Speaker: Chris Baldassano & Jamal Williams
Title: Using a Hidden Markov Model (HMM) to find temporal structure in continuous naturalistic data
Bios: Chris is currently an Assistant Professor in the Psychology Department at Columbia University. He was an undergraduate in Electrical Engineering at Princeton University, received his PhD in Computer Science at Stanford University, and was a postdoc at the Princeton Neuroscience Institute. His lab’s research focuses on how knowledge about the world – including semantic knowledge, temporal structure, spatial maps, or schematic scripts – is used to understand and remember complex naturalistic experiences. By applying machine learning techniques to data from behavioral and neuroimaging experiments, his work aims to uncover how dynamic representations in the mind and brain during perception lead to the formation of event memories.
Jamal is a PhD student at the Princeton Neuroscience Institute. He is interested in how natural stimuli, such as music, relate to cognition and memory. He uses fMRI and machine learning techniques to find relationships between brain activity, stimulus information, and behavioral measures of memory performance.
Tutorial description: In this tutorial, we will explore how to apply a Hidden Markov Model (HMM) to neuroimaging data from music, stories, or movies. We will examine a) how the HMM clusters timepoints into events with consistent response patterns; b) how the clustering changes with the number of clusters; and c) how to connect the HMM-defined events to the stimulus. This technique is most useful for datasets with long (>30 sec) stimuli with dynamics at multiple timescales, allowing us to identify how neural responses are driven by rapidly- and/or slowly-changing stimulus features.