3 Review(s)

EEG is used in a variety of applications it still suffers from some limitations which affect the analysis of a signal. The EEG has low signal to Noise (SNR) ratio because it measures the brain activity mixed with artifacts. Hence filtering and noise reduction techniques should be used to extract actual brain signal out of raw EEG. The EEG is a non-stationary signal [30] which varies across the time. Hence classifier trained on limited amount of user data may poorly classify generalized data recorded for the same individual. The physiological differences between the individuals generate EEG signal with varying magnitudes. This seriously affects the performance of the algorithms which are developed to generalize across the subjects This in turn limits the use of EEG applications. To solve some of the problems mentioned above, domain specific approaches are widely used which has reduced the flexibility and generalization capability of the system.