Extracting EEG epoch

Discussions
,
1

Hello François,
I downloaded a free EEG dataset and imported it into the Brainstorm auditory stimulation protocol as raw data. After that, I extracted the epochs. I am working on correlated-source localization, and my dataset includes several event types such as oddball, response, oddball and response, etc.

My first question is: Which of these event categories should I extract and use as the input to my source-localization method in order to recover the correlated auditory sources? The dataset is related to the P300 paradigm.

My second question concerns the oddball condition. For example, the oddball category contains 28 epochs/files. To prepare the input for my algorithm, I assume that I should average these epochs into a single data matrix. However, I am not sure whether these 28 oddball epochs are phase-aligned in Brainstorm.
If they are not phase-aligned, what is the recommended way to align them before averaging?

For averaging, I used the Process > Average > Average files option in Brainstorm. I would appreciate it if you could confirm whether this is the correct procedure for this paradigm.

Thank you very much for your guidance.

2

If you want to analyze the P300 evoke-related potential, this should be present only for the oddball condition, and the P300 is related to the person's response to this non-common stimulus.

This alignment is given by the alignment of the oddball events in your data. A simple way to see if they are align could be to plot all the trials in a raster plot for electrodes Cz or Pz (where the major effect is expected). Instructions in here:

https://neuroimage.usc.edu/brainstorm/Tutorials/Epoching#Raster_plot

erpimage.gif
erpimage.gif898×348 125 KB

If they are not aligned, you would need to readjust the events before importing the epochs, so they correspond to the presentation of the auditory stimulus. This is only possible if the data includes an analog channel recording the audio stimulus. You can read mode in here:

https://neuroimage.usc.edu/brainstorm/Tutorials/StimDelays

@haniehsotudeh, please note that your dataset strongly resembles the dataset in our Get started tutorials. Please take a look to them so you can have insights on how to proceed with your analyses.

https://neuroimage.usc.edu/brainstorm/Tutorials#Get_started