Presaccadic Processes in the Generation of Pro and anti Saccades in Human Subjects—A Reaction-Time Study

Abstract

It has been widely acknowledged that the generation of an anti saccade, ie a saccade towards the direction opposite to that of a visual stimulus, requires the correct function of special brain structures. In the present study attempts were made to measure the time consumption of brain processes preceding the execution of pro and anti saccades. The saccadic eye movements of five adult human subjects were investigated in a series of combined pro/anti saccade tasks with the aid of the gap and the overlap paradigms. The type of trial—pro saccade and anti saccade—was defined by the structure of the stimulus. In some sessions the subjects were, in addition, preinformed about the actual command by a cue at different lead times before stimulus onset. Pro-saccade and anti-saccade trials were randomly intermixed in equal proportions. High error rates (>30% of all trials in some subjects) occurred in the test sessions without preinformative cues. These errors had long reaction times (∼200 ms), whereas the latencies of correct pro or correct anti saccades were even longer (∼350 ms). Analysis of the errors revealed that they were related to the situation in the previous trial: a correct response in the previous trial enhanced the chance of making a saccade of the same type in the actual trial by up to 30%. This pretrial effect occurred whether the actual trial was a pro-saccade or an anti-saccade command. With a cue lead time of 100 ms the numbers of errors decreased, but the latencies of the correct pro or anti saccades were about 70 ms longer than those obtained in the nonrandom control. With a 200 ms cue lead time the reaction times corresponded to those in the control condition. The results suggest that the situation in a given trial creates a kind of default program for the saccade preparation in the next trial. When a cue about the actual command is given early enough, the default program is overridden correspondingly. The perception of the cue and the programming of the corresponding saccade takes an additional 150 to 200 ms.