A non-parametric statistical test to compare clusters with applications in functional magnetic resonance imaging data

Area of Science:

• Neuroscience
• Statistics
• Medical Imaging

Background:

• Functional magnetic resonance imaging (fMRI) is crucial for neuroscience.
• Psychiatric disorders may be linked to neuronal clustering differences in the brain.
• Current methods for group comparisons in brain imaging often overlook cluster structure.

Purpose of the Study:

• To introduce a novel non-parametric statistical test, Analysis of Cluster Structure Variability (ANOCVA).
• To enable direct statistical assessment of differences in clustering structures between multiple groups.
• To identify features contributing to differential clustering.

Main Methods:

• Developed and applied the Analysis of Cluster Structure Variability (ANOCVA) non-parametric test.
• Validated ANOCVA through simulations.
• Applied ANOCVA to an fMRI dataset of children with ADHD and controls.

Main Results:

• ANOCVA successfully identified significant differences in brain clustering structures between ADHD patients and controls.
• The analysis highlighted specific brain regions potentially involved in ADHD pathophysiology not previously recognized.
• The method demonstrated generalizability beyond fMRI data.

Conclusions:

• ANOCVA provides a novel approach for comparing cluster structures across populations.
• The findings suggest distinct brain clustering patterns in ADHD, offering new research hypotheses.
• The method is applicable to diverse datasets requiring cluster structure comparison.