Aging effects on the neural coding of proactive and reactive cognitive control

This proposal explores the neural and psychological mechanisms that underlie the well-established declines in cognitive control function experienced even by healthy older adults. A clear consensus in the cognitive neuroscience of aging is that age-related cognitive control declines reflect neurobiological changes that occur in the functioning of the mid-brain dopamine system, interacting with targets located in the lateral prefrontal cortex (lPFC) and anterior cingulate cortex (ACC). Although a large neuroimaging literature has investigated such neurobiological changes, it has been limited in the ability to relate these to the key control mechanisms postulated by neurocomputational models, which are often framed in representational terms. The current proposal adopts an innovative experimental approach to this issue, by leveraging the methodology of representational similarity analysis (RSA), to examine the neural coding of cognitive control and how it changes with advancing age. Specifically, we utilize RSA to test the Dual Mechanisms of Control (DMC) theoretical framework, which postulate two distinct modes – proactive and reactive – by which cognitive control can be deployed. A primary claim of the DMC framework is that older adults exhibit clear impairments in the engagement of proactive control, but relative preservation of reactive control. The project directly tests this hypothesis, employing novel theoretically-optimized variants of the work-horse color-word Stroop paradigm, to experimentally doubly dissociate proactive and reactive control. The Stroop task is combined with RSA to examine the neural mechanisms associated with proactive and reactive control, comparing younger and older adults through an innovative multi-modal neuroimaging approach. Specifically, we conduct convergent and matched fMRI and EEG studies, with RSA used to bridge between the two modalities. This multi-modal approach enables a systematic and comprehensive test of the DMC framework, as it applies to cognitive aging, by capitalizing on the complementary strength of each method to provide both high spatial and temporal resolution. We exploit these strengths to test whether proactive and reactive control have distinct temporal

dynamic signatures, and involve anatomically dissociable neural mechanisms within the lPFC and ACC. Finally, we exploit cutting-edge RSA methods to identify single-trial brain-behavior relationships, providing the strongest tests possible regarding the explanatory power of the DMC framework for understanding the nature of age-related cognitive control decline. As such, the findings of this project will have high public health relevance, by revealing new neural markers that can be used as targets for future diagnostic and intervention efforts.