Dynamic configuration of large-scale cortical networks during an inhibitory task accounts for heterogeneity in attention-deficit/hyperactivity disorder traits

Jonah Kember,Carolynn Hare,Ayda Tekok-Kilic,William Marshall,Stephen Emrich,Sidney J. Segalowitz,Erin J. Panda
DOI: https://doi.org/10.1101/2021.08.04.455077
2021-08-05
Abstract:Abstract The heterogeneity of attention-deficit/hyperactivity disorder (ADHD) traits (inattention vs. hyperactivity/impulsivity) complicates diagnosis and intervention. Identifying how the configuration of large-scale functional brain networks during cognitive processing correlate with this heterogeneity could help us understand the neural mechanisms altered across ADHD presentations. Here, we recorded high-density EEG while 62 non-clinical participants (ages 18-24; 32 male) underwent an inhibitory control task (Go/No-Go). Functional EEG networks were created using sensors as nodes and across-trial phase-lag index values as edges. Using cross-validated LASSO regression, we examined whether graph-theory metrics applied to both static networks (averaged across time-windows: -500–0ms, 0–500ms) and dynamic networks (temporally layered with 2ms intervals), were associated with hyperactive/impulsive and inattentive traits. Network configuration during response execution/inhibition was associated with hyperactive/impulsive (mean R 2 across test sets = .20, SE = .02), but not inattentive traits. Post-stimulus results at higher frequencies (Beta, 14-29Hz; Gamma, 30-90Hz) showed the strongest association with hyperactive/impulsive traits, and predominantly reflected less burst-like integration between modules in oscillatory beta networks during execution, and increased integration/small-worldness in oscillatory gamma networks during inhibition. We interpret the beta network results as reflecting weaker integration between specialized pre-frontal and motor systems during motor response preparation, and the gamma results as reflecting a compensatory mechanism used to integrate processing between less functionally specialized networks. This research demonstrates that the neural network mechanisms underlying response execution/inhibition might be associated with hyperactive/impulsive traits, and that dynamic, task-related changes in EEG functional networks may be useful in disentangling ADHD heterogeneity.
What problem does this paper attempt to address?