Citation: Jiang, Q.; Zhang, Y.; Zheng,
K. Motor Imagery Classification via
Kernel-Based Domain Adaptation on
an SPD Manifold. Brain Sci. 2022, 12,
659. https://doi.org/10.3390/
brainsci12050659
Academic Editors: Amedeo
D’Angiulli and Pietro Aricò
Received: 27 March 2022
Accepted: 13 May 2022
Published: 18 May 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
Motor Imagery Classification via Kernel-Based Domain
Adaptation on an SPD Manifold
Qin Jiang
1
, Yi Zhang
2,3,
* and Kai Zheng
2
1
College of Computer Science and Technology, Chongqing University of Posts and Telecommunications,
Chongqing 400065, China; namy_jiang@hotmail.com
2
School of Advanced Manufacturing Engineering, Chongqing University of Posts and Telecommunications,
Chongqing 400065, China; zhengkai2001@163.com
3
Advanced Manufacturing and Automatization Engineering Laboratory, Chongqing University of Posts and
Telecommunications, Chongqing 400065, China
* Correspondence: zhangyi@cqupt.edu.cn; Tel.: +86-023-6248-0054
Abstract:
Background: Recording the calibration data of a brain–computer interface is a laborious
process and is an unpleasant experience for the subjects. Domain adaptation is an effective technology
to remedy the shortage of target data by leveraging rich labeled data from the sources. However,
most prior methods have needed to extract the features of the EEG signal first, which triggers another
challenge in BCI classification, due to small sample sets or a lack of labels for the target. Methods: In
this paper, we propose a novel domain adaptation framework, referred to as kernel-based Riemannian
manifold domain adaptation (KMDA). KMDA circumvents the tedious feature extraction process
by analyzing the covariance matrices of electroencephalogram (EEG) signals. Covariance matrices
define a symmetric positive definite space (SPD) that can be described by Riemannian metrics. In
KMDA, the covariance matrices are aligned in the Riemannian manifold, and then are mapped to
a high dimensional space by a log-Euclidean metric Gaussian kernel, where subspace learning is
performed by minimizing the conditional distribution distance between the sources and the target
while preserving the target discriminative information. We also present an approach to convert
the EEG trials into 2D frames (E-frames) to further lower the dimension of covariance descriptors.
Results: Experiments on three EEG datasets demonstrated that KMDA outperforms several state-of-
the-art domain adaptation methods in classification accuracy, with an average Kappa of 0.56 for BCI
competition IV dataset IIa, 0.75 for BCI competition IV dataset IIIa, and an average accuracy of 81.56%
for BCI competition III dataset IVa. Additionally, the overall accuracy was further improved by 5.28%
with the E-frames. KMDA showed potential in addressing subject dependence and shortening the
calibration time of motor imagery-based brain–computer interfaces.
Keywords:
EEG; brain–computer interfaces; domain adaptation; subspace learning; symmetric
positive definite matrices; Riemannian manifolds
1. Introduction
A brain–computer interface (BCI) provides a direct control pathway between the hu-
man brain and external devices, without relying on peripheral nerve and muscle systems [
1
].
BCIs have demonstrated potential in medical rehabilitation, education, smart homes, and
so on. Most non-invasive BCIs are based on EEG signals, and the neural response patterns
are decoded by well-designed algorithms, which can convert movement intentions into
computer commands to control external devices, such as a wheelchair [
2
], an artificial
limb [
3
], a spelling system [
4
,
5
], or a quadcopter [
6
]. Steady-state visual evoked potential
(SSVEP), P300, and motor imagery (MI) are widely studied neural response paradigms for
BCIs. SSVEP and P300 have shown breakthroughs in spelling applications [
4
–
6
], while MI
is prized for its simple stimulus paradigm design, and allows subjects to express motor
intention in a natural way [7].
Brain Sci. 2022, 12, 659. https://doi.org/10.3390/brainsci12050659 https://www.mdpi.com/journal/brainsci
