Abstract
When two static visual targets are sequentially presented, the saccades with the shortest latencies land on the location of the first target whereas the later saccades land closer to the second one. The transition between these two locations is not abrupt but gradual and continuous. When the targets are simultaneously presented, the saccades land on locations situated along a line connecting them (averaging). The process specifying the oculomotor goal was further explored in three monkeys by testing the saccades made in response to two centrifugal targets moving simultaneously in different directions. After the fixation of a central target for a variable interval, two trial conditions were intermixed. During the single-target trials, the target moved toward the periphery along the cardinal (horizontal, vertical) or oblique axis. During the double-target trials, the central target was replaced by two identical targets moving centrifugally with equal or different speeds. One target moved along an oblique axis while the other moved along a cardinal axis. The monkeys were free to track anyone of the two targets or anywhere between, and were rewarded on every trial. During the double-target trials, the first saccade endpoints were scattered in the field situated between the target paths. Although their distribution was biased toward the slower target, the orientation of the scatter depended upon the location and speed of the two targets: the endpoints were scattered between lines of isochronous target locations. Following the "averaging" saccade, the eye moved slowly in the same direction for duration which could exceed 300ms. The use of accelerating and decelerating targets revealed curved saccades, with early and later parts of their trajectory biased by the slowest target. This work demonstrates that the visual signals from two moving targets are processed in parallel with high spatiotemporal resolution, and that they continuously drive the oculomotor system.
Meeting abstract presented at VSS 2017