Classification of electroencephalography signals recorded during smelling of rosewater and lotus flower odors

The human brain, nerve center of command system, receives stimulus from the sense organs and sends these signals out to the muscles. There are many kinds of techniques about watching answer the brain for inputs coming from the sense organs. Functional magnetic resonance imaging, electrocorticography, magnetoencephalography and electroencephalography (EEG) techniques are frequently used to measure these signals, but EEG is the most widely used all of these techniques. Advantages such as easy acquisition, painless and low cost make EEG preferable. In this work, EEG signals recorded during smelling of rosewater and lotus flower odors were analyzed and classified. The features calculated and classified are skewness, kurtosis and second order derivation of variance of EEG signals. The EEG signals recorded in Swiss Federal Institute of Technology are from 5 healthy subjects in two different conditions; eyes open and eyes closed. The data are classified by k-nearest neighbor algorithm and the mean of classification accuracy rate is obtained as 97.31 % for the subject eyes open condition and 97.34% for the subject eyes closed. The results achieved with this work prove that the proposed method have great potential for classification the EEG signals.