0 Applied Advanced Classifiers for Brain Computer Interface

Since that Dr. Hans Berger discovered the electrical nature of the brain, it has been considered the possibility to communicate personswith external devices only through the use of the brain waves (Vidal, 1973). Brain Computer Interface technology is aimed at communicating with persons using external computerised devices via the electroencephalographic signal as the primary command source (Wolpaw, J.R.; et al., 2000), (Birbaumer, N; et al., 2000). In the first international meeting for BCI technology it was established that BCI “must not depend on the brain’s normal output pathways of peripheral nerves and muscles” (Wolpaw, J. R.; et al., 2002). The primary uses of this technology are to benefit persons with blocking diseases, such as: Amiotrophic Lateral Sclerosis (ALS), brainstem stroke, or cerebral palsy; or persons whom have suffered some kind of traumatic accident like for example paraplegic (E. Donchin and K. M. Spencer and R. Wijesinghe, 2000). Actually different types of classifications can be established for BCI technology, from the physiologic point of view BCI devices can be classified in exogenous and endogenous. The exogenous devices provide some kind of stimuli to the user and they analyse the user’s responds to them, examples of this class are devices based on visual evoked potential or P300 (E. Donchin and K. M. Spencer and R. Wijesinghe, 2000). On the contrary, the endogenous devices do not depend on the user’s respond to external stimuli, they base their operation in detecting and recognising brain-wave patterns controlled autonomously by the user, examples of this class are devices based on the desynchronisation and synchronisation of μ and β rhythms (Wolpaw, J. R.; et al., 2002), (Pfurtscheller et al., 2000a), (Pineda, J.A. et al., 2003). But in any case, independently of the classification criteria, an algorithm that detects, acquires, filters, learns and classifies the electroencephalographic signal is required in order to control an external device using thoughts, associating some mental patterns to device commands, as it is shown in the block diagram of Figure 1, (Kostov, A., 2000), (Pfurtscheller et al., 2000b). The first block is in charge of acquiring and amplifying the brain signal, allocating the electrodes on specific places on the scalp in case of using superficial electrodes, or inside the brain in case of using intracortical ones; in the second block the signal is sampled, quantified and codified at periodic intervals of time in order to digitalize it, to simplify the following phases the digitalised signal may be filtered, for example to reduce the noise level obtaining a better SNR signal or identifying and processing artifacts. After this, in oder to obtain a set 2