Wilcoxon signed-rank test with Bonferroni correction

Discussions
1

Hello,

I am working on the following topic: "Estimation of brain networks from high resolution EEG".

My field of study includes 4 healthy subjects to determine the location of significant brain regions on the cortical surface in the resting state on 5 networks: dorsal attention network (DAN), vental attention network ( VAN), the visual network ( VIS), and the auditory network ( AUD) or other.

After importing the MRI and EEG signals on brainstorm, I defined the location of the electrodes, I created the forward model, then I created the noise and data covariance matrices then I downsampled it to atlas (Dessikan/killiany).

I exported to Matlab the MN EEG (Dessikan/killiany), and I calculated the connectivity matrix, this resulted in a 68 × 68 symmetrical connectivity matrix for each epoch (i have 40 epochs in total, 10 epoch for subject), after that I calculated the 4 metrics Betweenness Centrality, Strength, Clustering Coefficient, Vulnerability.

Those four network measures were calculated for each node of the 40 estimated networks.

Then i have to apply the Wilcoxon signed-rank test with Bonferroni correction (68 × 4 multiple comparisons) on each measure of each node to determine the location of the significant brain regions and see in which region belongs (DAN, VAN, VIS, AUD or other.

I found the signrank function in matlab, but this Wilcoxon test with Bonferroni correction is applied between 2 vectors of the same size to find the p-value, and I don't know how to use it in my case.

Could you help me?

Thanks in advance :slight_smile:

3

then I downsampled it to atlas (Dessikan/killiany).

Using the process Downsample to atlas is not recommended.
Instead, select the original source maps and use the scout options directly for the connectivity processes.

I found the signrank function in matlab, but this Wilcoxon test with Bonferroni correction is applied between 2 vectors of the same size to find the p-value, and I don't know how to use it in my case.

I'm sorry, this is a generic Matlab function that we are not using in Brainstorm. I'm not sure we'll be able to help you with this. You might have better chance asking a more generic community, like Matlab Central: https://www.mathworks.com/matlabcentral/answers

@pantazis @Sylvain ?

4

Thank you for your feedback

5

Hi, Im trying to do something similar.

Following the Brainstorm tutorial, they propose the following:

In the context of a group analysis: compute the same connectivity measure for each subject, for two different experimental conditions, or between an active state and a baseline. Then run a non-parametric paired permutation test to compare the two conditions across subjects. More information in the Statistics tutorial.

However its very short in the description and I have a lot of doubts proceding... Could you add your script so we could help each other?

Thanks

6

If you want to run any of the statistical tests available in Brainstorm, then you can just follow the Statistics tutorials to learn how to use the interface, you don't have anything to script.