Subtle visual change in a virtual environment induces heterogeneous remapping systematically in CA1, but not CA3

Distinct Neural Mechanisms for Heading Retrieval and Context Recognition in the Hippocampus During Spatial Reorientation

“…Here, we observe that information about geometry and features is integrated in area CA1 and that feature-sensitivity in distinct CA1 cell groups differentially contributes to context recognition and heading retrieval. Our data is in agreement with recent findings showing that CA1 subpopulations with distinct remapping properties differentially participate in context recognition vs. perception of cue alterations (Shin et al, 2022), suggesting that CA1 is the ideal substrate to integrate cognitive processes for efficient performance.…”

“…For example, a recent study examined responses in rat area CA1 after animals were exposed to a rich virtual spatial environment where levels of fog were manipulated, making the appearance of the context look different across trials. This study demonstrated that while a small subpopulation of CA1 cells discriminated contexts with or without fog by shifting the cells' preferred firing location, similarly to the context effects we observe in FS cells, most neurons did not alter their firing location but responded to distinct levels of fog through rate remapping (Shin et al, 2022). Based on these previous results, we hypothesized that information about features and geometry could be integrated in the active network, including both FS and FI cells, through rate changes.…”

“…In the current study, we attempted to close this knowledge gap by recording from hippocampal area CA1 while mice performed an extended version of the two-context reorientation task. We focused on CA1 because different place cell subpopulations within this region have been shown to be involved in detecting distinct aspects of contextual change (Shin et al, 2022), a characteristic that would be essential to simultaneously code context recognition and heading retrieval. First, we investigated if geometry-based heading retrieval persisted over time in a paradigm where animals were attending to non-geometric features for context recognition.…”

Predictive Sequence Learning in the Hippocampal Formation

“…Existing computational frameworks (Stachenfeld et al, 2017;Whittington et al, 2020;Mc-2 Namee et al, 2021;George et al, 2021) have successfully modeled cognitive functions of the hippocampus and reproduced the statistics of place cell under various task conditions. However, these models do not provide an implementation of these cognitive functions based on neural mechanisms or account for the distinct encoding and firing properties of neurons in CA3, CA1, and the dentate gyrus (DG) (Dong et al, 2021;Shin et al, 2022;Lee et al, 2004b,a).…”

“…Existing computational frameworks (Stachenfeld et al, 2017; Whittington et al, 2020; McNamee et al, 2021; George et al, 2021) have successfully modeled cognitive functions of the hippocampus and reproduced the statistics of place cell under various task conditions. However, these models do not provide an implementation of these cognitive functions based on neural mechanisms or account for the distinct encoding and firing properties of neurons in CA3, CA1, and the dentate gyrus (DG) (Dong et al, 2021; Shin et al, 2022; Lee et al, 2004b,a). For example, CA1 neurons are more responsive to unexpected signals than neurons in other hippocampal areas (Kumaran & Maguire, 2006; Knight, 1996; Duncan et al, 2012) and their activity decays over a time scale of weeks in familiar environments, faster than neuron in other subregions (Fig.S1).…”