Dear Michael,
thank you very much for the information and your advice! This is very helpful.
Best regards,
Else
________________________________________
Von: MOUTHON Michael [[log in to unmask]]
Gesendet: Donnerstag, 28. Mai 2020 16:22
An: Schneider, Else; [log in to unmask]
Betreff: RE: [SPM] 3-way interaction in SPM
Hi Else,
Donald MCLAREN wrote several posts concerning the use of Flexible factorial for a mix design (between and within factors as it is your case). The Flexible factorial is not suitable for this type of analysis because it generates a single error instead of a between and within error term (for example, see https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1504&L=SPM&P=R117152).
My advice would be to study the interaction of your within factors at the subject level (first level analysis or also call fixed-effect analysis) and to use a two-sample t-test at the group level (second level analysis or also called random effect analysis) to compare the interactions between the two groups.
Have a nice Day
Michael Mouthon
-----Original Message-----
From: SPM (Statistical Parametric Mapping) <[log in to unmask]> On Behalf Of Else
Sent: jeudi, 28 mai 2020 15:50
To: [log in to unmask]
Subject: [SPM] 3-way interaction in SPM
Dear SPM-experts,
I am analysing longitudinal fMRI data (pre- and post-test) in two groups (treatment/placebo group) using SPM. For each group, I have a task consisting of three different conditions and I am interested in whether there are any changes between the two groups in the three task conditions over time, i.e. the three-way interaction of Group X Condition X Time. I used the flexible factorial design with the factors Group, Conditions and Time to address the described three way interaction. However, I am not able to enter the according interaction term in the SPM batch. Generally, I am not sure whether this is a good approach to address the interaction between Group X Condition X Time and I am wondering whether there are other ways to test for this interaction.
Any thoughts and comments are highly appreciated!
Thanks,
Else
|