AFNI/SUMA for Mass Univariate Correlation of Cortical Thickness with psychological scores

AFNI Message Board
1

Hi All,
I have to admit, I believe I solved this very problem 5 to 8 years ago, but I think I forced the solution. Anyhow, what I want to do is correlate (ala 3dttest+++) a psychological score with cortical thickness across 50 subjects and obtain a group .gii or .nii or whatever group file that I can then threshold and find ROIs from the cortical thickness analysis. I have done this with FreeSurfer, but as I remember their parcellation can issues with significant clusters that happen to cross over parcel boundaries.
My specific question is: What files from SUMA make spec process (which made the folder and contents in the attached .png) would I use for a basic cortical thickness mass univariate correlation analysis? What would the subject files be and are there different levels of smoothing, etc in the files made by SUMA Make Spec process? Does anyone have any suggestions? The files made by that command are rather opaque, am I not finding a basic primer on SUMA?

Thanks,
Tim

filemadeSUMAMakespec.jpg
filemadeSUMAMakespec.jpg1898×1069 605 KB

2

Hello,

For thickness analysis you want the std.141.?h.thickness files.

AFNI board post about this.

Original blog post about this.

Happy to answer more questions as this is my primary workflow for Freesurfer analyses.

3

Adding on to Peter’s response, you can convert from the niml.dset format to other text formats with ConvertDset. Also see our calculation of thickness and comparisons with FreeSurfer here.

https://afni.nimh.nih.gov/pub/dist/HBM2018/OHBM_2018_Thickness.pdf

4

Deja vu. Peter thank you so much for finding things I should have been able to easily find myself. I’m only slightly embarrassed.

Thank you both so very much!!