Functional connectome harmonics and dynamic connectivity maps of the preadolescent brain

Resting-state brain activity is increasingly recognized as highly dynamic, with transient patterns of synchronized communities mapping onto canonical resting-state networks (RSNs). While dynamic functional connectivity has been studied in adults and clinical groups, little is known about these processes during preadolescence. We applied Functional Connectome Harmonics (FCH) and Leading Eigenvector Dynamics Analysis (LEiDA) to resting-state fMRI (rs-fMRI) data from 6,624 children aged 9–10 in the Adolescent Brain Cognitive Development (ABCD) Study. FCH captures brain activity as wave-like patterns shaped by functional connectivity, whereas LEiDA identifies recurrent dynamic states of phase synchronization. Both brain metrics and demographic variables were harmonized across 21 acquisition sites using ComBat, and correlations between brain measures and demographics were examined (N = 4,453), focusing particularly Triponderal Mass Index (TMI). The first six connectome harmonics correlated with known functional gradients (r > 0.2, p < 0.05), and LEiDA states mapped robustly onto the seven canonical RSNs (r > 0.4, p < 0.0001). Partial least squares regression showed that a small set of harmonics could predict LEiDA centroids, linking the two approaches. FCH-derived metrics (energy, power) correlated with demographic factors (r > 0.03), while LEiDA measures (fractional occupancy, dwell time, Markov Chain transition probabilities) were significantly associated with the same factors (r > 0.04, p < 0.05 after False Discovery Rate correction), in both original and harmonized datasets. Focusing on TMI, ElasticNet models predicting it revealed that rs-fMRI features performed best in the original dataset, whereas in the harmonized dataset LEiDA metrics improved prediction, and combining all imaging features yielded the highest performance. This study provides the first large-scale application of FCH and LEiDA in preadolescents, linking functional connectome harmonics with dynamic brain states and demonstrating their sensitivity to developmental and demographic factors. These findings open the way to better understand how functional brain dynamics relate to cognition and health in children.