SACCADE-RELATED ACTIVITY IN MONKEY SUPERIOR COLLICULUS .1. CHARACTERISTICS OF BURST AND BUILDUP CELLS

1. In the monkey superior colliculus (SC), the activity of most saccad e-related neurons studied so far consists of a burst of activity in a population of cells at one place on the SC movement map. In contrast, recent experiments in the cat have described saccade-related activity as a slow increase in discharge before saccades followed by a hill of activity moving across the SC map. In order to explore this striking d ifference in the distribution of activity across the SC, we recorded f rom all saccade-related neurons that we encountered in microelectrode penetrations through the monkey SC and placed them in categories accor ding to their activity during the generation of saccades. 2. When we c onsidered the activity preceding the onset of the saccade, we could di vide the cells into two categories. Cells with burst activity had a hi gh-frequency discharge just before saccade onset but little activity b etween the signal to make a saccade and saccade onset. About two third s of the saccade-related cells had only a burst of activity. Cells wit h a buildup of activity began to discharge at a low frequency after th e signal to make a saccade and the discharge continued until generatio n of the saccade. About one third of the saccade-related cells studied had a buildup of activity, and about three fourths of these cells als o gave a burst of activity with the saccade in addition to the slow bu ildup of activity. 3. The buildup of activity seemed to be more closel y related to preparation to make a saccade than to the generation of t he saccade. The buildup developed even in cases when no saccade occurr ed. 4. The falling phase of the discharge of these saccade-related cel ls stopped with the end of the saccade (a clipped discharge), shortly after the end of the saccade (partially clipped), or long after the en d of the saccade (unclipped). 5. Some cells had closed movement fields in which saccades that were substantially smaller or larger than the optimal amplitude were not associated with increased activity. Other c ells tended to have open-ended movement fields without any peripheral border; they were active for all saccades of optimal direction whose a mplitudes were equal to or greater than a given amplitude. We found bo th types of movement fields at all movement field eccentricities studi ed within the SC. 6. The activity of cells with open-ended movement fi elds did not result from the smear of the visual target as it swept ac ross the retina during a saccade because the discharge of the cell was still present when saccades were made in the dark to remembered rathe r than visual targets. The activity of these cells was also not due to the occurrence of corrective saccades because the activity was visibl e whether or not there was one. 7. In penetrations through the interme diate layers of the SC, we usually found cells with a burst of activit y and those with closed movement fields to Lie more dorsally than thos e with a buildup of activity and open-ended movement fields. 8. We als o compared the activity of the saccade-related cells with the activity of fixation cells located in the rostral pole of the SC. We found tra nsition between saccade-related cells with open-ended movement fields and fixation cells. Cells within this transition zone were tonically a ctive during fixation but also discharged during small contraversive s accades. These fixation cells were encountered deeper in the intermedi ate layers, at the same level as the cells with open-ended movement fi elds and buildup of activity. We propose that fixation cells form a ro stral extension of the layer of cells with a buildup of activity. 9. W e conclude that these characteristics of the saccade-related cells ove rlap sufficiently to allow us to place the cells into two groups. Burs t cells have a high-frequency burst occurring immediately before sacca des and no buildup of activity; the majority have clipped activity at the end of the saccade and usually have closed movement fields. In con trast, buildup cells show activity beginning with the signal to make a saccade that continues until the generation of the saccade; the major ity have partially clipped activity at the end of the saccade and have open-ended movement fields. Because we encountered the cells with bur st activity and closed movement fields more dorsally than we did cells with buildup activity and open-ended movement fields, we hypothesize further that the burst and buildup cells can be regarded as separate f unctional sublayers with the burst layer on top and the buildup layer below. The buildup cells are similar to the saccade-related cells in t he cat SC, but the burst cells may be an added feature of the primate SC.