Frequency domain normalizations of EEG signals

Şavklıyıldız A., İrim Y.

The 6th Mediterranean International Conference of Pure & Applied Mathematics and Related Areas (MICOPAM 2023), Paris, France, 23 - 27 August 2023, pp.87-91, (Summary Text)

  • Publication Type: Conference Paper / Summary Text
  • City: Paris
  • Country: France
  • Page Numbers: pp.87-91
  • Akdeniz University Affiliated: Yes

Abstract

In this paper, commonly used frequency domain normalizations and their effects on electroencephalography (EEG) signal are examined. Additionally, a novel normalization method is introduced to decrease between subject variability in frequency domain. This method basically uses the affine transformation to warp the frequency domain. Alpha band peak frequency and it’s near-harmonics are prominently active during eyes closed EEG recordings. However, these peak values can vary from person to person. Normalizing the frequency plane by mapping these peaks to fixed frequency positions on the frequency axis can reduce the variation caused by these oscillations. Thus, group analyzes can be evaluated in standardized time-frequency spectra.