External Email - Use Caution        

        External Email - Use Caution        

Hi Kersten,

Thank you for your detailed reply and sorry for this delayed response.

Yes the top row in my figure (attached again) was produced by using the '--olab' argument with the 'mri_surfcluster' command and then loading each cluster from a different label file. I attempted to follow your suggestion and adjust the label 'threshold' in Freeview to minimum value in my significance map (-5), but this did not display the missing regions - the surface remained the same as row 1 of the attached figure.

To provide you with a little more detail about my method - am using 'mri_surfcluster' with the '--olab' argument to derive all of my significant clusters of 100mm^2 or more. I am then using 'mri_label2label' to map each label files (from the 'mri_surfcluster' command) back onto the images for each of my individual subjects, and then generate subject level stats for each cluster using 'mris_anatomical_stats'. At the moment I am concerned that the subject level stats I have derived from my label files do not accurately represent each significant cluster in totality as regions appear to be missing even when i adjust the minimum threshold. As an alternative, when using 'mri_surfcluster', can I substitute the '--o', '--ocn', or '--oannot' arguments for '--olab', and still perform my 'mri_label2label' and 'mris_anatomical_stats' steps?

Kind regards,

Bronwyn Overs

Research Assistant

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
M 0411 308 769 T +61 2 9399 1725

neura.edu.au 

Twitter | Facebook | Subscribe

---

From: "Kersten Diers, DZNE" <Kersten.Diers@dzne.de>
To: "Freesurfer support list" <freesurfer@nmr.mgh.harvard.edu>
Sent: Thursday, September 5, 2019 2:13:50 AM
Subject: Re: [Freesurfer] Discrepancy between thresholded surface and cluster labels

        External Email - Use Caution        

Hi,

I suppose what we are seeing in the top row of your figure are Freesurfer labels, right? I.e. you probably used the '--olab' argument with the 'mri_surfcluster' command, and are loading each cluster from a different label file?

If my speculation is correct, then what you observe could be due to the 'threshold' setting in freeview, i.e. this little box just below the label box in the left part of the GUI.

This threshold is by default set to zero, but lowering it to some sufficiently negative value (i.e. the minimum value in the 'sig' map) should also display those regions that seem to be missing at the moment.

At least this is what I observed when I tried to reproduce your observation, i.e. I could create (and reverse by setting another threshold) similar 'cut-outs' in the cluster maps.

In addition, you could also load the ouput files that are produced by the '--o', '--ocn', or '--oannot' arguments of the 'mri_surfcluster' command. I would guess that they don't show this pattern of missing areas.

Hope this helps,

Kersten

On Di, 2019-09-03 at 14:00 +1000, Bronwyn Overs wrote:

        External Email - Use Caution        

Dear Freesurfer mailing list,

I am trying to get to the bottom of a discrepancy between two surface based images generated from the output of an LME analysis. The surfaces in row 1 of the attached image were generated by loading the four FDR significant clusters that were >100mm^2 in the cluster summary file.  The surfaces in row 2 of the attached image were generated by loading the significance map generated for the same fixed effect (saved from the LME), and thresholding at the FDR corrected p-value. The question I have is why are there such large swatches of the cortical surface in the row 2 images that are labelled as log10 p = 5, when they are not included in the most significant and largest clusters that were found in the cluster summary file?

Kind regards,

Bronwyn Overs

Research Assistant

Neuroscience Research Australia
Margarete Ainsworth Building
Barker Street Randwick Sydney NSW 2031 Australia
M 0411 308 769 T +61 2 9399 1725

neura.edu.au 

Twitter | Facebook | Subscribe

_______________________________________________
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