Development and validation of consensus clustering-based framework for brain segmentation using resting fMRI.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Data Science

Background:

• Resting-state functional magnetic resonance imaging (rs-fMRI) is widely used for brain region segmentation.
• Current clustering methods are sensitive to algorithm choice, initialization, and cluster number determination.

Purpose of the Study:

• To develop a robust framework for brain region segmentation using rs-fMRI.
• To address the sensitivity of clustering methods to algorithmic parameters and initialization.

Main Methods:

• A novel consensus clustering evidence accumulation (CC-EAC) framework was developed.
• Combined multiple clustering algorithms (K-means, hierarchical, spectral) and evaluated objective criteria for optimal cluster number selection.
• Extensive computer simulations and experimental rs-fMRI data were used.

Main Results:

• The CC-EAC framework, using K-means and hierarchical clustering with the probabilistic Rand index, accurately identified the number of clusters in simulated data.
• Reliably detected functional subdivisions in key brain regions including the supplementary motor area, insula, intraparietal sulcus, angular gyrus, and striatum in experimental rs-fMRI data.

Conclusions:

• The proposed CC-EAC framework offers a robust method for brain region segmentation using rs-fMRI.
• It accurately determines stable functional clusters and is resilient to initialization and parameter choices, outperforming conventional approaches.