Permanent Note Created: 01-20-2022 13:39
-
When setting up a [[Linear Regression|regression model]], if the outcome variable is a brain image and the predictor variables are not brain images, then a rapid version of regression can be used. The design matrix may inverted once and then repeatedly multiplied by the data at each [[voxel]].
-
I wonder if bootstrapping repeatedly can take advantage of the premultiplication routine used to speed up regression with brain imaging.