A Computational Cognitive Neuroscience Framework for Attentional Control Traits and States
Abstract
Approved for Public ReleaseSummary/Abstract:The ability to flexibly control attention is a fundamental component of human higher mental function. It allows us to focus and sustain information processing in a goal-directed manner, prioritizing information that is relevant to our goals while suppressing distraction from irrelevant information. A central, but poorly understood, aspect of attention control (AC) is its highly variable nature. Indeed, AC appears to fluctuate in a state-like manner, characterized by the occurrence of lapses and breakdowns, due to a range of factors, such as mind-wandering. There is also clear individual variation, marked by pronounced differences among people in AC utilization. This source of variation appears to reflect trait-like characteristics in both cognitive capacity and skill, but also may include affective and motivational components. The presence of both state and trait-related AC variation has critical implications for performance in real-world situations, particularly in the demanding and high-pressure task environments encountered by military personnel. Some individuals are more successful than others in such situations, but evenhigh performers are subject to instances of failure or break-down during critical periods (#choking under pressure#). Consequently,the development of effective training strategies and personalized interventions to enhance AC capacity is of high DoD relevance. Therefore, a key challenge for the field is to develop an integrative account of the causes, sources, and primary mechanisms that giverise to AC variation. In this project, we describe a research program to develop and test a comprehensive account of AC variation. At the core of this account is a multi-level neurocomputational architecture that spans from the neural coding level to cognitive task models to observable behavior. We will test neurocomputational model predictions with multi-modal neuroimaging data, combining fMRI and EEG. These data will be analyzed with cutting-edge multivariate pattern classification approaches that bridge between the modalities, and characterize activity in terms of neural coding patterns that can be closely linked to model representations. Conversely, the model is linked to observable behavioral performance data with Hierarchical Bayesian Cognitive Models that provide more accurate and precise measurements. These models appropriately partition AC variation obtained from choice and reaction time data into trait-related, condition-related, and trial-by-trial fluctuation components. We leverage these new modeling approaches to estimate AC function in unified task environments that are more flexible, portable, and engaging, and which better simulate the real-world, complex and high-pressure situations faced by DoD personnel. Finally, we integrate the modeling work and unified task environments to develop neurocomputationally-informed training strategies that enhance AC function, benchmarking these against current gold-standard measures. The modeling and experimental work inform each other in a reciprocally interactive and iterative fashion through the projectperiod: the model simulations will drive new experimental investigations and training approaches, and the results of training studies will provide new targets for modeling and experimentation. If our model-informed training strategies prove to be effective, they will strongly validate the utility of neurocomputational approaches for revealing the core mechanisms of AC variation. More generally, the proposed project represents exactly the kind of use-inspired basic and applied research that is critically needed to scale-upneurocomputational and cognitive models of AC, so that they can be of direct use to the defense sector.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 11, 2023
- Source ID
- N000142312792
Entities
People
- Todd S. Braver
Organizations
- Office of Naval Research
- United States Navy
- Washington University in St. Louis