Abstract
Attending to a specific feature, such as a color or direction of motion enhances the response of neurons that are selective to that feature, even outside the focus of spatial attention(Saenz, Buracas, & Boynton, 2002). This spread of activity across space can act to facilitate perception of behaviorally relevant stimuli. However, when multiple features are relevant, what is the effect on behavior and brain activity? We tested this by comparing behavioral performance and fMRI responses while observers viewed fields of overlapping moving dots (up and down) at two locations (left and right of fixation). Subjects performed a divided attention task in the MRI scanner by detecting target events within one field on each side of fixation. Attention could be directed to fields moving the same direction on both sides, or to different directions. We found that performance was better when dividing attention to the same direction on both sides than to different directions, replicating Saenz et al. (2003). We then used an inverted encoding model approach on fMRI responses (IEM: Brouwer and Heeger 2009; Foster et al. 2017) to examine the effect of attention on neuronal subpopulations selective to specific directions of motion at specific locations. In areas V1, V2, V3 and MT we found larger responses to the attended direction than the unattended direction when subjects attended to the same direction on both sides. However, when subjects attended to different directions, no overall difference was found in the response across the four fields. These results can be explained by a natural extension of the normalization model of attention to include a spatial spread of feature-based attention gain which serves to mutually enhance responses when attention is directed to the same direction on both sides, but suppresses responses when attention is directed to opposite directions of motion.
Acknowledgement: NIH EY12925