Individual differences in (dis)honesty are represented in the brain’s functional connectivity: Robust out-of-sample prediction of cheating behavior

Abstract

Abstract Many of the economically most costly forms of unethical behavior such as tax evasion, stock manipulations or movie and music piracy relate to the moral domain of (dis)honesty, in which unethical behavior is not targeted at a clearly identifiable victim. While large individual differences in (dis)honesty are evident, the neurocognitive determinants of this heterogeneity remain elusive. We combined connectome-based predictive modelling (CPM) on resting state functional connectivity patterns with a novel experimental task, which measures spontaneous and voluntary cheating inconspicuously, to investigate how these task-independent neural patterns shape our (dis)honest choices. Our analyses revealed that functional connectivity in a network of regions, including the dorsolateral prefrontal cortex and the inferior frontal gyrus, commonly linked to cognitive control processes, but also the medial prefrontal cortex and temporal pole, associated with self-referential thinking, and the caudate nucleus, linked to reward processing, are of central importance in promoting honesty. In a leave-one-out cross-validation analysis, we show that this neural model can reliably and accurately predict how much an unseen participant will cheat on our task. Participants who cheated the most, also scored highest on several impulsivity measures, which highlights the ecological validity of our task. Notably, when comparing neural and self-report measures, the neural measures were found to be significantly more important in predicting cheating. Our findings suggest that a person’s dis(honest) decisions depend on how well the self-referential thinking network is functionally connected to the control and reward networks.