Weighted group sparse functional connectivity modeling for Alzheimer's disease identification

Constructing functional connectivity/interactions networks enables better understanding of pathological underpinnings of neurological disorders. Functional connectivity network, as a simplified representation of those structural or functional interactions, has been widely used for diagnosis and classification of neurodegenerative diseases. Motivated by recent progress in functional connectivity network analysis for Alzheimer's disease (AD). In this paper, we propose a novel adjusted cosine similarity weighted group sparse (ACS-WGS) representation method for brain functional network construction. Then, we extract functional connectivity as features from the constructed connectivity networks. Finally, we use support vector machine (SVM) for classification. The experimental results on ADNI2 dataset demonstrate outperform performance. Furthermore, compared with the conventional connectivity network-based methods, our proposed method have further improved the classification performance and discover changed functional connectivity, which could be important for disease diagnosis.