Not necessarily. It could have been in the FS analysis. You can run mri_glmfit with --kurtosis (a hidden option). This will create a kurtosis.mgh and pkurtosis.mgh. Kurtosis is a measure of gaussianity of the residuals (larger being less gaussian). The pkurtosis the the sig (-log10(p)) of the probability of seeing the kurtosis under the null (ie, the residuals are gaussian). So you can see if one model gives you better kurtosis than another.
doug
On 06/03/2013 01:20 PM, Fotiadis, Panagiotis wrote:
I see. Yes from the one hand the results that I got make a lot more sense in the second type of analysis, compared to the first one, however, there are some somewhat important difference between the two analyses. If there is an issue with the data, do you believe that it would be in the acquisition? Because I checked the reconstructions and they look fine.
Thanks, Panos ________________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Douglas Greve [greve@nmr.mgh.harvard.edu] Sent: Monday, June 03, 2013 11:24 AM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Group Analysis Question
I would not expect them to be identical but similar. If they change a lot there may be an issue with your data. doug
On 6/3/13 11:11 AM, Fotiadis, Panagiotis wrote:
Hey Doug,
I have run a group analysis on my data with two different ways. (The input is cortical thickness.) In the first analysis, I specified 4 groups (Diseased_Male, Diseased_Female, Healthy_Male, and Healthy_Female) whereas in the second, just two groups (Diseased, Healthy). One of the comparisons that I did was the age slope between Diseased and Healthy. Therefore in the first analysis, I set the contrast vector to be [0 0 0 0 0.5 0.5 -0.5 -0.5] and in the second analysis I set it to [0 0 1 -1]. However, the results that came up were somewhat different. For instance there were clusters in the second analysis that were significant but were not significant in the first analysis. Shouldn't the results be the same?
Thank you, Panos _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer