Time-Frequency Analysis Pipeline in Brainstorm

Discussions
, ,
1

Hello everyone,

I would like to get your opinion on my time-frequency analysis pipeline to make sure that it is correct and that I am not making any mistakes.

Here are the steps I followed:

  1. I first filtered my signals in BrainVision Analyzer, then imported the data into Brainstorm.
  2. For the time-frequency analysis, I applied a Hilbert transform on my trials (at the source level Unconstrained), with the following steps:

pipeline
pipeline523×705 14.3 KB

  1. Normalization of the results (ERS/ERD)
  2. Statistical tests (permutation test Wilcoxon), because my data are not normally distributed for comparisons between conditions.

I really want to make sure that this pipeline is coherent, that the order of the steps is correct, and that there are no methodological errors in my approach.

Thank you very much for your feedback and advice!

Best regards,
Aurélien

2

The description on how the Hilbert transform is computed was not added in your post.

I sounds good.
This is for the across subjects, right?

For inter-subject, you may not want to use normalization, and perform the test on magnitude (root square of power) instead.

3

image
image428×587 13.8 KB

Thank you for your response. You will find the Hilbert transform parameters in the screenshot.
For my comparisons, I'm contrasting the activity between different conditions for each subject individually. At the end i'm performing statistics across subjects to see a statistical significant difference.

4

You may want to compute the measure Magnitude (sqrt(power)) since the statistical tests are more commonly done in magnitude rather than power.