You are here

Brain Decoding via Clustering Ensemble and Graph Representation


The general objective is the development of new brain decoding methods that use the spatial relation information in fRMI signals: (1) the development of a new feature selection method for brain decoding based on clustering ensemble algorithms that preserve the spatial relation among voxels and yield a consensus parcelation of the brain image; (2) the design of a graph representation of the brain image that captures all the information in a given parcelation; (3) the design of graph kernels able to measure the semantics inherit in the graph structure and with a low computational cost in its evaluation process.



EU Marie Curie
Research topics: