Schedule of Events | Symposia

Reactivation of cortical representations predicts vividness and precision of objects’ color and location

Poster Session C - Sunday, March 30, 2025, 5:00 – 7:00 pm EDT, Back Bay Ballroom/Republic Ballroom

Xinhao Wang¹ (xinhao.wang@hu-berlin.de), Simon Davis², Roberto Cabeza^1,2; ¹Humboldt University of Berlin, ²Duke University

Several representational fMRI studies have shown that the similarity of multivariate activation patterns during encoding and retrieval (Encoding-Retrieval Similarity, ERS) predict memory performance. These studies typically focused on whether a particular item is encoded (item memory). In the current ERS study, in contrast, we examined relational memory for specific object features, color and location, and measured also memory precision. Participants encoded everyday objects presented in an arbitrary color and location, and later they were presented with the objects in greyscale in the center of the screen and asked to recall either the original color or location. For the target feature, subjects first rated the vividness of their feature memory and then reported the feature along a continuous scale. We calculated the ERS for each pair of encoding and retrieval trials. Mixed-effect linear models were used to assess the relation between ERS and memory quality on a trial-by-trial basis. Overall, successful feature memory was predicted by ERS in visual cortex and lateral prefrontal cortex. Interestingly, we found a dissociation between subjective vividness and objective precision: whereas vividness was predicted by ERS in prefrontal and temporal cortices (color>location), precision was predicted by ERS in visual cortex (location>color). To our knowledge, this is the first study to investigate the relation between ERS and relational memory for object features. We observed a dissociation between the neural substrates for vividness and precision, and a further dissociation between color and location information. Taken together, the results clarify the neural mechanisms of relational memory for object features.

Topic Area: LONG-TERM MEMORY: Episodic