Neural network of superiority illusion predicts the level of dopamine in striatum
In evaluating the personality attributes and performance of the self, people are inclined to view themselves superior to others, a phenomenon known as superiority illusion (SI). This illusive outlook pervades peoples thoughts, creating hope for the future and promoting mental health. Although a specific cortico-striatal functional connectivity (FC) under dopaminergic modulation was previously shown to be implicated in SI, the underlying whole-brain mechanisms have remained unclarified. Herein, to reveal the neural network subserving individuals SI, we conducted a data-driven, machine-learning investigation to explore the resting-state FC network across the whole brain. Using the locally-acquired resting-state functional magnetic resonance imaging data (n = 123), we identified a set of 15 FCs most informative in classifying individuals with higher-versus lower-than-average levels of SI in evaluating positive trait words (area under the curve [AUC] = 0.81). Among the 15 FCs, the contribution level to the classification was 11% by the previously-highlighted cortico-striatal FC alone, but 60% by the encompassing cortico-limbico-striatal network cluster. A newly-identified, cortico-thalamic FC and another FC cluster also demonstrated substantial contribution. The classification accuracy was generalized into an independent cohort (n = 36; AUC = 0.73). Importantly, using the same set of 15 FCs, we achieved prediction on an individuals level of striatal dopamine D2 receptor availability (Pearson correlation, r = 0.46, P = 0.005). This is the first successful identification of the whole-brain neural network that simultaneously predicts the behavioral manifestation and molecular underpinning of an essential psychological process that promotes well-being and mental health. Significance StatementSuperiority illusion (SI) is a basic self-referential framework that pervades peoples thoughts and promotes well-being and mental health. An aberrant form of SI has been reported in psychiatric conditions such as depression. Our hypothesis-free, data-driven investigation revealed the spatially-distributed neural network that for the first time achieved prediction on an individuals levels of SI and the striatal dopaminergic transmission simultaneously. In principle, this multiple-biological-layer framework can be applicable to any behavioral trait to establish a link with its underlying neural network and neurochemical properties, which could quantitatively present the relation of its aberrant form with the pathophysiology of neuropsychiatric disorders. Future clinical research may aid in deriving a diagnostic biomarker for examining the related behavioral and neurochemical characteristics within individuals.