PHD2018-14

Understanding large-scale spatial navigation using virtual reality, multi-site electrophysiological recordings in the hippocampus and computational modeling

Host laboratory and collaborators

Jérôme Epsztein / INMED / jerome.epsztein@inserm.fr

Hervé Rouault / CPT / herve.rouault@gmail.com

Scientific background

Animals can flexibly navigate their environment by computing different paths to similar targets. This ability is thought to rely on an internal cognitive map. To date, spatial navigation has been well studied in rodents navigating small, un-cued laboratory environments. However, the mechanisms of large-scale navigation in more complex, cue-rich environments are far less understood. Place cells in the hippocampus, a key structure for spatial navigation, are pyramidal neurons which fire action potentials whenever an animal is at a specific position within its environment. Interestingly, the resolution of place cells spatial code differs between the dorsal and ventral parts of the hippocampus. During large-scale navigation, different part of the environment could be coded at different spatial resolution through “nested hierarchies” of coarse and fine grain coding. This hypothesis is difficult to test in regular sized laboratory environments. In this PhD project, we propose to take advantage of newly developed virtual reality systems for rodents to study large-scale spatial navigation. The project will combine behavioral analyses using virtual reality for rodents, multisite recordings of neuronal activity in dorsal and ventral parts of the hippocampus and computer modeling.

PhD Objectives

The PhD project will have three major objectives: Aim1: Adapt and further develop an existing virtual reality setup to create a large-scale navigation task for rats. Aim2: Perform multi-electrode (i.e. silicon probe) recordings of dorsal and ventral hippocampal neurons activity during this task. Specifically, we will ask how the dorsal and ventral parts of the hippocampus interact during such navigation. Aim3: Build a Continuous Attractor Neuronal Network (CANN) model fed by experimental data to decipher specific computational requirements for large-scale spatial navigation.

Proposed approach 

Experimental: The project will involve the development of a navigation task for rats in custom designed large-scale naturalistic environments using a virtual reality system for rodents already in use in the Epsztein lab. Places within the environment will differ in complexity and availability of external sensory cues (visual, tactile). In a second step, multi-site (silicon probe) recordings of neuronal activity will be performed in both the dorsal and ventral hippocampus which provide precise and coarse spatial coding, respectively.

Computational/theoretical: The place cell firing patterns are thought to result from the recurrent interactions in the CA3 layer of the hippocampus (among others) resulting in a continuous attractor network. Models also integrate the interactions between grid cells for path integration (update of the firing map without local cues) and place cells. This project proposes to revisit these models for large-scale navigation where recorded activities are more diverse and put stronger constraints on cell connectivities. We will explore how several scales (coarse vs. refined) of spatial coding can interact and modify existing continuous attractor models of place cells.

PhD student’s expected profile

The recruited PhD student will have a biology background with a keen interest in spatial navigation and computational neuroscience (familiarity and ease with a modern programming language is strongly recommended).

X