Dynamic Time Warping-based distance for constructing individual brain networks from FDG-PET Images: Application to Alzheimer’s Disease

This study presents a novel method for constructing individual brain networks from static FDG-PET images by integrating Kernel Density Estimation and Dynamic Time Warping algorithms. The effectiveness of the proposed method is validated using the ADNI dataset By comparing individual graph metrics across different diagnostic groups, including Cognitive Normal (CN), stable mild cognitive impairment (sMCI), progressive mild cognitive impairment (pMCI), and Alzheimer’s disease (AD), disruptions in information transmission within the FDG-PET-based brain network of AD subjects are confirmed. Our findings provide evidence for the effectiveness of KDE and DTW techniques in constructing individual brain networks from FDG-PET images. The proposed method shows promise for application in other brain disorders, and enabling personalized diagnostics.