Hi all,
I'm looking for a method to correlate Reaction Times (RTs) and beta activity measured with Time-Frequency analysis (complex Morlet). Is there any way to implement this calculation on brainstorm (e.g., uploading all RTs file or how to properly export time-freuqency files externally).
I'd like to do this analysis on time-frequency maps of significance t points (calculated using the cluster-based permutation).
Many thanks to everyone.
Hello,
You won't be able to compute these correlations in Brainstorm. It's difficult to implement generic tools like this in the program because everybody needs something slightly different.
But this is something easy to script directly in Matlab, because there are only a few variables per trial, and not more than a few hundred trials.
Therefore I think you should work on finding a solution to extract your measures of interest from the Matlab files. If you have already computed non-parametric tests on your time-frequency maps and you can visualize the results you want to correlate, you should be able to export to Matlab or to text files your values of interest.
- First run the process "Extract > Apply statistics threshold" to your stat results, in order to get thresholded TF maps: https://neuroimage.usc.edu/brainstorm/Tutorials/Statistics#Convert_statistic_results_to_regular_files
- Then, you could try using the process "Extract > Extract values" to keep only one value (maybe an average over a short time window and a specific frequency band?).
- To use this value in your scripts, you can simply load the .mat file with the Matlab function load(), or use the popup menu "File > Export to Matlab" or "Export to file".
Does this help?
Dear Francois,
many thanks for your help: your suggestions are definitely useful.
I will try this procedure.
Best,
Gianvito