Quantifying and Characterizing Dynamic Mechanisms of Cognitive Control

The human cognitive system is remarkably flexible and can rapidly adjust to shifting task demands and perceptual inputs. This flexibility is achieved through cognitive control, which comprises the goal-directed modulation of attention and behavior. Cognitive control is often measured by proxy, via speed and accuracy on behavioral tasks that are thought to access the brain networks of interest. However, there is much that remains unknown about how task-related neural subprocesses—visual perception, attention allocation, cognitive control, decision making—work together to produce behavior. My dissertation work uses computational modeling approaches to decompose behavior into its constituent mechanisms, with the overarching goals of understanding behaviors from the vantage point of neural mechanisms and quantifing those mechanisms in a meaningful way. I begin by introducing the flanker task as a viable means of tracking cognitive control on immediate timescales via a deep brain stimulation clinical case study. Next, I present competing theories for how and when cognitive control engages on a within-trial basis, and the mathematical details of generative models that instantiate each theory in a testable way. Using the best-fitting model across three experiments, I further investigate individual differences in the allocation of spatial attention. Finally, I package my findings into a tool for measuring and comparing mechanisms associated with cognitive control within- and between subjects, and demonstrate Bayesian-inspired methods for identifying meaningful cognitive changes through time.