1. The purpose of this study was to describe the response properties of neurons in the supplementary motor area (SMA), including the supplementary eye fields (SEF) of three rhesus monkeys (Macaca mulatta) performing visually guided eye and forelimb movements. Seven hundred thirty single units were recorded in the dorsomedial agranular cortex while monkeys performed a go/no-go visual tracking task. The unit activity associated with rewarded, task-related movements was compared with that associated with unrewarded, spontaneous movements executed in the intertrial interval or when the task was not running. A number of neuronal response types were identified. 2. Sensory cells were characterized by their response to the visual and/or auditory target stimuli combined with no discharge associated with eye or forelimb movements. New information was provided about the receptive fields of the visual cells; they varied in size and, although many included the ipsilateral hemifield, they tended to emphasize the contralateral. A significant proportion of the visually responsive cells had receptive fields restricted to within 8-degrees of the fovea. The response latency was relatively long (> 90 ms) and variable. 3. Preparatory set cells were activated from the appearance of the target until the presentation of the go/no-go cue. This subpopulation ceased firing 50-100 ms before the movement was initiated. These cells tended to respond best in relation to contralateral movements. The response latency was similar to that of the sensory cells, although some of these units began to discharge in anticipation of predictable target presentations. These neurons were not active before unrewarded, spontaneous saccades. 4. Sensory-movement cells comprised the largest population of neurons identified in SMA. They were active from the appearance of the target until after the execution of the saccade. These neurons tended to respond preferentially in association with contraversive saccades. The latency of response to the target was significantly longer than that of the sensory cells. There was a large amount of variability in the time to reach the peak level of activation, and this population of units generally became inactivated shortly after the saccade was initiated. Although there were counter-examples, most sensory-movement cells responded equally in association with visually and auditory guided movements. In addition, these neurons were not active in relation to self-generated eye movements made during the intertrial intervals. 5. Pause-rebound cells were identified by their suppression at the appearance of the target and subsequent discharge associated with the saccade. These units tended to respond preferentially to contralateral targets. Although the onset of the suppression tended to be of sufficiently short latency to be considered anticipatory, the onset of the saccade-related burst did not occur until shortly before the eye movement was initiated. 6. Presaccadic movement neurons were identified by an absence of any modulation associated with the targets combined with a discharge beginning as much as 300 ms before and decaying after the saccade was initiated. Many of these units were omnidirectional, but those that responded preferentially in relation to saccades in one direction usually preferred contraversive eye movements. Furthermore, these neurons were activated only for the task-related saccades and not for spontaneous saccades. 7. Postsaccadic movement cells were recorded infrequently. Detailed analysis of their properties was, therefore, not possible. 8. Eye position cells were characterized by a statistically significant orbital dependence to their modulation. Such units discharged before saccades of different directions and amplitudes that brought the eye to a broadly specified orbital position. These neurons were also active in relation to pursuit eye movements that ended in the particular position. In addition, once the eyes were at the specific angle of gaze, these neurons exhibited a sustained discharge that ceased before any eye movement away from that position. 9. Forelimb movement cells discharged before and during the reaching movements. Even though each of the three monkeys used his right arm to perform the task, these neurons were identified in both hemispheres. The forelimb movement cells typically responded preferentially for movements in one direction. Although a small number of these neurons were studied, the distribution of the preferred directions of movement was different for the units recorded in the left and right hemispheres. 10. No-go-specific neurons were modulated specifically when the monkeys withheld movements in no-go trials. These units tended to be activated in association with not moving into the contralateral hemifield. The response latency of these units was significantly longer than that of the sensory cells. 11. A rough somatotopic arrangement was evident in SMA. Eye movement-related cells tended to be found rostral to forelimb movement-related cells. Units that were active during mouth movements were encountered in between the eye and forelimb regions. The sensory and preparatory set cells appeared to be distributed over the entire region explored. 12. These data provide new information about the involvement of the SMA and its constituent SEF in gaze control. On the basis of the different neuron classes that were identified, it appears that SEF/SMA provides a variety of signals for use in the sensorimotor integration that underlies volitional skeletal and oculomotor movements.
