Abstract:
Traditionally, a great many studies of visual attention have used reaction time measures (either with manual button presses or saccadic eye movements) to make inferences about the locus and time course of attentional allocation. One classic example of such studies is the Posner cueing paradigm (Posner 1980), in which subjects maintain fixation and a cue is presented on one side or the other of space; a post-cue target appearing at different times and locations is used to elicit a reaction time and map the spatial and temporal development of cue-induced changes in internal brain state. However, tasks with prolonged fixation inevitably involve fixational eye movements, like microsaccades. Since microsaccades are the same as saccades, and are therefore associated with peri-movement changes in internal brain state, an imperative question we should ask is: how much of performance changes in tasks like Posner cueing may actually be attributable to peri-movement changes in vision associated with microsaccades? And, if this turns out to be a real, plausible possibility, can we predict, on a trial-by-trial basis, when and where microsaccades can occur, and therefore when and where performance changes in Posner cueing might be expected to take place? In order to investigate these questions, we started our Study I, which is a combined study including modeling simulations and behavioral psychophysics. Based on a minimalist model of oculomotor generation (microsaccades) without any other factors (i.e. knowledge about where attention is “supposed” to be allocated), we successfully simulated attentional effects and replicated all detailed observations in the classic Posner cueing paradigm. This means that from a theoretical perspective, classic concepts in cognitive neuroscience like “attentional capture (AC)” and “Inhibition of return (IOR)” become the outcomes of peri-microsaccadic enhancement or suppression of neural visual sensitivity. We next turned to the question of why microsaccades might be modulated in Posner cueing at all; can we predict when and where microsaccades should be seen? In Study II, we experimentally controlled instantaneous foveal motor error during the presentation of peripheral cues. Post-cue microsaccadic oscillations were severely disrupted, suggesting that microsaccades in Posner cueing occur for oculomotor control over foveal motor error and not necessarily because they form a “dirty” read-out of covert attention, as commonly assumed. We then went one step further in Study III, in which we delved deeper into the mechanisms for fixational eye position dynamics, and how they dictate when microsaccades occur (and therefore when performance changes in Posner cueing might be expected). We discovered a new phenomenon of “express microsaccades” that were highly precise in time and direction. We used this discovery to refine our understanding of why microsaccades might be triggered during Posner cueing, showing that there is an oculomotor “set point” that is very systematically modulated at different times after cue onset, and that the instantaneous relationship between eye position and this set point is sufficient to explain when and where microsaccades would be observed. Overall, our work takes a classic phenomenon in cognitive neuroscience, covert attention as studied with Posner cueing, and significantly recasts it from a completely different perspective related to the highly detailed workings of the oculomotor system during the simple act of gaze fixation. Our work has significant implications on potential neural correlates of covert visual attention and fixational eye position dynamics in the brain.
Attention
Print-on-Demand