Neurobiologie intégrative des comportements flexibles dans des modèles animaux par imagerie computationnelle : application en conditions normales et pathologiques – FLEXNEURIM

The three teams involved in this project have complementary skills and methodological approaches. They already collaborate on different aspects which constitute the founding of the present project. Their common interest is to propose, in mice model, to extract the most significant behavioural pattern common to different pathologies 'lack of flexible behaviour- and to break it down into smaller components 'control of impulsive choice, evaluation of reward value, response selection, ability to switch between options. For that, we have already developped behavioural tasks specific to mice and obtained preliminary data for a role of the prefrontal cortex (team 1), as it is the case in rats, human and non human primates and of nicotinic (team 1) and neuroligin receptors (teams 1 & 3, published common work). The use of a prototypical task, social interaction paradigm, in different conditions provide us with a robust framework for the study of flexible behaviours with at request modulation of different kind of motivations. We now want to capitalize on recent advances at two levels to further comprehend the neural basis and modality of function and dysfunction of high cognitive processes: 1) recent advances in the field of computer vision and signal processing will allow to automate the analysis of recorded movies documenting the behaviours of mice in different setups where a two-animal distinction is technically difficult 2) we have complemented our behavioural tools by methods to measure both motor and cognitive impulsivity in mice. Our goals will be to: 1) extract automatically (meta)data describing the spatio-temporal trajectories of animals and the timing of behavioural events such as approaching and sniffing at specific target areas or inter animal contacts 2) to establish a complete system for recording and analysing animal movement in 3D+t. This will imply the design of a complete acquisition system using several cameras (vertical and horizontal) that give access by reconstruction to the full 3D scene view. This enhanced view of the scene will open completely new opportunities in the behavioural analysis arena as it will give access to the true geometry of the animals' movement. 3) combine specific video and audio (meta)data and store the resulting analytical information in a database for later retrieval. A key aspect of this project is therefore to develop various data analysis and image processing algorithms necessary for the tracking of mice and the subsequent correlation of video and audio data. Indeed, ultrasonic vocalization may be a neglected but important feature of mice social interaction in specific situations that we will integrate to our corpus of data. We will thus decipher building blocks of highly complex behaviours with new automated methods 4) correlate these building blocks during normal functions (normal mice) or dysfunctions (in mice with selective brain lesions, knockout for nicotinic or neuroligin receptors) with the expression of early genes in specific brain regions 5) to establish the possibility of behavioural markers 'such as impulsivity- to be predictive for other behavioural dysfunction 'such as impaired behavioural flexibility- and to correlate these behavioural markers to neurobiological markers 'gene and receptor expression. In animals lacking some specific synaptic or scaffolding proteins, we will dissect precisely which subcomponents of complex behaviours may be altered and which are preserved. These complementary approaches will give to behavioural neuroscience precise tools for testing complex behaviours in mice and tools for analysing finely such behaviours, therefore capturing much more than gross deficits. Our project brings together different specialists from different fields in neurosciences, genetic and image processing sharing the view that an integrated, flexible and synergistic methodological framework is the only way to arrive at a comprehensive understanding of the interplay between animal behaviour and molecular events. This understanding is necessary for the fundamental study of neural functionning and for animal modeling of pathological conditions.