Abstract
[Background] Endogenous temporal attention is voluntary deployment of attention at specific moments, whereas temporal expectation is based on the certainty of when events occur. Previous experimental designs either did not distinguish expectation from attention or studied temporal attention only with constant temporal expectation, when there is no uncertainty. With temporal certainty, temporal attention benefits visual performance. Whether and how temporal uncertainty about stimulus timing modulates the benefits of endogenous temporal attention remains unknown. Microsaccades, small fixational eye movements are suppressed with expectations, and more so with temporal attention. [Goal] To investigate benefits of temporal attention with varying temporal expectations and corresponding microsaccade patterns. [Methods] We adapted a two-interval temporal attention experimental protocol to investigate endogenous temporal attention with varying temporal expectations. Observers performed a 2AFC orientation discrimination task. Two oriented Gabors were presented sequentially at the foveal location. An auditory valid precue instructed participants to attend to the first (T1) or second (T2) target, or a neutral precue instructed participants to attend to both targets. There was some temporal uncertainty for the target onset, which varied between 1200-1600 ms after precue onset for T1, with the most probable time point being 1400 ms, and 1650 ms for T2. Target was indicated by an auditory response cue. [Results] Accuracy was higher for valid than neutral cues. Higher benefits were found for both T1 and T2 when they appeared earlier than expected, as compared to later than expected. Microsaccades were suppressed more when attending to a specific moment than in the neutral condition. [Conclusions] Specific moments are selected and prioritized such that the performance is improved at the attended moments even when there is uncertainty about when the stimuli appeared. These performance improvements were more pronounced when stimuli were presented earlier, as compared to later than the expected average time.