The voluntary control of gaze implies the ability to make saccadic eye movements specified by abstract instructions, as well as the ability to repress unwanted orientating to sudden stimuli. Both of these abilities are challenged in the antisaccade task, because it requires subjects to look at an unmarked location opposite to a flashed stimulus, without glancing at it. Performance on this task depends on the frontal/prefrontal cortex and related structures, but the neuronal operations underlying antisaccades are not understood. It is not known, for example, how excited visual neurons that normally trigger a saccade to a target (a prosaccade) can activate oculomotor neurons directing gaze in the opposite direction. Visual neurons might, perhaps, alter their receptive fields depending on whether they receive a pro- or antisaccade instruction. If the receptive field is not altered, the antisaccade goal must be computed and imposed from the top down to the appropriate oculomotor neurons. Here we show, using recordings from the supplementary eye field (a frontal cortex oculomotor centre) in monkeys, that visual and movement neurons retain the same spatial selectivity across randomly mixed pro- and antisaccade trials. However, these neurons consistently fire more before antisaccades than prosaccades with the same trajectories, suggesting a mechanism through which voluntary antisaccade commands can override reflexive glances.

. Neuronal activities were recorded in the supplementary eye field (SEF) of 3 macaque monkeys trained to perform antisaccades pseudorandomly interleaved with prosaccades, as instructed by the shape of a central fixation point. The prosaccade goal was indicated by a peripheral stimulus flashed anywhere on the screen, whereas the antisaccade goal was an unmarked site diametrically opposite the flashed stimulus. The visual cue was given immediately after the instruction cue disappeared in the immediate-saccade task, or during the instruction period in the delayed-saccade task. The instruction cue offset was the saccade gosignal. Here we focus on 92 task-related neurons: visual, eye-movement, and instruction/fixation neurons. We found that 73% of SEF eye-movement-related neurons fired significantly more before antisaccades than prosaccades. This finding was analyzed at 3 levels: population, single neuron, and individual trial. On individual antisaccade trials, 40 ms before saccade, the firing rate of eye-movementrelated neurons was highly predictive of successful performance. A similar analysis of visual responses (40 ms astride the peak) gave less-coherent results. Fixation neurons, activated during the initial instruction period (i.e., after the instruction cue but before the stimulus) always fired more on antisaccade than on prosaccade trials. This trend, however, was statistically significant for only half of these neurons. We conclude that the SEF is critically involved in the production of antisaccades.

In addition to cells specifically active with visual stimuli, saccades, or fixation, the supplementary eye field contains cells that fire in precise temporal relationship with the occurrence of reward. We studied reward-related activity in two monkeys performing a prosaccade/antisaccade task and in one monkey trained in memory prosaccades only. Two types of neurons were distinguished by their reciprocal firing pattern: reward-predicting (RP) and reward-detecting (RD). RP neurons linearly increased their firing as early as 150 ms before saccade onset until the occurrence of reward, at which time they abruptly ceased firing. In contrast, RD neurons fired in phase with reward delivery, even when its duration was varied and when it was repeated at different frequencies. RD discharges were little affected or unaffected by the position of a visual cue that briefly anchored the goal at the onset of reward. The complementary firing patterns of the RP and RD neurons could provide a feedback mechanism necessary for learning and performing the task.

The antisaccade task requires a subject to make a saccade to an unmarked location opposite to a flashed stimulus. This task was originally designed to study saccades made to a goal specified by instructions. Interest for this paradigm surged after the discovery that frontal lobe lesions specifically and severely affect human performance of antisaccades while prosaccades (i.e., saccades directed to the visual stimulus) are facilitated. Training monkeys to perform antisaccades was rarely attempted in the past, and this study is the first one that describes in detail the properties of such antisaccades compared with randomly intermingled prosaccades of varying amplitude in all directions. Such randomization was found essential to force the monkeys to use the instruction cue (pro- or anti-) and the location cue (peripheral stimulus) provided within a trial rather than to direct their saccades to the location of past rewards. Each trial began with the onset of a central fixation target that conveyed by its shape the instruction to make a pro- or an antisaccade to a subsequent peripheral stimulus. In one version of the task, the monkey was allowed to make an immediate saccade to the goal; in a second version, the saccade had to wait for a go signal. Analyses of the accuracy, velocity, and latency of antisaccades compared with prosaccades were performed on a sample of 7,430 pro-/antisaccades in the "immediate saccade" task (delayed saccades suffering from known distortions). Error rates fluctuated approximately 25%. Results were the same for the two monkeys with respect to accuracy and velocity, but they differed in terms of reaction time. Both monkeys generated antisaccades to stimuli in all directions, at various eccentricities, but antisaccades were significantly less accurate and slower than prosaccades elicited by the same stimuli. Interestingly, saccades to the stimulus could be followed by appropriate antisaccades with no intersaccadic interval. Such instances are here referred to as "turnaround saccades." Because they occurred sometimes with a long latency, turnaround saccades did not simply reflect the cancellation of an early foveation reflex. Consistent with human data, latencies of one monkey were longer for antisaccades than for prosaccades, but the reverse was true for the other monkey who was trained differently. In summary, this study demonstrates the feasibility of providing a subhuman primate model of antisaccade performance, but at the same time it suggests some irreducible differences between human and monkey performance.

These results provide support at the single-neuron level for the role of the medial frontal cortex in the temporal organization and planning of movements in humans.