Abstract
Our remarked ability to identify faces is mediated by a specialized neural network comprising of visual as well as limbic and prefrontal regions. The Fusiform Face Area (FFA) is a key region in this network, showing consistent stronger response to visual faces than to non-face visual stimuli. Evidence collected over the years suggests that face selectivity in the FFA becomes more pronounced during childhood along with increased face identification proficiency. However it is unknown whether visual exposure is necessary for the development of this selectivity. Visual-to-Auditory Sensory Substitution Devices (SSD) convey information that is usually perceived visually via the auditory modality in a shape preserving manner, thus can be used in the investigation of face perception in the absent of sight. Using the EyeMusic Visual-to-Auditory SSD that encodes color information as well as information regarding shape, we developed a training program designed to teach blind persons the identification of colorful cartoon faces. 8 congenitally blind participants that took part in the program were able to correctly identify the faces, tell apart up-right from inverted orientation, and discriminate familiar faces from un-familiar faces. In order to identify the neural mechanism mediating auditory identification of faces we conducted an fMRI experiment that included auditory stimuli of faces from 4 categories: familiar faces, inverted faces, unfamiliar faces, and scrambled faces. Using a region-of-interest analysis we show selective activation to faces compared with scrambled faces in the left FFA. Whole-brain analysis shows selective response to faces in the bilateral ventral visual cortex. Our results suggest that face selectivity in the ventral visual stream develops without visual exposure. Additionally, while considered a visual area, our results suggest that FFA responds to facial stimuli conveyed via the auditory modality as well.
Meeting abstract presented at VSS 2016