EAGER: Neural Behavioral Analysis (NBA) Pipeline for Behavior and Neural Activity Analysis in Autism

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

Naturalistic behaviors are external reflections of brain's internal integration of bottom up processes that mediate inputs sent to the brain, and top down processes that mediate appropriate responses determined by the brain. Measuring behavioral data without concurrent neuronal activity monitoring would only provide an incomplete picture of brain function. One bottleneck in the field is that these behavioral data and neuronal activity data are typically collected separately under different experimental paradigms and subsequently analyzed with different analytical pipelines. It is, therefore, impractical to infer mechanistic correlation between behavior and neural activity using these existing pipelines. A system that enables simultaneous collection of behavior and neuronal activity data followed by integrated decoding of these two types of data would be a breakthrough that offers unique opportunities to explore behavior and its governing neuron activity pattern. This project will take advantage of a clinically relevant mouse model of autism, to develop a novel machine learning based pipeline for simultaneous decoding of behavioral and neuronal activity data. By providing an novel and integrated data analytic toolkit which enables analysis of high-dimensional large data sets of behaviors and neural activities in an autistic mouse model, this project will delineate the temporal and spatial pattern of neural activities underlying social deficits and sensory abnormities which might provide a novel mechanistic link between those two keys symptoms of autism. Besides providing a powerful and versatile toolkit, this project will help fill in the critical knowledge gap between brain circuit functional changes and social behavioral deficits in autism, with the potential of strong impact on other psychiatry disorders, such as schizophrenia and bipolar disorders. By developing a novel machine learning based pipeline to enable simultaneous analysis of animal behaviors and neuronal activities, this project will focus on addressing the following three research challenges: (1) concurrently collecting large data sets of mouse social behaviors and neural activity in different brain areas using intravital calcium imaging, which will be used to establish and optimize the machine learning-based neural behavioral analysis pipeline, (2) validating the neural behavioral analysis pipeline in an autistic mouse model, (3) using the neural behavioral analysis pipeline to interrogate mouse behavior and calcium imaging data from different brain areas and infer causal relation between neural activity pattern and autism-like behavior traits. Using unbiased machine learning algorithms to extract videotaped behavioral and neural imaging data in a high-throughput manner, this project will be able to make sense of neural circuit data in the context of complex behavior deficits. Successful execution of the proposal will establish a general "computational behavior-neural function" framework capable of identifying the hierarchy of social deficits and sensory abnormalities in autistic mice, which will provide a powerful tool to the field to untangle complex animal behavior and neuronal activity pattern for mechanistic exploration. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.