Extraction and classification of Electroencephalogram signals

Brain Computer Interface creates a communication path way between brain and outside world. Brain signals are recorded and processed to translate Electroencephalogram activity to an external command. Brain signals recorded from the scalp or from inside the brain, enable users to control a variety of applications. This capability can be very useful for the patient, suffering from severe motor disorder. Efficient working of a Brain Computer Interface widely depends upon the signal processing methodology applied for feature extraction and classification of Electroencephalograms. The efficiency of a versatile signal processing framework proposed in this work has been determined for Electroencephalogram signals in terms of feature extraction and classification. The power spectral density of rhythmic components of the Electroencephalogram signals extracted using IEEE standard 1057 four-parameter sine wave fit algorithm, is calculated using Welch method and classification of the Electroencephalogram signals is done using Non-linear Support Vector Machine model.