Introduction: One of the methods for investigating brain activity is called functional magnetic resonance imaging (fMRI), and research has shown that it has great potential for use in clinical applications. However, some of the inconsistent findings reported by several research place some limitations on fMRI. The absence of accepted and standardized techniques for evaluating fMRI data is one of the potential causes of the problem. To solve this issue, a standardized parcellation model is desirable.
Aims: In this paper, we evaluated the performance of a novel parcellation framework called the Node Vitality Model (NVM) for fMRI image region of interest definition using the anatomical, functional, and network features of the brain.
Materials and Methods: The model was evaluated using both real data made up of 50 images of the human brain and simulated data created using standard graph methods. Measures of segregation using clustering, resilience using global efficiency, and integration using assortativity were the metrics used to assess the vitality of the brain nodes.
Results: According to the findings, assortativity varied between 0.0022 and 0.1394, clustering varied between 0.5267 and 0.9083, and global efficiency varied between 0.5172 and 0.9167. Only 80 of the 132 nodes taken into consideration in the majority rule's final analysis were found to be significant, and this information was used to construct a brain network. The resulting graph was then used to re-parcellate the brain network using a reverse Engineering approach.
Conclusion: This study showed that the node vitality model has good promise for parcellating fMRI data considering anatomical, functional and network features of the brain.