Hi,
I was trying to normalize amplitude envelope correlation (AEC) and PAC plots relative to baseline.
In case of AEC, it is possible to do subtraction, but z transformation is not working.
PAC: it is possible to do both subtraction and z transformation but it is opening only for topoplots, not at electrode level.
Would it be possible to make them work?
Thanks
The Z-score normalization you are trying to use here computes the average and standard deviation over a baseline time period within the selected file, and normalizes the rest of the file with these values.
AEC and PAC are both results that do not have a time dimension. It is therefore not possible to compute a mean and standard deviation to normalize the data with.
If you are not familiar with this type of normalization, please refer to the introduction tutorials. If there are tutorials you haven't followed yet, they might contain information that would help you understand the files and methods in Brainstorm.
https://neuroimage.usc.edu/brainstorm/Tutorials/SourceEstimation#Z-score
https://neuroimage.usc.edu/brainstorm/Tutorials/TimeFrequency#Normalized_time-frequency_maps
https://neuroimage.usc.edu/brainstorm/Tutorials/TutPac#Further_analysis_of_PAC_files
Thanks Francois, got your point.
What about PSD normalization in a similar way, again there is no time dimension but z transformation is functioning for PSD.
Also, what about PAC subtraction?
again there is no time dimension but z transformation is functioning for PSD.
If it works, it is probably dividing the file by itself and the output makes no sense. You need to make sure you understand the methods you use before applying transformations randomly to your files. Brainstorm is a research program, that gives more freedom to the user than the average clinical or acquisition software, the counterpart is that you need to be more aware of what you are doing.
You could Z-score your data across recording sessions or trials (if you have a lot) or across subjects (for group analysis), but you'd need to script it yourself as Brainstorm cannot do that for you, and first of all, you need to have a purpose for it. We propose normalization procedures for the PSD in the resting state tutorial.
https://neuroimage.usc.edu/brainstorm/Tutorials/RestingOmega#Power_maps
For the analysis of the spectrum, you can also refer to this tutorial:
https://neuroimage.usc.edu/brainstorm/Tutorials/Fooof
Also, what about PAC subtraction?
@Soheila @Sylvain?
PS: I added a warning for baselines shorter than 3 samples (not an error because it might make sense in some contexts). Keep on reporting your doubts on your experiments, this is how the interface becomes safer and safer for everybody 
I am just interested in getting relative values.
e.g There are 3 condition a, b, c. I would like to get b/a vs c/a, because a is control.
That is why I am selecting all file in a timing box.
You the process2 tab with FilesA=b and FilesB=a, for many trials or subjects and run non-parameteric tests (or simple differences to start with)
OK, thanks.
