Electroencephalography (EEG)-based Brain–computer interface (BCI) technology allows a user to control an external device without muscle intervention through recorded neural activity. Ongoing research on BCI systems includes applications in the medical field to assist subjects with impaired motor functionality (e.g., for the control of prosthetic devices). In this context, the accuracy and efficiency of a BCI system are of paramount importance. Comparing four different dimension reduction techniques in combination with linear and nonlinear classifiers, we show that integrating these methods in a BCI system results in a reduced model complexity without affecting overall accuracy.
Dimension Reduction Techniques in a Brain–Computer Interface Application
Cozza F.;Galdi P.;Serra A.;Tagliaferri R.
2020-01-01
Abstract
Electroencephalography (EEG)-based Brain–computer interface (BCI) technology allows a user to control an external device without muscle intervention through recorded neural activity. Ongoing research on BCI systems includes applications in the medical field to assist subjects with impaired motor functionality (e.g., for the control of prosthetic devices). In this context, the accuracy and efficiency of a BCI system are of paramount importance. Comparing four different dimension reduction techniques in combination with linear and nonlinear classifiers, we show that integrating these methods in a BCI system results in a reduced model complexity without affecting overall accuracy.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
