help > How to set up contrasts for 3 groups and 3 observations/subject?
Showing 1-3 of 3 posts
Nov 28, 2021 05:11 PM | Remy Cohan
How to set up contrasts for 3 groups and 3 observations/subject?
Dear NBS developers,
I have extracted connectivity matrices from 30 subjects (10 x iTBS, 10 x cTBS and 10 x sham). Each subject was scanned three times; one scan on day 1, and two scans on day 2: pre/post. Multi echo data were preprocessed and denoised using ME-ICA, and then using an in-house script subject-specific stimulation sites timeseries were appended to the original timeseries as a new column (Data were labeled based on Schaefer 200 atlas).
Given that I used freesurfer's functions for individualized parcellation and timeseries extraction, each subject (at each time point) has two .txt files corresponding to each hemisphere. (Please see screenshot; some subjects have left hemisphere stimulation site, and some in the right).
I have two questions:
1- How can I get NBS to read two .txt files per participant (left and right hemisphere)?
2- Given that my .txt files are arranged alphanumerical and NBS reads them in the exact same order, what is the best way to set up my model if I am comparing connectivity changes between and within groups (timepoints)? (if a .mat file is required for this, could you please help me understand how it is done?)
Thank you in advance.
Remy
I have extracted connectivity matrices from 30 subjects (10 x iTBS, 10 x cTBS and 10 x sham). Each subject was scanned three times; one scan on day 1, and two scans on day 2: pre/post. Multi echo data were preprocessed and denoised using ME-ICA, and then using an in-house script subject-specific stimulation sites timeseries were appended to the original timeseries as a new column (Data were labeled based on Schaefer 200 atlas).
Given that I used freesurfer's functions for individualized parcellation and timeseries extraction, each subject (at each time point) has two .txt files corresponding to each hemisphere. (Please see screenshot; some subjects have left hemisphere stimulation site, and some in the right).
I have two questions:
1- How can I get NBS to read two .txt files per participant (left and right hemisphere)?
2- Given that my .txt files are arranged alphanumerical and NBS reads them in the exact same order, what is the best way to set up my model if I am comparing connectivity changes between and within groups (timepoints)? (if a .mat file is required for this, could you please help me understand how it is done?)
Thank you in advance.
Remy
Nov 28, 2021 11:11 PM | Andrew Zalesky
RE: How to set up contrasts for 3 groups and 3 observations/subject?
Hi Remy,
Each measurement should be linked to a single connectivity matrix. Therefore, the connectivity matrices for the left and right hemispheres will need to be combined before entering them into the NBS. Otherwise, you can also run the analysis independently for the l and r hemispheres.
Using individualized parcellations and multi-echo data is perfectly fine and this won't impact the NBS.
The fact that the matrices are read in alphanumeric order should not place any constrains on the model. This has no bearing on the model.
Check out the NBS Manual and previous posts on this forum for design matrix examples. You may first want to establish the hypotheses that you want to test.
The design matrix will most likely include main effects for the three conditions, main effect of pre vs post, a column for each subject to model within-subject effects and possible interactions. I am not sure if the first baseline scan would be modelled.
best,
Andrew
Originally posted by Remy Cohan:
Each measurement should be linked to a single connectivity matrix. Therefore, the connectivity matrices for the left and right hemispheres will need to be combined before entering them into the NBS. Otherwise, you can also run the analysis independently for the l and r hemispheres.
Using individualized parcellations and multi-echo data is perfectly fine and this won't impact the NBS.
The fact that the matrices are read in alphanumeric order should not place any constrains on the model. This has no bearing on the model.
Check out the NBS Manual and previous posts on this forum for design matrix examples. You may first want to establish the hypotheses that you want to test.
The design matrix will most likely include main effects for the three conditions, main effect of pre vs post, a column for each subject to model within-subject effects and possible interactions. I am not sure if the first baseline scan would be modelled.
best,
Andrew
Originally posted by Remy Cohan:
Dear NBS developers,
I have extracted connectivity matrices from 30 subjects (10 x iTBS, 10 x cTBS and 10 x sham). Each subject was scanned three times; one scan on day 1, and two scans on day 2: pre/post. Multi echo data were preprocessed and denoised using ME-ICA, and then using an in-house script subject-specific stimulation sites timeseries were appended to the original timeseries as a new column (Data were labeled based on Schaefer 200 atlas).
Given that I used freesurfer's functions for individualized parcellation and timeseries extraction, each subject (at each time point) has two .txt files corresponding to each hemisphere. (Please see screenshot; some subjects have left hemisphere stimulation site, and some in the right).
I have two questions:
1- How can I get NBS to read two .txt files per participant (left and right hemisphere)?
2- Given that my .txt files are arranged alphanumerical and NBS reads them in the exact same order, what is the best way to set up my model if I am comparing connectivity changes between and within groups (timepoints)? (if a .mat file is required for this, could you please help me understand how it is done?)
Thank you in advance.
Remy
I have extracted connectivity matrices from 30 subjects (10 x iTBS, 10 x cTBS and 10 x sham). Each subject was scanned three times; one scan on day 1, and two scans on day 2: pre/post. Multi echo data were preprocessed and denoised using ME-ICA, and then using an in-house script subject-specific stimulation sites timeseries were appended to the original timeseries as a new column (Data were labeled based on Schaefer 200 atlas).
Given that I used freesurfer's functions for individualized parcellation and timeseries extraction, each subject (at each time point) has two .txt files corresponding to each hemisphere. (Please see screenshot; some subjects have left hemisphere stimulation site, and some in the right).
I have two questions:
1- How can I get NBS to read two .txt files per participant (left and right hemisphere)?
2- Given that my .txt files are arranged alphanumerical and NBS reads them in the exact same order, what is the best way to set up my model if I am comparing connectivity changes between and within groups (timepoints)? (if a .mat file is required for this, could you please help me understand how it is done?)
Thank you in advance.
Remy
Dec 2, 2021 09:12 PM | Remy Cohan
RE: How to set up contrasts for 3 groups and 3 observations/subject?
Dear Andrew,
Thank you for your response. I was able to append all my connectivity matrices (per subject) in one .txt file. I should be able to make it work after figuring out my design matrix.
Thank you very much.
Remy
Thank you for your response. I was able to append all my connectivity matrices (per subject) in one .txt file. I should be able to make it work after figuring out my design matrix.
Thank you very much.
Remy