Spontaneous brain regional dynamics contribute to generalizable brain-behaviour associations

Spontaneous brain activity is fundamental to understanding the neural basis of inter-individual differences, making its characterization central to brain-wide association studies. While inter-regional coupling patterns have been extensively studied, intra-regional dynamics remain largely unexplored. Here, analysing data from four neuroimaging cohorts (ages 8-82 years; N = 30,148), we extracted ~5,000 time-series features from resting-state haemodynamic signals across 271 brain regions, offering a comprehensive characterization of intra-regional dynamics. We identified a reliable subset that serves as an individual-specific 'barcode', capturing multifaceted dynamic dimensions that stably reflect inter-individual variation across datasets. These barcodes linked nonlinear autocorrelations in unimodal regions to substance use traits and random walk dynamics in higher-order networks to general cognitive abilities. Importantly, these brain-behaviour associations generalized across life stages and populations, with substance use showing age-specific variation and cognition exhibiting consistent patterns across age groups. This work advances large-scale, generalizable brain-wide association studies by highlighting the potential of intra-regional dynamics.