A fundamental question in neuroscience is to understand how the brain represents the environment. Many studies approach this question by finding correlates between stimuli and neural activity, but this can only offer a partial picture: brains do not care about representing the outside world, but rather about turning perceptions into actions. Consequently, in this project we hypothesize that the representations of the environment in mice brains are modulated by the behavior of the animal.
Previous studies have shown that the representation of space in HPC is not fixed : the density and precision of place cells (cells that encode a specific location) changes depending on multiple factors such as the presence of rewards, other animals, rich sensory stimuli or multiple available paths . However, how these are modulated by behavior is not known yet. In line with our idea that the behavior of animals affects how they represent the environment we propose that:
Hippocampal place cell distribution is modulated by behavior, with more unpredictable behavior leading to denser, more precise place fields, and stereotypical, predictable behavior leading to fewer, imprecise place fields.
To make this hypothesis concrete and testable, we collected neural activity data from the hippocampus (HPC) of mice while they perform two different behavioral tasks in the same environment (Fig. Left: Maze indicating the two different behavioral areas. Right: Animal trajectory for a goal-directed and explorative behavior). The goal of this thesis will be to study how the representation of space, visible in HPC, depends on the behavior of the animal, and if time allows, on the choices and evidence accumulated by the animal, which is visible on HPC.
The project entails analyzing animal behavior using standard models such as the Kalman filter, survival analysis, and vector field maps. You will aim to analyze learning performance and then evaluate how the neural activity, which is recorded using calcium imaging, correlates with behavior, characterized by the predictability of the movement. You will employ dimensionality reduction, decoding techniques, and approaches from information theory, potentially incorporating machine learning methodologies.
We are searching for independent Master students who possess a strong coding background (Matlab or Python) with a commitment to good coding practices. Experience with data analysis would be a big plus.
- Ziv Y, Burns LD, Cocker ED, Hamel EO, Ghosh KK, Kitch LJ, El Gamal A, Schnitzer MJ. Long-term dynamics of CA1 hippocampal place codes. Nat Neurosci. 2013 Mar;16(3):264-6. doi: 10.1038/nn.3329 Masaaki Sato, Kotaro Mizuta, Tanvir Islam,
- Masako Kawano, Yukiko Sekine, Takashi Takekawa, Daniel Gomez-Dominguez, Alexander Schmidt, Fred Wolf, Karam Kim, Hiroshi Yamakawa, Masamichi Ohkura, Min Goo Lee, Tomoki Fukai, Junichi Nakai, Yasunori Hayashi, Distinct Mechanisms of Over-Representation of Landmarks and Rewards in the Hippocampus,Cell Reports, 2020. doi: https://doi.org/10.1016/j.celrep.2020.107864.
Additional Interesting Reviews and Publications
Supervisors: Roman Boehringer (roman (at) ethz.ch), Pau Aceituno (pau (at) ini.ethz.ch)