Abstract
In the past, some electroencephalogram (EEG) models have been suggested in order to analyze the EEG. Almost all these EEG models have rather complicated structures and include a great many parameters. They are defective in that it is difficult to find the correspondence between the model parameters and the findings of the doctors.
In this paper, we propose a simple EEG model expressed by the sinusoidal waves with the Markov process amplitude, which overcomes these defects. This EEG model can aptly represent the features of the EEG in terms of a few parameters, and it is easy to find the intuitive correspondence between the model parameters and the findings of the doctors. To determine the parameters of the EEG model, the method for optimally approximating the spectrum with a distribution function in a frequency domain is used. In the latter half of this paper, the validity of the EEG model it both a frequency domain and a time domain is shown, and this EEG model can then be applied to quantitatively express the features of the EEG. Some examples are investigated herein.
