This thesis presents a comprehensive computational framework for the numerical modeling of electromagnetic (EM) field interactions with the human brain, addressing the urgent need for high-precision dosimetry in the context of expanding radiofrequency exposure. The research adopts a multi-scale hierarchical approach, systematically transitioning from idealized models to patient-specific anatomical representations. Initial cranial and cerebral geometries were derived from high-resolution MRI datasets and processed through a rigorous 3D Slicer segmentation pipeline. These data were subsequently imported into a Finite Element Method (FEM) environment (COMSOL Multiphysics), where a progression of geometric complexity, ranging from single-domain ellipsoids to intricate two-domain phantoms, was evaluated to isolate the physical phenomena governing field penetration, namely resonance, shielding, and dielectric absorption.