Improving session-to-session transfer performance of motor imagery-based BCI using adaptive extreme learning machine

Abstract

Non-stationarity of electroencephalograph (EEG) data from session-to-session transfer is one of the challenges for EEG-based brain-computer interface systems, which can inversely affect their performance. Among methods proposed to address non-stationarity, adaptation is a promising method. In this study, an adaptive extreme learning machine (AELM) is proposed to update the initial classifier from the calibration session by using chunks of EEG data from the evaluation session whereby the common spatial pattern (CSP) algorithm is used to extract the most discriminative features. The effectiveness of the proposed algorithm is on motor imagery data collected from 12 healthy subjects during a calibration session and an evaluation session on a separate day. The results from the proposed AELM were compared with non-adaptive ELM and SVM classifiers. The results showed that AELM was significantly better (p=0.03). Moreover, the results also showed that accumulating the evaluation session data and using them for adapting the classifier will significantly improve the performance (p=0.001). Hence, the proposed AELM is effective in addressing the non-stationarity of EEG signal for online BCI systems.

Comparative analysis of strategies for feature extraction and classification in SSVEP BCIs

“…The canonical approach is to use the method of least squares, presented in Section 7.3. The ELM is an interesting option in the context of BCI in view of the simplicity of its associated training process and of its inherent regularization properties [30,31].…”

Binary Classification on ECoG Signals Using Optimized Extremely Learning Machine

“…Defects in the application of SVM can be successfully overcame by OELM for its high accuracy and fast speed. The BCI system can generate better results with OELM than with other state-of-the-art methods when analyzing and processing ECoG signals [12]. Seen from Table 7, ELM is comparable with OELM in speed, however OELM runs much fast than SVM by a factor up to thousands, whether in training or testing module.…”

A review on transfer learning for brain-computer interface classification

“…However, this hypothesis may not be true in practical appli cations. Atieh et al [14] proposed an adaptive extreme learning machine (ELM) to address non-stationarity of EEG data from one session to another. Adaptive ELM algorithm first trains ELM on the EEG data from prior sessions, then uses the leaned ELM classifier to test data from the test session.…”

“…Unsupervised Adaptive LDA I updates the common mean at time t, fL(t), by the rule as Eq. (14) fL(t) = (1 -7])fL(t -1) + 7]x(t) (14) where 7] denotes the update coefficient. Unsupervised Adap tive LDA II updates both the common mean fL and the inverse matrix of global covariance matrix by Eq.…”