Comparative analysis of classification techniques for motor imagery based BCI

The main principle behind EEG-based brain computer interfaces (BCI) is the recording and accurate classification of EEG signals during imagination of different types of motor movements. The changes in the neural activity effected by motor imagery are a lot similar to those induced by actual movement. Common features, e.g., band power values, present in the single EEG trials are extracted by suitable methods for classification using SVM, neural networks or ensemble classifiers. The classifiers yield different efficiencies and are compared to find the optimal technique for same number of features. The neural net techniques proved to be the most efficient.