TY - GEN
T1 - A Data Augmentation Framework for Decoding Movement Attempt in Stroke Patients Based on EEG
AU - Xu, Xiangyu
AU - Bi, Luzheng
AU - Wang, Xinyi
AU - Fei, Weijie
AU - Wang, Jiarong
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - Stroke is a leading cause of disability and death globally, impacting the quality of life for patients. Brain-computer interfaces (BCIs) offer promising potential for motor recovery. However, their performance is often limited by the scarcity of experimental data. This paper proposes a novel data augmentation framework based on an improved Adversarial Augmentation Network (AAN) to enhance motor attempt decoding from electroencephalogram (EEG) signals in stroke patients. The AAN integrates an attention-U-Net generator and a Domain-Adversarial Neural Network (DANN)-based discriminator to address sample scarcity. Experimental results show that the model improves decoding accuracy by 3.75% to 7.91% across five subjects, with an augmentation scale of 1x achieving optimal performance. This work provides a robust solution for motor attempt decoding and advances its potential application in rehabilitation systems.
AB - Stroke is a leading cause of disability and death globally, impacting the quality of life for patients. Brain-computer interfaces (BCIs) offer promising potential for motor recovery. However, their performance is often limited by the scarcity of experimental data. This paper proposes a novel data augmentation framework based on an improved Adversarial Augmentation Network (AAN) to enhance motor attempt decoding from electroencephalogram (EEG) signals in stroke patients. The AAN integrates an attention-U-Net generator and a Domain-Adversarial Neural Network (DANN)-based discriminator to address sample scarcity. Experimental results show that the model improves decoding accuracy by 3.75% to 7.91% across five subjects, with an augmentation scale of 1x achieving optimal performance. This work provides a robust solution for motor attempt decoding and advances its potential application in rehabilitation systems.
KW - Brain-Computer Interface (BCI)
KW - Data Augmentation Framework
KW - Motor Attempt Decoding
KW - Stroke Rehabilitation
UR - https://www.scopus.com/pages/publications/105031918133
U2 - 10.1109/ICUS66297.2025.11294085
DO - 10.1109/ICUS66297.2025.11294085
M3 - Conference contribution
AN - SCOPUS:105031918133
T3 - Proceedings of 2025 IEEE International Conference on Unmanned Systems, ICUS 2025
SP - 13
EP - 17
BT - Proceedings of 2025 IEEE International Conference on Unmanned Systems, ICUS 2025
A2 - Song, Rong
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2025 IEEE International Conference on Unmanned Systems, ICUS 2025
Y2 - 18 September 2025 through 19 September 2025
ER -