Spectral subtraction denoising improves accuracy of Slow Cortical Potential based brain-computer interfacing

Abstract

Spectral subtraction denoising as a new preprocessing block for Slow Cortical Potential(SCP)-based brain computer interface is proposed. This method adaptively reduces the noise attached to the signal using spectral subtraction denoising. We show that this technique provides better performance when low number of electrodes is used. This suggests its potential application in portable BCI real time applications where low power and minimum weight are required for practical use. Here, the classification accuracy was used as performance measure. The developed method provides relatively better performance compared to the more commonly used wavelet shrinkage signal denoising.

A novel approach of decoding EEG four-class motor imagery tasks via scout ESI and CNN

“…EEG has become more and more popular because of its advantages such as low cost, easy portability and high temporal resolution. EEG signals contain sensorimotor rhythms (SMRs) [1][2][3], slow cortical potentials [4][5][6], P300 [7][8][9], and steady-state visual evoked potential [10][11][12]. Due to its spontaneity and high application value, much attention is paid to SMRs.…”