TY - JOUR
T1 - Multiview Contrastive Learning for Unsupervised Domain Adaptation in Brain-Computer Interfaces
AU - Asgarian, Sepehr
AU - Wang, Ze
AU - Wan, Feng
AU - Wong, Chi Man
AU - Liu, Feng
AU - Mohsenzadeh, Yalda
AU - Wang, Boyu
AU - Ling, Charles X.
N1 - Publisher Copyright:
© 1963-2012 IEEE.
PY - 2024
Y1 - 2024
N2 - Domain adaptation has gained significant attention to address the nonstationarity problem in electroencephalography (EEG) data in motor imagery (MI) classification. In MI classification, domain adaptation addresses cross-session variations, enhancing the classifier's generalization capabilities. However, the existing methods have struggled to effectively capture both temporal and spatial features, resulting in limited classification accuracy. To tackle this issue, we propose the multiview adversarial contrastive network (MACNet). The proposed MACNet simultaneously learns spatial and temporal features in two different views: Euclidean and Riemannian. Furthermore, we introduce a multilevel domain mix-up technique to enhance domain alignment at both signal and embedding levels. The proposed MACNet method is evaluated on three public datasets. It achieves an accuracy of 83.79% on the BCI Competition IV dataset, 80.00% on the open source brain-machine interface (OpenBMI) dataset, and 85.83% on the sensorimotor rthythms (SMR) dataset that outperforms previous methods in cross-session transfer learning.
AB - Domain adaptation has gained significant attention to address the nonstationarity problem in electroencephalography (EEG) data in motor imagery (MI) classification. In MI classification, domain adaptation addresses cross-session variations, enhancing the classifier's generalization capabilities. However, the existing methods have struggled to effectively capture both temporal and spatial features, resulting in limited classification accuracy. To tackle this issue, we propose the multiview adversarial contrastive network (MACNet). The proposed MACNet simultaneously learns spatial and temporal features in two different views: Euclidean and Riemannian. Furthermore, we introduce a multilevel domain mix-up technique to enhance domain alignment at both signal and embedding levels. The proposed MACNet method is evaluated on three public datasets. It achieves an accuracy of 83.79% on the BCI Competition IV dataset, 80.00% on the open source brain-machine interface (OpenBMI) dataset, and 85.83% on the sensorimotor rthythms (SMR) dataset that outperforms previous methods in cross-session transfer learning.
KW - Electroencephalography (EEG)
KW - multiview contrastive
KW - neural network
KW - Riemannian neural network
KW - transfer learning
KW - unsupervised domain adaptation (UDA)
UR - http://www.scopus.com/inward/record.url?scp=85185368822&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85185368822&partnerID=8YFLogxK
U2 - 10.1109/TIM.2024.3366285
DO - 10.1109/TIM.2024.3366285
M3 - Article
AN - SCOPUS:85185368822
SN - 0018-9456
VL - 73
SP - 1
EP - 10
JO - IEEE Transactions on Instrumentation and Measurement
JF - IEEE Transactions on Instrumentation and Measurement
M1 - 2509410
ER -