Abstract
Dynamic objects are ubiquitous in our visual environment. In previous work we identified a network of brain regions recruited during the perception of dynamic objects, and provided evidence that many regions within higher-level and retinotopic visual cortex encode invariant information about dynamic objects (Pyles & Tarr, 2013, 2014). However the structural connectivity of this network remains largely unknown. Here we combine fMRI with diffusion-weighted imaging and deterministic fiber-tracking to investigate the white-matter connectivity between these areas. Dynamic object-selective regions were identified with a new fMRI localizer using short animations of moving, articulating novel objects, contrasted with phase scrambled versions of the same animations. Subjects also participated in a diffusion spectrum imaging scan using a 257 direction sequence. During viewing of dynamic objects we observed the recruitment of large regions of occipito-temporal cortex, substantially overlapping with LOC and hMT+ (designated dynamic-LOC) as well as regions of parietal cortex. These functional regions as well as results from retinotopy and a MT/MST localizer were used as regions of interest for deterministic fiber-tracking to map the white matter connections between these areas. Additionally we investigated the connectivity of these regions to the rest of the brain in an unconstrained analysis. Results show short range connections between nearby areas selective for dynamic objects, as well as connectivity to retinotopic cortex. Fiber streamlines originating in the large dynamic-LOC region showed longer range connections to regions of anterior temporal lobe and frontal lobe. Finally, we explicate the relationship of these tracts to the inferior longitudinal fasciculus, and the vertical occipital fasciculus (a tract connecting dorsal and ventral cortex (Yeatman et al., 2014)). A better account of the structural connectivity of this network and its relation to other major tracts in visual cortex will improve our understanding of its functional organization and inform future research.
Meeting abstract presented at VSS 2015