Metabolic and Structural Insights of Cerebellar Dysfunction in Spinocerebellar Ataxia Type 12.

Spinocerebellar ataxia 12 (SCA12) is a progressive degenerative neurological disorder, primarily characterized by impaired coordination and balance. To investigate the correlation between proton (1H) magnetic resonance spectroscopy (MRS) and structural imaging indices in patients with SCA12. T1-weighted MRI, DTI, and single voxel MRS point resolved spectroscopy (PRESS) in the left hemispheric cerebellum were acquired using a 3-T MR scanner in 40 SCA12 patients and 25 healthy controls. Correlations between metabolites, gray and white matter volume of lobules, fractional anisotropy (FA), and clinical, nonclinical, and genetic data were examined. Three machine learning algorithms (KNN, LDA, and SVM) were used to analyze the metabolic feature differences between SCA12 and HC groups. Significant decreases in choline (Cho [GPC (glycerophosphocholine) + PCh (phosphocholine)]) and N-acetyl aspartate (NAA) levels, along with increases in myo-inositol ratios to creatine, FA, and white matter volume values (p < 0.05), were observed in the cerebellum of the SCA12 group compared to healthy controls. Positive correlations were observed between NAA levels and cerebellar lobule volume, the SPM IQ score with the right crus II in the SCA12 group. The International Cooperative Ataxia Rating Scale (ICARS) score showed a negative correlation with white matter and specific cerebellar lobules. Disease duration and cytosine, adenine, and guanine (CAG) repeat length were negatively correlated with right lobule VIIIB, lobule IX, and left lobule X. Machine learning algorithms achieved an accuracy of over 95% in MRS data, and 88.89% in volumetric data. MRS, VBM, and DTI techniques reveal neuronal degeneration in SCA12 compared to healthy individuals.