Dear SPM community,
I am examining resting-state functional connectivity in two groups of subjects; unfortunately, the groups were acquired with two different TRs and numbers of slices at the same rate of #slices acquired/TR. My understanding is that between-group comparison would not be valid, because the scanning parameters would pose as confounds.
Does it make any sense to attempt to remove the additional slices from the subjects with longer TRs, assuming they are removed from parts that don't overlap with brain? My initial thinking was that this would effectively reduce the longer TRs to the time of the shorter TRs; however, upon further thought, this might be completely impractical in part due to the way the pulse sequence works (I apologize for using vague language). Another approach is not to perform a between-group comparison at the second level, but to treat the variables that define the groups as continuous variables perhaps...
Are there any other ways that one could reasonably account for different TRs?
Many, many thanks,
Ted
|