Finding Community Modules for Brain Networks Combined Uniform Design With Marine Predators Algorithm

Abstract

Brain networks are made up of multiple neural units that are interconnected and form neural unit modules. The precise and efficient identification of these modules is crucial for detecting diseases and providing targeted therapy. A novel algorithm, named UMPA, has been developed to analyze brain networks and obtain their neural unit modules. UMPA combines the strengths of uniform design and Marine Predators Algorithm (MPA) to achieve uniformly scattered feasible solutions across the vector domain and quickly find the optimal solution. It is the first method to integrate MPA and uniform design for identifying unit modules in brain networks. The performance of UMPA has been tested on 29 TD resting-state functional MRI brain networks, and the results demonstrate that it outperforms the other five methods in terms of modularity and is comparable in terms of conductance. Furthermore, a comparative analysis of UMPA and MPA indicates that uniform design contributes to the enhancement of UMPA. UMPA has been applied to a variety of brain networks, including those of healthy individuals and those of patients with neurological disorders. The results show that UMPA can accurately identify the neural unit modules of both healthy and diseased brain networks. UMPA has also been used to analyze brain networks in different states, such as the resting-state and task-state brain networks of healthy individuals. The results demonstrate that UMPA can accurately identify the unit modules of brain networks in different states. In conclusion, UMPA is a promising algorithm for the accurate and efficient identification of neural unit modules in brain networks, which has potential applications in disease detection and targeted therapy.

Keywords

• Marine predators algorithm
• uniform design
• brain network
• neural unit module