Yongcheng Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yongcheng Li has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
- Guanglin Li
Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China
7 shared publications - Frank Kulwa
Southern Medical University
4 shared publications Pengrui Tai
Shenzhen Institutes of Advanced Technology (SIAT), Chinese Academy of Sciences (CAS), Shenzhen, Guangdong, China
4 shared publications- Oluwarotimi Williams Samuel
University of Derby
3 shared publications - Chenying Xu
Southern University of Science and Technology, Shenzhen, Guangdong, China
2 shared publications Doreen S. Sarwatt
University of Science and Technology Beijing, Beijing, China
2 shared publications
Latest Publications
- Open Access
- 15 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Vol. 34 (15344320) , 2752-2761
A Novel Sliding Mode Differentiator-Based Feature for EMG-Based Hand Gesture Characterization
Motor intent (MI)-based muscle computer interfaces (MCIs) have been widely explored for prosthetic control as a means of restoring functionality to the lost limb in amputees. However, persistent issues remain, such as insufficient robustness of the current features, the presence of coherent noise, and the low spatial resolution of electromyography (EMG) sensors. Subsequently, these results in decreased movement characterization performance. Therefore, this study introduces a novel feature extraction technique that utilizes a Sliding Mode Differentiator (SMD) to extract unique patterns from EMG signals, followed by the deployment of symmetric positive definite matrices (SPD) to efficiently leverage the spatial-temporal properties of the EMG signal. The average classification results of 98.7±3.0% and 97.9±5.2% for 21 able-bodied subjects and 15 amputees respectively, suggests an improvement in accuracy for characterizing 13 hand gestures, thereby outperforming other state-of-the-art feature methods. Further, the channel optimization analysis shows that the number of channels can be reduced by 75% (from 24 channels to 6 channels) without compromising the performance of the proposed technique. This justifies the potential of the proposed technique in both high-density and sparse-density EMG electrode configurations. Additional analysis of the performance of the techniques in the presence of noise indicates that the proposed method can significantly outperform other methods. Therefore, the findings of this study have the potential to significantly improve the control performance of prostheses, rehabilitation assistive robots, and hand gesture-related games.
- 15 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Vol. 34 (15344320) , 2752-2761
Open AccessA Novel Sliding Mode Differentiator-Based Feature for EMG-Based Hand Gesture Characterization
Motor intent (MI)-based muscle computer interfaces (MCIs) have been widely explored for prosthetic control as a means of restoring functionality to the lost limb in amputees. However, persistent issues remain, such as insufficient robustness of the current features, the presence of coherent noise, and the low spatial resolution of electromyography (EMG) sensors. Subsequently, these results in decreased movement characterization performance. Therefore, this study introduces a novel feature extraction technique that utilizes a Sliding Mode Differentiator (SMD) to extract unique patterns from EMG signals, followed by the deployment of symmetric positive definite matrices (SPD) to efficiently leverage the spatial-temporal properties of the EMG signal. The average classification results of 98.7±3.0% and 97.9±5.2% for 21 able-bodied subjects and 15 amputees respectively, suggests an improvement in accuracy for characterizing 13 hand gestures, thereby outperforming other state-of-the-art feature methods. Further, the channel optimization analysis shows that the number of channels can be reduced by 75% (from 24 channels to 6 channels) without compromising the performance of the proposed technique. This justifies the potential of the proposed technique in both high-density and sparse-density EMG electrode configurations. Additional analysis of the performance of the techniques in the presence of noise indicates that the proposed method can significantly outperform other methods. Therefore, the findings of this study have the potential to significantly improve the control performance of prostheses, rehabilitation assistive robots, and hand gesture-related games.
- Open Access
- 17 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 430-435
Exploring the Cortical Responses and Brain Network Changes in Multisensory Motor Learning with EEG
- 17 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 430-435
Open AccessExploring the Cortical Responses and Brain Network Changes in Multisensory Motor Learning with EEG
- Open Access
- 27 February 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 12 (9) , eaea3355
Inflamed vessel–anchored release of H 2 across the blood-brain barrier for ischemic stroke neuroprotection
Ischemic stroke followed by reperfusion urgently requires safe and efficient cytoprotective strategies, a need still unmet by current pharmacotherapies. Nanotechnology holds promise for improved drug delivery to the brain, yet the efficacy of nanomaterials crossing the blood-brain barrier (BBB) is quite limited, and long-term intracranial retention of nanomaterials may provoke neurotoxicity. Leveraging the anti-inflammatory, BBB-crossing, and biosafe properties of hydrogen (H 2 ), we develop an inflamed vessel–targeted/anchored H 2 –producing system by modifying ZrSi 2 nanoparticles with a P-selectin–binding peptide (ZSNP), mimicking P-selectin/P-selectin glycoprotein ligand–mediated innate immune recruitment. Rather than relying on nanoparticle penetration into the brain parenchyma, this design enables ZSNP to anchor at the BBB vasculature, where it locally and continuously generates H 2 via hydrolysis. The released H 2 traverses the BBB, exerting cytoprotection through antioxidant and immunomodulatory mechanisms that coordinate multicellular recovery processes. Furthermore, ZSNP promotes microglia-mediated angiogenesis and neurogenesis, guides axonal projections along neovascular trajectories, and facilitates microglia-neuron interaction via the noncanonical Wnt/Ca 2+ pathway. This reconstruction of the neurovascular network supports the reintegration of functional neural circuits, leading to structural and functional recovery that surpasses the effects of edaravone. By enabling sustained H 2 release at the BBB interface without requiring nanoparticle intracranial accumulation, this strategy represents a promising and low-burden neuroprotective approach for ischemic stroke.
- 27 February 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 12 (9) , eaea3355
Open AccessInflamed vessel–anchored release of H 2 across the blood-brain barrier for ischemic stroke neuroprotection
Ischemic stroke followed by reperfusion urgently requires safe and efficient cytoprotective strategies, a need still unmet by current pharmacotherapies. Nanotechnology holds promise for improved drug delivery to the brain, yet the efficacy of nanomaterials crossing the blood-brain barrier (BBB) is quite limited, and long-term intracranial retention of nanomaterials may provoke neurotoxicity. Leveraging the anti-inflammatory, BBB-crossing, and biosafe properties of hydrogen (H 2 ), we develop an inflamed vessel–targeted/anchored H 2 –producing system by modifying ZrSi 2 nanoparticles with a P-selectin–binding peptide (ZSNP), mimicking P-selectin/P-selectin glycoprotein ligand–mediated innate immune recruitment. Rather than relying on nanoparticle penetration into the brain parenchyma, this design enables ZSNP to anchor at the BBB vasculature, where it locally and continuously generates H 2 via hydrolysis. The released H 2 traverses the BBB, exerting cytoprotection through antioxidant and immunomodulatory mechanisms that coordinate multicellular recovery processes. Furthermore, ZSNP promotes microglia-mediated angiogenesis and neurogenesis, guides axonal projections along neovascular trajectories, and facilitates microglia-neuron interaction via the noncanonical Wnt/Ca 2+ pathway. This reconstruction of the neurovascular network supports the reintegration of functional neural circuits, leading to structural and functional recovery that surpasses the effects of edaravone. By enabling sustained H 2 release at the BBB interface without requiring nanoparticle intracranial accumulation, this strategy represents a promising and low-burden neuroprotective approach for ischemic stroke.
- 1 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 73 (2) , 742-755
A NMF-Based Non-Euclidean Adaptive Feature Extraction Scheme for Limb Motion Pattern Decoding in Pattern Recognition System
Feature extraction is a crucial step in electromyogram (EMG)-based pattern recognition systems for decoding motor intents. However, despite the existence of numerous proposed techniques for feature extraction, their decoding performances have remained relatively low. Furthermore, these techniques are often evaluated without taking into account the drift between the training and test datasets. This study proposes a feature extraction scheme that operates in an unsupervised manner to address these limitations. This approach focuses on reducing drift between the training and test sets by utilizing feature adaptation based on non-negative matrix factorization (NMF) and Riemann operations. Additionally, we minimize drift by aligning the distribution of the test data with that of the training set. The results demonstrate that the proposed feature extraction technique exhibits significantly higher performance (p < 0.05) in decoding motor intent for 13 hand and finger movements, achieving an average accuracy of 99.91 ± 0.35% for amputee participants and 99.99 ± 0.02% for able-bodied participants. We also conducted further investigations to assess the effectiveness of the proposed feature scheme against varied signal-to-noise ratios (SNRs). These investigations revealed that our technique outperforms other feature extraction techniques in terms of decoding performance, even in the presence of varied SNRs. Overall, the findings show that the proposed feature extraction technique can effectively enhance the reliability and robustness of EMG control systems in both clinical and commercial applications.
- 1 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Biomedical Engineering
- Vol. 73 (2) , 742-755
A NMF-Based Non-Euclidean Adaptive Feature Extraction Scheme for Limb Motion Pattern Decoding in Pattern Recognition System
Feature extraction is a crucial step in electromyogram (EMG)-based pattern recognition systems for decoding motor intents. However, despite the existence of numerous proposed techniques for feature extraction, their decoding performances have remained relatively low. Furthermore, these techniques are often evaluated without taking into account the drift between the training and test datasets. This study proposes a feature extraction scheme that operates in an unsupervised manner to address these limitations. This approach focuses on reducing drift between the training and test sets by utilizing feature adaptation based on non-negative matrix factorization (NMF) and Riemann operations. Additionally, we minimize drift by aligning the distribution of the test data with that of the training set. The results demonstrate that the proposed feature extraction technique exhibits significantly higher performance (p < 0.05) in decoding motor intent for 13 hand and finger movements, achieving an average accuracy of 99.91 ± 0.35% for amputee participants and 99.99 ± 0.02% for able-bodied participants. We also conducted further investigations to assess the effectiveness of the proposed feature scheme against varied signal-to-noise ratios (SNRs). These investigations revealed that our technique outperforms other feature extraction techniques in terms of decoding performance, even in the presence of varied SNRs. Overall, the findings show that the proposed feature extraction technique can effectively enhance the reliability and robustness of EMG control systems in both clinical and commercial applications.
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Vol. 34, 3759-3769
The DMD-Based Non-Euclidean Descriptor for Limb Movement Decoding in Pattern Recognition System
Accurate decoding of motor intent in electromyogram (EMG)-based pattern recognition systems heavily depends on effective feature extraction. Although numerous techniques have been introduced, their performance remains suboptimal. Many fail to utilize spatial dependencies across EMG channels and overlook the distributional shift between training and test data. To overcome these challenges, this study introduces an unsupervised feature extraction method. It employs Dynamic Mode Decomposition (DMD) to isolate meaningful spatiotemporal patterns associated with hand gestures, followed by enhanced spatial representation through the Sample Covariance Matrix (SCM). To address dataset drift, the method aligns test data within the same non-Euclidean space as the training data. Experimental results show a significant improvement (p < 0.05) in decoding 13 distinct hand and finger gestures, with average accuracies of 99.89 ± 0.28% for amputees and 99.90 ± 0.31% for non-disabled subjects. Importantly, the method retains over 97% accuracy even when the number of EMG channels is reduced from 24 to just 3, suggesting its potential across both dense and sparse sensor configurations. This may be useful for low-cost prosthetic designs, where reducing the number of electrodes can decrease hardware complexity and cost. Thus, the proposed technique shows promise for improving the performance, reliability, and adaptability of EMG-based control systems, with potential relevance to clinical and commercial applications after further real-time and clinical validation.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Vol. 34, 3759-3769
Open AccessThe DMD-Based Non-Euclidean Descriptor for Limb Movement Decoding in Pattern Recognition System
Accurate decoding of motor intent in electromyogram (EMG)-based pattern recognition systems heavily depends on effective feature extraction. Although numerous techniques have been introduced, their performance remains suboptimal. Many fail to utilize spatial dependencies across EMG channels and overlook the distributional shift between training and test data. To overcome these challenges, this study introduces an unsupervised feature extraction method. It employs Dynamic Mode Decomposition (DMD) to isolate meaningful spatiotemporal patterns associated with hand gestures, followed by enhanced spatial representation through the Sample Covariance Matrix (SCM). To address dataset drift, the method aligns test data within the same non-Euclidean space as the training data. Experimental results show a significant improvement (p < 0.05) in decoding 13 distinct hand and finger gestures, with average accuracies of 99.89 ± 0.28% for amputees and 99.90 ± 0.31% for non-disabled subjects. Importantly, the method retains over 97% accuracy even when the number of EMG channels is reduced from 24 to just 3, suggesting its potential across both dense and sparse sensor configurations. This may be useful for low-cost prosthetic designs, where reducing the number of electrodes can decrease hardware complexity and cost. Thus, the proposed technique shows promise for improving the performance, reliability, and adaptability of EMG-based control systems, with potential relevance to clinical and commercial applications after further real-time and clinical validation.
- 26 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Robotics and Bionics
- Vol. 7 (3) , 1248-1262
fNIRS-Based Action Detection for Lower Limb Amputees in Sit-to-Stand Tasks
Traditional transfemoral lower-limb prostheses often overlook the intuitive neuronal connections between the brain and prosthetic actuators. This study bridges this gap by integrating a functional near-infrared spectroscopy (fNIRS) into real-time lower-limb prosthesis control with preliminary clinical tests on the above-knee amputee, enabling a more reliable volitional control of the prosthesis. Cerebral hemodynamic responses were measured using a 56-channel fNIRS headset, and lower-limb kinematics were recorded with a optical motion capture system. Artifacts in fNIRS were mitigated using short-separation regression, and eight features of the fNIRS data were extracted. ANOVA revealed the means, slope, and entropy as top-performing features across all subjects. Among eight classifiers tested, k-nearest neighbor (KNN) emerged as the most accurate. In this study, we recruited eleven healthy subjects and one unilateral transfemoral amputee. Classification rates surpassed 97% for all classes, maintaining an average accuracy of 99.86±0.01%. Notably, the amputee exhibited higher precision, sensitivity, and F1 scores than healthy subjects. Maximum temporal latencies for healthy subjects were 120.00±49.40 ms during sit-down and 119.09±45.71 ms during stand-up, while the amputee showed maximum temporal latencies of 90 ms and 190 ms, respectively. This study marks the first application of action detection in sit-to-stand tasks for transfemoral amputees via fNIRS, which underscores the potential of fNIRS in neuroprostheses control.
- 26 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Robotics and Bionics
- Vol. 7 (3) , 1248-1262
fNIRS-Based Action Detection for Lower Limb Amputees in Sit-to-Stand Tasks
Traditional transfemoral lower-limb prostheses often overlook the intuitive neuronal connections between the brain and prosthetic actuators. This study bridges this gap by integrating a functional near-infrared spectroscopy (fNIRS) into real-time lower-limb prosthesis control with preliminary clinical tests on the above-knee amputee, enabling a more reliable volitional control of the prosthesis. Cerebral hemodynamic responses were measured using a 56-channel fNIRS headset, and lower-limb kinematics were recorded with a optical motion capture system. Artifacts in fNIRS were mitigated using short-separation regression, and eight features of the fNIRS data were extracted. ANOVA revealed the means, slope, and entropy as top-performing features across all subjects. Among eight classifiers tested, k-nearest neighbor (KNN) emerged as the most accurate. In this study, we recruited eleven healthy subjects and one unilateral transfemoral amputee. Classification rates surpassed 97% for all classes, maintaining an average accuracy of 99.86±0.01%. Notably, the amputee exhibited higher precision, sensitivity, and F1 scores than healthy subjects. Maximum temporal latencies for healthy subjects were 120.00±49.40 ms during sit-down and 119.09±45.71 ms during stand-up, while the amputee showed maximum temporal latencies of 90 ms and 190 ms, respectively. This study marks the first application of action detection in sit-to-stand tasks for transfemoral amputees via fNIRS, which underscores the potential of fNIRS in neuroprostheses control.
- Open Access
- 22 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Vol. 33 (15344320) , 1586-1599
A Multi-Level Integrated EEG-Channel Selection Method Based on the Lateralization Index
The importance of optimizing channel selection for portable brain-computer interface (BCI) technology is increasingly recognized. Effective channel selection reduces computational load and enhances user experience by making BCI systems more comfortable and easier to use. A significant challenge lies in reducing the number of electrodes without compromising decoding accuracy. Although some methods have been proposed in previous studies, these often increase computational load and overlook the importance of channel selection across different subjects. Therefore, we propose a novel Multi-level Integrated EEG-Channel Selection method based on the Lateralization Index (MLI-ECS-LI). This method leverages the lateralization index in selecting important channels and can achieve the channel selection for the cross-tasks and the cross-subjects scenarios. To evaluate the effectiveness of the proposed method, the time and frequency domain features from selected channels were extracted. Three widely used classifiers, Least Squares Support Vector Machine (LSSVM), Random Forest (RF), and Support Vector Machine (SVM) were used to classify movement types based on these features. Compared to the conventional condition (C1-C6), the average decoding accuracies across 21 healthy subjects demonstrated an improved performance of 6.6%, 4.9%, 6.9% (LSSVM); 3.8%, 2.8%, 4.5%(RF); and 7.6%, 5.6%, 9.2%(SVM) via using the channels selected from the conditions of the single task, the cross-tasks, and the cross-subjects scenarios, respectively. These results demonstrated the potential of the proposed method in improving the utility of the portable Motor Imagery Brain-Computer Interface (MI-BCI) and effectiveness in practical applications.
- 22 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Systems and Rehabilitation Engineering
- Vol. 33 (15344320) , 1586-1599
Open AccessA Multi-Level Integrated EEG-Channel Selection Method Based on the Lateralization Index
The importance of optimizing channel selection for portable brain-computer interface (BCI) technology is increasingly recognized. Effective channel selection reduces computational load and enhances user experience by making BCI systems more comfortable and easier to use. A significant challenge lies in reducing the number of electrodes without compromising decoding accuracy. Although some methods have been proposed in previous studies, these often increase computational load and overlook the importance of channel selection across different subjects. Therefore, we propose a novel Multi-level Integrated EEG-Channel Selection method based on the Lateralization Index (MLI-ECS-LI). This method leverages the lateralization index in selecting important channels and can achieve the channel selection for the cross-tasks and the cross-subjects scenarios. To evaluate the effectiveness of the proposed method, the time and frequency domain features from selected channels were extracted. Three widely used classifiers, Least Squares Support Vector Machine (LSSVM), Random Forest (RF), and Support Vector Machine (SVM) were used to classify movement types based on these features. Compared to the conventional condition (C1-C6), the average decoding accuracies across 21 healthy subjects demonstrated an improved performance of 6.6%, 4.9%, 6.9% (LSSVM); 3.8%, 2.8%, 4.5%(RF); and 7.6%, 5.6%, 9.2%(SVM) via using the channels selected from the conditions of the single task, the cross-tasks, and the cross-subjects scenarios, respectively. These results demonstrated the potential of the proposed method in improving the utility of the portable Motor Imagery Brain-Computer Interface (MI-BCI) and effectiveness in practical applications.