Similarity in functional connectome architecture predicts teenage grit (2023)
Overall, the analysis is consisted of two parts. If you need FC stability or similarity measures, you only have to run 1. Stabsim_Grit_FC.R! You can find more detailed explanations about the procedures in the manuscript.
Code for analysis are written in R (4.2.1) and Python (3.9.12).
1. StabSim_Grit_FC.R
Here, we calculate FC stability and similarity features using fMRI data from multiple conditions :)
[requirement] functional connectome matrices
(1) Bring in the data: bring the FC data and vectorize it
(2) Connectome stability: calculate within-subject FC stability (cross_movie stability as an example)
(3) Connectome similarity: calculate between-subject FC similarity (movieDM similarity as an example)
- calculate mean stability and mean similarity and bind up all the calculated FC measures. output (filename(1))
(4) Brain-Behavior: conduct multiple linear regression and get partial correlation
2. StabSim_Grit_IS-RSA.ipynb
We use FC similarity measure calculated above as an input for brain metric in IS-RSA :) Codes are adapted from Dartbrains tutorial: https://naturalistic-data.org/content/Intersubject_RSA.html
[requirement] FC similarity measure calculated in 'StabSim_Grit_FC' and behavior scores
(1) Get ready: python dependencies, get FC similarity and behavior data
(2) IS-RSA: run IS-RSA with AnnaK framework and do permutation test
(3) Network lesion: run the same process only with within-network functional connectivities according to Shen_268 atlas and then bring it back to IS-RSA
Behavioral data and neuroimaging data utilized in this code can be accessed through: http://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/.