On 07/07/2013 01:36 PM, jh kim wrote:
Dear FreeSurfer experts...
Please be patient with a newbie's questions. For the first time, I'm trying to analyze between-group differences in cortical thickness (controls vs. patients) by using FS 5.1.0 and QDEC. I've read through the online manual and mailing lists; however, I cannot figure out the following queries.
- I selected 'group' (con, pat) as a fixed factor (gender is not
applicable because all are females), and demeaned age as a continuous covariate. After analyzing using DODS option, several questions are shown in the Display GUI panel. Does the question "Does the average thickness differ between con and pat?" mean between-group difference of cortical thickness after *controlling for the effect of age*?
Yes, it controls for all your other factors.
- Should I include average cortical thickness of the left or right
cortex (calculated from the command, mris_anatomical_stats -l lh.cortex.label -f subid.txt subid lh)as a nuisance variable? I know this issue has been addressed many times before, and it is generally recommended not to include as a nuisance factor. However, in case of statistically significant difference in mean cortical thickness of one hemisphere (unpaired two sample t-test by SPSS), is it still not necessary to include it as a nuisance factor? From the similar points of view, how about volume and surface area analyses (include or not total cortical volume and total surface area as a nuisance factor, respectively) ?
I don't have a strong opinion about removing the global mean thickness, but I think others (Mike Harms?) recommend to do so. For volume and surface area, you probably want to normalize by the ICV.
- I found several significant clusters that are corrected for
multiple comparisons using Monte Carlo simulation (corrected P < 0.01) in the bottom of Display panel. I assume that this procedures take 10,000 permutations by default. Am I right?
It is 10k iterations of generating white noise (not a permutation test unless you specified that explicitly).
- Last question, how can I extract individual value (cortical
thickness per subject) from the significant clusters? I want to do this 'all-at-once'. Please let me know the command.
There should already be a file in the output folder called something.y.ocn.dat that has a row for each input subject and a column for each cluster. doug
Apology for bothering you. Any comments would be greatly appreciated.
Thanks. Kim
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer