Muse internal smoothing or de-noising method



Consider the following process which the Muse device uses to calculate the absolute band power:

S1: Capture raw EEG values
S2: Carry out FFT
S3: Use the FFT values to consolidate the absolute values for various waves (alpha, beta etc.)

I just wanted to understand the following:

  1. What are the smoothing and de-noising techniques (if any) that are carried out in the above process flow?
  2. At which step is the de-noising applied?
  3. Are techniques like ICA, PCA, Wavelet de-noising used?



Hi chhetri,

The EEG is filtered first by the analog electronics which impose a simple bandpass filter characteristic. Then there is some digital low pass filtering in the firmware, as well as a digital notch filter centred around the mains power frequency (50Hz or 60Hz depending on region). Special techniques like PCA, ICA, or wavelet denoising are not applied to the raw EEG signal.