A generalizable neural signature of perceived loneliness

Social interactions are fundamental to human well-being. Much of what people invest time, energy, and money in revolves around connecting with others. Yet many people still feel lonely, increasing their risk for poor sleep quality, depression, and mortality. Identifying underlying mechanisms of loneliness is key to developing efficient interventions. Here, we used neuroimaging to investigate how multivariate patterns of functional connectivity between and within brain systems relate to loneliness. We developed a neural signature of loneliness from distributed patterns of resting-state functional connectivity in the Human Connectome Project (HCP) dataset. We then demonstrate the generalizability of this neural signature in predicting individual differences in loneliness in an independent sample of young adult students. Our results reveal that functional connectivity between sensory systems and higher-order process networks (frontoparietal control and default networks) is a key feature of how loneliness is instantiated in the brain. Finally, this neural signature appears to be specific for loneliness, as it was not predictive of other closely related measures of social connection or negative affect. Overall, this work offers a neural marker that could be used across new datasets, and suggests key brain systems to target with future interventions.