Overall localization performance of both monkeys.

<p>Means and standard deviations (in deg) of gaze-saccade endpoint distritbutions of azimuth (top row) and elevation (bottom row) responses with respect to the target location (origin of <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0047606#pone-0047606-g006" target="_blank">Fig. 6</a>) for the first (left) and second (right) target flashes. A positive (negative) mean indicates a target overshoot (undershoot) in that component. Gaze shifts tended to slightly undershoot the target. Comparisons for a statistical difference between distributions were made between the same target components and for the same animal, based on a KS test. Static and dynamic double-steps had significantly more endpoint variability than single-step responses (p<0.05) in the majority of cases. The same holds for dynamic vs. static double steps. Endpoint scatter of monkey O saccades was larger than for monkey M (p<0.001).