Christoph,
"Greater misalignment induces greater artefacts by applying the realignment parameters during reslicing."
Yes, but putting the sessions in different first-level models doesn't fix that problem, but rather merely postpones it.
"I agree with the extra statistical care for repeated measures somebody needs to take on second level."
Part of my reluctance is related to my disagreement with the way repeated measures are handled by SPM, which is a separate topic. As outlined in "ANOVAs and SPM"
link http://www.fil.ion.ucl.ac.uk/~wpenny/publications/rik_anova.pdf
there's the partitioned variance method and pooled variance method. IMHO the pooled variance method (the one commonly used by the SPM community) is incorrect (because it gets df counting wrong), though that appears to be a minority opinion. On the other hand, if I recall correctly, there was a thread on the listserv devoted to the topic of the main effect of group which implicitly showed that the pooled variance method was indeed faulty.
If one tries to use the partitioned variance method in the context of SPM, one could still do it modelling the sessions separately, but the extra labor involved in using ImCalc makes it disadvantageous.
These are all pretty nuanced points; I'm definitely not making a strong claim that putting each day in a separate subject-level model is a grave error.
Best wishes,
Stephen J. Fromm, PhD
Contractor, NIMH/MAP
(301) 451--9265
________________________________________
From: Christoph Berger [[log in to unmask]]
Sent: Monday, January 03, 2011 10:46 AM
To: [log in to unmask]; Fromm, Stephen (NIH/NIMH) [C]
Subject: Re: Design matrix for each or all subject(s)?
Dear Stephen,
I don’t wanted to say, that’s impossible to realign sessions from different days and to put them into one model in a first level design, but I was thinking that the probability of misalignment of sessions of different days will be greater than for sessions that were scanned in quick succession. Greater misalignment induces greater artefacts by applying the realignment parameters during reslicing.
I agree with the extra statistical care for repeated measures somebody needs to take on second level. But we take usually only the main effects from first level on the second level and model effects and interactions between groups conditions and treatments on the second level. The handling of contrasts is also easy in this case and there exist also a good tutorial for flexible factorial models by Jan Gläscher and Darren Gitelman.
http://www.sbirc.ed.ac.uk/cyril/downloads/Contrast_Weighting_Glascher_Gitelman_2008.pdf
kind regards,
Christoph
|