Defining a Connectome-Based Predictive Model of Attentional Control in Aging

Abstract With advancing age, declines in the executive control of attention are accompanied by shifts in the functional topology of brain networks. However, there is increasing recognition of the considerable individual variability in the extent and types of attentional deficits that older adults exhibit, with results from neuroimaging investigations paralleling behavioral heterogeneity. Emerging computational methods leverage whole-brain functional connectivity to predict individual-level behaviors. These approaches are well-suited to the cognitive aging context, as they may elucidate configurations of functional connections that best explain group- and individual-level differences across older adults. Two independent samples of neurologically and psychiatrically healthy older adults were used to separately derive a predictive model of attentional control and test the model’s external validity. Here we show that despite challenges posed by heterogeneity in these aging samples, select functional connections carried meaningful variance, allowing for successful prediction of attention in a novel sample of older individuals.