PHD2023-03

PHD2023-03

Behavioral grammar and synaptic rules of the social brain

Host laboratory and collaborators

Pascale CHAVIS - Institut de Neurobiologie de la Méditerranée / pascale.chavis@inserm.fr

Jean-Marc FREYERMUTH - Institut de Mathématiques de Marseille (ALEA team) /  jean-marc.freyermuth@univ-amu.fr

Abstract

We aim to develop a quantitative framework for one of the cardinal brain output, the social behavior, and to understand how the rules of the “social brain” are transformed in psychiatric disorders (e.g., autism, schizophrenia). Interconnected structures in the social brain (e.g., prefrontal cortex, ventral hippocampus, accumbens) generate adapted behaviors. We postulate that 1/ complex behaviors are the product of elementary behavioral modules (« syllable ») organized by a grammar into meaningful sequences to generate appropriate actions, and 2/ neuropsychiatric diseases distort this grammar. We will leverage state of the art statistical modeling and learning technics to dissect natural social behavior in mice, neuronal ensemble activity in freely-moving mouse and in genetic model of psychiatric disorders. We will decipher how complex behavior specified by the genetic context is (dis)constructed from identifiable components and (dis)organized in a predictable manner.

Keywords

Social Brain - Naturalistic behavior - Neuronal activity - The Reeler Mouse model - Data fusion - Big data - Machine learning - Probabilistic modeling

Objectives

Three main objectives will be addressed in healthy and in the reeler diseased mouse model: 1/ characterize the underlying modular structure of spontaneous and naturalistic behavior in mice; 2/ identify specific neural representations for syllables and grammar in interconnected structures of the social brain (prefrontal cortex, ventral hippocampus, accumbens) during behavior; 3/ infer the predictable rules of a behavioral "grammar" specified by the genetic context (wild-type vs reeler).

Proposed approach (experimental / theoretical / computational)

Experimental: Multimodal description of brain function and activity in health and disease.
We will combine Live Mouse Tracker, a 3D imaging deep learning-based method to continuously monitor groups of freely interacting mice with DeepSqueak, a Machine Vision tool for bioacoustic identification of ultrasonic communications. During behavior, the underlying neural representations will be objectified by recordings neural ensemble activity (joint photometry of neurotransmitter and calcium indicators).

Computational: In order to identify elementary behavioral syllables, we will build a machine learning pipeline for data fusion. The pretreatment stage of our multimodal data will involve sparsity inducing representation of signals, dimension reduction technics and topological data analysis to seek for salient features. Then, we will infer the syllables organization and concomitantly to activity of neuronal circuits of the social brain using probabilistic modelling based on hierarchical Dirichlet process mixture.

Interdisciplinarity

The project operates at the very frontiers of (i) basic research aiming to gather fundamental new knowledge in the field of neurosciences and neurodevelopmental brain disorders and (ii) modern statistical, machine learning methods and information theory, and clearly goes beyond the current state-of-the-art in these research areas.
Because of its multidisciplinary nature and its heuristic approach, the anticipated findings will generate a comprehensive set of experimental data of interdisciplinary character and provide a significant impact in understanding fundamental brain functions, as well on health care and public.

 

Expected profile

Robust theoretical background in Neurosciences.
Strong attraction for computation, statistics and big data.
Strong collaborative and positive team spirit.

Is this project the continuation of an existing project or an entirely new one? In the case of an existing project, please explain the links between the two projects

This is an entirely new project buiding up onto the previous discoveries of both supervisors laboratories.

2 to 5 references related to the project

  • De Chaumont F. et al. (2019). Nat Biomed Eng., 3(11):930-942
  • Jossin Y. (2020) Biomolecules, 26;10(6):964
  • Chazal F. and Michel B. (2021). Frontiers in Artificial Intelligence, 4:667963
  • Tong M. et al (2020). Information Fusion, 57, 115-129
  • Liang P. et al (2010). In T. O'Hagan and M. West (Eds.), The Handbook of Applied Bayesian Analysis, Oxford University Press.

3 main publications from each PI over the last 5 years

CHAVIS Pascale

  • Labouesse M.A., Lassalle O., Richetto J., Iafrati J., Weber−Stadlbauer U., Notter T., Geschwind T., Pujadas L., Soriano E., Reichelt A.C., Labouesse C., Langhans W., Chavis P.* and Meyer U*. * Shared senior authorship. Hypervulnerability of the adolescent prefrontal cortex to nutritional stress via reelin deficiency. Molecular Psychiatry, 2017, 22(7):961
  • Silva-Hurtado T.J., Giua G., Lassalle O., Murphy M.N., Wager-Miller J., Mackie K., Manzoni O.J. and Chavis P. Cannabinoid during adolescence phenocopies reelin haploinsufficiency in prefrontal cortex synapses. bioRxiv 2021.12.22.473793; doi: https://doi.org/10.1101/2021.12.22.473793
  • Iezzi D., Caceres-Rodriguez A., Chavis P*. and Manzoni OJJ*. * Shared senior authorship. In utero exposure to cannabidiol disrupts select early-life behaviors in a sex-specific manner. Translational Psychiatry, 2022, in press

 

FREYERMUTH Jean-Marc:

  • Rey A., Fagot J., Mathy F., Lazartigues L., Tosatto L., Bonafos G., Freyermuth J-M., Lavigne F. Learning higher-order transitional probabilities in nonhuman primates. Cognitive Science, 2022, 46(4).
  • Aston J., Autin F., Claeskens G., Freyermuth J.-M. & Pouet C. Minimax optimal procedures for testing the structure of multidimensional functions. Applied and Computational Harmonic Analysis, 2019, 46, 288-31.
  • Farouj Y., Freyermuth J-M., Navarro L., Clausel M., Delachartre P. Hyperbolic Wavelet-Fisz denoising for a model arising in Ultrasound Imaging. IEEE Transactions on Computational Imaging, 2017, 3(1), 1-10.