HIPPOCAMPAL 3D CODING IN MICE: VOLUMETRIC PLACE REPRESENTATIONS

Navigation is inherently three-dimensional, yet hippocampal spatial coding is still predominantly studied in two dimensions. With the increasing use of naturalistic 3D behavioral paradigms in rodents, including vertical locomotion and climbing, we asked how hippocampal CA1 ensembles encode 3D space during repeated free exploration, and whether changes in illumination reshape these representations. Adult C57BL/6 mice explored a multi-level 3D maze for 30 min per session across 7 days in different illumination conditions. The maze included segments sharing the same (x,y) coordinates at different heights, and vice versa, enabling dissociation of horizontal and vertical components of space. 3D trajectories were reconstructed from top-view camera recordings using DeepLabCut and custom scripts through the integration of maze topological constraints. CA1 calcium activity was recorded with a head-mounted miniscope. Using an information-theoretic framework to disentangle mixed selectivity, we quantified tuning to 3D position (spherical place fields), 2D position (cylindrical place fields), z-position (planar place fields). We detected substantial fractions of 2D- and 3D-tuned neurons, a smaller fraction of height-selective cells, and additional space- and speed-related selectivities. Illumination did not produce consistent changes in exploration and did not induce systematic reorganization of the formed 3D cognitive map after the switch. Together, these results demonstrate that CA1 populations jointly encode volumetric space during naturalistic 3D exploration, and these cognitive maps are robust to changes in illumination. The work was supported by Non-Commercial Foundation for Support of Science and Education "INTELLECT".