Critical Dynamics in the Association Cortex Predict Higher Intelligence in Typically Developing Children

Abstract

Neuronal network models have indicated that the so-called critical dynamics facilitate efficient information processing, while criticality disruptions were linked to neuropathology through excitation/inhibition (E/I) imbalances. However, there is limited empirical evidence for a relationship between critical brain dynamics and cognition in healthy children and adolescents. Here, we investigate how these dynamics relate to intelligence in a developing cohort. We recorded eyes-open resting EEG in 128 children (6–19 years, 72 female) and quantified near-critical dynamics in the alpha-band using functional excitation/inhibition ratio ( f E/I) and in nonoscillatory activity using the 1/ f aperiodic exponent of the power spectrum. We devised models relating intelligence to f E/I and 1/ f exponent across seven Yeo7 functional brain networks ranked from lower-order sensorimotor to higher-order association networks. We observed significant correlations between f E/I and 1/ f exponent and IQ in association cortices, in contrast to sensorimotor cortices. Children in the high-IQ group had f E/I ratios closer to the theoretical critical value of 1 in association cortices compared with the low-IQ group. The association–sensorimotor axis rank moderated the associations between 1/ f exponent and IQ, these associations decreasing on a gradient across the hierarchy of the Yeo7 networks. Age and rank moderated the f E/I–IQ association, with the association–sensorimotor effect size gradient most visible in adolescents. Together, the results suggest that individual variation in criticality-sensitive biomarkers in association networks may be linked to IQ differences in an age-dependent manner, consistent with the hypothesis that developmental modulation of critical dynamics across the cortical hierarchy may support more efficient cognitive processing.