OVERLAPPING COMMUNITY DETECTION OF ADHD CHILDREN'S BRAIN NETWORK BASED ON BAYESIAN PRIOR NON-NEGATIVE MATRIX FACTORIZATION

In order to explore the differences in brain function activities between ADHD children and normal children under visual stimulation, a study is carried out on the functional overlapped communities of the two groups of children's brain function networks. Task-state fMRI data of ADHD children and normal children was obtained, and data preprocessing was performed. Adaptive sparse representation was used to construct brain function networks. Bayesian prior-based non-negative matrix factorization (NMF) method was used. By presetting different numbers of overlapping communities, the brain function networks of the two groups of children were tested for overlapping communities. The experimental results show that the brain function overlap ratio index of ADHD children is 10.7%, which is slightly lower than that of normal children, indicating that ADHD children have a lower brain function coordination efficiency in the task, and the frontal lobe-amygdala-occipital lobe network of ADHD children is connected abnormally. The overlapping community values of the two groups of children are classified as characteristics, and the classification accuracy is higher than that of the traditional method, reaching 96.6%.