> A couple of related questions concerning the modeling/control for reaction
> time differences between conditions in an event-related design. If I want to
> include RT as a covariate to control for common RT effects across 4
> conditions (rather than use condition-specific parametric modulations), but
> not turn my analysis into a parametric analysis (such as in Will Penny's
> example:
> http://www.fil.ion.ucl.ac.uk/~wpenny/datasets/face-rep/SPM2/README-SPM2.txt)
> , could I do the following:
>
>
>
> a) create a model where I only have one even-type (all events), and
> employ RTs as a parametric modulation of that regressor; then extract the RT
> regressor from the design matrix and enter it as a user-specified regressor
> in my original model?
>
>
>
> b) How about directly modeling each event as having the duration of its
> associated RT (i.e. event durations would vary between ca. 500-1000ms,
> rather than being modeled as stick functions)? Is there a strong a priori
> reason not to analyze data in this way? And finally, would these 2 analyses
> possibly produce the same results?
Tobias -
For your question b, see:
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind04&L=spm&P=R555203&I=-1
For your question a, for which I don't see the rationale unless you are simply
wondering how to covary RT effects that are common across conditions (rather
than separately for each condition), this can be done with parametric modulations
as described on this thread:
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind04&L=spm&P=R199168&I=-1
Rik
---------------------------------------------------------
DR RICHARD HENSON
MRC Cognition & Brain Sciences Unit
15 Chaucer Road
Cambridge, CB2 2EF
England
EMAIL: [log in to unmask]
URL: http://www.mrc-cbu.cam.ac.uk/~rik.henson
TEL +44 (0)1223 355 294 x522
FAX +44 (0)1223 359 062
MOB +44 (0)794 1377 345
---------------------------------------------------------
|