They may or may not give the same result. You have two different models, and the addition of the two other conditions affects the estimates for the original 3 (also affects efficiency). You should be using the model that you believe accurately reflects your data.
Cameron Ellis wrote:
Hi Don, yes, I figured this was the case. When I run an roi analysis using both sets of data I do indeed differrent results for % signal change, which makes sense because each condition is compared to baseline and my baselines are different.but shouldn't creating a contrast of two conditions give the same relative result? Cameron
Every time point not specified as a condition would be counted as baseline, right? So does that mean then, that you have a different baseline for your 3 and 5 condition analyses?
From: "Cameron Ellis" cameo@nmr.mgh.harvard.edu To: freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] activation patterns and par files Date: Thu, 18 Oct 2007 14:50:08 -0400 (EDT)
Hi all, I have run two different group analyses in stable3. All of the analysis parameters between the two are identical except that one uses a paradigm file containing three conditions and the other uses a paradigm file containing five conditions. The first three conditions in the analysis using the five condition paradigm file are identical to those in the three condition paradigm files, meaning the timing and definitions are exactly the same. The five condition paradigm file just contains additional conditions that were not originally defined, as additions to the three condition paradigm files.
The problem is that the activation patterns I receive for contrasts using only the first three conditions of the five condition paradigm files are different from the activation patterns of contrasts using the three condition paradigm files. I do not know why this would be. Is there something about not defining all events that would lead to differences of this sort? I have visually inspected all of my paradigm files so I am sure the only diffences between them are the additional 2 conditions in the five condition files. I have attached images of the differences and below are the paths to both analyses as well as the commands run to produce the 5 condition results. The three condition commands were identical except for calling the different analysis. Any help or insight would be much appreciated! Cameron
3condition: /space/troy/3/users/retrievalnewest/Controlssh+smvsnh/bold/retrievalSPMHRF/sphsmooth-rfx/sh+smvsnh
5condition: /space/troy/3/users/retrievalnewest/Controlssh+smvsnh/bold/retrievalSPMHRF_5cond/sphsmooth-rfx/sh+smvsnh
mkanalysis-sess.new -analysis retrievalSPMHRF_5cond -TR 2.5 -paradigm paradigmmacro5 -designtype event_related -spmhrf 0 -nconditions 5 -timewindow 25 -prestim 5 -funcstem fmcsm -taumax 30 -polyfit 2 -inorm
selxavg-sess -sf sessidall -df sesspar -analysis retrievalSPMHRF_5cond
mkcontrast-sess -contrast -sh+smvsnh -analysis retrievalSPMHRF_5cond -ncond 5 -a 1 -a 2 -c 3
func2sph-sess -sf sessidall -df sesspar -analysis retrievalSPMHRF_5cond
sphsmooth-sess -smoothsteps 20 -analysis retrievalSPMHRF_5cond -insphdir sph -outsphdir sphsmooth -hemi rh -sf sessidall -df sesspar
isxavg-re-sess -analysis retrievalSPMHRF_5cond -contrast sh+smvsnh -group Controlssh+smvsnh -space sph -spacedir sphsmooth -hemi rh -sf sessidcontrolssh+smvsnh -df sesspar
<< 3cond.jpg >>
<< 5cond.jpg >>
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
More photos; more messages; more storage--get 5GB with Windows Live Hotmail. http://imagine-windowslive.com/hotmail/?locale=en-us&ocid=TXT_TAGHM_migr...
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer