Abstract
We have studied the neural correlates in frontal eye fields and superior colliculus for saccadic behavior in monkey, employing a choice response task. Saccades to visually similar color-coded targets were elicited by the color change of a foveal fixation mark. Here we describe the saccadic choice behavior in these animals as a function of the number of alternative targets in the array and the spatial discordance between actual and cue locations in the array. Arbitrary color-location associations are built up during training and held in long term memory. On each trial the correct location of the saccade target is signaled by the central cue. During the experiment the locations of the colored stimuli are rotated to new locations, while the same color cues from training are delivered centrally. In the limiting case where the array is a single alternative target (1NA condition), no entry into associative memory was needed; regardless of the color-association discordance, performance was at or near ceiling. However, saccade latencies were significantly longer than those where a delayed saccade was made to a single, neutrally colored target, as if the animals continued to decode the central color cue. In the 4NA condition performance with no rotation of the array was > 95%. Performance degraded smoothly and symmetrically for both ipsilaterally and contralaterally rotated discordances. Error direction varied as a function of color-location discordance. The pattern of errors suggests these learned associations have broad spatial tuning, but the tuning does not extend across angular separations of 90°.
T32EY014536, R01EY08060, Rachel C. Atkinson Fellowship