This study investigates age- and gender-related variations in the volume of the Substantia Nigra (SN) using 7T structural MRI data ultra-high-resolution from the publicly available ATAG dataset. The dataset comprises T1-weighted MP2RAGE brain scans from 53 healthy individuals, stratified into young (n = 30), middle-aged (n = 14), and elderly (n = 9) groups. After standardized preprocessing, including skull stripping, spatial normalization, and application of the ATAG atlas, a 3D U-Net model enhanced with DenseNet-style encoding blocks was employed for SN segmentation. Model performance was assessed using Dice Similarity Coefficient, IoU, Precision, and Recall, with stratified 3-fold cross-validation ensuring robustness across age groups. Post-processing refined binary segmentation masks for volumetric analysis. Descriptive statistics showed a mean SN volume of 589.56 ± 29.74 mm3. While Pearson and Spearman correlations revealed no significant association between continuous age and SN volume, one-way ANOVA indicated a significant group effect (p = 0.0058), with elderly individuals showing significantly reduced SN volumes compared to younger cohorts. Regression analysis confirmed the significance of age group and gender as predictors, explaining 37.5% of the variance in SN volume (R2 = 0.375). These findings suggest that SN atrophy becomes pronounced in late adulthood, following a non-linear age trajectory, and underscore the utility of categorical age modeling for detecting neurodegenerative patterns.