Related Experiment Video
Updated: Apr 30, 2026
05:33
How to Calculate and Validate Inter-brain Synchronization in a fNIRS Hyperscanning Study
Published on: September 8, 2021
6.5K
Multiple testing in fMRI: an empirical case study on the balance between sensitivity, specificity, and stability.
Joke Durnez1, Sanne P Roels, Beatrijs Moerkerke
1Department of Data Analysis, Ghent University, H. Dunantlaan 1, 9000, Ghent, Belgium.
Biometrical Journal. Biometrische Zeitschrift
|May 9, 2014
Summary
Bonferroni thresholding offers more stable results in functional magnetic resonance imaging analysis than Benjamini-Hochberg, despite similar sensitivity and specificity. Stability is crucial for reliable brain imaging findings.
Area of Science:
- Neuroscience
- Cognitive Psychology
- Statistics
Background:
- Functional Magnetic Resonance Imaging (fMRI) visualizes brain activation using statistical tests.
- Current methods control error rates but overlook result stability.
- Comparing statistical procedures requires consistent error rate requirements.
Purpose of the Study:
- Compare Bonferroni and Benjamini-Hochberg procedures for fMRI data analysis.
- Evaluate a novel stability-inclusive procedure (Gordon et al.).
- Assess the stability of thresholding methods using real and simulated data.
Main Methods:
- Simulation study comparing Bonferroni and Benjamini-Hochberg.
- Implementation and simulation of Gordon et al.'s bootstrap-based procedure.
- Stability assessment using bootstrapping and real fMRI data (face recognition).
Main Results:
- Bonferroni and Benjamini-Hochberg offer similar sensitivity/specificity balances.
- Benjamini-Hochberg demonstrates less stability than Bonferroni.
- Gordon et al.'s method improves stability over Benjamini-Hochberg but is computationally intensive and not superior to Bonferroni.
- Bonferroni yields the most stable results in real face recognition data.
Conclusions:
- Bonferroni provides more stable thresholding for fMRI data compared to Benjamini-Hochberg.
- Stability should be a key consideration in selecting statistical procedures for neuroimaging.
- The Bonferroni procedure is recommended for its stability in fMRI analysis.
