FMRI Study Statistical Methods Review

I need an independent statistical-methods review of a prospective human fMRI study before participant recruitment and funding. The scientific protocol, primary estimand, decision rules, and simulation code already exist. This is not a request for general data analysis, manuscript statistics, or a redesign of the neuroscience. The review should focus on: * validating a repeated-measures within-subject primary estimand and variance/standardization approach; * assessing possible residual timing/proximity bias in a counterbalanced design; * reviewing prespecified GO / KILL / INCONCLUSIVE decision rules at a meaningful effect boundary; * evaluating false-negative / false-KILL operating characteristics; * assessing whether a bounded-bias decision rule is statistically defensible when the residual bias magnitude is not empirically identifiable; * reviewing fixed-N enrollment versus one possible prespecified interim analysis; * reviewing effect-blinded technical QC rules to ensure they do not introduce optional stopping or outcome-driven changes; * independently inspecting supplied Python Monte Carlo simulation code and key operating-characteristic results. I already have a compact review packet, a focused list of statistical questions, and reproducible simulation code. Primary deliverable: A concise written methodological review suitable for incorporation into the prospective Statistical Analysis Plan, including: 1. any statistical defects or unresolved assumptions; 2. recommendations for the final decision framework; 3. recommendation on fixed-N versus the proposed interim design; 4. recommendation on handling residual bias and false-KILL risk; 5. confirmation or criticism of the effect-blinded QC framework. No participant data have been collected.

Preferred qualifications

* PhD in biostatistics, statistics, quantitative psychology, psychometrics, or a closely related field; * strong experience with prospective experimental design; * repeated-measures / within-subject methods; * Monte Carlo simulation and operating-characteristic analysis; * bias / sensitivity analysis; * preferably sequential or group-sequential design experience; * ability to review Python statistical simulation code; * fMRI / neuroimaging experience is helpful but not mandatory if the statistical-methods background is strong. Please include 1–3 concrete examples of prior work involving at least two of the following: * simulation-based operating characteristics; * sequential/group-sequential designs; * sensitivity analysis for unidentified or partially identified bias; * advanced repeated-measures experimental design. This is a bounded review. I am not seeking ongoing statistical support or a mandatory video call. A brief follow-up discussion can be arranged only if needed to clarify the written recommendations.