Neural Mechanisms of Cognitive Meta-Flexibility

Duke University, Durham NC

Investigators

Linked publications & trials

Abstract

Project Summary/Abstract Adaptive behavior requires the ability to keep a current task set in mind and shield it from distraction (cognitive stability), as well as to update task sets in response to changing requirements (cognitive flexibility). Cognitive stability and flexibility are complementary but opponent processing modes, as greater stability comes at the cost of lower flexibility, and vice versa. Importantly, neither a stable nor a flexible mental state is inherently beneficial; rather, it is the adaptation of one’s flexibility level to varying environmental demands – or meta-flexibility - that produces optimal cognition. Accordingly, adopting a contextually optimal set point in the stability/flexibility trade-off is considered a challenging (and poorly understood) “cognitive control dilemma”, and this ability is severely impaired in many psychiatric disorders. Therefore, understanding the neural mechanisms that mediate meta-flexibility is of great significance to both basic and clinical brain science. However, whereas individual differences and externally (e.g., drug-) induced changes in flexibility set points have been studied with some success, it is not presently known how the brain produces strategic meta-flexibility, i.e., learning to adapt one’s flexibility level to changing contexts. While a budding behavioral literature shows that people strategically adapt their readiness to switch tasks in line with changes in contextual switch-likelihood, no study to date has examined the neural mechanisms underlying these dynamic, learned changes in cognitive flexibility. Therefore, the overall goal of this proposal is to foster a new understanding of strategic meta-flexibility. We triangulate this goal via 3 strategic aims: in aim 1 we delineate the neural loci and mechanisms of meta-flexibility by combining proactive (Study 1) and reactive (Study 2) cognitive flexibility learning protocols with functional magnetic resonance imaging (fMRI) and a suite of cutting-edge analysis approaches, including brain state “pinging” and representational similarity (RSA), multivoxel pattern (MVPA), and dynamic functional connectivity (dFC) analyses. In aim 2, we then move on to determine the learning processes that guide the above mechanisms of strategic shifts in switch-readiness. Here, we use model-based fMRI, multivariate measures of neural memory reinstatement, and fMRI-guided model-based transcranial magnetic stimulation (TMS) to test how both “incremental” reinforcement learning (RL) (Studies 3 and 4) and episodic reinstatement (Study 5) may contribute to guiding strategic meta-flexibility. Having established the basic brain mechanisms of strategic meta-flexibility, in aim 3 we then examine its potential clinical significance as a transdiagnostic cognitive endophenotype, by leveraging large-scale online data collection to relate individual differences in strategic meta-flexibility to variance in clinically relevant self-report measures (Study 6). We thus lay the foundation for a potential computational psychiatry of cognitive meta-flexibility. In sum, this proposal is the first systematic research program into neurocognitive mechanisms of strategic meta-flexibility, the ability to adapt switch-readiness to suit changing contexts. This innovative project will significantly enhance our understanding of how task-switching is regulated, and help identify potential failure modes of meta-flexibility in health and disease.