Highly accurate local functional fingerprints and their stability

The neural underpinnings of individual identity reflected in cognition, behavior, and disease remain elusive. Functional connectivity profiles have been used as a "fingerprint" with which an individual can be identified in a dataset. These established functional connectivity fingerprints generally show high accuracy but are still sensitive to mental states. A truly unique, and especially state-independent, neural fingerprint will shed light on fundamental intra-individual brain organization. Moreover, a fingerprint that also captures inter-individual differences in brain-behavior associations will provide the necessary ingredients for the development of biomarkers for precision medicine. With resting-state and task fMRI-data of the Human Connectome Project and the enhanced Nathan Kline Institute sample, we show that the local functional fingerprint, and especially regional homogeneity (ReHo), is 1) a highly accurate neural fingerprint, 2) more stable within an individual regardless of their mental state (compared to the baseline functional connectome fingerprint), and 3) captures specific inter-individual differences. Our findings are replicable across parcellations as well as resilient to confounding effects. Further analyses showed that the attention networks and the Default Mode Network contributed most to individual "uniqueness". Moreover, with the OpenNeuro.ds000115 sample, we show that ReHo is also stable in individuals with schizophrenia and that its stability relates to intelligence subtest scores. Altogether, our findings show the potential of the application of local functional fingerprints in precision medicine.