Kernel classification of connectomes based on earth mover's distance between graph spectra
This work addresses the problem of improving phenotype prediction from brain networks for medical diagnosis, though it appears incremental as it builds on existing spectral methods with a specific kernel adaptation.
The paper tackled predicting autism spectrum disorder from structural brain connectomes by using earth mover's distance between graph spectra as a kernel for SVM classification, achieving an AUC of 0.71, which outperformed simpler graph embedding methods.
In this paper, we tackle a problem of predicting phenotypes from structural connectomes. We propose that normalized Laplacian spectra can capture structural properties of brain networks, and hence graph spectral distributions are useful for a task of connectome-based classification. We introduce a kernel that is based on earth mover's distance (EMD) between spectral distributions of brain networks. We access performance of an SVM classifier with the proposed kernel for a task of classification of autism spectrum disorder versus typical development based on a publicly available dataset. Classification quality (area under the ROC-curve) obtained with the EMD-based kernel on spectral distributions is 0.71, which is higher than that based on simpler graph embedding methods.