Deficient functional MRI selectivity and connectivity in developmental prosopagnosia is specific to face regions


Xian Li, Joseph Arizpe, David Rothlein, Michael Esterman, and Joseph DeGutis. 10/20/2020. “Deficient functional MRI selectivity and connectivity in developmental prosopagnosia is specific to face regions.” In Vision Sciences Society, 11th ed., 20: Pp. 827-827. Journal of Vision. Publisher's Version


Developmental prosopagnosia (DP) is characterized by severe facial recognition deficits, though it is currently debated whether only face-specific neural mechanisms are disrupted. A recent fMRI study found widespread decreases in category-selective (e.g., scene, body) regions in DPs (Jiahui et al., 2019), suggesting disruption may not be face-specific. To further evaluate neural deficits in DP, we scanned 23 DPs and 23 controls in task-based fMRI (dynamic localizer of faces/scenes/objects/bodies) and resting-state fMRI. In the task-based fMRI, DPs exhibited reduced face-selectivity across ventral face areas (anterior temporal lobe-ATL, fusiform face area-FFA, and occipital face area-OFA). This category-selectivity reduction was not found for scenes, bodies or objects in their respective regions, suggesting that the deficit of DP is limited to faces. Moreover, only the face area selectivity correlated with face recognition ability (average face-selectivity across OFA/FFA r= .37, p<.05), whereas scene, body, and objects area selectivity did not (all p’s>.4). This shows that our measure of face-selectivity reflects behavioral performance on face identification. In the resting-state fMRI, DPs showed reduced functional connectivity (FC) across most region pairs within the face network. While FC within the body network showed some mild reduction, this reduction was not observed for FC within the scene or object network. Moreover, only the FC in the face network correlated with face recognition ability (average face-network FC r=.46, p<.001), while scene, body, and objects area FC did not (all p’s>.13). This suggests that only resting connectivity between face regions is related to face recognition ability. Interestingly, we do not find a significant correlation between face-selectivity and face-network FCs, indicating they explain independent variance in face-recognition ability. Together, the fMRI selectivity and FC results, as well as their association with behavioral performance across all examined categories, suggests that DPs’ functional neural deficits are largely limited to the faces.