The hippocampus is essential for forming long-lasting episodic memories. Yet, patterns of hippocampal activity associated with an individual event continue to change with experience and learning. What consequence do these changes have for the underlying memory representations? In rodents, the phenomenon of hippocampal pattern change has been termed `remapping' and has most typically been reported in place cells. This allows for remapping to be directly related to spatial coding of the environment. In humans, however, it is less clear how reconfiguration of hippocampal activity patterns (remapping-like phenomena) translates to specific changes in memory representations. Potentially, changes in hippocampal activity patterns reflect changes in the specific features that, when bound together, comprise a memory. The goal of the proposed research is to gain insight into how and why hippocampal activity patterns (and memories more generally) change with learning. We will address this by developing and leveraging innovative and highly-sensitive pattern-based fMRI methods that map specific features of a memory to patterns of cortical activity. In particular, we will use these methods to reconstruct images from memory and to read out the semantic components of memories. This will allow us to measure how individual features of a memory change with learning and to test whether cortically-expressed feature changes are predicted by remapping-like phenomena in the hippocampus. We will specifically target feature representations in lateral parietal cortex, motivated by accumulating evidence that lateral parietal cortex actively represents the contents of memory and is functionally coupled with the hippocampus. The proposed research will bridge rodent and human models of memory while introducing conceptual approaches and analysis methods that have the potential to significantly advance the field. Because the specific brain regions that will be targeted in the research (the hippocampus and lateral parietal cortex) are frequently implicated in disease- and stroke-related memory impairments, the proposed research will also support critical foundational knowledge that has the potential to guide clinical interventions.
The hippocampus and lateral parietal cortex have each been implicated in successful remembering and are also common targets of disease, pathology and stroke. However, it is not well understood how these regions interact in support of memory. We will test a theoretical framework in which hippocampal activity patterns adaptively change with learning and these changes dictate the specific features that are reinstated in lateral parietal cortex when a memory is retrieved.
