By "parameter estimates will be computed for each individual trial", you don't mean for each of the 216, right? You should just model the 6 conditions, possibly as a factorial design.
________________________________________
From: SPM (Statistical Parametric Mapping) [[log in to unmask]] On Behalf Of Alla Yankouskaya [[log in to unmask]]
Sent: Saturday, November 06, 2010 1:50 PM
To: [log in to unmask]
Subject: [SPM] Matrix design (individual trials)
Dear SPMers,
I am going to run fMRI experiment (event-related design, 216 trials (6 faces, three with target information and required response "target present", three nontarget with no response) random; interstimulus 3-7 s random). I am looking at functional connectivity in procession target information.
I would like to employ a standard general linear model approach for estimating type/face-specific activity, but adapting a model such that separate parameter estimates will be computed for each individual trial and then they will be used as the dependent data in a correlation analysis.
Any suggestion about how to create such matrix in SPM-8 will be greatly appreciated. Is there any rational way to do it ? (without editing each covariate manually).
Thank you in advance.
Alla Yankouskaya
[log in to unmask]<mailto:[log in to unmask]>
|