School of Automation, Guangdong Key Laboratory of IoT Information Technology, Guangdong University of Technology, Guangzhou 510006, People's Republic of China.
Physiol Meas. 2020 Aug 11;41(7):075007. doi: 10.1088/1361-6579/aba07b.
Brain-computer interfaces (BCIs) are aimed at providing a new way of communication between the human brain and external devices. One of the major tasks associated with the BCI system is to improve classification performance of the motor imagery (MI) signal. Electroencephalogram (EEG) signals are widely used for the MI BCI system. The raw EEG signals are usually non-stationary time series with weak class properties, degrading the classification performance.
Nonnegative matrix factorization (NMF) has been successfully applied to pattern extraction which provides meaningful data presentation. However, NMF is unsupervised and cannot make use of the label information. Based on the label information of MI EEG data, we propose a novel method, called double-constrained nonnegative matrix factorization (DCNMF), to improve the classification performance of NMF on MI BCI. The proposed method constructs a couple of label matrices as the constraints on the NMF procedure to make the EEGs with the same class labels have the similar representation in the low-dimensional space, while the EEGs with different class labels have dissimilar representations as much as possible. Accordingly, the extracted features obtain obvious class properties, which are optimal to the classification of MI EEG.
This study is conducted on the BCI competition III datasets (I and IVa). The proposed method helps to achieve a higher average accuracy across two datasets (79.00% for dataset I, 77.78% for dataset IVa); its performance is about 10% better than the existing studies in the literature.
Our study provides a novel solution for MI BCI analysis from the perspective of label constraint; it provides convenience for semi-supervised learning of features and significantly improves the classification performance.
脑-机接口（BCI）旨在为人类大脑与外部设备之间提供一种新的通信方式。BCI 系统的主要任务之一是提高运动想象（MI）信号的分类性能。脑电图（EEG）信号广泛用于 MI BCI 系统。原始 EEG 信号通常是非平稳的时间序列，具有较弱的类特性，从而降低了分类性能。
非负矩阵分解（NMF）已成功应用于模式提取，为数据表示提供了有意义的方法。然而，NMF 是无监督的，无法利用标签信息。基于 MI EEG 数据的标签信息，我们提出了一种新的方法，称为双约束非负矩阵分解（DCNMF），以提高 NMF 在 MI BCI 上的分类性能。该方法构建了一对标签矩阵作为 NMF 过程的约束，以使具有相同类标签的 EEG 在低维空间中具有相似的表示，而具有不同类标签的 EEG 尽可能具有不同的表示。因此，提取的特征具有明显的类特性，这对 MI EEG 的分类是最优的。
本研究在 BCI 竞赛 III 数据集（I 和 IVa）上进行。所提出的方法有助于在两个数据集上实现更高的平均准确率（数据集 I 为 79.00%，数据集 IVa 为 77.78%）；其性能比文献中现有的研究方法要好 10%左右。
我们的研究从标签约束的角度为 MI BCI 分析提供了一种新的解决方案；它为特征的半监督学习提供了便利，并显著提高了分类性能。
