The ability of the mammalian brain to generate internal representations of the outside world have been extensively studied in the past, yet it is not well understood how a structure central to the formation and recall of memory, the hippocampus, enables both a generalized and yet flexible encoding of external features. In particular, it is not well understood how the encoding properties of neural populations in the hippocampus emerge and refine during early-life brain development.
To overcome this knowledge gap, I sought to investigate the encoding of spatial features in the CA3 region of mouse hippocampus while animals repeatedly explore linear environments. For this, I labelled neural population in CA3 with a fluorescent calcium indicator and trained animals to move head-fixed under a two-photon microscope on a linear treadmill, and to explore familiar and novel haptic environments comprised of distinct tactile cues.
By measuring the activity changes of 100s of cells simultaneously, I could identify a subset of CA3 neurons following a generalized, or rigid, encoding scheme with their spatial activity profile being tuned to the general structure of the treadmill task, while another subset of neurons followed a plastic encoding scheme, showing spatial tuning and activity remapping to specific haptic features on the treadmill. This suggests that both encoding schemes are employed in parallel during explorative behaviors, with rigid neurons encoding task structure and plastic neurons encoding specific features.
Current work focuses on the formation, update, and refinement of rigid–plastic encoding schemes in the developing brain. For this, infant (and adult) mice are trained to move through virtual reality environments, to overcome limitations of movement of very young animals on the haptic treadmill apparatus, and to enable spatial manipulations of the virtual reality environment that are impossible in real-world tasks (e.g. teleportations).
