Subjects with long-standing obstructive sleep apnea (OSA) show gray matter volume loss in autonomic, cognitive, and mood regulatory sites, as evaluated with high-resolution T1-weighted, voxel-based-morphometry (VBM) procedures.^1,2 VBM procedures require partitioning gray, white, and cerebrospinal fluid tissue by probability maps, which are used for voxel-by-voxel comparisons. However, this procedure has limited sensitivity from the inherent limited range of probability values, and is not suitable to detect subtle chronic/acute gray or white matter changes; thus, the procedure is unable to differentiate acute from chronic tissue pathology. Although mean diffusivity procedures can differentiate such acute vs chronic differences, those methods have poor spatial-resolution, and require specialized image processing skills. Tissue texture is a measure that can quantify local changes in intensity patterns/gray level values in an image. That measure uses high-resolution T1-weighted images, which are better suited to detect gray level values due to higher spatial-resolution and better gray and white matter contrast over other brain structural techniques. A texture feature, entropy, which measures the extent of homogeneous or randomness in tissue can be examined. Higher entropy values are associated with higher randomness, or lower homogeneity, and vice versa. The procedure has been used to assess neural changes in different conditions, including normal aging, with increasing age associated with higher entropy, and thus, the technique may be useful to examine tissue injury types in OSA subjects.³ Our aim was to examine regional entropy values, using high-resolution T1-weighted images, in newly-diagnosed, treatment-naive OSA over controls, to determine whether the measure can detect acute tissue changes in the condition. We hypothesized that regional brain entropy values would be decreased in OSA over controls, suggesting acute tissue damage in those areas.The entropy value at a given voxel from the bias-corrected high-resolution T1-weighted images can be computed using the following equation: $$$ E = -\sum_{i=1}^{N} p_{i}log(p_{i})$$$by defining a 3×3×3 volume of interest (VOI). Where, N is number of gray values in the VOI, and $$$p_{i}$$$ is probability of occurrence of i^th gray value in the VOI.We assessed 56 recently-diagnosed, treatment-naïve OSA [age, 48.5±8.6 years; body-mass-index, 31.1±6.1 kg/m²; 14 female; apnea-hypopnea-index (AHI), 35.9±22.9 events/hour], and 91 controls (age, 46.7±9.0 years; body-mass-index, 24.8±3.5 kg/m²; 35 female). All OSA subjects were diagnosed via overnight polysomnography with at least moderate severity (AHI≥15). Control subjects were healthy, without any medications that might alter brain tissue. Brain imaging data were collected using a 3.0-Tesla MRI scanner. Two high-resolution T1-weighted image series were collected using a MPRAGE pulse sequence (TR=2200 ms; TE=2.2, 2.34 ms; inversion-time=900 ms; flip-angle=9°; matrix-size=256×256, 320×320; FOV=230×230 mm; slice-thickness=0.9, 1.0 mm). Both high-resolution T1-weighted image series were realigned, averaged, bias corrected, and using bias corrected images, entropy values were calculated. We normalized whole-brain entropy maps to Montreal Neurological Institute (MNI) space (Fig. 1A), and smoothed (Gaussian kernel, 10 mm). High-resolution T1-weighted images of a control subject were also normalized to MNI space, and used as background images. The smoothed entropy maps were compared between groups using ANCOVA [covariates: age, gender; SPM12, family wise error (FWE) corrected p<0.05, cluster size 20 voxels]. Brain clusters with significant entropy differences between groups were overlaid onto background images for structural identification.No significant differences in gender (p=0.10) or age (p=0.23) emerged between groups. However, OSA subjects showed significantly higher body-mass-indices (p<0.0001) over controls. White matter regions showed lower entropy values than gray matter sites, indicating more homogenous regions (Fig. 1A). Multiple brain areas in OSA showed decreased entropy values, showing acute tissue changes in those regions, compared to control subjects (Fig. 1B, p <0.05, FWE corrected). Brain sites in OSA subjects that showed decreased entropy values included the bilateral basal ganglia (Fig. 1Ba,1Bb), insular, frontal, parietal, and occipital regions (Fig. 1Be,1Bf,1Bj), mid cingulate cortex and cingulum bundle (Fig. 1Bf), cerebellar peduncles (Fig. 1Bg), bilateral basal forebrain (Fig. 1Be), left hippocampus and amygdala (Fig. 1Bc), bilateral internal and external capsules (Fig. 1Ba,1Bb), hypothalamus (Fig. 1Be), and periaqueductal gray matter (Fig. 1Bd).Regional entropy values are significantly decreased in newly-diagnosed, treatment-naive OSA over controls, suggesting predominantly acute tissue pathology in those sites. During acute changes, tissue becomes more organized, randomness decreases, and thus, entropy values are reduced. Brain areas with reduced entropy values in OSA are localized in critical motor, autonomic, respiratory, cognitive, and neuropsychologic control sites; these functions are deficient in the condition. A range of processes, including hypoxic and ischemic processes, repeated arousals, hypercarbia, and inflammatory mechanisms can contribute to the acute tissue damage. The findings suggest that regional entropy examination can demonstrate acute tissue changes in the recently-diagnosed OSA subjects.