Abstract
Express saccades are the shortest latency eye movements known (70–110ms). However, previous work in our lab has shown that when elicited by the appearance of a spatially extended visual stimulus, express saccades have vectors influenced by a visuomotor set. In this case, the visuomotor set was established by an instruction to make a saccade to the left or right side of the stimulus array (Edelman et al, 2007). Remarkably, such a spatial “Cartesian instruction” does not increase saccade latency. Here, we extend these experiments in two ways. First, we examined whether a visuomotor set expressed in polar coordinates, established by an instruction to make a saccade to the nearer or farther of two visual stimuli arranged along an isodirectional line, is as effective as an instruction expressed in Cartesian coordinates. Second, we examined whether Cartesian instructions are effective when they are contingent, such that, for example, saccades are instructed to be made to the leftmost of two spots of light separated horizontally if they appear in the upper visual field and to the rightmost spot if they appear in the lower visual field. Subjects viewed a CRT display controlled by Vision Shell software routines while eye position was monitored by videooculography at 500 frames/s (Eyelink II, SR Research). A traditional gap task (gap duration: 150–200 ms) was used to facilitate express saccade production. First, we found that a visuomotor set established by a polar coordinate instruction influenced express saccade vector almost as much as Cartesian instructions, though there was a strong bias to make a saccade to the closer of the two targets. Second, we found that contingent Cartesian instructions had much less of an effect on saccade vector as non-contingent Cartesian instructions.
Supported by: NIGMS GM00816-28 (SCORE) and NIH/NCRR 5G12 Rr03060 (RCMI).