Abstract
Theories of rhythmic perception propose that the environment is sampled in cycles, with alternating states of enhanced or diminished perceptual sensitivity at an attended location. These cycles involve periodic adjustments in functional communication between cognitive control areas and either the sensory or motor regions (Fiebelkorn et al., 2019). Previous research has shown robust rhythmic sampling with exogenous spatial cues that automatically direct attention (e.g., Landau & Fries, 2012). However, it remains unclear whether the same applies to endogenous, voluntary, attention, which relies on different neural mechanisms, and how spatial uncertainty affects the rhythmic process. To investigate, we conducted an experiment that combined elements of perceptual decision making and spatial attention tasks. In each trial, participants maintained central fixation while viewing two gratings displayed at either side in their periphery. A centrally presented cloud of red and green dots served as a cue, with the red-to-green ratio of the colored dots indicating the location of an upcoming target with 80% validity. Spatial uncertainty was manipulated by varying the ratio of the colored dots. After cue offset, the target, a brief near-threshold orientation change, appeared at one of the two grating locations at a randomly selected cue-target interval (CTI) from 300 to 1100 ms. We observed a significant endogenous attention effect, where target detection performance (accuracy and reaction time) was better at the cued vs. uncued location. This effect diminished with increased spatial uncertainty. When analyzing performance as a function of the CTI, we found a strong behavioral oscillation in the theta band and an anti-phase relationship between cued and uncued locations. This finding supports our hypothesis that endogenous attentional sampling is a rhythmic process. Importantly, preliminary data revealed a larger behavioral oscillation under high versus low spatial uncertainty, indicating an increasing reliance on periodic functional connectivity reweighting to resolve increasing uncertainty.