ANOCVA

ANOCVA performs non-parametric statistical comparisons of clustering structures in R to assess differences in clustering and per-ROI contributions across populations for biological datasets such as fMRI-derived brain region clusters.

Key Features:

• Non-Parametric Statistical Testing: ANOCVA implements non-parametric methods to compare clustering structures without distributional assumptions.
• Comparative Analysis Across Populations: It tests whether a set of regions of interest (ROIs) are equally clustered between two or more populations.
• ROI Contribution Significance Testing: The tool evaluates the statistical significance of each ROI's contribution to observed differences in clustering structures.
• R-Based Implementation: The method is implemented in R.
• Applicability to Biological Entities: ANOCVA operates on clustering representations of brain regions or other biological entities.

Scientific Applications:

• Neuroimaging and fMRI analysis: ANOCVA has been applied to functional magnetic resonance imaging (fMRI) data to examine brain activity clusters and the organization of neural networks.
• Autism (ABIDE) study: Applied to the Autism Brain Imaging Data Exchange (ABIDE) Consortium dataset (896 individuals: 529 controls, 285 with autism spectrum disorder), ANOCVA detected significant differences in clustering structures between controls and ASD subjects (p < 0.001), indicating atypical organization within domain-specific brain modules.
• Biomarker and connectivity comparison: The method can identify ROIs that contribute to clustering differences and support comparisons of neural connectivity patterns across populations for biomarker discovery.

Methodology:

Researchers input data representing the clustering of ROIs across different populations; ANOCVA, implemented in R, conducts non-parametric tests to compare clustering structures and determines statistical significance; it identifies which ROIs significantly contribute to observed clustering differences using a significance threshold (e.g., p < 0.05).