Poster F1, Tuesday, March 28, 8:00 – 10:00 am, Pacific Concourse
Reconstructing Changes in the Spatial Deployment of Attention According to Environmental Statistical Structure
Anthony W. Sali1, Tobias Egner1; 1Duke University
Individuals adjust their readiness to shift spatial attention, referred to here as attentional flexibility, according to the statistical properties of the environment (e.g. Sali et al., 2015; JEP:LMC). However, the precise neural mechanisms underlying learned attentional flexibility remain unknown. To this end, we combined functional magnetic resonance imaging (fMRI) with a task requiring participants to hold or shift covert spatial attention among rapid serial visual presentation (RSVP) streams in response to embedded visual cues. Critically, the likelihood of receiving a shift or hold attention cue was predicted by RSVP stream location. As anticipated, the behavioral cost in target-detection RT associated with shifting relative to holding attention was largest in the context associated with frequent hold attention cues and decreased as a function of shift cue likelihood, documenting that attentional flexibility settings were sculpted by statistical task structure. To track the neural processes underlying this learned attentional flexibility, we trained an fMRI encoding model of spatial representation with data from a separate flashing checkerboard task. Inversion of the training weights allowed us to reconstruct maps of spatial selectivity throughout visual and parietal cortices for each location-defined context in the temporal epochs preceding and following each cue presentation. Using a sliding temporal window, we found that shift-cue likelihood systematically modulated the amplitude and spatial breadth of reconstructed representations across location-defined contexts. These results suggest that contextual learning of shift likelihood dynamically modulates spatial attentional priority settings in the visual brain.
Topic Area: ATTENTION: Spatial