Abstract
Introduction: Early blind individuals show BOLD responses to a wide variety of tactile and auditory stimuli within occipital and ventro-occipital regions that are thought to be primarily visual in sighted individuals. These responses are widespread for sound vs. silence comparisons (e.g. Lewis et al., 2010), though specific task comparisons suggests 'functional specificity': e.g. selectively enhanced responses to motion in hMT+ and to Braille in the visual word form area. One possible source of these selective auditory responses is auditory belt and parabelt regions .These areas contain complex feature detectors tuned for specific combinations of behaviorally relevant frequencies that may play a crucial role in abstract, higher level sound representations. Our goal was to compare responses to naturalistic stimuli within auditory belt and parabelt (Moerel et al. 2013) and occipital cortex across early blind and sighted individuals. Methods: Using fMRI we measured responses to 56 3s naturalistic auditory scenes in 3 early blind and 4 sighted individuals. Subjects performed a trivially easy one-back task: reporting when an auditory scene was repeated. For each voxel, we found the frequency tuning profile that best predicted that voxel's BOLD response across the set of stimuli using either pRF modeling (Dumoulin et al., 2008) or a close variant previously used to estimate auditory frequency tunig (Moerel et al., 2015) in belt and parabelt regions. Results: Robust BOLD responses to sound vs. silence and/or sound vs. noise comparisons were found within primary and secondary auditory areas across all individuals. For both subject groups we could estimate frequency tuning within auditory areas using naturalistic stimuli. Surprisingly, despite robust responses in auditory cortex, we saw little evidence of auditory responses within occipital regions in early blind individuals. One possibility is that recruitment of occipital cortex only occurs when subjects are actively engaged in parsing auditory stimuli.
Meeting abstract presented at VSS 2017