Neuroimaging Sample Size Calculator

Purpose

Traditional power analysis (e.g., using G*Power for a t-test) fails for neuroimaging because it cannot account for the massive multiple comparisons problem, spatial correlation structure, or the multi-level nature of neuroimaging inference. Neuroimaging requires simulation-based approaches that generate synthetic datasets, apply the full analysis pipeline including multiple comparison correction, and estimate power as the proportion of simulations detecting the effect.

A competent programmer without neuroimaging training would use standard power formulas and dramatically overestimate the power of a whole-brain analysis. They would not know that cluster-extent thresholds, random field theory corrections, and spatial smoothness all affect the effective number of tests, nor that pilot-data-based simulation is the gold standard for neuroimaging power analysis. This skill encodes the domain-specific methodology for simulation-based sample size planning.

When to Use This Skill

• Planning sample size for a new fMRI, EEG, or MEG study
• Conducting power analysis for a grant application or registered report
• Estimating required N when pilot data or published effect size maps are available
• Choosing between whole-brain and ROI-based analysis based on power constraints
• Evaluating the statistical adequacy of a proposed or completed study

Research Planning Protocol

Before executing the domain-specific steps below, you MUST: