Interindividual variability in functional connectivity predicts delay discounting during intertemporal choice

During intertemporal choice tasks, in which individuals are required to choose between an immediate smaller reward and a larger delayed reward, future outcomes are usually devaluated as a function of the delay until their receipt increases, a phenomenon known as temporal discounting. Because temporal discounting represents a highly idiosyncratic function that shows large variations across individuals, but also a relative stability over time (Kirby, 2009), it has been fruitfully exploited as model for the study of the neural representation of subjective value. Based on fMRI task-evoked activity during intertemporal choice tasks, at least two neural accounts have been proposed: one assuming a single valuation system for both immediate and delayed rewards (Kable and Glimcher, 2007) and one assuming two different neural networks (β and δ systems), the former responding to immediate rewards but less sensitive to the value of future rewards, and the latter able to evaluate future rewards and thus responsible for “patient” choices (McClure et al., 2004). Given the centrality of intertemporal choice in human decision-making, however, one of the currently most critical challenges is to define a specific neural marker able to predict discounting behavior independently of task-evoked activity. Here we tested whether resting-state functional connectivity MRI (fc-MRI) could be used as a reliable predictor of discounting behavior measured several months later. In particular, we studied 25 healthy subjects to test whether interindividual differences in functional connectivity strength within and between key nodes of the single and dual valuations systems, as described in previous studies, significantly accounted for interindividual differences in discounting rate. Moreover, we systematically explored the relationship between discounting behavior and individual trait differences in impulsivity, as measured by the BIS- 11 Italian Adult Version (Caprara et al., 1993). Our results suggest that discounting behavior can be reliably predicted by both the internal correlation of the single valuation system (within-network fc- MRI) and by its correlation with the “patient” δ system of the dual valuation system (between-network fc-MRI). Notably, since the δ system includes regions involved in cognitive control and episodic future thinking, our findings support an alternative view according to which such “control” regions exert a modulatory effect on regions of the so-called single valuation system that are more directly involved in reward evaluation, thus influencing the process of selection between alternatives. Finally, we also found that discounting behavior is completely independent from trait impulsivity, thus contrasting with the widely held idea that preference or selection of immediate reward during intertemporal choice is a direct consequence of impulsive behavior. In summary, our findings indicate that interindividual variability in functional connectivity within and between critical nodes of task-evoked neural networks associated with intertemporal choice is able to predict delay discounting behavior measured a long time afterwards, independently of impulsivity