Functional connectivity null distribution
Shinozuka2023 presented a good way of creating null distributions for comparing against baseline in pharmaco-fMRI studies:
- Calculate FC b/w randomly selected windows of e.g. Human Connectome Project BOLD timeseries
- Identify pairs of significantly correlated ROIs
- Compute alignment b/w pairs of ROIs and Yeo network (if performing resting state networks analysis)
- Permute k (e.g. 1000) times, generating histogram of aggregate FC matrices effectively constituting a null distribution