Improved optimization algorithm for human brain structural connectivity with functional connectivity using dynamic mean-field model

Abstract

Recent studies have demonstrated that structural connectivity (SC) derived from the diffusion imaging method has relations with functional connectivity (FC) measured by functional magnetic resonance imaging (fMRI). Therefore, it has motivated many studies which establish computational models to describe this relationship. A critical problem is that the diffusion imaging method is hard to measure anatomical links between two brain hemispheres. Luckily, functional connectivity contains much interhemispheric information and there are many computational models that build bridges between SC and FC. Here, we use the dynamic mean-field (DMF) model and the hemodynamic model with the help of a modified iterative procedure to optimize SC. Compared with studies in the past, our study get a further improvement of SC. After optimization, not only interhemispheric links are generated but the strength of existing links are also redefined. What's more, the fitting between empirical FC and simulated FC (come from the simulation with the optimal SC) is further improved. The results of the optimization of SC with models may shed light on the human brain structure-function relation and clinical problems.