Autism spectrum disorder (ASD) is characterized by heterogeneous phenotypes classically encompassing three domains — difficulties with social interactions, language communication challenges and an overrepresentation of stereotyped behaviors. Mouse models of ASD are effective tools for preclinical studies, but current assays used to quantify the behavioral hallmarks of ASD are cumbersome and often lack reproducibility across labs.
Vivek Kumar and colleagues at the Jackson Laboratory have developed and validated a novel assay that quantifies social and motor behaviors ethologically over long periods using machine learning1-4. The process allows for the measurement of behaviors in groups of animals in an automated and unbiased way with little human intervention. The methods are highly scalable and reproducible over time and have good validity for different locations and testers, thus enabling reproducible preclinical studies.
Kumar’s lab plans to apply these methods to quantitate the appearance and progression of social and motor behavior differences in four mouse models, including:
(1) Fmr1 knockout (JR# 003025)
(2) B6129S-Del(7Slx1b-Sept1)4Aam/J (JR# 013128)
(3) Shank3B knockout (JR# 017688)
(4) B6.129(Cg)-Cntnap2tm1Pele/J (JR# 028635)
They will further develop new behavior classifiers that are relevant for ASD models and share the data with other research groups. This project will greatly facilitate future ASD animal studies, particularly high-throughput interventional studies, by developing a simple, automated, reproducible and ethological behavior analysis pipeline for the ASD research community.