Sorry, I can't remember whether the qdec correction for mult comp
included an option to specify that there were two hemispheres. If not,
you can take the cluster p-value that you get from qdec and multiply it
by 2 (bonferonni correction for 2 hemis), this is mentioned in Greve and
Fischl, 2017, NI. False positive rates in surface-based anatomical analysis
On 6/22/2020 11:07 AM, xunzhang_hmrrc wrote:
External Email - Use Caution
Thanks for your timely response.
To assess the two group differences in cortical thickness, whole-brain
vertex-wise analyses were performed using the graphical user interface
of FreeSurfer known as QDEC. We used a general linear model to compare
the thickness between the patients and healthy controls with age and
sex as covariates. The Monte Carlo Null-Z Simulation was conducted to
control for multiple comparisons (cluster-wise corrected P< 0.05).
But, as we all know, QDEC analyses have always be done for the two
hemispheres separately, so the multiple comparisons are also conducted
for two hemispheres separately, which, the reviewer assumed, is
different from multiple comparisons for the whole brain? So, could you
please provide some perspective and citation about the validation for
such design of QDEC (i.e., the right and left hemisphere were tested
separately)?
------------------ Original ------------------
*From:* "Douglas N. Greve"<dgr...@mgh.harvard.edu>;
*Date:* Mon, Jun 22, 2020 10:17 AM
*To:* "freesurfer"<freesurfer@nmr.mgh.harvard.edu>;
*Subject:* Re: [Freesurfer] QDEC analyses
How you did you do your analysis? As an exploratory/map-based
analysis? Or as an ROI analysis? How did you correct for multiple
comparisons? Generally, in a map-based analysis, the correction for
multiple comparisons incorporates the fact that there are two
hemispheres being tested.
On 6/20/2020 10:04 PM, xunzhang_hmrrc wrote:
External Email - Use Caution
Hello Freesurfer developers,
Recently, I have used QDEC of freesurfer 6.0.0 to examine the
cortical thickness differences between two groups, and the manuscript
has been submitted to a journal. One reviewer raised a question -"The
manuscript stated that the right and left hemisphere were tested
separately. Could the authors provide an explanation for this? I
assume that testing over 68 (all ROIs) vs testing separately left and
right (34 ROIs) would have an effect?" Unfornately, I don't know how
to responde to this reviewer, so I would like to ask for your help. I
just followed the tutorials of QDEC
(http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/QdecGroupAnalysisV6.0),
and I found QDEC analyses have always be done for the two hemispheres
separately. So, could you please provide some perspective and
citation about the validation for such design of QDEC?
Thank you in advance for your assistance,
Best regards,
------------------------------------------------------------------------
Xun
_______________________________________________
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
