false discovery rate
Also written as: FDR — False Discovery Rate
Statistical correction method controlling the expected proportion of false positive findings.
Full Definition
False Discovery Rate (FDR) is a statistical method for controlling the expected proportion of incorrectly rejected null hypotheses among all rejected hypotheses when performing multiple comparisons. In neuroimaging, FDR correction is commonly applied when testing thousands of voxels or connections simultaneously to maintain statistical validity while being less conservative than family-wise error rate corrections. The Benjamini-Hochberg procedure is the most widely used FDR method, offering a good balance between Type I error control and statistical power. FDR correction is particularly useful in exploratory neuroimaging analyses where some false discoveries may be acceptable.
Usage
Usage note: Often abbreviated as 'FDR' in methods sections; distinguish from 'false positive rate' which has different meaning.
In Context
- "Multiple comparisons were corrected using false discovery rate at q < 0.05." — Statistical methods section
- "The analysis revealed 450 significantly activated voxels after false discovery rate correction." — fMRI activation study results