JRG/3: The Architecture of Attentional Processes in Active Vision

Head and eye movements shape our perception of the visual world. With no apparent effort, they bring relevant events and objects into view and select relevant data from a rich visual environment. Swift movements of our eyes and heads (saccades), however, also entail several costs. Objects that have fixed places in the world rapidly slip across the retina several times per second. We will use a combination of motion-tracking, visual psychophysics, and computational modelling to investigate the role of attention in visual perception before natural eye-head gaze shifts, and reveal their significance for transsaccadic continuity. Studies in which head movements were minimized have shown that spatial attention shifts to the targets of saccades, and that the retinocentric distribution of attention is updated just before the saccade, to continuously cover the relevant locations in space. Natural eye-head gaze shifts have not yet been examined and the role of feature-based attention remains unspecified. This research program will complete the picture of the attentional basis of active vision. Our aims are to (1) understand the dynamics of spatial attention before unconstrained eye-head gaze shifts, (2) characterize feature-based attention before gaze shifts, (3) isolate the respective roles of spatial and feature-based attention in perceptual and behavioural continuity across gaze shifts, and (4) present a quantitative theoretical model that accounts for the established attentional processes at the systemic level.