Abstracts

Rationale: Deficits in face processing and memory are well documented in patients with temporal lobe epilepsy. This is thought to be due to disruption of processes in hippocampus and medial temporal lobe (MTL). These are the primary regions investigated in face memory studies, however, several recent studies have shown evidence of activation in retrosplenial cortex (RSC) in response to familiar faces. The exact contribution of the RSC to human cognition remains largely unclear. Past studies have focused primarily on its role in navigation and spatial processing and its role in face memory is poorly understood. Here, we investigated the contributions of RSC to the memory of faces in epilepsy patients using intracranial recordings. Methods: We studied the responses to human faces from 236 electrodes in RSC and 347 in MTL from a large patient cohort (n = 66) who underwent neurosurgical implantation of intracranial electrodes for localization of seizure foci in intractable epilepsy. To test activity in these regions, we used a famous face identification task, naming celebrities from photos of their face and a visual discrimination task between faces based on gender. To establish category selectivity, we also investigated responses to famous landmarks, common objects, face parts and scrambled faces. Broadband gamma activity (BGA; 70-150Hz) was measured in these electrodes to index local processing. Results: Our results showed that the right RSC has a distinct region in sub-parietal sulcus that shows strong selectivity for faces. The MTL displayed an anterior-to-posterior gradient of selectivity with entorhinal regions showing face selectivity and parahippocampal cortex displaying selectivity to landscapes. The retrosplenial face area showed sensitivity to the ability of the patients to identify to the faces, with correctly identified faces showing significantly greater BGA, an effect comparable to that seen in the MTL. This identification effect was also apparent even when the patients were not performing a naming task but instead discriminated between faces. We compared the timing of the emergence of this identification effect between RSC and MTL within patients who had electrodes in both regions. This showed no significant difference in timing of the distinction between correctly and incorrectly identified faces between the two regions. Conclusions: Our novel findings of the role of the RSC in the cortical network for identification have implications for the functional impairments in face processing seen in MTL epilepsy. Plasticity of this region may determine the extent of recovery following resections. This finding is also significant as RSC is one of the earliest regions to show hypometabolism in Alzheimer’s Disease and may explain the deficits seen in the early stages of the disease which have previously been attributed to MTL dysfunction. Funding: NIH National Institute on Deafness and Other Communication Disorders DC014589