Dear FSL list,
I am trying to replicate an analysis involving the percentage of residual errors in the signal after full GLM fitting because I want to match individuals from 2 groups on this index within some ROIs
The original analysis used the ResMS.img generated by SPM as an index of residual error (and divided it by overall signal amplitude averaged across the whole timeseries). However, because levels are split in FSL I'm not sure which residuals image would be equivalent to the ResMS.img. Would the stats/sigmasquareds.nii.gz images that I find in my first level feat dirs be the most appropriate? Would it be correct to simply average these values across multiple runs within one subject for some voxels of interest?
many thanks,
Tessa
|