Extracting spatial-temporal characteristics from Dynamic Connectivity Network with rs-fMRI Data for AD Classification

Resting-state functional magnetic resonance imaging (rs-fMRI) based dynamic functional connectivity (dynamic FC) networks have been used to better comprehend the functioning of the brain, and have been used to early stage (i.e., mild cognitive impairment, MCI). Deep learning (e.g., convolutional neural network, CNN) approaches have recently been used to analyze dynamic FC networks, and they outperform classic machine learning methods. The sequence information of temporal properties from dynamic FC networks is largely ignored in previous investigations. To that aim, we propose a neural network based on CNN and TCN model for extracting spatial and temporal features from dynamic FC networks using rs-fMRI data for brain disease categorization in this research. The efficiency of our suggested technique in binary classification tasks is demonstrated by experimental findings on 134 ADNI individuals.