The Journal of Neurophysiology Vol. 80 No. 3 September 1998, pp. 1562-1566
Copyright ©1998 by the American Physiological Society

RAPID COMMUNICATION

Movement Sequence-Related Activity Reflecting Numerical Order of Components in Supplementary and Presupplementary Motor Areas

William T. Clower and Garrett E. Alexander

Department of Neurology, Emory University School of Medicine, Atlanta, Georgia 30322

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Clower, William T. and Garrett E. Alexander. Movement sequence-related activity reflecting numerical order of components in supplementary and presupplementary motor areas. J. Neurophysiol. 80: 1562-1566, 1998. The supplementary motor area (SMA) and presupplementary motor areas (pre-SMA) have been implicated in movement sequencing, and neurons in SMA have been shown to encode what might be termed the relational order among sequence components (e.g., movement X followed by movement Y). To determine whether other aspects of movement sequencing might also be encoded by SMA or pre-SMA neurons, we analyzed task-related activity recorded from both areas in conjunction with a sequencing task that dissociated the numerical order of components (e.g., movement X as the 2nd component, irrespective of which movements precede or follow X). Sequences were constructed from eight component movements, each characterized by three spatial variables (origin, direction, and endpoint). Task-related activity recorded from 56 SMA and 63 pre-SMA neurons was categorized according to both the epoch (delay, reaction time, and movement time) and the spatial variable or component movement with which it was associated. All but one instance of task-related activity was selective for one of the spatial variables (SV-selective) rather than for any of the component movements themselves. Of 110 instances of SV-selective activity in SMA, 43 (39%) showed significant effects of numerical order. The corresponding incidence in pre-SMA, 82 (71%) of 116, was substantially higher (P < 0.00001). No effects of numerical order were evident among the hand paths, movement times, or electromyographic activity associated with task performance. We concluded that neurons in SMA and pre-SMA may encode the numerical order of components, at least for sequences that are distinguished mainly by that aspect of component ordering.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Both the supplementary motor area (SMA) and presupplementary motor area (pre-SMA) have been implicated in the control of movement sequencing (Chen et al. 1995; Halsband et al. 1994; Mushiake et al. 1991; Picard and Strick 1997; Tanji and Shima 1994). Neuronal activity related to specific sequences was first demonstrated in SMA by Mushiake et al. (1990). Subsequently, Tanji and Shima (1994) reported sequence-related activity in SMA that reflected the relative ordering of movement components. An SMA neuron might discharge during movement X if and only if X were preceded, or followed, by Y (Tanji and Shima 1994). Such activity might be said to reflect the relational order among sequence components. Movement sequences may also be distinguished by the numerical order of specific components, i.e., by their ordinal positions (1st, 2nd, etc.), independent of their relational ordering. For example, movement X may be the second of three components, irrespective of which movements precede, or follow, X. Here we present evidence that the numerical order of specific components may also be represented by movement sequence-related activity of SMA and pre-SMA neurons.

	METHODS

Abstract Introduction Methods Results Discussion References

A juvenile macaque (Macaca nemestrina, 5 kg) was trained to position a cursor on a video display by moving a joystick with the right hand. The left arm was lightly restrained. A neckplate obscured the monkey's view of its working forelimb. Trials began with illumination of one of four potential targets (each a 2 × 2-cm white square) arrayed in the form of a diamond (Fig. 1, A and B). Once the cursor was aligned with this "home" target (Fig. 1A1), the others were also illuminated (Fig. 1A2). Dimming (100 ms) of one of the home-adjacent targets served as an instruction stimulus (IS) designating the first target, and thus the clockwise (CW) or counterclockwise (CCW) orientation, of the required movement sequence (Fig. 1A2). The monkey then executed a delayed sequence of target captures, proceeding around the diamond along the instructed CW or CCW path. Each movement was preceded by a delay (600-1,200 ms), during which the cursor remained at the target just captured (Fig. 1, A3, A5, and A7) until the latter's color changed from white to yellow as a nondirectional trigger stimulus (TS), prompting the next target capture (Fig. 1, A4, A6, and A8). Once the required sequence was completed, a fourth delay (600-1,200 ms) was followed by a catch stimulus (CS) that appeared the same as a TS (Fig. 1, A9 and A10). The monkey signified its assessment that the three-component sequence was complete by not initiating a fourth movement following the CS. After a 900-ms post-CS delay (Fig. 1A11), correct performance was rewarded with a drop of juice (Fig. 1A12).

View larger version (42K): [in this window] [in a new window]   FIG. 1. A: subject's display over the course of 1 trial. Numbered arrows represent sequence components; filled arrows, current; open arrows, next/previous. Squares represent targets; filled circles represent cursor positions. For this trial, the bottom target is home (HT) (1). Once HT has been captured (HTC), an instruction stimulus (IS) specifies the required sequence orientation (2). Trigger stimuli (TS1-TS3), preceded by delays (pre-TS1-TS3), prompt the component movements (Mvt1-Mvt3) (3-8). Another delay (pre-CS) (9) precedes the catch stimulus (CS) (10). Following a post-CS delay (11), a liquid reward is dispensed (REW) (12). B: time line of events, epochs, and responses. C: numerical order of spatial variables and component movements (numbered arrows) was dissociated across the 8 sequences, as were the spatial variables themselves. Components of each sequence are coded with the same fill as the originating home target. Origins, endpoints: T = top; B = bottom; R = right; L = left. Directions: DL = down/left; DR = down/right; UL = up/left; UR = up/right.

Single cell activity was recorded from left SMA and pre-SMA with the use of conventional recording techniques (Alexander and Crutcher 1990). Each cell was tested with 8-10 repetitions of all eight trial types (Fig. 1C), presented in pseudorandom order. Receptive fields were mapped with tactile and proprioceptive stimuli, and effects of local microstimulation were assessed (40 µA, biphasic 200 µs negative/positive pulses, 330 Hz × 50-100 ms) (Alexander and Crutcher 1990; Luppino et al. 1991; Matsuzaka et al. 1992; Mitz and Wise 1987). Task-related activity was recorded from muscles of the working shoulder, elbow, and wrist with pairs of Teflon-coated stainless steel wires. Electromyographic (EMG) activity was amplified, rectified, integrated (20-ms bins), and analyzed in the same manner as the neural data.

Hand/joystick position was sampled at 200 Hz. Hand paths were plotted (as in Fig. 2) and visually compared for possible order effects with the use of computer overlays. For quantitative comparisons, we used the mean midpoint deviation (perpendicular distance between cursor position and midpoint of the intertarget line) as a measure of trajectory curvature (Wolpert et al. 1994). These and similar comparisons of reaction times (RTs) and movement times (MTs) by ordinal position were performed with the Tukey multiple comparison procedure for estimating simultaneous 95% confidence intervals.

View larger version (37K): [in this window] [in a new window]   FIG. 2. A: numerical order effect, type A. Delay-specific activity on trials where point B was the endpoint of the 3rd movement. Rasters and peri-event histograms are aligned with TS1-TS3 (dashed lines, arrowheads). Ticks represent action potentials. Hand trajectories are shown beside each raster. Abbreviations as in Fig. 1. Scales, horizontal in milliseconds, vertical in spikes/s. B: numerical order effect, type B. Reaction time-specific activity on trials in which L was the endpoint of either the 1st or 2nd (not the 3rd) movement. Conventions as in A, except here the activity is aligned with movement onsets (Mvt1-Mvt3).

Task-related activity was first categorized both by epoch (task interval in which it occurred) and by the associated movement or movement variable. Trial-by-trial discharge rates were computed for three epochs; delay (TS - 300 ms  TS), RT (TS  movement onset), and MT (movement onset  target capture). Four target locations and two sequence orientations yielded eight types of component movements and eight distinct movement sequences (Fig. 1C). Each component movement entailed three spatial variables: target of origin (T, top; B, bottom; R, right; and L, left), direction (DL, down/left; DR, down/right; UL, up/left; UR, up/right), and targeted endpoint (T, B, R, L). Spatial variables were dissociated across the eight sequences, as were their ordinal positions and those of the component movements.

A cell might show task-related activity in more than one epoch. Each instance of epoch-specific activity was tested for relatedness to one of the spatial variables or component movements with three two-way analyses of variance (ANOVAs): origin × orientation, endpoint × orientation, and direction × orientation ( = 0.01). Activity was considered selective for a spatial variable if in the corresponding ANOVA there was a main effect for that variable and no interaction with sequence orientation. This implied selective dependence on that variable independent of the other two (the 3 spatial variables being dissociated across sequence orientations) and ruled out a selective relation to one of the component movements themselves (each of the 8 component movements being confined to sequences of a single orientation). Activity considered selective for one of the eight component movements was identified by the conjunction of main effectsplus interactions with sequence orientationfor all three of that movement's associated spatial variables.

Activity found selective for a spatial variable or component movement was then tested with a one-way ANOVA for ordinal position ( = 0.01). If a main effect was present, planned comparisons were made of mean firing rates associated with the three ordinal positions. These were ranked by magnitude and designated accordingly: µ₁ for the highest, µ₂ for the median, and µ₃ for the lowest mean firing rate. Two linear contrasts, µ₁ - µ₂ = 0 and µ₂ - µ₃ = 0, resulted in three categories of numerical order effects: type A (µ₁ > µ₂ = µ₃), type B (µ₁ = µ₂ > µ₃), and type AB (µ₁ > µ₂ > µ₃). Pearson's ² goodness-of-fit test ( = 0.05) was used to compare the relative frequencies of categorized activity.

	RESULTS

Abstract Introduction Methods Results Discussion References

Each component movement was executed in a comparable manner irrespective of its ordinal position. No order effects were evident among the hand paths or MTs of the component movements or in the task-related EMG activity recorded from 12 muscles of the working shoulder, elbow, and wrist. Mean MTs (±SD) for first, second and third sequence components were 278 ± 26, 280 ± 31, and 279 ± 29 ms, respectively. Mean RTs for second and third sequence components (396 ± 134 and 435 ± 124 ms, respectively) were comparable, but each was shorter than that for 1st components (591 ± 172 ms).

Task-related activity was recorded from two mesial foci in the left superior frontal gyrus, one caudal and one rostral to the coronal plane traversing the genu of the arcuate sulcus (ASg) (He et al. 1995; Picard and Strick 1997). In the caudal focus (ASg  6 mm  ASg  2 mm), corresponding to the arm area of SMA (Alexander and Crutcher 1990; Luppino et al. 1991; Matsuzaka et al. 1992; Mitz and Wise 1987), activity was often driven by somatosensory stimuli applied to the forelimb, and microstimulation evoked forelimb movements at low threshold (50 ms, 25 µA). In the rostral focus (ASg + 3 mm  ASg + 7 mm), corresponding to pre-SMA, somatosensory driving was absent, and microstimulation evoked forelimb movements only rarely and at high threshold (100 ms, 35-40 µA).

Task-related recordings from 56 SMA and 63 pre-SMA neurons included 227 examples of epoch-specific activity that could be categorized and thus identified across different sequences by their selectivity for a particular movement or spatial variable. All but one showed selectivity for one of the three spatial variables (SV-selective activity) rather than one of the eight component movements.

Numerical order effects were evident in both cortical areas (Table 1), but the incidence (per instance of SV-selective activity) was higher in pre-SMA than in SMA (82/116 vs. 43/110; ² = 22.8, df = 1, P = 1.76 × 10^-6). For both areas, the incidence of order effects was comparable for all three spatial variables (order-by-SV, Table 1), although variables direction and endpoint were represented more frequently than was origin (SV). In SMA, order effects were more common during the delay than during the RT or MT epochs (order by epoch), whereas the proportion of SV-selective activity was the same for all three epochs (epoch). Conversely, in pre-SMA order effects were equally common in all epochs, but SV-selective activity was more prevalent during the delay.

Figure 2A illustrates a type A numerical order effect. This pre-SMA neuron showed delay-specific, endpoint-selective activity before movements that ended at the bottom target, but only if the movement constituted the third sequence component (µ₁ > µ₂ = µ₃, µ₁ = 3rd component). Figure 2B illustrates a type B numerical order effect. This pre-SMA neuron showed RT-specific, endpoint-selective activity before movements that ended at the left target, but only if the movement constituted either the first or second (not the 3rd) sequence component (µ₁ = µ₂ > µ₃, µ₃ = 3rd component). Note that in each case the epoch-specific activity was related to a single spatial variable rather than a single component movement, and the order effect was not relational, i.e., it was independent of the specific components (both movements and spatial variables) that preceded and followed.

Table 2 shows the distribution of numerical order effects in terms of the specific ordinal positions associated with highest (µ₁) and lowest (µ₃) mean firing rates. As indicated in the table, type AB order effects could not be classified in terms of any single ordinal position where mean firing rates were highest or lowest; instead, they could be viewed as composites (or superpositions) of specific type A and type B order effects.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

More than one-third of the epoch-specific activity sampled from SMA and more than two-thirds of that from pre-SMA were sequence related, reflecting the numerical order of specific components. Sequence-related activity reflecting the relational order among components was demonstrated previously in SMA (Tanji and Shima 1994). It is unclear whether both aspects of component ordering might also be represented within pre-SMA. However, the higher incidence of numerical order effects observed in this region suggests that pre-SMA may have a preferential role in representing the numerical order of sequence components. Both the numerical effects observed here and the relational effects observed by Tanji and Shima (1994) could have been determined in part by differences in the behavioral tasks employed in each study. Each would naturally have been most sensitive in detecting factors that were varied systematically, numerical order in this study, relational order in their study. If the learning of new motor tasks results in corresponding changes in the behavior-correlated activity of motor and premotor neurons (Aizawa et al. 1993; Mitz et al. 1991), then task differences might also influence the observed proportions of neurons representing numerical versus relational order.

The absence of such effects in task-related hand paths, MTs, and EMG activity indicates that numerical order effects observed in SMA and pre-SMA were unlikely to result from ordinal position-dependent differences in the way movements were executed. On the other hand, RTs for first-component movements did differ consistently from those for second- and third-component movements. This particular pattern, first =/ second = third, was characteristic of only 2 of the 12 subtypes of numerical order effects (i.e., Table 2, A1 and B3). Some of this activity might conceivably involve processes underlying motor preparation for an entire three-component sequence.

Finally, although not the intended focus of our study, we did find it noteworthy that virtually all of the task-related activity in both regions showed selectivity for specific spatial features of the subject's limb movements (origin, direction, and endpoint) rather than for the underlying movements themselves. This could not be attributed to any bias favoring detection of SV-selective activity. Activity selective for one of the component movements should if anything have been more easily detected, being necessarily limited to sequences of a single orientation (Fig. 1C). The implication that movement variables (rather than specific movements) may be represented preferentially in SMA and pre-SMA is consistent with growing evidence that both regions operate at relatively high levels within the network of cortical motor fields (Matsuzaka and Tanji 1996; Tanji 1994).

	ACKNOWLEDGEMENTS

  Present address for W. T. Clower: Vision et Motricite, INSERM Unite 94, 69500 Bron, France.

	FOOTNOTES

  Address for reprint requests: G. E. Alexander, Dept. of Neurology, WMB 6000, Emory University School of Medicine, 1639 Pierce Dr., Atlanta, GA 30322.

  Received 12 December 1997; accepted in final form 14 May 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• AIZAWA, H., INASE, M., MUSHIAKE, H., SHIMA, K., TANJI, J. Reorganization of activity in the supplementary motor area associated with motor learning and functional recovery. Exp. Brain Res. 84: 668-671, 1993.
• ALEXANDER, G. E., CRUTCHER, M. D. Preparation for movement: neural representations of intended direction in three motor areas of the monkey. J. Neurophysiol. 64: 133-150, 1990.[Abstract/Free Full Text]
• CHEN, Y.-C., THALER, D., NIXON, P. D., STERN, C. E., PASSINGHAM, R. E. The functions of the medial premotor cortex. II. The timing and selection of learned movements. Exp. Brain Res. 102: 461-473, 1995.[Medline]
• HALSBAND, U., MATSUZAKA, Y., TANJI, J. Neuronal activity in the primate supplementary, pre-supplementary and premotor cortex during externally and internally instructed sequential movements. Neurosci. Res. 20: 149-155, 1994.[Medline]
• HE, S.-Q., DUM, R. P., STRICK, P. S. Topographic organization of corticospinal projections from the frontal lobe: motor areas on the medial surface of the hemisphere. J. Neurosci. 15: 3284-3306, 1995.[Abstract]
• LUPPINO, G., MATELLI, M., CAMARDA, R. M., GALLESE, V., RIZZOLATTI, G. Multiple representations of body movements in medial area 6 and adjacent cingulate cortex: an intracortical microstimulation study in macaque monkey. J. Comp. Neurol. 311: 463-482, 1991.[Medline]
• MATSUZAKA, Y., AIZAWA, H., TANJI, J. A motor area rostral to the supplementary motor area (presupplementary motor area) in the monkey: neuronal activity during a learned motor task. J. Neurophysiol. 68: 653-662, 1992.[Abstract/Free Full Text]
• MATSUZAKA, Y., TANJI, J. Changing directions of forthcoming arm movements: neuronal activity in the presupplementary and supplementary motor area of monkey cerebral cortex. J. Neurophysiol. 76: 2327-2342, 1996.[Abstract/Free Full Text]
• MITZ, A. R., GODSCHALK, M., WISE, S. P. Learning-dependent neuronal activity in the premotor cortex: activity during the acquisition of conditional motor associations. J. Neurosci. 11: 1855-1872, 1991.[Abstract]
• MITZ, A. R., WISE, S. P. The somatotopic organization of the supplementary motor area: intracortical microstimulation mapping. J. Neurosci. 7: 1010-1021, 1987.[Abstract]
• MUSHIAKE, H., INASE, M., TANJI, J. Selective coding of motor sequence in the supplementary motor area of the monkey cerebral cortex. Exp. Brain Res. 82: 208-210, 1990.[Medline]
• MUSHIAKE, H., INASE, M., TANJI, J. Neuronal activity in the primate premotor, supplementary, and precentral motor cortex during visually guided and internally determined sequential movements. J. Neurophysiol. 66: 705-718, 1991.[Abstract/Free Full Text]
• PICARD, N., STRICK, P. L. Activation on the medial wall during remembered sequences of reaching movements in monkeys. J. Neurophysiol. 77: 2197-2201, 1997.[Abstract/Free Full Text]
• TANJI, J. The supplementary motor area in the cerebral cortex. Neurosci. Res. 19: 251-268, 1994.[Medline]
• TANJI, J., SHIMA, K. Role for supplementary motor area cells in planning several movements ahead. Nature 371: 413-416, 1994.[Medline]
• WOLPERT, D. M., GHAHRAMANI, Z., JORDAN, M. I. Perceptual distortion contributes to the curvature of human reaching movements. Exp. Brain Res. 98: 153-156, 1994.[Medline]

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society