Interfaz Cerebro Computador Controlada por Sincronización y Desincronización Relacionada a Eventos en Sujetos no Entrenados

摘要

A Brain Computer Interface (BCI) is a system that allows the brain signal translation into commands for the control of external devices or communication. The research in BCI systems has increased in the past few years because it provides patients in motor disability situation with an alternative channel of communication, so their autonomy and independence which is reflected in the improvement of their quality of life. The increase or decrease of the brain signal power after certain kind of stimuli is known as event related synchronization or event related (de)synchronization (ERS/ERD), respectably. The μ and β rhythms are brain signals that present event related ERS/ERD before, during and after a motor task. This is why the μ and (β rhythms can be used as a control signal of a BCI. This paper explains the implementation of a BCI that detects the (de)synchronization of the μ and β rhythms through a feature extraction algorithm called Common Spatial Pattern and a pattern recognition technique called Support Vector Machine. The final application of the system is the control of the horizontal position of a computer cursor. The EEG pattern recognition system was verified by means of signals previously acquired by another BCI research groups and by signals acquired by us. The achieved results were satisfactory; during the training phase classification up to 90% was obtained and the system was able to successfully control the cursor movement up to 67% in average. Future studies must be made to the system In order to determine the error sources so it becomes a computer independent electronic system that can be easily carried out, for example, in a wheel chair.