Analysis of Repetitive Flash Stimulation Frequencies and Record Periods to Detect Migraine Using Artificial Neural Network

Abstract

Different kind of methods has been applied to detect the migraine by using flash stimulation. Especially frequency analysis of EEG signal is the most preferred method to detect the migraine by using flash stimulation. Different flash stimulation frequencies at wide frequency range have been used in migraine detection. But the effects of these flash stimulation frequencies and the most effective frequency can be determined by analyzing these frequencies separately. Since each stimulation frequency has been implemented in different time periods, it is necessary to determine the time period to detect magnitude increase in migraine patients. The aim of this study is to determine the most effective flash stimulation frequency and time duration to detect the migraine. In this study, we analyzed the flash stimulation frequencies and time duration separately for detecting migraine. Performance of each flash stimulation frequency has been determined to detect the migraine by analyzing the power spectrums obtained under 2 Hz, 4 Hz and 6 Hz and artificial neural network has been used to determine the which data has a superior performance. Afterwards we analyzed the 2 s, 4 s, 6 s, 8 s and 10 s of flash stimulation periods separately by observing the power spectrums and the results are verified by using artificial neural network. As a result of this study we proposed the 4 Hz of flash stimulation frequency is the most effective frequency and 8 s time period is necessary to detect the migraine at the beta band of EEG's T5-T3 channel.