Dr. Ivan Grahek - Neural and Computational Mechanisms of Flexible Goal Engagement

As we accomplish hundreds of goals each day, we constantly adjust how we engage with each goal, and seamlessly transition between different goals. Achieving some goals requires high levels of focus and cognitive effort, while reaching other goals is less taxing. Cognitive control processes guide goal engagement and goal switches by adjusting information processing strategies in accordance with our current goal. In this talk I will outline the neurocomputational mechanisms that determine levels of goal engagement, as well as the ability to flexibly transition between goals. I will first present a series of experiments which demonstrate that processing of goal-relevant information depends on learned components of motivation. This work shows that people dynamically adjust how they engage with the current task (e.g., how much attention they deploy) based on the current expected value of investing cognitive control. I will then present a model which specifies the dynamics through which people adjust levels of cognitive control, and the ensuing costs. Empirical evidence for this model shows substantial inertia in control adjustments, which produces performance costs in situations that demand frequent changes in of control levels. Together, this work demonstrates that cognitive control is a set of dynamical processes, rather than a fixed ability. Cognitive control dynamics are governed by motivation and learning, and control adjustments come at a cost caused by inertia in adjusting control states. I will argue that this dynamical perspective on cognitive control helps us better understand the computational and neural basis of flexible goal engagement, as well as disorders of goal-directed cognition and behavior.