Abstract
The hippocampus plays a critical role in spatial and episodic memory. Mechanistic models predict that hippocampal subfields have computational specializations that differentially support memory. However, there is little empirical evidence suggesting differences between the subfields, particularly in humans. To clarify how hippocampal subfields support human spatial and episodic memory, we developed a virtual reality paradigm where participants passively navigated through houses (spatial contexts) across a series of videos (episodic contexts). We then used multivariate analyses of high-resolution fMRI data to identify neural representations of contextual information during recollection. Multi-voxel pattern similarity analyses revealed that CA1 represented objects that shared an episodic context as more similar than those from different episodic contexts. CA23DG showed the opposite pattern, differentiating between objects encountered in the same episodic context. The complementary characteristics of these subfields explain how we can parse our experiences into cohesive episodes while retaining the specific details that support vivid recollection.
Highlights
The hippocampus plays a critical role in spatial and episodic memory
Recognition memory performance was indexed by evaluating responses to new and old items in the object recognition test completed during MRI scanning
Remember rates varied by spatial context, t(22) = 3.33, p = 0.003, d = 0.21, such that recognition was slightly lower for items that had been studied in the gray house
Summary
The hippocampus plays a critical role in spatial and episodic memory. Mechanistic models predict that hippocampal subfields have computational specializations that differentially support memory. Most mechanistic models suggest that the hippocampal subfields play complementary roles in spatial and/or episodic memory[3,4,5,6,7] These models generally predict large differences between neural coding in CA1 and CA3, between-subfield differences at the level of single units in rodents and in overall activity in human fMRI studies have been relatively modest. Both CA1 and CA3 have been implicated in representations of temporal[8,9] and spatial[10,11] contextual information. We tested whether multi-voxel patterns elicited during item recognition carried information about the associated spatial (house) or episodic (same house and video) context
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
