Abstract
Since the electroencephalogram (EEG) is used as a supporting method in the diagnosis of cerebral damage in a medical field, it is important to grasp the features of the EEG. At present, doctors read the EEG in the time domain qualitatively.
In this paper, we propose a method for optimally approximating power spectrum with distribution function. In this method, we assume a distribution function which is similar in form to the power spectrum of the EEG, and calculate the parameters of this distribution function by minimizing the square error of the area between the power spectrum of the EEG and the distribution function using the extremum search method. Then the features of the power spectrum of the EEG can be expressed by the parameters of the distribution function. In order to show the effectiveness of this method, using the parameters of the distribution function we calculate the feature parameters of the EEG which are closely related to the findings of doctors, and we judge the EEG quantitatively using these parameters. The results by this method agree approximately with the judgements of doctor.
