Dear SPM'ers,
 
to complement Marco question, I would propose an alternative way, but I am not sure if it is fully correct :
 
1) at 1st level, compute separately every individual contrasts lets say A : [Cond1 vs Cond2] AND B : [Cond3  vs Cond4]
2) forward the resulting A and B *con*.img at the 2nd level using Basic models -- multiple regression without the constant term design
3) estimate the effect of A inclusively or exclusively masked by B
 
Practically, it works (SPM propose the masking option) , but theoretically ? Did the SPM experts agree with the procedure ?
 
Philippe Peigneux
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]On Behalf Of Marco Tettamanti
Sent: Wednesday, May 16, 2001 5:20 PM
To: [log in to unmask]
Subject: masking and RFX analysis

Dear SPMers,
there have been a couple of similar questions on this topic a while ago:
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0005&L=spm&P=R1286&D=0
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0006&L=spm&P=R7873&D=0
but I couldn't find no clear answers to them.
I also would like to perform some masking of contrasts at a second level RFX analysis. Could someone please specifiy whether the following procedure is correct:
1) At the first level, compute for every single subject the statistical map of say contrast x (Condition 1 vs Condition 2) masked with contrast y (Conditions 3 vs Condition 4).
2) Write filtered the maps obtained by step 1.
3) Perform a second level RFX analysis (e.g. one-sample t-test) on the images obtained by step 2.
 
Thank you a lot for your help!
Best wishes,
 
Marco Tettamanti
Istituto Scientifico San Raffaele
Medicina Nucleare
Via Olgettina 60
I-20131 Milano
Italia
tel. 0039 - 02 - 26 42 34 60
fax. 0039 - 02 - 21 71 75 58
[log in to unmask]">email:[log in to unmask]