772. Altered Control Energy of Brain Functional Integrated and Segregated States in Attention-Deficit/Hyperactivity Disorder

Abstract

Abstract Background Attention-deficit/hyperactivity disorder (ADHD) has been linked to disrupted balance between brain functional integration and segregation. However, most resting-state fMRI studies assume stable brain activity and do not quantify the energetic cost required to sustain or transition between large-scale network states. Brain control theory offers a principled framework to estimate the minimum control energy needed to maintain or shift brain states, potentially providing a mechanistic marker of impaired neural regulation in ADHD. Aims & Objectives To characterize abnormalities in brain control energy underlying functional integration and segregation states in children with ADHD. Method Resting-state fMRI data were obtained from the ABCD cohort (N = 3,828; ADHD = 211, controls = 3,617). Integration and segregation states were identified and dynamic state metrics (fraction time, dwell time, and state transition frequency) were computed. Using an optimal linear control model based on participants' static functional connectivity, we estimated control energy for: (i) maintaining the integration state, (ii) maintaining the segregation state, and (iii) transitions between integration and segregation states. Linear mixed-effects models were used to compare ADHD and controls at whole-brain and network levels. Results No significant group differences were observed in dynamic metrics of integration/segregation. In controls, fraction time and dwell time were significantly higher in the integration state than in the segregation state, whereas this within-group difference was not evident in ADHD. Compared with controls, children with ADHD showed reduced whole-brain control energy required to maintain the integration state (t = −2.28, P = 0.022) and to transition from segregated to integrated states (t=-2.31, P = 0.021). At the network level, reduced control energy for maintaining integration was observed in the ventral attention network (t = −2.70, PFDR = 0.021), cingulo-opercular network (t=-3.30, PFDR = 0.012), sensorimotor network (t = −2.33, PFDR = 0.048), visual network (t = −2.94, PFDR = 0.019), and auditory network (t = −2.82, PFDR = 0.019). Controls showed greater control energy for maintaining integration than segregation (t = 4.39, P = 1.18×10-5), while no such difference was found in ADHD. Discussion & Conclusions Although standard dynamic indices (fraction time, dwell time, and transition frequency) did not differentiate ADHD from controls, control-theoretic modeling revealed a more mechanistic alteration: children with ADHD required less energy to maintain an integrated state and to drive segregation-to-integration transitions. This dissociation suggests that conventional state-occupancy metrics may overlook abnormalities in the energetic constraints that govern how brain states are stabilized and reconfigured. The network-specific pattern-spanning ventral attention, cingulo-opercular control systems and sensory networks supports a systems-level account in which ADHD involves weakened capacity to sustain globally integrated configurations that are typically supported by attention and control circuitry. From a mechanistic perspective, reduced integration-related energy may reflect altered controllability of large-scale functional interactions, potentially contributing to inefficient coordination of distributed processing demands. Overall, these findings highlight brain control energy as a complementary marker to dynamic FC measures for quantifying dysregulated state regulation in ADHD and motivate future longitudinal and task-based studies to test developmental trajectories, symptom relevance, and the stability of these energetic signatures across contexts.