Graph-structured data classification based on spectral methods and the generalized likelihood ratio test

An application to Alzheimer's disease diagnosis from PET image data

1. Introduction

Brain imaging data often exhibit structured relationships between regions. Instead of treating features independently, this project models PET data for each subject as a signal on a graph, where nodes represent anatomical brain regions and edge weights encode similarity between regions across subjects. The classification task is formulated as a hypothesis test between two graph models (healthy versus diseased), using spectral properties of graph signals.

References

  1. Chenhui Hu, Jorge Sepulcre, Keith A. Johnson, Georges E. Fakhri, Yue M. Lu, and Quanzheng Li, "Matched signal detection on graphs: Theory and application to brain imaging data classification", NeuroImage (2016)
Published on April 24, 2026, last update on April 24, 2026