Brain parcellation based on information theory.

•A whole-brain hierarchical parcellation method is presented.•The method can be applied to both structural and functional connectivity matrices.•The mutual information is employed as a measure of brain structure, which the method maximizes.•The method hierarchically reduces the complexity of the networks preserving the global structure.•Experiments are done with synthetic and human connectivity networks and compared with the k-means method showing a high robustness of the method.