Hi Lauri, the z we compute is based on the transform from a p-value, not the r-value. This is often referred to as a "pseudo-z". The pseudo-z depends on the degrees of freedom. Now that you mention it, it would probably be better to use the r-to-z transform to compute something closer to a real z. Historically, I wrote the software to compute the pseudo-z a long time ago, the pcc was done just a few years ago.
doug
On 05/14/2018 06:26 PM, Lauri Tuominen wrote:
Dear freesurfers, I’m using mri_glmfit to estimate the correlation between a time series from a seed region and all the voxels in the brain. For one subject I get values in the pcc.nii.gz around .4 in my region of interest (not seed) but the z-scores are extremely high, i.e. > 4. Am I doing something totally wrong? My understanding was that z = arctanh(r).
Thank you so much again! Lauri Tuominen
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer