Testing the Robustness of Newly Acquired Face Memory: An fMRI Study

Abstract

Numerous faces are encountered on daily basis. While some remain novel, others may hold the propensity of becoming familiar. In a series of studies, we have shown that multiple exposures coupled with sufficient variation in expression and pose can help transform the identity of a novel face into a perceptually familiar one (Shyi & He, 2011; Shyi & Lin, 2014; Cheng & Shyi, 2014). However, the mechanisms for transforming a changeable aspect (e.g., expression) of a face into invariant features for identity is not clear. Moreover the robustness of these newly acquired face memory is also unknown. We addressed these questions by first training participants to learn a set of novel faces via the combination of multiple exposures with variations in expression. Their recognition memory for the newly learned faces were then tested twice consecutively while functional scanning of their brains was carried out. Behavioral results shown that the accuracy for both the original and generalized images decreased from the first to the second test, while the false alarm for distractor images maintained at the same level, suggesting that the memory strength for both original and generalized images decayed between the two recognition tests. Brain imaging results on the other hand showed that the extent of memory decay for the original images was positively correlated with the decrement of brain activity in both the bilateral posterior and middle region of superior temporal sulcus (STS). In contrast, the memory decay for the generalized images exhibited no such correlations. These findings suggest that (a) as a region selective for processing changeable aspect of faces, STS can help create invariant features for identity to support recognition memory, and (b) the neural basis underlying generalized face images differs from those underneath the original images.

Meeting abstract presented at VSS 2017