Electroencephalographic activity patterns distinguish interdigitated default and frontoparietal networks

Abstract

The brain is comprised of distributed functional networks that operate on diverse time scales. Evidence from precision fMRI shows that canonical networks, including default (DN), frontoparietal control (FPCN), and dorsal attention (dATN), each fractionate into multiple finely interdigitated networks. However, as fMRI is blind to fast processes underlying neural communication, there remains a need to neurophysiologically characterize individual-level network fractionation. To address this need, we conducted a simultaneous EEG-fMRI study, including multiple sessions per individual, resting-state and task conditions, and carbon wire loop sensors for optimized artifact correction. Using precision network mapping, we discovered dissociable electrophysiological correlates of blood-oxygen-level-dependent activity for two networks within the boundaries of the canonical DN. One (DNa) related more positively with alpha and high-beta power, while another (DNb) related more negatively with theta and low-beta power. Univariate and machine learning analyses further revealed electrophysiological dissociations between networks within the canonical FPCN (FPCNa versus FPCNb) but not dATN. The network-specific electrophysiological activity patterns were reliable across sessions, only modestly variable between resting-state and demanding task conditions, and observable for neighboring nodes of different networks across multiple cortical zones. Our findings provide insight into rhythmic neural processes that fine-grained networks may engage to support divergent cognitive operations.