In this paper we present a novel approach to group-wise multi-modal community detection, i.e. identification of coherent sub-graphs across multiple subjects with strong correlation across modalities. This approach is based on joint diagonalization of two or more graph Laplacians aiming at finding a common eigenspace across individuals, over which spectral clustering in fewer dimension is then applied. The method allows to identify common sub-networks across different graphs. We applied our method on 40 multi-modal structural and functional healthy subjects, finding well known sub-networks described in literature. Our experiments revealed that detected multi-modal brain sub-networks improve the consistency of group-wise unimodal community detection.
Joint laplacian diagonalization for multi-modal brain community detection
Sona, Diego
2014-01-01
Abstract
In this paper we present a novel approach to group-wise multi-modal community detection, i.e. identification of coherent sub-graphs across multiple subjects with strong correlation across modalities. This approach is based on joint diagonalization of two or more graph Laplacians aiming at finding a common eigenspace across individuals, over which spectral clustering in fewer dimension is then applied. The method allows to identify common sub-networks across different graphs. We applied our method on 40 multi-modal structural and functional healthy subjects, finding well known sub-networks described in literature. Our experiments revealed that detected multi-modal brain sub-networks improve the consistency of group-wise unimodal community detection.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
