Date of Completion
Physiology and Neurobiology
Behavioral Neurobiology | Cognitive Neuroscience | Computational Neuroscience | Life Sciences | Neuroscience and Neurobiology
Learning by watching others, or observational learning, is important for social development and survival. However, not much is known about the brain mechanisms underlying this type of learning. Since the 1960s, observational learning has been widely studied in humans, but developing and analyzing experiments for animals has been challenging. Here, I explore observational learning using a novel paradigm while performing an analysis that involves tracking the rats using an active learning paradigm called DeepLabCut. In this novel paradigm, customized operant conditioning chambers are used for the rats to observe and learn from another animal repeatedly on multiple trials each day. The task is automated for the rats which creates less animal handling and bias. Using light cues, rats are able to associate the location to which they are asked to go to. This paradigm gives an observer rat multiple new observation trials, allowing for more power in the experiment. This analysis allows researchers to quantify behavior in neuroscience and is an efficient method for tracking estimation. Using videos gathered from the animals in the paradigm, detailed markers can be used to assist with the computer tracking and provide quantitative behaviors such as how well the rats are observing. Incorporating DeepLabCut into an observational learning model is a novel method that provides much more power in such tasks. Developing this animal model is a powerful method to assess observational learning and will lead to a better understanding of the neuronal circuits involved in social learning. Here, we found that rats learned the location of a food reward by observing a conspecific through the modality of vision, rather than other cues such as smell. Performance was affected by the angle at which observers were facing the demonstrator, how often they were facing the demonstrator, and the distance to the demonstrator rat. Performance did not seem to be affected by the social inducing drug oxytocin nor its antagonist, Atosiban. The current paradigm allows for repeated trials of observational learning and allows for a method of behavioral quantification for neurobiological studies of social learning.
Shao, Thomas, "Using Machine Learning to Conduct a Detailed Behavioral Analysis in an Appetitive Social Learning Task" (2020). Honors Scholar Theses. 694.