Extract power per electrodes ROIs (frontal, temporal, etc.)

Hello :slight_smile:
I have two questions concerning my analyses of EEG data.

  1. I imported my channel file and added info in the "group" column to identify different ROIs, as shown in the picture.

    I would like to extract the power values in the gamma band (35-45 Hz) for the entire length of the signal. Is there a procedure I could use to average the value per group of channels (i.e., my ROIs)? (instead of having a value per channel)
  2. I did not understand how to properly adjust/select some of the options in the "extract value" pipeline editor. I could not find examples of how/why one should select different parameters. I also noticed some of them can be selected at the same time. Could you provide me with an overview of it? That would be very helpful. Thank you so much!

Hi @AnnalisaPalmisano, here are my thoughts;
Q1. Brainstorm's frequency decomposition methods such as FFT and PSD (Welch's method) allow the user to input the sensor types or names as shown below. So you could perhaps write down the names of the sensors that correspond to your group? (not sure if you can call the group name specifically).
And if you click on the "Edit" button you can specify different frequency bands - Brainstorm has its default frequency band names and what frequencies they represent but you can adjust them as you wish.

Q2. Extract value process

  • Compute absolute values; this computes and extracts the absolute values (i.e. positive values of the data) of the recording
  • Average selected time window; this averages the data from the specified time window by the user (if you select all file it will average the entire file)
  • averaged selected signals; this averages all the specified signals that the user has specified
  • the multiple options are there so that you can compute things like; average selected time window AND return the absolute values of the averaged time window (i.e. you can combine the different options to get what you want)
  • concatenate signals; if you have multiple sensors specified (e.g. Fpz, Fz, Pz and Oz) and select this option Brainstorm will stack all the sensors on top of each other with time on the x-axis
  • concatenate time; if you have multiple epochs of data (regardless of how many sensors you've specified) and select this option Brainstorm will merge all the epochs together in time (i.e. along the x-axis). That is, if you have 5 X 500 ms epochs with 32 channels of EEG data and select this option Brainstorm will return 1 X 2500 ms recording with 32 channels of EEG data merged side to side along the x-axis.

I think that's it - I'm sure the Brainstorm team will add additional info to the above - hope that made sense.

1 Like

Hi @mcp0228,
Thank you for the clarification! Your feedback was very helpful.