Improving Accuracy of Event-Related Potentials Classification by Channel Selection Using Independent Component Analysis and Least Square Methods

Improving Accuracy of Event-Related Potentials Classification by Channel Selection Using Independent Component Analysis and Least Square Methods

Wenxuan Li, Mengfan Li, Wei Li
DOI: 10.4018/IJSSCI.2016070101
(Individual Articles)
No Current Special Offers


This paper proposes a method for achieving a high performance of N200 and P300 classification by applying independent component analysis to select the channels, which deliver brain signals with large N200 and P300 potentials and small artifacts. In this study, the authors find out the relationship between the highest accuracy and the weights of the independent components and use this relationship to predict the optimal channels of each individual subject. They compare five channel selection methods: the ICA-based method and the curve-fitting-based method proposed in this paper, the amplitude-based method, the experiential optimal 8 channel combination and all 30 channel combination methods. The comparative studies show that the ICA-based method achieves an average accuracy of 99.3% across four subjects, which is superior to the other four methods.
Article Preview

1. Introduction

A mind-controlled system constructs a bridge between human brain and external devices for communication and control (Wolpaw, 2000; Lebedev, 2006; Ortiz-Rosario, 2013). This technology extracts some specific features of user’s brain activity that related to their intention and afterwards transfers those features to control commands (Wang et al., 2013). Event-related potential (ERP) (Wei and Luo, 2010) includes two kinds of potentials that have been used for a mind-controlled humanoid robot: N200 potential is a negative deflection that occurs at post-stimulus 180~325ms (Li et al., 2014a) and P300 potential is a large positive deflection that occurs at post-stimulus 200~800ms (Li et al., 2014b). N200 and P300 have been applied into many areas, such as the control of a robotic arm (Palankar, 2008), an internet browser (Mugler, 2008), a humanoid robot (Li et al., 2013a), and a wheelchair (Puanhvuan and Wongsawat, 2012).

Achieving high classification accuracy is one of the key issues in N200-based and P300-based control systems. The high classification accuracy is closely related to the channel combination for effective feature extraction (Shahriari and Erfanian, 2011). Therefore, it is important to determine the channels with obvious ERP features for classification. Several research groups have investigated the channels where N200 and P300 potentials occur frequently, e.g., the channels often selected for exploring P300 distribution are Fz, Cz, Pz, Oz, P3, P4, P7 and P8 (Hoffmann et al., 2008) because the P300 potential mainly appears from these channels; the channels selected for examining N200 distribution are usually P3 and/or P7 because the amplitudes of the N200 potential in these channels are large (Hong, 2009). However, there exist individual differences in inducing brain signals, e.g., latency, amplitude and scalp distribution. To choose the channels for ERP feature extraction correctly by considering individuals would be able to significantly to improve the classification accuracy. The average classification accuracy across all the subjects may be low if every subject adopts the same channel combination for classification because they may produce their highest amplitudes in different channels. Decreasing the impact of individual differences on processing brain signals is a wise way to improve the classification accuracy. Some research groups have taken the individual differences into consideration when extracting brain signal features. The work (Hong, 2009) selected the channels with the largest N200 amplitude for classification. The work (Zhang, 2010) proposed an algorithm based on Discrete Particle Swarm Optimization (DPSO) to select the optimal channel combination of P300.

Complete Article List

Search this Journal:
Volume 15: 1 Issue (2023)
Volume 14: 4 Issues (2022): 1 Released, 3 Forthcoming
Volume 13: 4 Issues (2021)
Volume 12: 4 Issues (2020)
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing