NACS Seminar: "Multiple solutions in functional neuroanatomy: A case study of fear"
Friday, October 13, 2023
1103 Bioscience Research Building
Speaker: Ajay Satpute (Northeastern University)
Title: "Multiple solutions in functional neuroanatomy: A case study of fear"
Abstract: What is the neural basis of fear? Or memory, or mentalizing? In cognitive neuroscience, there are well-established methods to address such questions using functional magnetic resonance imaging (fMRI). Experiments are designed to evoke instances of fear in comparison to other states, and univariate or multivariate analyses are performed to capture a mind-brain mapping. However, what if the basic assumptions underlying these experimental designs and analytical approaches are incorrect? A growing body of both theoretical and empirical work suggests that a surplus of neural pathways may underlie a given psychological state or behavior. This many-to-one or many-to-many relationship is referred to as “multiple solutions” or “degeneracy” in computational biology wherein, for example, different combinations of genes may underlie the same phenotype. In this talk, I investigate multiple solutions in functional neuroanatomy using fear as a case study owing to its pivotal position in emotion theory (i.e., as either a putative basic emotion or a psychological construction). First, I present findings showing that the neural and peripheral autonomic correlates of fear profoundly depend on the person and the situation. These findings challenge assumptions of uniformity as are often implied in so-called neural or autonomic signatures models of emotion. Second, I present a new analytical framework that enables researchers to test for multiple solutions in task-based fMRI data. The analytical approach models whole brain states as they vary across participants and trials during a task. In doing so, the approach provides opportunities to reason about how whole brain states relate with psychological phenomena of interest in terms of single or multiple solution(s) models.