FriSem

Speaker: Arielle Keller, Graduate Student (Williams Lab, Neuroscience program) John Cocjin, Graduate Student (Lee Lab, Bioengineering program) Akshay Jagadeesh, Graduate Student (Gardner Lab, Psychology program), Stanford UniversityThe Neural Basis of Category Selective AttentionSelective attention is a critical function that allows the brain to focus on relevant information while ignoring distraction. However, the top-down influence of attention on visual category representations remains to be fully understood. To explore the effect of attention on category representations in high-level visual cortex and to determine the sources of this top-down modulation, we combine data-driven and hypothesis-driven approaches. Using a geodesic spotlight search across the entire brain in conjunction with a multi-voxel pattern analysis (MVPA) decoding approach, we are able to (i) identify a set of brain regions that represent category information and (ii) determine that attention affects the encoding of category information. Additionally, we use this method to determine the most informative scale of neural representation for studying the encoding of categories. Our results indicate that (1) category representations are encoded at an intermediate scale between that of a single voxel and an anatomical region, (2) unattended categories are still decodable from cortex, though attention enhances category representations, and (3) attention enhances representations most in the parietal cortex. In a complementary approach, we utilize a novel method for extracting top-down activity from bottom-up activity in the BOLD signal. Using this technique, we examined whether regions in the intraparietal sulcus (IPS) are a potential source of top-down attentional modulation of category-selective regions in ventral temporal cortex (VTC). We find that there is a positive correlation between the residuals in the IPS and functionally-defined category selective ROIs in the VTC. Notably, the right IPS is more highly correlated with VTC than the left IPS. We also use this method to show that attention and active ignoring often involve similar top-down modulation, and explore the extent to which top-down activity is influenced by category preference. Together, our combined approaches provide insights to the neural mechanisms of selective attention to objects and their modulatory role on distributed information.