Improving Statistical Efficiency, Methodological Coverage, and Contributor Onboarding in NiMARE

This project enhances NiMARE’s performance and developer accessibility through a dual-phase optimization strategy. The first phase modernizes the contributor onboarding framework by updating development environment documentation and Git workflows to reflect current best practices, ensuring a lower barrier to entry for new developers. The core technical phase implements a fixed-width confidence interval stopping rule for ALE and MKDA workflows. By integrating an adaptive framework based on the Wilson Score Interval, the system dynamically prunes voxels as p-value convergence reaches statistical decisiveness. This approach significantly reduces computational overhead and execution time while maintaining Type I error control. Deliverables include an extensive benchmarking against standard 10,000-iteration runs, and—progress permitting—the expansion of NiMARE’s methodology through Signed Difference Mapping (SDM) and Publication Bias Diagnostics (Fail-safe N).