Prediction of Brain Functional Connectivity Based on Bayesian Model

Aiming at the complex spatiotemporal correlation of longitudinal neuroimaging data, this paper proposes a brain function connection prediction method based on spatial correlation and dynamic features. The method uses a conditional autoregressive model to capture the spatial correlation between adjacent voxels and combines it with the dynamic Bayesian model to overcome the existing brain function connection prediction method, considering only time correlation, predicting a single time period, and avoiding simultaneous consider the defects of voxel-level connections. This paper selects the data in the ADNI database to verify the prediction method. The experimental results show that the proposed method can obtain robust and accurate brain function connection prediction results in AD group and control group, thus achieving the implementation of mental illness. Early intervention helps doctors make clinical decisions and prevent serious consequences of mental illness.