Combining anatomical and functional networks for neuropathology identification: A case study on autism spectrum disorder

•A Graph Signal Processing framework for the diagnosis of brain diseases.•The framework is illustrated for the diagnosis of autism spectrum disorder (ASD).•A study based on the analysis of resting-state fMRI as graph signals.•Functional and structural connectivity are merged to build discriminative features.•The features are favorably used as inputs of an explainable machine learning model.