Classification of electroencephalography signals recorded during olfaction of some spices

Responses to different information from the sensory organs of the brain can be analyzed by various brain imaging techniques. Among these techniques, electroencephalography (EEG) is widely used because it is performed at a low cost without additional equipment and because of noninvasive method. In recent years, EEG signals recorded during olfaction based studies have been tried, but the response of the human brain to different odors has not been fully proven due to differences in experimental outputs and lack of odor use. In this study, to improve the use of limited odor types, EEG data recorded during the smelling of 4 natural oils (mint, clove, thyme, rosemary) obtained by one %100 cold printing were used. After the EEG data set recorded from 3 subjects are preprocessed with unit change or minimum-maximum normalization, statistically based features were extracted from the signal. Then the dual combinations of these oils were classified with k-nearest neighborhood method and a 6 classification results were obtained for each subject. We calculated the average 72.66%, 72.27% and 70.40% SD for each subject. It shows that the proposed method will be used clinically to successfully determine the loss or lack of odor in subjects.