Abstract
Despite advances in transcranial electrical stimulation, existing electric field focusing methods continue to struggle with achieving both precise spatial targeting and stable optimization outcomes, limiting their clinical utility. To overcome these limitations, we introduce SHE, a Scalable High-Equalization optimization framework that that integrates knowledge partitioning with model grouping to achieve unprecedented focusing precision and solution stability. Specifically, SHE decomposes the optimization space into specialized subspaces using anatomical prior knowledge, then employs a multi-objective strategy with our proposed Best Selection mechanism that preserves Pareto-optimal solutions through iterative normalization. To rigorously evaluate focusing performance, we introduce two quantitative metrics, focusing precision (F1PDI ) and spatial barycenter (A2T, Av2T), which capture both intensity distribution and spatial concentration. Extensive experiments across four stimulation modalities and five cortical depths demonstrate that that SHE consistently achieves over 10% improvement in focusing precision while reducing solution variance by more than 88% compared to existing methods. Notably, SHE maintains consistent performance when simultaneously stimulating multiple regions of interest.