On 11/03/2018 02:20 AM, Vikas Bandalli wrote: > > External Email - Use Caution > > Dear freesurfer team. > > I am running an analysis betweene two groups Normal and Patients. > Here is an example of my particular design > diagnosis :- Normal vs Bipolar > Co-variate :- Age > Nuisance factor :- Total_Intra cranial volume > > I would please like to know a couple of things related to QDEC > 1.I am looking for difference in volume between patients and controls > and the analysis. For correcting for multiple comparisons I ran Monte > carlo - Z simulation at threshold of 1.3 > My understanding is the analysis overall runs a two-tailed t-test > first and then when we select to run MC-z simulation it iterates for > multiple comparisions (by repeating the saem say 1000 times ) and > gives out 'clusterwise p-value' > Is my understanding correct? It synthesizes a z-field, then smoothes it based on the smoothness of the residual of the real data, then thresholds, finds the max cluster size. It then repeats 1000 times. It then computes the p-value of your cluster based on the number of times it saw a cluster of the same size or bigger out of the 1000 trials. Note: for a threshold of 1.3, we recommend that you do permutation instead since the cluster p-value is very liberal using the MCZ simulation. > > 2.I would like to extract T-values for the same . Are we getting > cluster wise T-value similar to cluster wise p-value ? > in the terminal The cluster pvalue is not based on a t-test. You can get t-value from the original voxel-wise analysis. Look in the output folder under each contrast. You will find F.mgh and gamma.mgh. t = sign(gamma)*sqrt(F) (assuming it is a t-contrast and not an F) > > 3. Where would I find t-values for the analysis ??? See above > > 4.Are there any methods to extract T-values in qdec or freesurfer . If > so,please specify. > > I am enclosing screenshots from my analysis for the same . Kindly > request you to please have a look please help me out with this ?? > > Thanking You, > > Yours sincerely , > Dr.Vikas Bandalli > > -- > 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
