Unsupervised analysis of mouse social behaviour reveals how hyperactivity impacts social aspects in a mouse model of autism

Unsupervised classification of animal behaviour triggers growing interest. Tracking algorithms are allowing more and more efficient recording and thus produce increasing amount of data to be classified. In this context, we propose two levels of analysis to describe the behaviour of mouse models of neuropsychiatric disorders. We studied mice mutated in Shank2, a gene associated with autism spectrum disorders. We compared them with their wildtype littermates, regarding the proportions of behavioural events classified along time, and also the thresholds used to define these events. The developed algorithm allowed to understand to what extent hyperactivity could explain social deficits in Shank2 mutant mice.