. I extracted my source data using -> extract--> scouts time series for 4 regions of interest using the desikan-Killiany scout and created my matrix file of 4 scouts with source data points per millisecond. I then moved my source matrix into 'files to process' clicked on ;process user matrix' and ran a PSD welch to obtain power values
Simpler solution: select the source file and run the PSD process and select the scouts you want. You don't need to extract the scouts time series explicitly first.
Reference tutorial for resting state power analysis (using dSPM):
https://neuroimage.usc.edu/brainstorm/Tutorials/RestingOmega
My sources are on the x axis and my mean Hz values are on the y. I've never seen numbers so small in my life.
sLORETA values are small, and discussed in the tutorials and some forum posts listed in the tutorial:
- https://neuroimage.usc.edu/brainstorm/Tutorials/SourceEstimation#Source_estimation_options_.5BTODO.5D
- https://neuroimage.usc.edu/brainstorm/Tutorials/SourceEstimation#Additional_documentation
The PSD process computes the power per frequency bin, output in (signal units^2/Hz): you get squared very small values (1e-10 ^ 2 = 1e-20).
Anyway, if i didnt, are these absolute power values?
Power, hence strictly positive values.
Do these values require further computation to get them into "normal" looking units if they are not already in abs units?
Start from a different statistic of the min norm maps (eg. dSPM or Z-score), or normalize the power values with respect to a baseline.
For my next step i just want to average across frequency bands and obtain my mean values for Delta, theta etc....
This is done easily with the PSD process directly (see the tutorial referenced above).
Or use the process "Group in time or frequency bands".