Distribution of attention and parallel saccade programming in antisaccades.

The aim of our study was to investigate the spatiotemporal distribution of attention during the programming of pro- and antisaccades. The visual display consisted of a central fixation dot and six circularly arranged peripheral squares. In each trial one of the targets was briefly highlighted and participants were either instructed to shift their gaze to the cued square (prosaccade task) or to the diagonally opposed square (antisaccade task). At the end of the trial they reported the orientation of a visual probe that had appeared in one of the six squares and indicated whether they had made a correct saccade or a saccade in the wrong direction. In line with the results of previous studies, saccade latencies were longer and error rates were substantially higher in the antisaccade condition as compared to the prosaccade condition. The analysis of discrimination performance in the cue-saccade interval revealed that in the prosaccade condition attention was allocated exclusively to the saccade goal. In contrast, attention in the antisaccade condition was allocated both to the cued location and to the correct antisaccade goal. This was the case both before correct antisaccades and erroneous prosaccades and the amount of attention at the cue was predictive of the occurrence of errors. Interestingly, more than half of the erroneous prosaccades were not recognized. Unrecognized errors had smaller amplitudes and were corrected faster than recognized errors. In conclusion, our results indicate that the goal in the antisaccade task is selected through a competitive process between reflexive attention at the saccade cue and voluntary attention at the antisaccade goal. The occurrence of very short intersaccadic intervals between unrecognized erroneous prosaccades and corrective saccades provides evidence that these saccades were programmed in parallel. Meeting abstract presented at VSS 2015.