How to Extract PSD Results Per Scout for Group Comparison in Brainstorm

Discussions
,
1

Hi, I need some help regarding group comparison for PSD data at the source level (using scouts). In my case, I am working with resting-state EEG (approximately 200 seconds of recording duration, 19 channels). I performed source reconstruction using minimum norm imaging with sLORETA (constrained). On this source reconstruction, I computed the Power Spectral Density (PSD) for the entire signal using a 5-second window with 50% overlap. For regions of interest, I used the Desikan-Killiany scouts with the scout function set to "mean". Then, I standardized the PSD by calculating the relative power and extracted the Magnitude.

First of all, is this procedure correct, or is there anything I should do differently?

Next, I performed an equal t-test (as shown in the image) to compare the PSD between two independent groups at the source level. However, when I export the results to the MATLAB workspace, I get a matrix of size 15002x1x4 (the 4 corresponds to the 4 frequency bands). What I want is a result of size 68x1x4 (one per scout).

How can I achieve this?

Thank you!

image
image363×701 21 KB

2

It sounds good. A relevant details that is missing from you description is whether the Scout function was applied before or after the PSD.
https://neuroimage.usc.edu/brainstorm/Tutorials/TimeFrequency#Scouts

You are right the expected result should be of the size [68 x 1 x 4]. You may want to double-check that:

  • The PSD files have the expected size [nScouts x 1 x nBands]
  • You are using those scout-wise PSD files, rather than the source-wise PSD files