Topics Covered
Covariates
Add nuisance variables to your model to isolate the signal you care about.
Residual Analysis
Extract residuals, check for patterns, and use them as "cleaned" variables downstream.
Confound Control
Strategies for dealing with age, sex, motion parameters, and other common confounds in neuroimaging.
Key Concepts
- Covariates vs. predictors of interest — what belongs in the model and why
- Partial correlations: removing the influence of a third variable
- Residualizing approach: fitting a model, extracting residuals, then analyzing them
- Common confounds in neuroscience: head motion, age, scanner drift
Homework
- Fit a model controlling for age and sex as covariates
- Compute partial correlations and compare with zero-order correlations
- Residualize a dependent variable and re-run your analysis on the residuals