nuisance regressor
Variables included in neuroimaging statistical models to account for unwanted sources of signal variance.
Full Definition
Nuisance regressors are covariates included in neuroimaging statistical models specifically to remove unwanted variance from the data that could otherwise confound results of interest. In neuroinformatics, these regressors typically account for physiological artifacts (cardiac, respiratory), motion effects, scanner drift, and other non-neural sources of signal variation. Common nuisance regressors include motion parameters, white matter and CSF signals, and physiological monitoring data. The inclusion of appropriate nuisance regressors is crucial for improving statistical sensitivity and reducing false positives in functional connectivity and activation analyses. Careful selection and preprocessing of nuisance regressors significantly impacts the validity of neuroimaging results, making this a critical consideration in analysis pipeline design.
Usage
Usage note: Clearly distinguish from regressors of interest in methods descriptions.
In Context
- "Six motion parameters were included as nuisance regressors in the first-level GLM analysis." — Statistical model specification
- "Nuisance regressors derived from white matter and CSF time series reduced spurious correlations." — Preprocessing methodology