Hi,
I'm trying to detrend an fMRI signal by fitting a GLM with confound
regressors only (like outliers, sessions and motion parameters) and
taking the residual images. The problem is that the residual images,
in contrast to original timeseries, are saved with double data type
and are very big (that's why I guess, by default, they are deleted by
SPM) - 1,3GB for 340 timepoints. This makes the processing quite time
consuming. Does anyone have an idea how to get around this? Thanks in
advance.
Chris
|