You have 4 classes with 3 variables, so there are 4*(1+3)=16 items in
your beta vector. In general, the betas are arranged
Class1offset Class2offset ... Class1Var1Slope Class2Var1Slope ...
Class1Var2Slope Class2Var2Slope ...
If you want the mean of the APOE Status No, then you would average
beta1 and beta3 (ie, contrast [0.5 0 0.5 0 ... 0]). If you want the
difference between the Yes and the No, then you would subtract the avg
of beta1 and beta3 from beta2 and beta 4 ([0.5 -0.5 0.5 -0.5 0 0 ...]).
If you want just the effect of age, then you would average betas 9-12.
If you are trying to interpret the offsets, then I would make sure that
there is not an interaction between group and covariate. If that is the
case, then you can use DOSS instead of DODS in which case the covariate
betas just reduce down to a single slope, which is easier to interpret.
On 8/20/2020 2:11 PM, Rizvi, Batool wrote:
External Email - Use Caution
Dear FreeSurfer experts,
After running the longitudinal pipeline with QDEC, I needed
beta-values of the associations between my main continuous variable
and the significant cortical cluster, as well as of all the nuisance
factors/categorical variables. Since the summary file doesn’t list
beta values, I attempted running this command to find beta coefficients:
mri_segstats --i beta.mgh --annot fsaverage lh
lh-Avg-long.thickness-spc-Parietal_WMH-Cor/mc-z.abs.th20.sig.ocn.annot
--excludeid 0 --avgwf avgwf.dat
From this, I got this output file listing beta values for 1 cluster
(entorhinal cortex) that was related to my main continuous IV
(parietal WMH). I’m struggling to understand how to still compute 1
beta-coefficient per variable from this list. I understand that this
list, based on my fsgd file, can be split into 4 sets of numbers, but
I’m not sure what do next.
avgwf.dat:
0.93879
-8.80748
0.38452
3.13252
-0.09493
-0.13107
-0.09960
-0.31087
-0.02291
0.03127
-0.02417
-0.07858
0.13087
1.80749
0.32535
0.64947
Below is how my fsgd file was structured:
GroupDescriptorFile 1
Title ParietalWMH_age_gender_APOe_EC_left
MeasurementName long.thickness-spc
Class APOE_statusNO-genderF
Class APOE_statusYES-genderF
Class APOE_statusNO-genderM
Class APOE_statusYES-genderM
Variables Parietal_WMH age_years Avg_ECthickness
Input W38284_base APOE_statusNO-genderF 0.732010 85.830000 3.125000
.. (only listing the 1^st Input line here)
My contrast file (C.dat) looks like below:
0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0
So basically, I would like to get beta coefficients for the below
variables:
* Parietal WMH
* Age
* Gender
* APOE status
* Average EC thickness (baseline values)
Thank you so much in advance for your help!
Batool
_______________________________________________
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
