Anomaly Detection for Structural and Functional Connectivity in Glioma Patients

Authors: Colpo M.; Pollitt R.; Leemans A.; Cecchin D.; Corbetta M.; Bertoldo A.; De Luca A.

Keywords

anomaly detection; brain tumor; functional connectivity; glioma; integration; single subject; structural connectivity; variational autoencoder.

Summary

This study uses a variational autoencoder (VAE) trained on healthy multimodal data to integrate structural (SC) and functional (FC) connectivity and detect abnormalities in glioma patients. The model reconstructs normative patterns, enabling identification of patient-specific deviations. Results show that FC captures alterations distant from the tumor, while SC is more affected locally. FC alterations align more closely with combined FC+SC changes. These findings may support personalized diagnosis and treatment planning.

For further details, here is the link for the open access publication.