
This research outlines a multi-stage approach to analyze brain imaging data: (1) Data processing and environment setup: An Ubuntu system is deployed on a server, and fMRI data is processed using HCP standard preprocessing tools to establish a multimodal brain imaging analysis environment. (2) High-order topological representation model construction: Building upon the existing HYBRID method, this study will optimize the model structure, focusing on improving the hyperedge weight learning process. By introducing new constraint mechanisms and optimization strategies, the aim is to enhance the reliability and interpretability of high-order brain network representations and construct a more robust framework for analyzing high-order interactions between brain regions. (3) Multi-dimensional functional validation: Using the rich cognitive behavioral data from HCP, the association between high-order hyperedges and multi-dimensional cognitive features such as executive function, working memory, and emotion processing will be systematically validated. Stability analysis and empirical validation: Based on HCP multi-timepoint scanning data, the reproducibility and stability of the model output will be evaluated and validated in an independent dataset.