clearly, the action of the left hand, ipsilateral to the recorded neurons, also produced a substantial, but weaker modulation, and this was the case for both the M1 and SMA. Equivalent histograms from the right hemisphere (not shown) revealed a more pronounced activity change around movement onset, reflecting a prominent involvement of the right hemisphere in initiating the sequence with the contralateral leading hand. Again, activity modulated with the ipsilateral righthand task was weak, but clearly present in both areas. The apparent similarity of activity profiles

between the M1 and SMA populations was supported by a correlation analysis (BIM task). Temporally corresponding values of each bin in the M1 and SMA histograms were subjected to linear correlations, separately for the left and right hemispheres. The correlation coefficients were rleft =0.936 and rright =0.815. The visual impression of ‘‘different’’ activity patterns exhibited by neurons of the left and the right hemispheres was accordingly expressed in relatively low correlation coefficients between right and left histograms of M1 neurons (rr vs l =0.615) and of SMA neurons (rr vs l =0.574). DISCUSSION

Previous support for the bimanual coordination hypothesis of the SMA in subhuman primates was derived from single-unit experiments4,61 and from a lesion study.3 The fact that, in the present study, we found a proportion of SMA neurons in which the pattern of activity was potentially determined by the requirement of coordinated activity of the two hands to achieve a goal could be interpreted as supporting the bimanual coordination hypothesis. However, despite actively searching for such cells in a demanding bimanual task, their number was small, raising

670

O. Kazennikov et al.

the question of the extent to which the SMA is involved in bimanual coordination. This issue is discussed further below. Problems with the bimanual coordination hypothesis First, it should be noted that the large majority of the recorded neurons was well lateralized, either exclusively with the contralateral effectors or with only a minor involvement during execution of the ipsilateral RIGHT or LEFT task. Such neurons were frequent among both motor areas, as also became evident in the population histograms exhibiting relatively modest but significant modulations with the ipsilateral tasks in both the SMA and M1. These neurons therefore do not point to a special involvement of the SMA in bimanual coordination; minor activity changes during ipsilateral task performance are more likely to reflect spurious movements and/or postural adjustments of the resting arm which may be difficult to detect, even in electromyogram recordings. With regard to the completely lateralized neurons, a subclass had somewhat different degrees of activation (but the same pattern) in the BIM task as compared to the contralateral unimanual tasks. This could be taken to indicate that a higher activity in the BIM task is due to the extra workload imposed by bimanual coordination. However, we believe that this is unlikely, because the activity was often also less pronounced in the BIM task and also because the distribution was similar in the two motor areas. The differences are better explained by subtle differences in parameters of execution. We argue that the role of the SMA in bimanual coordination might be best predicted by neurons showing specific association with the bimanual task. In the second category of neurons, a first subclass was only activated during bimanual action, while others were silent during such coordination, despite the fact that they were active during unimanual movement. This confirms that such cells, previously only described for simple finger keypress movements,61 are also found in the SMA during complex, goal-directed bimanual coordination. Although we were particularly searching for such neurons, their number was small in both areas (2% in the M1 and 5% in the SMA). A second subclass of neurons in this category was characterized by similar activity changes in BIM, LEFT and RIGHT tasks, which match a group of cells identified by Tanji et al.61 Such activity patterns might represent a higher-order coding for grasping or reaching, independent of the executive limb. Activity by neurons with such a bilateral relationship might lead to mirror-type movements and/or stabilization of postural support by the trunk. A bilateral involvement of neurons would be useful for controlling simultaneous and similar movements of the two hands, albeit at a relatively low level of coordination.

Consistent with this are observations in human brain-imaging studies that the SMA is also activated ipsilaterally during unilateral task performance, although to a lesser extent than contralaterally.29 We reason, however, that the mere presence of neurons with activity related to contralateral, bimanual as well as to ipsilateral movements is a much weaker argument for a bimanual coordinative role of such a neuron population, since there is no specificity in their association with bimanual movements. Moreover, neurons with an ipsilateral involvement appear to be widely distributed. For example, it was reported by Rizzolatti et al.52 that in area F5 of the lateral premotor cortex ‘‘...almost the totality of neurons fired during motor acts performed with either hand’’. Ipsilateral projections from the M1 to the spinal cord have long been known to exist,45,46 and were considered to be chiefly involved with actions of proximal and trunk muscles. More recently, a cluster of M1 neurons was also discovered by Aizawa et al.,1 who concluded that ‘‘...this efferent zone between the traditional face and digit areas of M1 is a site utilized for the execution of bilaterally organized hand movements’’. Similarly, Gentilucci et al.18 reported that in area 4 (=F1) of one monkey ‘‘...22 out of 37 recorded neurons fired during movements of both hands’’. They further commented that it is not unlikely that ‘‘the number (of bimanual neurons) in this rostral location of area 4 has been underestimated in the past’’. Functional imaging techniques in humans also revealed ipsilateral activations with various unimanual motor tasks in a region of the postcentral sulcus,14 the insula,31 the premotor cortex35 and even the primary motor cortex, especially of the dominant hemisphere.40 Contrasting with results of the present single neuron study, which provides only poor support for the bimanual hypothesis of the SMA, are neurological case studies. Laplane et al.43 were the first to report on patients with unilateral lesions in the mesial frontal cortex who had difficulties in coordinating both hands. Similar dys-coordination was subsequently confirmed by a number of authors.13,23,42,56 A special case is the load-lifting paradigm when the active hand is unloading the postural load-bearing hand. This action requires an anticipatory stabilization of the postural arm. It was found that lesions in the mesial frontal cortex impaired this coordinated anticipatory stabilization.65 More complex bimanual disturbances, such as the alien hand syndrome,19 apraxia,66 mirror movements7 and intermanual conflict,32 have also been reported. However, the latter lesions were either poorly documented or very large, involving a number of adjoining areas, such as cingulate motor areas, premotor or mesial prefrontal areas, and frequently also the corpus callosum (for review, see also Ref. 68). Unilateral resections of the SMA for therapeutic reasons are usually followed by a transient contralateral poverty of voluntary movements and of spontaneous speech, but no case among

Role of SMA neurons in bimanual coordination

28 recently operated patients was reported to have had postoperative difficulties in bimanual coordination.71 On the other hand, cortical lesions of structures outside the SMA, such as the posterior parietal cortex, the frontal lobe and even the temporal lobe, as well as of the cerebellum and the basal ganglia, may also interfere with bimanual coordination.5,6,16,20,25,27,34,44,58,64,70 In contrast to the human data cited above, and despite the conclusion generally drawn from the lesion experiments by Brinkman3 that this structure is vital for bimanual coordination, when considered overall, lesion experiments in monkeys do not add much in support of the bimanual hypothesis. The mirror-like movements observed by Brinkman3 occurred only in the second postoperative week, when the minor contralateral symptoms, known to occur after a unilateral SMA lesion, had already disappeared. In a study parallel to the present one, we have re-investigated the effects of unilateral and bilateral lesions of the SMA in monkeys performing the same bimanual drawer pull-andpick task.36 We could confirm the occurrences of secondary changes in task performance described by Brinkman,3 which we interpreted as adaptive strategies for compensating immediate, but transient, postoperative deficits. Bimanual coordination, quantified in terms of synchronization and temporal correlation between the hands was, however, preserved, both in the two unilaterally lesioned cases and in one bilaterally lesioned monkey. In the latter this was true, even in the face of dramatic slowing and increased variance of the individual limb movements. Recently, Kermadi et al.39 used the technique of reversible inactivation of the SMA by means of multiple microinfusions of muscimol, thus avoiding the emergence of adaptive changes that often occur after irreversible lesions. The experiments were repeated in several sessions, also including simultaneous microinfusions into the SMA of both hemispheres. The effects of SMA dysfunction were assessed in two monkeys taught to perform a similar bimanual drawer pulland-grasp task as used in the present study. The striking result was that the monkeys did not initiate the bimanual task if not prompted by a go signal, whereas the bimanual synergy was performed normally when it was triggered by an external signal; the authors concluded ‘‘...that the SMA is not necessary for bimanual coordination’’. Other studies in SMAlesioned monkeys were not designed to assess bimanual coordination, and subtle deficits may have been missed. However, it seems clear that the rather dramatic bimanual coordination problems that have been described in neurological patients (discussed above) would have been noticed. In none of the reports (except in Ref. 3) on the effects of irreversible unilateral and bilateral SMA lesions in monkeys9,12,62,63 (sometimes enlarged to also include additional cortical areas24,47,51) was a deficit in bimanual coordination mentioned. In summary, the available data, including the present results, together

671

provide only weak support for the idea that the SMA is of pre-eminent importance in controlling bimanual coordination. It should be noted, however, that the vast majority of SMA neurons was recorded in the ‘‘SMA proper’’; we can therefore not exclude that a higher proportion of ‘‘pre-SMA’’ neurons is tuned for the control of bimanual coordination. Alternative mechanism for bimanual coordination In the literature on bimanual coordination, a number of implicated mechanisms has been discussed, dealing with self-organizing dynamic patterns38,59 and with dynamics of coupled oscillators.22 For the control of temporal order in movement sequences, a popular idea is that feedforward proprioceptive signals are used for appropriate timing (see Ref. 33 for review). For example, there is indirect evidence from deafferented patients2,21 and from mechanical perturbation studies10,11 that signals generated by the moving limb are indeed necessary for proper temporal goal achievement. Using a bimanual task, Flanders and Cordo15 were able to demonstrate that ‘‘...subjects used kinesthetic input from the right arm to produce accurate increases or decreases in voluntary motor commands to the muscle of the left elbow’’. The authors interpreted this as ‘‘a reparametrization of new motor commands’’. In human studies, it has also been reported57 that somatosensory signals are rapidly conducted also to the ipsilateral somatosensory cortex. Accordingly, in monkeys bilateral receptive fields have been found in somatosensory cortex neurons30 and SMA neurons (present study). In our bimanual task, vigorous proprioceptive ‘‘cue’’ signals could arise when the left arm started to move or when it touched the drawer handle. Alternatively, cueing by proprioceptive afferents during movement of the leading arm and updating of the command delivered to the other arm could be a continuous dynamic process. Thus, as an alternative proposition to the original SMA hypothesis, bimanual goal synchronization may be achieved by feedforward transmission of proprioceptive signals from the leading arm to the ipsilateral cortex. This would allow for updating the timing of the delayed arm, perhaps with reinforcement by a transcallosal transfer of an efference copy.17 CONCLUSIONS

The data reported here confirm that there are neurons with activity patterns consistent with their having a special role in bimanual coordination, since their activity patterns were related to the bimanual task rather than to movements of the individual limbs. However, the number of specifically bimanually related neurons in the SMA was very low, despite the task involving a high degree of temporal linkage of the two hands. On the basis of the present single-unit data, it thus seems unlikely that bimanual

672

O. Kazennikov et al.

coordination is an exclusive or even a major function of the SMA. The case for self-initiation of movements as a principal function of the SMA is stronger.36,39,48 Evidence is accumulating to suggest instead that distributed and interconnected cortical networks of the two hemispheres, including the SMA, are engaged in the temporal and spatial coordination of the two limbs, in varying combinations, depending on the particular task. Previous evidence also points to an important role of proprioceptive feedforward signals for coordinating the two hands, as for any goal-directed multi-joint coordination.

Acknowledgements—We are grateful to A. Schwarz and J. Corpataux for taking excellent care of the animals, to A. Gaillard and B. Aebischer for the mechanical and electronic construction of the drawer manipulandum, and to V. Moret for histological work (all at the Institute of Physiology in Fribourg). This work was supported by grants from the Swiss National Science Foundation (to M.W.: grants 31-27569.89, 31-36183.92, 4038-044053; to E.M.R.: 3128572.90, 31-43422.95, 3130-025138), and by financial support from the INTAS program (Brussels) and the Novartis Foundation for the collaboration with O.K. (Russian Academy of Science, Moscow). B.H. was supported by a Research Training Fellowship from the New Zealand Neurological Foundation and Medical Research Council.

REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.

Aizawa H., Mushiake H., Inase M. and Tanji J. (1990) An output zone of the monkey primary motor cortex specialized for bilateral hand movement. Expl Brain Res. 82, 219–221. Bard C., Paillard J., Lajoie Y., Fleury M., Teasdale N., Forget R. and Lamarre Y. (1992) Role of afferent information in the timing of motor commands: a comparative study with a deafferented patient. Neuropsychologia 30, 201–206. Brinkman C. (1984) 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. 4, 918–929. Brinkman C. and Porter R. (1979) Supplementary motor area in the monkey: activity of neurons during performance of a learned motor task. J. Neurophysiol. 42, 681–709. Brown R. G., Jahanshahi M. and Marsden C. D. (1993) The execution of bimanual movements in patients with Parkinson’s, Huntington’s and cerebellar disease. J. Neurol. Neurosurg. Psychiat. 56, 295–297. Castiello U. and Bennett K. M. B. (1997) The bilateral reach-to-grasp movement of Parkinson’s disease subjects. Brain 120, 593–604. Chan J. L. and Ross E. D. (1988) Left-handed mirror writing following right anterior cerebral artery infarction: evidence for nonmirror transformation of motor programs by supplementary motor area. Neurology 38, 59–63. Chen D. F., Hyland B., Maier V., Palmeri A. and Wiesendanger M. (1991) Comparison of neural activity in the supplementary motor cortex and in the primary motor cortex in monkeys performing a choice-reaction task. Somatosensory Motor Res. 8, 27–44. Chen Y.-C., Thaler D., Nixon P. D., Stern C. E. and Passingham R. E. (1995) The functions of the medial premotor cortex. II. The timing and selection of learned movements. Expl Brain Res. 102, 461–473. Cody F. W. J., Schwartz M. P. and Smit G. P. (1990) Proprioceptive guidance of human voluntary wrist movements studied using muscle vibration. J. Physiol. 427, 455–470. Cordo P. J. (1990) Kinesthetic control of a multijoint movement sequence. J. Neurophysiol. 63, 161–172. Coxe W. S. and Landau W. M. (1965) Observations upon the effect of supplementary motor cortex ablation in the monkey. X. Brain 88, 763–772. Dick J. P. R., Benecke R., Rothwell J. C., Day B. L. and Marsden C. D. (1986) Simple and complex movements in a patient with infarction of the right supplementary motor area. Mov. Disord. 1, 255–266. Faillenot I., Toni I., Decety J. and Gre´goire M.-C. (1997) Visual pathways for object-oriented action and object recognition: functional anatomy with PET. Cerebral Cortex 7, 77–85. Flanders M. and Cordo P. J. (1989) Kinesthetic and visual control of a bimanual task: specification of direction and amplitude. J. Neurosci. 9, 447–453. Fleminger S. and Lishman W. A. (1994) Bimanual motor performance in controls and patients. J. Neurol. Neurosurg. Psychiat. 57, 126. Geffen G. M., Jones D. L. and Geffen L. B. (1994) Interhemispheric control of manual activity. Behav. Brain Res. 64, 131–140. Gentilucci M., Fogassi L., Luppino G., Matelli M., Camarda R. and Rizzolatti G. (1988) Functional organization of inferior area 6 in the macaque monkey: I. Somatotopy and the control of proximal movements. Expl Brain Res. 71, 475–490. Goldberg G. and Bloom K. K. (1990) The alien hand sign. Localization, lateralization and recovery. Am. J. phys. med. Rehab. 69, 228–238. Gordon A. M., Ingvarsson P. E. and Forssberg H. (1997) Anticipatory control of manipulative forces in Parkinson’s disease. Expl Neurol. 145, 477–488. Gordon A. M. and Soechting J. F. (1995) Use of tactile afferent information in sequential finger movements. Expl Brain Res. 107, 281–292. Grossberg S., Pribe C. and Cohen M. A. (1997) Neural control of interlimb oscillations. 1. Human bimanual coordination. Biol. Cybern. 77, 131–140. Halsband U., Ito N., Tanji J. and Freund H.-J. (1993) The role of premotor cortex and the supplementary motor area in the temporal control of movement in man. Brain 116, 243–266. Hartje W. and Ettlinger G. (1973) Reaching in light and dark after unilateral posterior parietal ablations in the monkey. Cortex 9, 346–354. Hausmanowa-Petrusewicz I. (1959) Interaction in simultaneous motor functions. Archs Neurol. Psychiat., Chicago 81, 173–187. He´caen H. (1978) Les apraxies ide´omotrices, essai de dissociation. In Du Controˆle Moteur a` l’Organisation du Geste (eds He´caen H. and Jeannerod M.), pp. 343–358. Masson, Paris.

Role of SMA neurons in bimanual coordination 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56.

673

Horstink M. W. I. M., Berger H. J. C., Van Spaendonek K. P. M., Van den Bercken J. H. L. and Cools A. R. (1990) Bimanual simultaneous performance and impaired ability to shift attention in Parkinson’s disease. J. Neurol. Neurosurg. Psychiat. 53, 685–690. Hummelsheim H., Bianchetti M., Wiesendanger M. and Wiesendanger R. (1988) Sensory inputs to the agranular motor fields: a comparison between precentral, supplementary-motor and premotor areas in the monkey. Expl Brain Res. 69, 289–298. Ikeda A., Lu¨ders H. O., Burgess R. C. and Shibasaki H. (1992) Movement-related potentials recorded from supplementary motor area and primary motor area. Role of supplementary motor area in voluntary movements. Brain 115, 1017–1043. Iwamura Y., Iriki A. and Tanaka M. (1994) Bilateral hand representation in the postcentral somatosensory cortex. Nature 369, 554–556. Jahanshahi M., Jenkins H., Brown R. G., Marsden C. D., Passingham R. E. and Brooks D. J. (1995) Self-initiated versus externally triggered movements. 1. An investigation using measurement of regional cerebral blood flow with PET and movement-related potentials in normal and Parkinson’s disease subjects. Brain 118, 913–933. Jason G. W. and Pajurkova E. M. (1992) Failure of metacontrol: breakdown in behavioural unity after lesion of the corpus callosum and inferomedial frontal lobes. Cortex 28, 241–260. Jeannerod M. (1991) The interaction of visual and proprioceptive cues in controlling reaching movements. In Motor Control: Concepts and Issues (Proceedings of a Dahlem Conference, 1989) (eds Humphrey D. R. and Freund H. J.), pp. 277–291. Wiley, Chichester. Jeannerod M., Decety J. and Michel F. (1994) Impairment of grasping movements following a bilateral posterior parietal lesion. Neuropsychologia 32, 369–380. Kawashima R., Yamada K., Kinomura S., Yamaguchi T., Matsui H., Yoshioka S. and Fukuda H. (1993) Regional cerebral blood flow changes of cortical motor areas and prefrontal areas in humans related to ipsilateral and contralateral hand movement. Brain Res. 623, 33–40. Kazennikov O., Hyland B., Wicki U., Perrig S., Rouiller E. M. and Wiesendanger M. (1998) Effects of lesions in the mesial frontal cortex on bimanual coordination in monkeys. Neuroscience 85, 703–716. Kazennikov O., Wicki U., Corboz M., Hyland B., Palmeri A., Rouiller E. M. and Wiesendanger M. (1994) Temporal structure of a bimanual goal-directed movement sequence in monkeys. Eur. J. Neurosci. 6, 203–210. Kelso J. A. S. (1995) Dynamic Patterns—The Self-Organization of Brain and Behavior, pp. 1–334. MIT, Cambridge, MA. Kermadi I., Liu Y., Tempini A. and Rouiller E. M. (1997) Effects of reversible inactivation of the supplementary motor area (SMA) on unimanual grasp and bimanual pull and grasp performance in monkeys. Somatosensory Motor Res. 14, 268–280. Kim S.-G., Ashe J., Hendrich K., Ellermann J. M., Merkle H., Ugurbil K. and Georgopoulos A. P. (1993) Functional magnetic resonance imaging of motor cortex: hemispheric asymmetry and handedness. Science 261, 615–617. Kimura D. and Archibald Y. (1974) Motor functions of the left hemisphere. Brain 97, 337–350. Kornhuber H. H., Deecke L., Lang W., Lang M. and Kornhuber A. (1989) Will, volitional action, attention and cerebral potentials in man: bereitschaftspotential, performance-related potentials, directed attention potentials, EEG spectrum changes. In Volition Action (ed. Hershberger W. A.), Advances in Psychology, Vol. 10, pp. 107–168. Elsevier (North-Holland), Amsterdam. Laplane D., Talairach J., Meininger V., Bancaud J. and Orgogozo J. M. (1977) Clinical consequences of corticectomies involving the supplementary motor area in man. J. neurol. Sci. 34, 301–314. Leonard G., Milner B. and Jones L. A. (1988) Performance of unimanual and bimanual tapping tasks by patients with lesions of the frontal or the temporal lobe. Neuropsychologia 26, 79–91. Liu C. N. and Chambers W. W. (1964) An experimental study of the cortico-spinal system in the monkey (Macaca mulatta). The spinal pathway and pre-terminal distribution of degenerating fibers following discrete lesions of the pre- and postcentral gyri and bulbar pyramid. J. comp. Neurol. 123, 257–284. Matsunami K. and Hamada I. (1981) Characteristics of the ipsilateral movement-related neuron in the motor cortex of the monkey. Brain Res. 204, 29–42. Moll L. and Kuypers H. G. J. M. (1977) Premotor cortical ablations in monkeys: contralateral changes in visually guided reaching behavior. Science 198, 317–319. Passingham R. E. (1996) Functional specialization of the supplementary motor area in monkeys and humans. Adv. Neurol. 70, 105–116. Penfield W. and Jasper H. (1954) Epilepsy and the Functional Anatomy of the Human Brain. Little & Brown, Boston. Preilowski B. (1975) Bilateral motor interaction: perceptual–motor performance of partial and complete ‘‘split-brain’’ patients. In Cerebral Localization (eds Zu¨lch K. S., Creutzfeldt O. and Galbraith G. C.), pp. 115–132. Springer, Berlin. Rea G. L., Ebner T. J. and Bloedel J. R. (1987) Evaluations of combined premotor and supplementary motor cortex lesions on a visually guided arm movement. Brain Res. 418, 58–67. Rizzolatti G., Camarda R., Fogassi L., Gentilucci M., Luppino G. and Matelli M. (1988) Functional organization of inferior area 6 in the macaque monkey. II. Area F5 and the control of distal movements. Expl Brain Res. 71, 491–507. Roland P. E., Larsen B., Lassen N. A. and Skinhoj E. (1980) Supplementary motor area and other cortical areas in organization of voluntary movements in man. J. Neurophysiol. 43, 118–136. Romo R. and Schultz W. (1992) Role of primate basal ganglia and frontal cortex in the internal generation of movements. III. Neuronal activity in the supplementary motor area. Expl Brain Res. 91, 396–407. Rouiller E. M., Babalian A., Kazennikov O., Moret V., Yu X.-H. and Wiesendanger M. (1994) Transcallosal connections of the distal forelimb representations of the primary and supplementary motor cortical areas in macaque monkeys. Expl Brain Res. 102, 227–243. Schell G. L., Hodge C. J. and Cacayosin E. (1986) Transient neurological deficits after therapeutic embolization of the arteries supplying the medial wall of the hemisphere including the supplementary motor area. Neurosurgery 18, 353–356.

674 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71.

O. Kazennikov et al. Schnitzler A., Salmelin R., Salenius S., Jousma¨ki V. and Hari R. (1995) Tactile information from the human hand reaches the ipsilateral primary somatosensory cortex. Neurosci. Lett. 200, 25–28. Steenbergen B., Hulstijn W., De Vries A. and Berger M. (1996) Bimanual movement coordination in spastic hemiparesis. Expl Brain Res. 110, 91–98. Swinnen S. P., Dounskaia N., Verschueren S., Serrien D. J. and Daelman A. (1995) Relative phase destabilization during interlimb coordination: the disruptive role of kinesthetic afferences induced by passive movement. Expl Brain Res. 105, 439–454. Tanji J. (1994) The supplementary motor area in the cerebral cortex. Neurosci. Res. 19, 251–268. Tanji J., Okano K. and Sato K. C. (1988) Neuronal activity in cortical motor areas related to ipsilateral, contralateral, and bilateral digit movements of the monkey. J. Neurophysiol. 60, 325–343. Thaler D., Chen Y.-C., Nixon P. D., Stern C. E. and Passingham R. E. (1995) The functions of the medial premotor cortex. I. Simple learned movements. Expl Brain Res. 102, 445–460. Travis A. M. (1955) Neurological deficiencies following supplementary motor area lesions in Macaca mulatta. Brain 78, 174–198. Viallet F., Massion J., Massarino R. and Khalil R. (1987) Performance of a bimanual load-lifting task by parkinsonian patients. J. Neurol. Neurosurg. Psychiat. 50, 1274–1283. Viallet F., Massion J., Massarino R. and Khalil R. (1992) Coordination between posture and movement in a bimanual load lifting task: putative role of a medial frontal region including the supplementary motor area. Expl Brain Res. 88, 674–684. Watson R. T., Fleet W. S., Gonzales-Rothi L. and Heilman K. M. (1986) Apraxia and the supplementary motor area. Archs Neurol. 43, 787–792. Wiesendanger M., Rouiller E. M., Kazennikov O. and Perrig S. (1996) Is the supplementary motor area a bilaterally organized system? In Supplementary Sensorimotor Area (ed. Lu¨ders H. O.), pp. 85–93. Raven, New York. Wiesendanger M., Wicki U. and Rouiller E. M. (1994) Are there unifying structures in the brain responsible for interlimb coordination? In Interlimb Coordinaton: Neural, Dynamical and Cognitive Constraints (eds Swinnen S. P., Heuer H., Massion J. and Casaer P.), pp. 179–207. Academic, San Diego. Wise S. P. and Tanji J. (1981) Supplementary and precentral motor cortex: contrast in responsiveness to peripheral input in the hindlimb area of the unanaesthetized monkey. J. comp. Neurol. 195, 433–452. Wyke M. (1971) The effects of brain lesions on the learning performance of a bimanual co-ordination task. Cortex 7, 59–72. Zentner J., Hufnagel A., Pechstein U., Wolf H. K. and Schramm J. (1996) Functional results after resective procedures involving the supplementary motor area. J. Neurosurg. 85, 542–549. (Accepted 10 June 1998)