We did observe an asymmetry in the increase in error rates on anti-saccade trials, with short-duration SEF stimulation causing a larger increase in contralateral (Fig. 2A) vs. ipsilateral anti-saccade errors (Fig. 2B). A three-way repeated-measures analysis of variance (anova) of error rate across
the factors of task (pro- or anti-saccades), direction (contra- or ipsilateral to stimulation) and time of stimulation (including control trials) revealed significant effects of task and time of stimulation (P < 10−5), and significant two-way and three-way interactions between all factors (task and direction: P = 0.02; task and stimulation time: P < 10−5; direction and stimulation time: P = 0.003; task, direction and MG-132 supplier stimulation time: P = 0.03). Subsequent two-way repeated-measures anovas of error rates on pro- or anti-saccade trials revealed a far greater influence of stimulation time on anti-saccade vs. pro-saccade trials, suggesting that the three-way interaction between task, direction and stimulation is primarily
driven by the anti-saccade error rate. The filled symbols in Fig. 2 show data that differed significantly from the respective Selleck RG 7204 control trials (paired t-tests, Bonferroni-corrected for multiple comparisons), and the frequency histograms in Fig. 2C and D represent the change in error rate vs. control trials for pro- or anti-saccades for each stimulation interval. The greater impact of ICMS-SEF on anti-saccade error rate across our sample can be appreciated by gauging the degree of shift of these histograms away from zero (rightward shifts convey increases in error rate). Note also that the histograms shifts tend to be greater for contralateral vs. ipsilateral anti-saccade errors for the later stimulation intervals, emphasizing some degree of laterality to the change others in anti-saccade error rate. The influence of short-duration ICMS-SEF on RTs is shown in Fig. 3 in a similar fashion. As with error rates, the influence of ICMS-SEF on correct
RTs is highly dependent on the task, and on the timing of stimulation relative to cue presentation (Fig. 3). Short-duration ICMS-SEF during the fixation interval exerted only a minor effect on RTs, but exerted a much greater effect when delivered after cue onset on anti-saccade trials, progressively prolonging the RTs of correctly performed anti-saccades in either direction. Interestingly, short-duration ICMS-SEF had little effect on the RTs of contralateral pro-saccades, although we did observe a modest increase in the RTs for pro-saccades to an ipsilateral cue for later stimulation times. Finally, the RTs of anti-saccade errors displayed a dependency with saccade direction, becoming shorter for errors made to contralateral cues, and longer for errors to ipsilateral cues.