Comparison of sub-space based MUSIC and AR BURG methods in diagnosis of migraine by support vector machines

Migraine disease is a painful and throbbing brain disorders. Important studies have done about automatic diagnose method of migraine by biomedical equipment. In these studies Flash stimulation which is a triggering factor of migraine used as a triggering factor and EEG signals of migraine patient obtained results by flash stimulation is observed. Most important result of these flash stimulation based analyze method is magnitude increasing have observed at the 13–30Hz ranged frequency EEG signals of migraine patient. In this study performance of frequency domain transform methods which can be used for determining the magnitude increasing of migraine patient´s EEG signals are researched. In our study EEG signals obtained from both migraine patients and healthy subjects are transformed to frequency domain by using (AR) Burg method which is a Autoregressive based method and (MUSIC) method which is a Sub-space based method. Obtained results by using these methods are compared by observing the power spectral density (PSD) graphics and support vector machine classification method is used again for performance analysis of these methods again. According to obtained performance result of these methods which transform method is more useful for diagnosis of migraine is determined.