NiMARE Coordinate-based Meta-Regression

NiMARE Coordinate-based Meta-Regression implements a multi-group coordinate-based meta-regression framework that employs a generative spline-based spatial model regularised by a roughness penalty for analyzing functional MRI datasets.

Key Features:

• Flexible Meta-Regression: Enables meta-regression and inference for multiple groups of coordinate-based meta-analytic functional MRI datasets.
• Spline-Based Spatial Model: Employs a generative spline-based spatial model with flexible control over smoothness via roughness penalty regularisation.
• Parametric Inference: Provides valid parametric inference when all groups have at least 200 foci.
• Parametric Bootstrap Inference: Supports inference via parametric bootstrap for sparser datasets with fewer foci.

Scientific Applications:

• Multi-Group Neuroimaging Synthesis: Synthesizes findings from multiple neuroimaging studies to identify consistent patterns of brain activation across different groups.
• fMRI Focus Distribution Comparison: Compares foci distributions between groups of studies without requiring balanced sample sizes.
• Coordinate-Based Meta-Analysis: Conducts coordinate-based meta-analysis on functional MRI data involving multiple groups and covariates.

Methodology:

The framework employs a generative spline-based spatial model regularised by a roughness penalty. Extensive evaluations are conducted with simulations, and the method is demonstrated on real data. Parametric inference is used when all groups have at least 200 foci; sparser datasets require inference via parametric bootstrap.