Obstructive sleep apnea syndrome (OSAS) is a highly prevalent multifactorial disease leading to chronically fragmented sleep and repeated intermittent hypoxic episodes¹. Several comorbidities, including hypertension and other cardiovascular diseases have been reported in OSAS^2,3. As two of the most prominent risk factors of OSAS, obesity and older age are on the rise, the prevalence of the condition is expected to continue increasing. OSAS may lead to autonomic, cognitive, and affective abnormalities⁴. Although multiple brain sites are involved in the regulation of these symptoms, the thalamus, and midbrain are key structures that serve such functions through critical relays in nuclei^5,6, but the status of this region is unclear OSAS. Thalamus and midbrain are parts of the extended prefrontal neural systems which likely integrate higher order brain functions with more developmentally fundamental brain activities such as autonomic functions. Previous studies, mainly based on structural imaging, have shown brain injury in these areas^7,8 in patients with OSAS. However, to date there are only a limited number metabolite studies using only 1D magnetic resonance spectroscopy (MRS)⁹ in thalamus. Here, we examined neurochemical changes in the thalamus and midbrain of OSAS using compressed sensing (CS) based 4D echo-planar J-resolved spectroscopic imaging (EP-JRESI)¹⁰ and prior knowledge fitting (ProFit) algorithm¹¹ for metabolite quantification.We studied 15 patients (49.12±10.1 years) with a sleep-laboratory diagnosis of OSAS and 22 age-matched healthy controls (HC) (50.27±11.6 years). All data were collected on a Siemens 3T Trio-Tim MRI scanner. 3D high resolution T₁-weighted images for localization were collected followed by CS-based 4D EP-JRESI sequence sampling 25% of the raw signal using NUS. EP-JRESI was performed using the following parameters: TR/TE=1.5s/30ms, FOV=24x24cm², voxel resolution=3.37cm³, 256 bipolar echo pair, 64∆t₁ increments, averages=2. After postprocessing⁷ the resulting bandwidth along F1 and F2 was ±250Hz and 1190Hz respectively. The acquired undersampled data were reconstructed using a modified Split Bregman algorithm¹² which solves the unconstrained optimization problem, $$\arg min_{u} \|\triangledown\ u\|_1 +\frac{\lambda}{2} \|F_{p}u - d\|_2^2$$ where u is the reconstructed data, ∇ is the gradient operator, F_p is the undersampled Fourier transform, d is the under-sampled data, λ is a regularization parameter, and ||x||_n is the l_n norm. Reconstructed data were further post-processed with a custom MATLAB-based program¹⁰ and metabolite ratios with respect to creatine (Cr; 3.0ppm) (S/S_Cr) peak were calculated using the Modified Profit algorithm. The metabolite differences between OSAS patients and healthy controls were tested with analysis of covariance (ANCOVA) with age and gender as covariates using SPSS software. A p value < 0.05 was considered statistically significant.Figure 1(a) shows voxel locations on a T1-weighted axial brain MRI of a 50-year-old OSAS patient. A representative 2D J-resolved spectrum extracted from right thalamus region of the same subject and then CS reconstructed is shown in Figure 1(b). Figure 2 and Figure 3 show the metabolite ratios with respect to Cr in the right and left thalamus and midbrain regions of OSAS patients and healthy controls. We observed significantly increased mI in midbrain and bilateral thalamus. Significantly increased Glx/Cr, Glu/Cr was found in right thalamus and midbrain, and decreased tNAA/Cr, NAA/Cr in left thalamus and midbrain respectively. Thalamus showed significantly reduced tCho/Cr bilaterally. We also found significantly decreased GPC/Cr, increased Gln/Cr, Asc/Cr in right thalamus and increased Asc/Cr in midbrainOur findings of increased mI/Cr and reduced tNAA/Cr, NAA/Cr, tCho/Cr are in broad agreement with previous 1D MRS studies in the thalamus and other regions^9,13,14. Increased mI/Cr ratio may be a reflection of increased glial activation⁵ and reactive gliosis¹³, which could result in increased inflammatory action leading to more neuronal injury presumably additional to known repeated episodes of hypoxia in OSAS patients. Decreased NAA ratios are indicative of chronic neural injury in those regions, whereas reduced tCho ratios may result from loss of myelin lipids or dysfunction of phospholipid metabolism¹⁴. Increased Glx reflect damaging excitotoxic processes arising from intermittent hypoxias¹⁵. The finding of increased Asc in the right thalamus and midbrain has not been reported earlier in OSAS, and it is possibly due to oxidative stress reported in OSAS¹⁶. The combined chemical, functional and structural findings confirm injury and reorganization of the thalamus and midbrain in OSA, which likely contribute to clinical symptoms in the condition.Our findings using the CS-based EPJRESI are consistent with the known phenomenon of oxidative stress in OSAS. Neuronal injury to thalamus and midbrain may contribute to abnormal autonomic and neuropsychological functions in OSAS. Most of these metabolites can be manipulated through pharmacological approaches, and could serve as a biomarker of any possible intervention.