Abdul Halim Sapuan2025-07-082025-07-082025https://studentrepo.iium.edu.my/handle/123456789/33031This doctoral research examines how intensive cognitive training influences sexual dimorphism in brain structure, focusing on Huffaz—individuals who have memorized the entire Quran. Huffaz undergo sustained cognitive training, which may stimulate neuroplasticity and induce structural changes in the brain. This study addresses a critical gap in understanding how such culturally specific training modulates sex differences in brain morphology. Using voxel-based morphometry (VBM), voxel of interest (VOI) analyses, and machine learning classification, this study explores the interaction between cognitive training and biological sex. The primary objectives were to: (1) analyse sex differences in brain structure and assess the effects of Huffaz training using VBM, (2) investigate sexually dimorphic brain regions through VOI analysis with volume estimation and fractal analysis, and (3) develop machine learning models to classify sex based on brain structure while evaluating the impact of Huffaz training. A retrospective dataset of T1-weighted MRI scans from 47 healthy young adults (19 males, 28 females; aged 20-25) was used, including 23 Huffaz and 24 non-Huffaz. Both VBM and VOI analyses were applied across multiple brain atlases. Machine learning models—logistic regression, support vector machines, deep learning, and random forest—were trained to classify sex based on brain structure. VBM analysis revealed significant sexual dimorphism, with males exhibiting larger grey matter volumes in the cerebellum, fusiform gyrus, and temporal gyrus. After accounting for Huffaz status, additional regions such as the hypothalamus and amygdala were identified. VOI analysis confirmed larger male brain volumes, with the medial frontal gyrus reaching significance after controlling for Huffaz training. Fractal dimension analyses highlighted structural complexity variations, with males showing higher FDs in regions like the fusiform gyrus and superior occipital gyrus, while females exhibited higher values in the parahippocampal gyrus and anterior cingulate. Machine learning models demonstrated high accuracy in sex classification, with deep learning and random forest models achieving up to 92.86% accuracy when incorporating Huffaz status. These findings underscore the potential of AI in analysing complex brain structures and integrating cultural factors into sex classification models. This study enhances our understanding of how intensive cognitive training influences brain structure and sexual imorphism. It provides insights into the interaction between sex and culture in shaping brain plasticity, with implications for the development of sex-specific cognitive training programmes and therapeutic interventions.enOWNED BY STUDENTgender dimorphism;artificial intelligence;magnetic resonance imagingdoctoral thesis