Yes, you should always use some form of correction, though this might be restricted to a small-volume correction for regions of interest (rather than whole-brain), as long as that volume is defined independently (or from an orthogonal contrast). This might be tricky in your case if your ROIs are defined by your existing comparison of controls vs patients - are you able to define the ROIs in the control group alone, for the purposes of comparing the two patients? Or if you are actually interested in differences between the two patients, I guess it is fine to define ROIs by differences between controls and the average of the 2 patients (but not by comparison of controls with each of only one of the patients).
BW,R
-----Original Message-----
From: Ben Becker [mailto:[log in to unmask]]
Sent: 11 October 2011 11:56
To: [log in to unmask]; Rik Henson
Subject: Re: direct comparison of two subjects possible in spm?
Comparing the two subjects on the first level (each subject = one session / contrasts across sessions) - do I have to correct the results for multiple comparisons using FWE-correction?
Thanks & Best regards,
Ben
|