A multi-pronged approach to identifying functional subdivisions of the human pulvinar
63.427, Wednesday, 21-May, 8:30 am - 12:30 pm, Banyan Breezeway
Jason Fischer1, Nancy Kanwisher1; 1Department of Brain and Cognitive Sciences and McGovern Institute for Brain Research, Massachusetts Institute of Technology
The pulvinar nucleus of the thalamus is well-positioned to play an integrative role in vision, owing to its widespread connections with the visual cortex. However, the precise function of the pulvinar remains elusive, and likewise, its functional organization is still poorly understood. Here we adopted a multi-pronged approach to uncovering functional subdivisions within the human pulvinar. We collected fMRI data during resting state (in the absence of visual input or a task), and, separately, while subjects performed an object discrimination task designed to test the position selectivity of visual responses in the pulvinar. Within the resting state data, we used k-means clustering to parcellate the pulvinar into subregions with distinct resting signals. We then tested the functional characteristics of each subdivision in two ways: i) we measured position selectivity within each subdivision using multivariate pattern analysis on data from the object discrimination task, and ii) we computed the pattern of resting state correlations across the brain using each pulvinar subdivision as a seed. In preliminary data from four subjects, we found lateral, medial, and inferior clusters within the pulvinar that displayed distinct profiles of resting state activity. Within independent data, these clusters also showed differences in position selectivity, with the lateral cluster supporting the best decoding of object locations. Mirroring this pattern of position selectivity, the lateral pulvinar cluster was strongly correlated with early visual cortex in the resting state signal. These preliminary results point toward the presence of functionally distinct subregions within the human pulvinar and demonstrate an approach for parcellating the pulvinar based on a combination of resting state and task-related fMRI.