Dear Orhan,
I would suggest having four regressors in your design matrix: word_conflict, word_nonconflict, nonword_conflict, nonword_nonconflict. Then for each subject you create contrasts for your main effects and interactions. E.g. to look for the main effect of words - nonwords, you would enter the contrast [1 1 -1 -1], and for the main effect of conflict - nonconflict, you would enter the contrast [1 -1 1 -1]. Then for each contrast you can take the contrast image to the second level and do a one-sample t-test.
If you'd like all the contrasts generated for you, you could investigate the 'Factorial Design' function when specifying your GLM.
I would recommend reading the SPM Manual, and watching the SPM course videos on the General Linear Model (http://www.fil.ion.ucl.ac.uk/spm/course/video/#GLM) and on contrasts (http://www.fil.ion.ucl.ac.uk/spm/course/video/#Contrasts ).
Best,
Peter.
> -----Original Message-----
> From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]
> On Behalf Of orhan murat koçak
> Sent: 04 March 2014 23:34
> To: [log in to unmask]
> Subject: [SPM] first level analysis
>
> Hi everyone,
>
> we have two question about the first level analysis. We are proceeding
> an event related fmri study. The paradigm consists of two regressors;
> one is word (real word vs, nonword) and the other is conflict
> (conflict(+) vs, conflict(-)). Thus we have four conditions. The
> questions are, (1) for each subject, which statistics (one sample t-
> test, two samples t test or paired t test) proper to obtain the
> contrasts that will be used in the second level analysis, (2) how
> should contrast weights vector be entered.
>
> Apologize for these fundamental questions Regards
>
> Orhan
|