Benchmarks

The following notebooks benchmark and validate NiSpace’s statistical machinery. They serve two purposes: (1) transparency and reproducibility for users who want to understand the assumptions and performance of the methods; (2) regression tests that can be re-run after code changes.

These notebooks are under construction – more will follow in the future.

• Null map performance: false positive rate
• Group-label permutation: false positive rate
• Within-region correlation — false positive rate