Abstract
Tracking a moving visual target triggers a saccade which is followed by a slow eye movement whose velocity can match the target velocity only after several hours of practice. The notion of internal model has been proposed to account for this ability to foveate and maintain a moving target within the central visual field, in spite of visuomotor delays. To further investigate this hypothesis and develop potential oculomotor probes to confirm a neurophysiological substrate, we trained three macaque monkeys to track a target moving along four possible paths (one per visual quadrant) and tested the impact of two types of training on interceptive saccades toward a briefly moving target (100 ms). A first training consisted of tracking a continuously visible target (800ms). In the second training, the target was momentarily concealed 100 ms after its onset and reappeared 300 ms later. Our results show that long duration of target exposure increased the range of landing positions and times of interceptive saccades and that the first training did not change the landing of saccades toward a briefer target. During the second training, while the duration of target invisibility was gradually increased (from 100 to 300 ms) over consecutive training days, a slow eye movement (glissade) followed the interceptive saccade. After the training, the landing of interceptive saccades was changed. They were also followed by a post-saccadic glissade, but no statistically significant correlation was found between the position/time landing ratio of saccades and the post-saccadic speed. Our study shows that saccades are differently controlled depending upon the target visibility. The notion of internal model of target motion must be revisited to account for our results.