An application of wavelet analysis to the human cerebral evoked potential
木原, 章一 徳島大学
生田, 琢己 徳島大学
古田, 典子 徳島大学
吉松, 誠 徳島大学
大蔵, 雅夫 徳島大学
Adopting Wavelet analysis for the analysis of the waveform of EP (Evoked Potential), it was intended to confirm possibility of the subjective diagnosis for schizophrenia, epilepsy and healthy subject.
1. Each waveform of EPs (SEP, VEP and AEP) of 100 healthy male subjects (25.4±3.1 y.o.), 100 male schizophrenics (31.6± 10.4 y.o.) and 99 male epileptics (35.9± 13.1 y.o.) were subjected to the Wavelet analysis and the multiresolution analysis. As a result, the difference of Wavelet values among these subject groups were verified significantly. The Wavelet values was defined as the square means of each wavelet coefficients at each scale of multiresolution analysis.
2. The waveform of EPs of each subject were discriminated by use of discriminant method derived from Wavelet Function, in order to judge to which subject group the subject belongs, according to the wavelet values. Analyzing the waveform of EP, between 15~600msec of latency, the highest value of sensitivity, 89%, 87% and 88% was obtained for normal subject, schizophrenics and epileptics, respectively.
As the conclusion, the possibility of the subjective diagnosis for schizophrenia, epilepsy, and normal subjects, by Wavelet analysis of EP waveform, were verified possible.
sam_53_6_269.PDF 7.04 MB