When
Yuhua Yu
Postdoctoral Research Associate II
Neuroscience of Emotion & Thought (NET) Lab
University of Arizona
Decoding the Spontaneous Mind: unravel the dynamics of neural processes with latent brain states
Abstract: Cognition unfolds over time, with mental states dynamically shifting both spontaneously and in response to external contexts. Characterizing the dynamic processes requires computational approaches that move beyond static representations of brain activity. Latent brain state modeling has emerged as a powerful tool for studying neural dynamics across diverse contexts. With unsupervised machine learning methods like hidden Markov models, this modeling approach provides rich insights into the temporal dynamics of time-sustained cognition.
In this talk, I will present two studies to illustrate the general strategy and strengths of latent state modeling. The first is an fMRI study examining the neural dynamics of unconstrained thought, where participants verbalized their stream of consciousness. By associating the brain state dynamic with moment-to-moment think-aloud content, we provide new information on the neural substrates of spontaneous thought. The second study used electroencephalogram data to investigate metaphor generation, a goal-directed activity involving creative spontaneity. Despite differences in the task contexts and imaging modality, both studies underscore the critical role of transient brain dynamics in shaping cognition.