ekmolloy / fmri_test-retest

Documentation and MATLAB code for test-retest functional MRI studies.
4 stars 0 forks source link

How to compute ICC between two adjacency matrices? #1

Open Abrar-26 opened 9 months ago

Abrar-26 commented 9 months ago

Hi, I have two adjacency matrices ((adj.RS1 & adjRS2) computed as an output of functional connectivity (graph theory) analysis of a test-retest dataset from CONN toolbox. I would like to compute and plot the Intraclass Correlation Coefficient (ICC) to assess test-retest reliability between these two adjacency matrices. How can I perform ICC analysis and plot the output with ROI labels? Matrices are as given below in the screenshots (matrix 1=adjRS1). Thanks for your help.

Screenshot 2024-02-11 at 1 07 58 AM Screenshot 2024-02-11 at 1 08 02 AM Screenshot 2024-02-11 at 1 08 37 AM
ekmolloy commented 8 months ago

To replicate our analyses, you need to re-format your data. For our ICC function, the input is an n x k matrix where n is the number of ROIs and k is the number of scanning sessions. Each entry of the matrix M[i,j] is the rs connectivity for ROI i in scan j in relation to a seed ROI. It's not clear your current data works because it looks like it has been thresholded.

Abrar-26 commented 8 months ago

Thanks for your comment. Yes, it is thresholded as in the attached screenshot. The matrix represent averaged functional connectivity between each pair of ROIs 21x21 for 9 subjects (scans?) across one dataset, so as in the second dataset. Do you have any suggestion how to reformat data? I am quite beginner with MATLAB.

Note: "RRC matrices represent the level of functional connectivity between each pair of ROIs. Each element in an RRC matrix is defined as the Fisher-transformed bivariate correlation coefficient between a pair of ROI BOLD timeseries (CONN)"

Screenshot 2024-02-11 at 5 37 02 PM
ekmolloy commented 8 months ago

This software is over 10 years old now, and I no longer have access to a Matlab subscription, so I am unable to provide advice on reformatting data.