Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster D in Poster Session D: Wednesday, August 5, 2:00 – 3:45 pm, Kimmel Center, Shorin & Rosenthal Rooms

Encoding models reveal selective differences in face representations in developmental prosopagnosia

Florencia Martinez-Addiego¹, Daniel Stehr², Alish Dipani¹, Bradley Duchaine², Apurva Ratan Murty¹; ¹Georgia Institute of Technology, ²Dartmouth College

Presenter: Florencia Martinez-Addiego

People with developmental prosopagnosia (DP) struggle to recognize faces even though other aspects of their vision remain intact. This suggests that neural representations in DP are altered. Are these differences specific to the face-selective cortex and face stimuli, or a part of a broader visual processing deficit? To address this question, we collected high-quality fMRI data from 7 neurotypical and 3 DP subjects and used ANN-based encoding models to test for differences in neural representations. We find that model prediction scores in the fusiform face area (FFA) were consistently lower in DPs than in NTs. This reduction was not explained by noisier data, as prediction scores in the body-selective extrastriate body area (EBA) were comparable across groups. Category-level analyses further showed that the reduction in FFA predictivity was specific to face images, with no group differences for bodies, scenes, or objects, and no face-specific deficit in EBA. Together, these findings point to selective differences in face representations in DPs in the FFA and show how ANN-based models can be used to characterize representational differences in the brain.

Topic Area: Development, Individual Differences & Clinical Populations