Abstract
Several studies have shown that microsaccades index the direction of covert attention shifts. Yet, recently, this correlation has been questioned. In this study, we investigated this topic in a modified Posner task while recording eye movements with a high-precision eye-tracker. Participants (n=7, 13500 trials) were instructed to maintain fixation on a central square. Two differently colored peripheral circles (1-deg diam) simultaneously appeared 5 degrees to the right and to the left of the fixation point for 500ms. Then, the circles turned gray, while the central square took on one of the two colors, thereby serving as a cue for directing covert attention to the corresponding side (Cue window, 700-1000ms). The cue was valid in 80% of trials. A low-contrast target then appeared in one of the two circles, and participants had to release a button as soon as the target was detected. As expected, reaction times were modulated by trial validity (valid: 412ms vs invalid: 482ms; p<0.001). A fine grain temporal analysis on microsaccade directionality showed that, even though each participant was characterized by an idiosyncratic bias before cue onset, microsaccade direction on average followed the cue from 100 to 400ms after cue onset. However, directionality of microsaccades in this window did not affect reaction times, raising questions on the effectiveness of microsaccades as an index of covert attention.