The geometry of face memory

Abstract: The ability to recognize familiar visual objects is critical to survival. Neurons in inferotemporal (IT) cortex represent the percept of visual objects using a distributed axis code. However, the network code for the memory of visual objects remains elusive. Here, we measured neuronal responses to familiar and unfamiliar faces in two face patches, AM and PR. In both areas, familiar and unfamiliar faces were represented in distinct subspaces. The familiar face subspace was shifted relative to the unfamiliar face subspace at short latency and then distorted to increase neural distances between familiar faces at long latency. Our results suggest that memories of familiar faces are represented in IT and perirhinal cortex by a distinct long-latency code that is optimized to distinguish familiar identities.