Interindividual variability in functional connectivity predicts discounting behavior during intertemporal choice

During intertemporal choice tasks (IT), in which individuals are required to choose between an immediate small 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 (TD). Previous neuroimaging studies have proposed two neural accounts for TD: the first assuming two agonistic neural networks (β and δ systems) for the evaluation of immediate and delayed rewards, respectively; the second assuming a single system for the evaluation of both rewards. Given the great inter-individual variability of TD and its relevance for several disorders such as pathological gambling, one of the currently most critical challenges is to define a specific neural marker able to predict TD independently of task-evoked activity. To address this issue we submitted twenty-five healthy volunteers (mean age: 25.8) to: (1) three fMRI scans (3T) of resting state activity; (2) an IT computer-based task, in which they choose between an immediate amount (10€) and 7 (amounts: from 15 to 60€) x 6 (delays: from 7 to 180 days) future outcomes; (3) a BIS II questionnaire measuring trait impulsivity. The main connectivity analysis was conducted using a seed-based approach in which the BOLD time course was first extracted from a set of seed regions identified in previous fMRI studies and then a correlation coefficient was computed between two seeds timecourses or between the timecourse of a seed and those of all other brain voxels. The analysis of the predictive relationship between fc-MRI and behavior was conducted using linear regressions and correlation analyses. We also conducted a leave-one-region-out analysis, which is able to identify the specific weight of each region of such networks. Results showed that TD can be reliably predicted by both the internal correlation of the single system (beta = 0.44) and by its correlation with both β (beta = 0.12) and δ (beta = 0.09) systems. Moreover, all regions of the single and δ systems appeared crucial for the fc-behavior relation, while in the β system only the mPFC appeared crucial. The results suggest an alternative view according to which the δ system, which includes regions involved in cognitive control and episodic future thinking, exerts a modulatory effect on regions of the single valuation system that are more directly involved in reward evaluation, thus influencing the selection between alternatives. Finally, we also found that TD is independent of trait impulsivity, in contrast with the widely held idea that impulsivity has a key role in TD. Overall, our findings indicate that individual variability in functional connectivity within and between critical nodes of task-evoked neural networks associated with IT is able to predict discounting behavior measured a long time afterwards, independently of impulsivity.