Hi Doug, I just wanted to make sure that we were on the same page. I may have been a little unclear in what I would ideally like to do.
For #2 what I would like to do is take into account age slopes and total intracranial slopes for gender, and age slopes and total intracranial slopes for diagnosis. I was not implying that I wanted diagnosis and gender to interact. Assuming that you understood this previously could you please explain why you would model the matrix this way? I do not fully understand. Respectfully, -Tim >I did not get the same matrix. Here's how I would do it >Columns 1-16 model the intercepts for all your classes >Columns 17-20 model the ages for your gender-x-diagnosis subclasses >Columns 21-24 model the ICV for your gender-x-diagnosis subclasses On 07/06/2016 02:04 PM, Timothy Hendrickson wrote: > Freesurfer Support, > > Thank you for your support with my previous line of questions regarding > design matrix creation. > I have manually created another design matrix and want to ensure that I > designed it correctly. > > An example of my FSGD file is as follows below: > Class SITE 1-Male-Control > Class SITE 1-Male-PATIENT > Class SITE 1-Female-Control > Class SITE 1-Female-PATIENT > Class SITE 2-Male-Control > Class SITE 2-Male-PATIENT > Class SITE 2-Female-Control > Class SITE 2-Female-PATIENT > Class SITE 3-Male-Control > Class SITE 3-Male-PATIENT > Class SITE 3-Female-Control > Class SITE 3-Female-PATIENT > Class SITE 4-Male-Control > Class SITE 4-Male-PATIENT > Class SITE 4-Female-Control > Class SITE 4-Female-PATIENT > Variables age_at_scan total_intracranial_volume > study site levels = 1, 2, 3, 4 > gender levels = Male, Female > diagnosis levels = PATIENT, Control > age_at_scan = covariate age > total_intracranial_volume = covariate total intracranial volume > > What I would ideally like to do is: > > 1) Take into account offset differences amongst diagnosis, gender, and study > site. > > 2) Allowing a difference in age slope, and total intracranial volume slope > amongst the diagnosis and gender levels. > > 3) Modeling the age slope and total intracranial volume slope as the same for > the study site levels. > Let's image that the first participant is from site 1, Male, and control, and > is 12 and has a TIV of 30,000. > The second participant is from site 2, Female, and PATIENT, and is 14, and > has a TIV of 25,000. > My understanding of the design matrix would be as follows: > 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 0 0 0 30000 0 0 0 > 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 14 0 0 0 25000 > -Tim > > >Yes, create your matrix manually. > > > >Those matrix lines are not quite right. The ages are in the correct > >column, but you need a 1 somewhere in columns 1-8 to indicate the class > >(ie, site/dx) that the subject is in. > On 05/26/2016 12:57 PM, Timothy Hendrickson wrote: > > Hi Doug, > > > > Thank you for such a prompt response. Just to be clear you are > > recommending that I manually create the matrix file right? > > > > If so I want to ensure that I am understanding how to design the > > matrix file properly. > > > > Let's imagine that the first participant is a control and is 13 and > > the second is a patient and is 15. My understanding is that the matrix > > file would be as follows: > > 0 0 0 0 0 0 0 0 13 0 > > 0 0 0 0 0 0 0 0 0 15. > > > > -Tim > > Previous correspondences are below: > > > > You'll need a regressor for each of the 8 classes you describe below. > > You can use mri_glmfit to generate this (Xg.dat file) > > You'll need two more regressors for age, one for each diagnosis. If a > > subject (ie, row) is a control then the two values will be AGE 0. If the > > subject of the row is a patient, then the two values will be 0 AGE. You > > can then set up a Controls-Patients age (ie, interaction between dx and > > age) contrast like > > [0 0 0 0 0 0 0 0 1 -1] > > On 05/24/2016 02:30 PM, Timothy Hendrickson wrote: > > > > > > Freesurfer Support, > > > > > > I'd like to create a design matrix for a group analysis outside of the > > > DODS and DOSS models. I understand that in order to do this the -X > > > flag must be used. However, I have been unable to find examples of how > > > to do this. > > > > > > I am hoping to reveal a difference in thickness or gyrification > > > amongst a clinical population. The data set contains two factors: > > > diagnosis, and study site and one covariate: age. Diagnosis has two > > > levels: controls, and patients. Study site has four levels, one level > > > for each location the data has been collected from. > > > > > > What I would ideally like to do is: > > > > > > 1) Take into account offset differences amongst diagnosis and study site. > > > > > > 2) Allowing a difference in age slope amongst the diagnosis levels > > > > > > 3) Modeling the age slope as the same for the study site levels > > > > > > My FSGD file is designed as follows > > > > > > Class SITE 1-Control > > > Class SITE 1-PATIENT > > > Class SITE 2-Control > > > Class SITE 2-PATIENT > > > Class SITE 3-Control > > > Class SITE 3-PATIENT > > > Class SITE 4-Control > > > Class SITE 4-PATIENT > > > Variables age_at_scan > > > > > > study site levels = 1,2,3 and 4 > > > diagnosis levels = PATIENT and Control > > > age_at_scan = covariate age > > > > > > Any advice would be greatly appreciated. > > > > > > Respectfully, > > > > > > Tim > > > > > > -- > > > Timothy Hendrickson > > > Department of Psychiatry > > > University of Minnesota > > > Mobile: 507-259-3434 <tel:507-259-3434> (texts okay) > >
_______________________________________________ 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.
