Forum General Neuroscience and EEG Unit of Measurement for EEG data after FFT

Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
    Posts
  • #2693
    Misino Augusto
    Participant
      @arian-m-678578fc4f2f3

      When using EmotivPRO Analyzer, and applying FFT (Fast Fourier Transform), it returns the average power of each frequency band.

      BUT…

      What is the unit of measurement for that Average Power?
      Is it measured in µV or µV^2/Hz, or in dB ?

      And, the values… What range shold have? Positive and Negative numbers, or only Positive?

      Should one apply a method of normalization after applying FFT ??

       

      #2698
      Misino Augusto
      Participant
        @arian-m-678578fc4f2f3

        So… inspecting the page Band Power calculation | EmotivPRO Analyzer

        found a part of code:

        band = list(range(16, 25))
        band_power = np.sum(eeg_fft_square_3d[band, :, :], axis=0) / len(band)
        //Finally EmotivPRO Analyzer will output band power in decibels (dB):
        band_power_db = 10*np.log10(band_power))

        —————-

        They are using a variation of the Logarithmic transformation:
        P_dB = 10 * log_⁡10 ⁣( P / P_ref)

        Emotiv replaces the ” /P_ref ” with “/ len(band)”.

        On the paper Time-frequency analysis methods and their application in developmental EEG data;  it explains that “researchers generally use the decibel conversion … to baseline normalize raw total power values”.
        Meaning that “P_ref” is equal to a Baseline Activity.

        But Emotiv is using “len(band)”, where: band = list(range(16, 25)).

        In Python, len(list(range(16, 25))) returns 9…

        Maybe the code in the gitbook manual is just an example, but nevertheless I’m confused about this change (using the lenght of a list intead of a eeg reference).

        If someone from Emotiv or someone from the community with more experience can help me understand this, I will be very grateful.

      Viewing 2 posts - 1 through 2 (of 2 total)
      • You must be logged in to reply to this topic.
      Login