Hi Vikas, I can't quite understand what the problem is. Is it that the direction of the effect is not what you expect? That would have nothing to do with the correction for multiple comparisons. If you were not able to run in QDEC, what does the subject line of the email mean? doug
On 02/22/2016 06:07 AM, Vikas Bandalli wrote:
Dear Freesurfer Team,
I am facing a discrepancy in the result when I run a Monte Carlo Z simulation. I have a group with 3 levels(normal,Non-psychotic and psychotic).I tried running an analysis to c the variation in cortical thickness between the same groups in the QDEC GUI,but since iit was impossible to do so and reading a few similar instances on the Freesurfer mailing list,I decided to run the analysis in Command line using mri_glmfit after mris_preproc.The detailed instructions I followed was from the below mentioned page https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis However,Published literature with similar cases has ran the simulation using pairing method i.e, lets say in this case between normal-psychotic,normal-non-psychotic and psychotic-non-psychotic. I have done both the methods to compare,However,I have the opposite results and discrepancy in results. (If there was a negative correlation between normal-psychotic in GUI, there was a positive correlation between normal-psychotic in the Command line analysis). Kindly note that the simulation was run for MC-z 1.3(0.05) in both GUI and Command line. Please help me with this issue and which is the TRUE result.
-- Vikas B. R. Medical student Bangalore Medical College and Research Institute (BMCRI), Bangalore,India, Ph : +918904286825
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer