Obstructive sleep apnea (OSA) is accompanied with tissue injury and challenge-based functional deficits in multiple brain regions serving autonomic, affective, executive, sensori-motor, and cognitive functions.^1-4 Regional brain tissue injury in OSA also alters overall spontaneous functional connectivity (FC) and the network organization in the condition.⁵ Network-level approaches suggest that human brain networks are organized into modular systems, which are characterized by efficient integration of segregated brain regions through short paths, with low wiring costs, consisting of a few densely-connected core sites. In order to obtain better understanding on abnormal information flow resulting to various functional deficits, it is important to separately characterize roles of the individual vulnerable regions in terms of intra- and inter-modular information transfer, rather than investigating topological properties simply across the whole-brain in OSA subjects. Here, our aim was to investigate the modular reorganization of functional brain networks in OSA over healthy control subjects using FC and graph-theoretical analyses procedures. We hypothesized that OSA subjects would show abnormal intra- and/or inter-modular roles at multiple brain areas.We investigated 69 recently-diagnosed, treatment-naïve OSA [age, 48.3±9.2 years; body-mass-index (BMI), 31.0±6.2 kg/m²; 52 male; apnea-hypopnea-index (AHI), 35.6±23.3 events/hour] and 82 control subjects (age, 47.6±9.1 years; BMI, 25.1±3.5 kg/m²; 58 male). All OSA subjects had a moderate-to-severe diagnosis (AHI≥15 events/hour), no history of neurological illness or psychiatric disorders other than OSA condition. Control subjects were healthy, without any evidence of sleep disorders or neurological issues, and were recruited from the campus and neighboring area. All participants gave written informed consent before data acquisition and the protocol was approved by the Institutional Review Board. Brain imaging studies were performed using a 3.0-Tesla MRI scanner (Siemens, Magnetom Tim-Trio). Resting-fMRI data were acquired with an echo planar imaging-based sequence in the axial plane [TR=2000 ms; TE=30 ms; FA=90°; FOV=230×230 mm²; matrix size=64×64; slice thickness=4.5 mm; volumes=59]. High-resolution T1-weighted images were collected using a MPRAGE pulse sequence (TR=2200 ms; TE=2.2, 2.34 ms; FA=9°; FOV=230×230 mm²; matrix size=256×256, 320×320; slice thickness=0.9, 1.0 mm). Data were canonically preprocessed with SPM8. For resting-fMRI time series averaged across all voxels in each region based on 116 distinct sites defined by automated anatomical labeling atlas, we applied canonical resting-fMRI signal processing procedures, including removing confounding factors and band-pass filtering (0.009–0.08 Hz). We calculated individual whole-brain FC as a correlation map among 116 regions, and estimated group-level brain networks using one sample t-test (P<0.05, Bonferroni correction, two-tailed). For each group-level brain network, we conducted a signed modularity optimization, which assesses a partition maximizing total sum of connection weights within arbitrary subdivision relative to total sum of chance-expected connection weights, to cluster brain regions (i.e., to detect communities), calculated participation coefficients (PC) and within-module degree z-scores, and compared the metrics between groups. Higher PC for a node indicates higher importance of the node in inter-modular communication, while higher within-module degree z-score for a node implies higher importance of the node within the module. No differences in age (p=0.7) or gender (p=0.4) appeared between OSA and control subjects. However, BMI values were significantly higher in OSA vs controls (p<0.001). OSA subjects showed decreased PC in sensori-motor regions (Figure 1A, blue) and increased PC in the inferior frontal, limbic, lateral temporal, and lingual areas (Figure 1A, red) (FDR<0.05). Within-module degree z-scores were decreased in the lateral frontal/parietal/occipital and right amygdala/insula (Figure 1B, blue), and were increased in the several cerebellar (including most vermal regions), medial orbitofrontal, left supplementary motor, right temporal, and right parahippocampal regions (Figure 1B, red) (FDR<0.05).OSA subjects showed abnormal intra- and/or inter-modular communication roles in brain regions involved in autonomic, affective, executive, sensori-motor, and cognitive control. Although autonomic regions showed enhanced participation to communicate with other regions within the same modules, sensori-motor areas showed reduced participation to communication with brain sites in other modules. The right parahippocampal and middle temporal gyrus revealed enhanced level in both inter- and intra-modular communication with other regions.The findings suggest that dysfunctions associated with OSA subjects may be related to abnormal information flow by modular reorganization in the condition, which can be examined with modular reorganization assessment procedures.