Hi Doug,
thanks for your answer. My commands are:
mri_mcsim --o $FREESURFER_HOME/average/mult-comp-cor/fsaverage/lh/labelname \
--base mc-z \
--surface fsaverage lh \
--nreps 10000 \
--label /path/to/label/lh.labelname.label
mri_glmfit --C /path/to/contrast/contrast.mtx \
--fsgd /path/to/fsgd/fsgd_file.fsgd dods \
--glmdir glm_dir \
--y /path/to/data/concat_lh_fsaverage_smooth15.mgz \
--surf fsaverage lh white
mri_glmfit-sim --cache 2.000000 neg \
--cache-label labelname \
--cwpvalthresh 0.050000 \
--glmdir /path/to/glm_dir/glm_dir \
--overwrite \
--2spaces
(this is for a cross-sectional case. In a longitudinal, I would use the glm_dir
from a pseudo analysis and replace sig.mgh with the one from the matlab lme
analysis as you have explained to me in my last thread)
---
Because I run those commands embedded into a NiPype workflow, I had to modify
mri_glmfit-sim a little:
1) because (I don't know why) the data file in the glmfit log file was not the
correct one, mri_glmfit-sim was failing.
--> So I changed the script so that it takes an --y argument (as mri_glmfit) to
explicitly define the data file
2) because, when I only handed the glm_dir from the preceeding mri_glmfit step,
different calls to mri_glmfit-sim (e.g., different --cache-label s) were all
written into the same directory and overwriting each other.
--> So I changed the script to copy the directory specified with --glmdir to
the current folder and carry out all further operations on it
Here a link to my modified mri_glmfit-sim version
(https://dl.dropboxusercontent.com/u/255214/mri_glmfit-sim_nipype), I hope that
these changes should not be responsible for the clustering I am experiencing?!
Thanks a lot,
Janosch
> Can you send your mri_glmfit, mri_glmfit-sim, and mri_mcsim command lines?
>
> On 03/20/2015 02:24 PM, Janosch Linkersdörfer wrote:
>> Hi all,
>>
>> I have a question/problem related to restricting monte carlo cluster
>> correction to a reduced search space.
>>
>> I have:
>>
>> 1) created a label including the labels of regions I am interested in
>> 2) used mri_mcsim with the --label flag and the label created in 1) to
>> create
>> cached monte carlo simulations for the search space
>> 3) used mri_glmfit-sim with the --cache-label flag and the sim created in 2)
>>
>> The result is:
>>
>> - I definitively get more significant clusters than when I use the cortex
>> label (so there must have been a reduction in search space/no. of vertices),
>> but
>> - some of these clusters lie only partly inside the regions included in the
>> label
>> - some lie even completely outside of the label
>>
>> I thought that I should get only clusters within the borders of the label so
>> that vertices not in the label would not be considered.
>>
>> Thanks for your help,
>>
>> Janosch
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
