Click Here to register in advance for this webinar.
After registering, you will receive a confirmation email containing information about joining the webinar.

Abstract: The marmoset is emerging as a model organism for the study of neural disorders that affect complex natural behaviors such as autism and schizophrenia. Therefore, we aimed to establish a framework for quantifying marmoset behavior to aid in recognizing deviations from typical behavioral patterns. First, we collected months of video of 80 marmosets living in pairs in their home environments. Then, we trained a neural network to label the animals’ body parts in each image, and identified continuous and discrete features in the data. Next, we used a hidden Markov model (HMM) to identify behavioral states that were easily recognizable by human observers. Finally, we shifted the animals’ motivational state toward either a “socially-motivated state” or a “non-socially-motivated state” using two paradigms and measured the changes in behavioral state usage and transitions between states. Overall, this study provides a template for quantifying natural marmoset behavior, a necessary first step toward their use as a model organism.