Clustering Neurons in a Microcircuit by Firing Patterns

• Activity Number: 243
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Mental Health Statistics Section
• Author(s): Jordan Rodu* and Dylan Small and Shane T. Jensen
• Companies: Carnegie Mellon University and University of Pennsylvania and University of Pennsylvania
• Keywords: Hidden Markov model ; Dirichlet Process ; Neuroscience ; Autism

Abstract

Some scientists hypothesize that repetitive behavior, a hallmark of autism, is generated by altered organization of neuron microcircuitry within the prefrontal cortex. As part of testing this hypothesis, we must first understand the structure of neural microcircuits and how they are altered in mouse models of autism, in particular how neurons in the microcircuit cluster together in their firing patterns. We use a Dirichlet Process to cluster these neurons by their hidden state trajectories. Specifically, we assume that neuron firing rate is governed by a hidden Markov model, and that neurons within the same cluster follow the same hidden state path. As a simplifying assumption, we assume that HMM parameters are shared across clusters.