Causal Connectivity based Classification of Functional MRI data

Abstract

The study of the functional connectivity of the human brain has been of significant interest in the research community. Causal connectivity refers to the understanding of the causal relationship between the brain regions. Estimation of causal interactions using fMRI data is a challenge for computational neuroimaging. In this work, we have estimated task-specific and disease-specific causal interactions between the brain regions using fMRI data. Granger causality is used to find the causal relationship between different brain regions. The quantification of causal configurations between the brain regions is achieved using transfer entropy. The obtained transfer entropy values are used as features for the classification of fMRI data. The performance of the proposed method has been validated on StarPlus and ADNI fMRI data. It achieves an average classification accuracy of 97.3% for cognitive state classification. The proposed technique achieves 99% accuracy for classification of Alzheimer’s disease and Control Normal subjects, 97% accuracy while classifying Alzheimer’s Disease and Mild Cognitive Impairment subjects, and 95% accuracy while classifying control normal and Mild cognitive impairment subjects. The proposed framework achieves an improvement of 2% and 3% for classification of task-specific and disease-specific fMRI data when compared to the existing methods.