Pub Date : 2024-12-25DOI: 10.1109/TNSRE.2024.3522681
Ke Ma, Siwei Liu, Mengjie Qin, Stephan M Goetz
Motor-evoked potentials (MEPs) are among the few directly observable responses to external brain stimulation and serve a variety of applications, often in the form of input-output (IO) curves. Previous statistical models with two variability sources inherently consider the small MEPs at the low-side plateau as part of the neural recruitment properties. However, recent studies demonstrated that small MEP responses under resting conditions are contaminated and over-shadowed by background noise of mostly technical quality, e.g., caused by the amplifier, and suggested that the neural recruitment curve should continue below this noise level. This work intends to separate physiological variability from background noise and improve the description of recruitment behaviour. We developed a triple-variability-source model around a logarithmic logistic function without a lower plateau and incorporated an additional source for background noise. Compared to models with two or fewer variability sources, our approach better described IO characteristics, evidenced by lower Bayesian Information Criterion scores across all subjects and pulse shapes. The model independently extracted hidden variability information across the stimulated neural system and isolated it from background noise, which led to an accurate estimation of the IO curve parameters. This new model offers a robust tool to analyse brain stimulation IO curves in clinical and experimental neuroscience and reduces the risk of spurious results from inappropriate statistical methods. The presented model together with the corresponding calibration method provides a more accurate representation of MEP responses and variability sources, advances our understanding of cortical excitability, and may improve the assessment of neuromodulation effects.
{"title":"Extraction of three mechanistically different variability and noise sources in the trial-to-trial variability of brain stimulation.","authors":"Ke Ma, Siwei Liu, Mengjie Qin, Stephan M Goetz","doi":"10.1109/TNSRE.2024.3522681","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3522681","url":null,"abstract":"<p><p>Motor-evoked potentials (MEPs) are among the few directly observable responses to external brain stimulation and serve a variety of applications, often in the form of input-output (IO) curves. Previous statistical models with two variability sources inherently consider the small MEPs at the low-side plateau as part of the neural recruitment properties. However, recent studies demonstrated that small MEP responses under resting conditions are contaminated and over-shadowed by background noise of mostly technical quality, e.g., caused by the amplifier, and suggested that the neural recruitment curve should continue below this noise level. This work intends to separate physiological variability from background noise and improve the description of recruitment behaviour. We developed a triple-variability-source model around a logarithmic logistic function without a lower plateau and incorporated an additional source for background noise. Compared to models with two or fewer variability sources, our approach better described IO characteristics, evidenced by lower Bayesian Information Criterion scores across all subjects and pulse shapes. The model independently extracted hidden variability information across the stimulated neural system and isolated it from background noise, which led to an accurate estimation of the IO curve parameters. This new model offers a robust tool to analyse brain stimulation IO curves in clinical and experimental neuroscience and reduces the risk of spurious results from inappropriate statistical methods. The presented model together with the corresponding calibration method provides a more accurate representation of MEP responses and variability sources, advances our understanding of cortical excitability, and may improve the assessment of neuromodulation effects.</p>","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"PP ","pages":""},"PeriodicalIF":4.8,"publicationDate":"2024-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143541913","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Advancements in neuroscience and artificial intelligence are propelling rapid progress in brain-computer interfaces (BCIs). These developments hold significant potential for decoding motion intentions from brain signals, enabling direct control commands without reliance on conventional neural pathways. Growing interest exists in decoding bimanual motor tasks, crucial for activities of daily living. This stems from the need to restore motor function, especially in individuals with deficits. This review aims to summarize neurological advancements in bimanual BCIs, encompassing neuroimaging techniques, experimental paradigms, and analysis algorithms. Thirty-six articles were reviewed, adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The literature search result revealed diverse experimental paradigms, protocols, and research directions, including enhancing the decoding accuracy, advancing versatile prosthesis robots, and enabling real-time applications. Notably, within BCI studies on bimanual movement coordination, a shared objective is to achieve naturalistic movement and practical applications with neurorehabilitation potential.
{"title":"A Systematic Review of Bimanual Motor Coordination in Brain-Computer Interface","authors":"Poraneepan Tantawanich;Chatrin Phunruangsakao;Shin-Ichi Izumi;Mitsuhiro Hayashibe","doi":"10.1109/TNSRE.2024.3522168","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3522168","url":null,"abstract":"Advancements in neuroscience and artificial intelligence are propelling rapid progress in brain-computer interfaces (BCIs). These developments hold significant potential for decoding motion intentions from brain signals, enabling direct control commands without reliance on conventional neural pathways. Growing interest exists in decoding bimanual motor tasks, crucial for activities of daily living. This stems from the need to restore motor function, especially in individuals with deficits. This review aims to summarize neurological advancements in bimanual BCIs, encompassing neuroimaging techniques, experimental paradigms, and analysis algorithms. Thirty-six articles were reviewed, adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The literature search result revealed diverse experimental paradigms, protocols, and research directions, including enhancing the decoding accuracy, advancing versatile prosthesis robots, and enabling real-time applications. Notably, within BCI studies on bimanual movement coordination, a shared objective is to achieve naturalistic movement and practical applications with neurorehabilitation potential.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"266-285"},"PeriodicalIF":4.8,"publicationDate":"2024-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10816155","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142938185","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-24DOI: 10.1109/TNSRE.2024.3522121
Minyeong Hong;Suh-Yeon Dong;Roger S. McIntyre;Soon-Kiat Chiang;Roger Ho
Adult attention deficit hyperactivity disorder (ADHD), a prevalent psychiatric disorder, significantly impacts social, academic, and occupational functioning. However, it has been relatively less prioritized compared to childhood ADHD. This study employed a functional near-infrared spectroscopy (fNIRS) during verbal fluency tasks in conjunction with machine learning (ML) techniques to differentiate between healthy controls (N =75) and ADHD individuals (N =120). Efficient feature selection in high-dimensional fNIRS datasets is crucial for improving accuracy. To address this, we propose a hybrid feature selection method that combines a wrapper-based and embedded approach, termed Bayesian-Tuned Ridge RFECV (BTR-RFECV). The proposed method facilitated streamlined feature selection and hyperparameter tuning in high-dimensional data, thereby reducing the number of features while enhancing accuracy. HbO features from the combined frontal and temporal regions were key, with the models achieving precision (89.89%), recall (89.74%), F-1 score (89.66%), accuracy (89.74%), MCC (78.36%), and GDR (88.45%). The outcomes of this study highlight the promising potential of combining fNIRS with ML as diagnostic tools in clinical settings, offering a pathway to significantly reduce manual intervention.
{"title":"fNIRS Classification of Adults With ADHD Enhanced by Feature Selection","authors":"Minyeong Hong;Suh-Yeon Dong;Roger S. McIntyre;Soon-Kiat Chiang;Roger Ho","doi":"10.1109/TNSRE.2024.3522121","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3522121","url":null,"abstract":"Adult attention deficit hyperactivity disorder (ADHD), a prevalent psychiatric disorder, significantly impacts social, academic, and occupational functioning. However, it has been relatively less prioritized compared to childhood ADHD. This study employed a functional near-infrared spectroscopy (fNIRS) during verbal fluency tasks in conjunction with machine learning (ML) techniques to differentiate between healthy controls (N =75) and ADHD individuals (N =120). Efficient feature selection in high-dimensional fNIRS datasets is crucial for improving accuracy. To address this, we propose a hybrid feature selection method that combines a wrapper-based and embedded approach, termed Bayesian-Tuned Ridge RFECV (BTR-RFECV). The proposed method facilitated streamlined feature selection and hyperparameter tuning in high-dimensional data, thereby reducing the number of features while enhancing accuracy. HbO features from the combined frontal and temporal regions were key, with the models achieving precision (89.89%), recall (89.74%), F-1 score (89.66%), accuracy (89.74%), MCC (78.36%), and GDR (88.45%). The outcomes of this study highlight the promising potential of combining fNIRS with ML as diagnostic tools in clinical settings, offering a pathway to significantly reduce manual intervention.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"220-231"},"PeriodicalIF":4.8,"publicationDate":"2024-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10813598","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142938192","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-23DOI: 10.1109/TNSRE.2024.3521229
Tanying Su;Xin Tan;Xinyu Jiang;Xiao Liu;Bo Hu;Chenyun Dai
Electromyography (EMG) based Human-Computer Interaction (HCI) through wearable devices frequently encounter variability in body postures, which can modify the amplitude and frequency features of surface EMG (sEMG) signals. This variability often results in reduced gesture recognition accuracy. To enhance the robustness of sEMG-based gesture interfaces, mitigating the effects of body position variability is essential. In this paper, we proposed a Dynamic Balanced Single-Source Domain Generalization (DBSS-DG) transfer learning framework, which only used sEMG signal data from one posture as source domain for model training but can also generate good performance under different body postures as target domain. Validation was performed on the sEMG dataset from 16 subjects across four postures: standing, sitting, walking, and lying. With standing as the source domain, the model achieved gesture recognition accuracies of 90.79 ± 0.09%, 88.78 ± 0.06%, and 90.87 ± 0.1% for sitting, walking, and lying as the target domains, respectively, producing an average improvement of 4.71% over non-transfer learning approaches. Furthermore, the performance of our model exceeded that of many well-known single-source domain generalization methods, establishing its effectiveness in practical applications.
{"title":"A Dynamic Balanced Single-Source Domain Generalization Model for Cross-Posture Myoelectric Control","authors":"Tanying Su;Xin Tan;Xinyu Jiang;Xiao Liu;Bo Hu;Chenyun Dai","doi":"10.1109/TNSRE.2024.3521229","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3521229","url":null,"abstract":"Electromyography (EMG) based Human-Computer Interaction (HCI) through wearable devices frequently encounter variability in body postures, which can modify the amplitude and frequency features of surface EMG (sEMG) signals. This variability often results in reduced gesture recognition accuracy. To enhance the robustness of sEMG-based gesture interfaces, mitigating the effects of body position variability is essential. In this paper, we proposed a Dynamic Balanced Single-Source Domain Generalization (DBSS-DG) transfer learning framework, which only used sEMG signal data from one posture as source domain for model training but can also generate good performance under different body postures as target domain. Validation was performed on the sEMG dataset from 16 subjects across four postures: standing, sitting, walking, and lying. With standing as the source domain, the model achieved gesture recognition accuracies of 90.79 ± 0.09%, 88.78 ± 0.06%, and 90.87 ± 0.1% for sitting, walking, and lying as the target domains, respectively, producing an average improvement of 4.71% over non-transfer learning approaches. Furthermore, the performance of our model exceeded that of many well-known single-source domain generalization methods, establishing its effectiveness in practical applications.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"255-265"},"PeriodicalIF":4.8,"publicationDate":"2024-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10811947","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142938195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-23DOI: 10.1109/TNSRE.2024.3521286
Hsin-Hui Hsu;Yea-Ru Yang;Li-Wei Chou;Yung-Cheng Huang;Ray-Yau Wang
Subacute low back pain (sLBP) is a critical transitional phase between acute and chronic stages and is key in determining the progression to chronic pain. While persistent pain has been linked to changes in brain activity, studies have focused mainly on acute and chronic phases, leaving neural changes during the subacute phase—especially during movement—under-researched. This cross-sectional study aimed to investigate changes in brain activity and the impact of pain intensity in individuals with sLBP during rest and reaching movements. Using a 28-electrode EEG, we measured motor-related brain waves, including theta, alpha, beta, and gamma oscillations. Transitioning from rest to movement phases resulted in significant reductions (>80%) in mean power across all frequency bands, indicating dynamic brain activation in response to movement. Furthermore, pain intensity was significantly correlated with brain wave activity. During rest, pain intensity was positively correlated with alpha oscillation activity in the central brain area (r = 0.40, p <0.05). In contrast, during movement, pain intensity was negatively correlated with changes in brain activity (r = −0.36 to −0.40, p <0.05). These findings suggest that pain influences brain activity differently during rest and movement, underscoring the impact of pain levels on neural networks related to the sensorimotor system in sLBP and highlighting the importance of understanding neural changes during this critical transitional phase.
亚急性腰痛（sLBP）是急性和慢性阶段之间的关键过渡阶段，是决定慢性疼痛进展的关键。虽然持续的疼痛与大脑活动的变化有关，但研究主要集中在急性期和慢性期，对亚急性期（尤其是运动期间）的神经变化研究不足。本横断面研究旨在探讨sLBP患者在休息和伸展运动时脑活动的变化和疼痛强度的影响。使用28个电极的脑电图，我们测量了与运动相关的脑电波，包括θ、α、β和γ振荡。从休息阶段过渡到运动阶段导致所有频段的平均功率显著降低（约80%），表明大脑对运动的动态激活。此外，疼痛强度与脑电波活动显著相关。休息时疼痛强度与中央区α振荡活动呈正相关（r = 0.40, p <0.05）。相反，在运动过程中，疼痛强度与大脑活动变化呈负相关（r = - 0.36 ~ - 0.40, p <0.05）。这些发现表明，疼痛在休息和运动时对大脑活动的影响不同，强调了疼痛水平对sLBP中与感觉运动系统相关的神经网络的影响，并强调了理解这一关键过渡阶段神经变化的重要性。
{"title":"The Brain Waves During Reaching Tasks in People With Subacute Low Back Pain: A Cross-Sectional Study","authors":"Hsin-Hui Hsu;Yea-Ru Yang;Li-Wei Chou;Yung-Cheng Huang;Ray-Yau Wang","doi":"10.1109/TNSRE.2024.3521286","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3521286","url":null,"abstract":"Subacute low back pain (sLBP) is a critical transitional phase between acute and chronic stages and is key in determining the progression to chronic pain. While persistent pain has been linked to changes in brain activity, studies have focused mainly on acute and chronic phases, leaving neural changes during the subacute phase—especially during movement—under-researched. This cross-sectional study aimed to investigate changes in brain activity and the impact of pain intensity in individuals with sLBP during rest and reaching movements. Using a 28-electrode EEG, we measured motor-related brain waves, including theta, alpha, beta, and gamma oscillations. Transitioning from rest to movement phases resulted in significant reductions (>80%) in mean power across all frequency bands, indicating dynamic brain activation in response to movement. Furthermore, pain intensity was significantly correlated with brain wave activity. During rest, pain intensity was positively correlated with alpha oscillation activity in the central brain area (r = 0.40, p <0.05). In contrast, during movement, pain intensity was negatively correlated with changes in brain activity (r = −0.36 to −0.40, p <0.05). These findings suggest that pain influences brain activity differently during rest and movement, underscoring the impact of pain levels on neural networks related to the sensorimotor system in sLBP and highlighting the importance of understanding neural changes during this critical transitional phase.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"183-190"},"PeriodicalIF":4.8,"publicationDate":"2024-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10811937","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142918353","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-23DOI: 10.1109/TNSRE.2024.3520984
Ya Jiang;Kendi Li;Yuankai Liang;Di Chen;Mingkui Tan;Yuanqing Li
Amyotrophic lateral sclerosis (ALS) is a chronic, progressive neurodegenerative disease that mainly causes damage to upper and lower motor neurons. This leads to a progressive deterioration in the voluntary mobility of the upper and lower extremities in ALS patients, which underscores the pressing need for an assistance system to facilitate communication and body movement without relying on neuromuscular function. In this paper, we developed a daily assistance system for ALS patients based on a wearable multimodal brain-computer interface (BCI) mouse. The system comprises two subsystems: a mouse system assisting the upper extremity and a wheelchair system based on the mouse system assisting the lower extremity. By wearing a BCI headband, ALS patients can control a computer cursor on the screen with slight head rotation and eye blinking, and further operate a computer and drive a wheelchair with specially designed graphical user interfaces (GUIs). We designed operating tasks that simulate daily needs and invited ALS patients to perform the tasks. In total, 15 patients with upper extremity limitations performed the mouse system task and 9 patients with lower extremity mobility issues performed the wheelchair system task. To our satisfaction, all the participants fully accomplished the tasks and average accuracies of 83.9% and 87.0% for the two tasks were achieved. Furthermore, workload evaluation using NASA Task Load Index (NASA-TLX) revealed that the participants experienced a low workload when using the system. The experimental results demonstrate that the proposed system provides ALS patients with effective daily assistance and shows promising long-term application prospects.
{"title":"Daily Assistance for Amyotrophic Lateral Sclerosis Patients Based on a Wearable Multimodal Brain-Computer Interface Mouse","authors":"Ya Jiang;Kendi Li;Yuankai Liang;Di Chen;Mingkui Tan;Yuanqing Li","doi":"10.1109/TNSRE.2024.3520984","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3520984","url":null,"abstract":"Amyotrophic lateral sclerosis (ALS) is a chronic, progressive neurodegenerative disease that mainly causes damage to upper and lower motor neurons. This leads to a progressive deterioration in the voluntary mobility of the upper and lower extremities in ALS patients, which underscores the pressing need for an assistance system to facilitate communication and body movement without relying on neuromuscular function. In this paper, we developed a daily assistance system for ALS patients based on a wearable multimodal brain-computer interface (BCI) mouse. The system comprises two subsystems: a mouse system assisting the upper extremity and a wheelchair system based on the mouse system assisting the lower extremity. By wearing a BCI headband, ALS patients can control a computer cursor on the screen with slight head rotation and eye blinking, and further operate a computer and drive a wheelchair with specially designed graphical user interfaces (GUIs). We designed operating tasks that simulate daily needs and invited ALS patients to perform the tasks. In total, 15 patients with upper extremity limitations performed the mouse system task and 9 patients with lower extremity mobility issues performed the wheelchair system task. To our satisfaction, all the participants fully accomplished the tasks and average accuracies of 83.9% and 87.0% for the two tasks were achieved. Furthermore, workload evaluation using NASA Task Load Index (NASA-TLX) revealed that the participants experienced a low workload when using the system. The experimental results demonstrate that the proposed system provides ALS patients with effective daily assistance and shows promising long-term application prospects.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"150-161"},"PeriodicalIF":4.8,"publicationDate":"2024-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10811971","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142905892","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-23DOI: 10.1109/TNSRE.2024.3521583
Tianzhe Bao;Zhiyuan Lu;Ping Zhou
Recently, robot-assisted rehabilitation has emerged as a promising solution to increase the training intensity of stroke patients while reducing workload on therapists, whilst surface electromyography (sEMG) is expected to serve as a viable control source. In this paper, we delve into the potential of deep learning (DL) for post-stroke hand gesture recognition by collecting the sEMG signals of eight chronic stroke subjects, focusing on three primary aspects: feature domains of sEMG (time, frequency, and wavelet), data structures (one or two-dimensional images), and neural network architectures (CNN, CNN-LSTM, and CNN-LSTM-Attention). A total of 18 DL models were comprehensively evaluated in both intra-subject testing and inter-subject transfer learning tasks, with two post-processing algorithms (Model Voting and Bayesian Fusion) analysed subsequently. Experiment results infer that for intra-subject testing, the average accuracy of CNN-LSTM using two-dimensional frequency features is the highest, reaching 72.95%. For inter-subject transfer learning, the average accuracy of CNN-LSTM-Attention using one-dimensional frequency features is the highest, reaching 68.38%. Through these two experiments, it was found that frequency features had significant advantages over other features in gesture recognition after stroke. Moreover, the post-processing algorithm can further improve the recognition accuracy, and the recognition effect can be increased by 2.03% through the model voting algorithm.
{"title":"Deep Learning-Based Post-Stroke Myoelectric Gesture Recognition: From Feature Construction to Network Design","authors":"Tianzhe Bao;Zhiyuan Lu;Ping Zhou","doi":"10.1109/TNSRE.2024.3521583","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3521583","url":null,"abstract":"Recently, robot-assisted rehabilitation has emerged as a promising solution to increase the training intensity of stroke patients while reducing workload on therapists, whilst surface electromyography (sEMG) is expected to serve as a viable control source. In this paper, we delve into the potential of deep learning (DL) for post-stroke hand gesture recognition by collecting the sEMG signals of eight chronic stroke subjects, focusing on three primary aspects: feature domains of sEMG (time, frequency, and wavelet), data structures (one or two-dimensional images), and neural network architectures (CNN, CNN-LSTM, and CNN-LSTM-Attention). A total of 18 DL models were comprehensively evaluated in both intra-subject testing and inter-subject transfer learning tasks, with two post-processing algorithms (Model Voting and Bayesian Fusion) analysed subsequently. Experiment results infer that for intra-subject testing, the average accuracy of CNN-LSTM using two-dimensional frequency features is the highest, reaching 72.95%. For inter-subject transfer learning, the average accuracy of CNN-LSTM-Attention using one-dimensional frequency features is the highest, reaching 68.38%. Through these two experiments, it was found that frequency features had significant advantages over other features in gesture recognition after stroke. Moreover, the post-processing algorithm can further improve the recognition accuracy, and the recognition effect can be increased by 2.03% through the model voting algorithm.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"191-200"},"PeriodicalIF":4.8,"publicationDate":"2024-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10812756","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142918342","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the past decade, significant focus has been on electromyography (EMG) control of prostheses in transtibial amputees (TTAs). Reliable signal acquisition requires accurate EMG electrode placement. Conventional electrode placement methods are challenging due to altered post-surgical anatomy. This study investigated the application of ultrasound imaging for placement of EMG electrodes in TTAs. Four residual limb muscles, Tibialis Anterior (TA), Peroneus Longus (PL), Gastrocnemius Medial (GM), and Gastrocnemius Lateral (GL), were examined in 9 unilateral TTAs. Ultrasound was used to identify each muscle belly’s thickest part and fiber orientation. A Certified Prosthetist Orthotist (CPO) then performed palpation to identify muscle bellies, blinded to ultrasound findings. Distances between ultrasound- and palpation-identified spots were measured. EMG data were contrasted between methods in terms of root mean square (RMS) amplitude and signal-to-noise ratio (SNR). The results indicated that Ultrasound-guided placement produced slightly higher, though non-significant, signal amplitudes (p =0.06) and significantly higher SNR (p =0.04). Moreover, palpation misidentified muscles in four cases. In 72.2% of cases, the distance between ultrasound- and palpation-identified spots was more than 10 mm. The mean distance was the greatest for PL and GL. Relying on palpation to identify PL and TA in TTAs may provide irrelevant EMG due to erroneous placement. Using ultrasound imaging can avoid this and, in addition to accurate muscle identification, may improve signal amplitude and SNR. In conclusion, ultrasound imaging is a valuable tool for enhancing the accuracy of EMG electrode placement in TTAs, which may lead to better prosthetic control outcomes.
{"title":"Improving Electromyography Electrode Placement Accuracy in Transtibial Amputees: A Comparative Study of Ultrasound and Palpation Methods","authors":"Faranak Rostamjoud;Friðrika Björk Þorkelsdóttir;Atli Örn Sverrisson;Sigurður Brynjólfsson;Kristín Briem","doi":"10.1109/TNSRE.2024.3520720","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3520720","url":null,"abstract":"In the past decade, significant focus has been on electromyography (EMG) control of prostheses in transtibial amputees (TTAs). Reliable signal acquisition requires accurate EMG electrode placement. Conventional electrode placement methods are challenging due to altered post-surgical anatomy. This study investigated the application of ultrasound imaging for placement of EMG electrodes in TTAs. Four residual limb muscles, Tibialis Anterior (TA), Peroneus Longus (PL), Gastrocnemius Medial (GM), and Gastrocnemius Lateral (GL), were examined in 9 unilateral TTAs. Ultrasound was used to identify each muscle belly’s thickest part and fiber orientation. A Certified Prosthetist Orthotist (CPO) then performed palpation to identify muscle bellies, blinded to ultrasound findings. Distances between ultrasound- and palpation-identified spots were measured. EMG data were contrasted between methods in terms of root mean square (RMS) amplitude and signal-to-noise ratio (SNR). The results indicated that Ultrasound-guided placement produced slightly higher, though non-significant, signal amplitudes (p =0.06) and significantly higher SNR (p =0.04). Moreover, palpation misidentified muscles in four cases. In 72.2% of cases, the distance between ultrasound- and palpation-identified spots was more than 10 mm. The mean distance was the greatest for PL and GL. Relying on palpation to identify PL and TA in TTAs may provide irrelevant EMG due to erroneous placement. Using ultrasound imaging can avoid this and, in addition to accurate muscle identification, may improve signal amplitude and SNR. In conclusion, ultrasound imaging is a valuable tool for enhancing the accuracy of EMG electrode placement in TTAs, which may lead to better prosthetic control outcomes.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"133-139"},"PeriodicalIF":4.8,"publicationDate":"2024-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10810499","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142912518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The loss of speech function following a laryngectomy usually leads to severe physiological and psychological distress for laryngectomees. In clinical practice, most laryngectomees retain intact upper tract articulatory organs, emphasizing the significance of speech rehabilitation that utilizes articulatory motion information to effectively restore speech. This study proposed a deep learning-based end-to-end method for speech reconstruction using ultrasound tongue images. Initially, ultrasound tongue images and speech data were collected simultaneously with a designed Mandarin corpus. Subsequently, a speech reconstruction model was built based on adversarial neural networks. The model includes a pretrained feature extractor to process ultrasound images, an upsampling block to generate speech, and discriminators to ensure the similarity and fidelity of the reconstructed speech. Finally, both objective and subjective evaluations were conducted for the reconstructed speech. The reconstructed speech demonstrated high intelligibility in both Mandarin phonemes and tones. The character error rate of phonemes in automatic speech recognition was 0.2605, and tone error rate obtained from dictation tests was 0.1784, respectively. Objective results showed high similarity between the reconstructed and ground truth speech. Subjective perception results also indicated an acceptable level of naturalness. The proposed method demonstrates its capability to reconstruct tonal Mandarin speech from ultrasound tongue images. However, future research should concentrate on specific conditions of laryngectomees, aiming to enhance and optimize model performance. This will be achieved by enlarging training datasets, investigating the impact of ultrasound tongue imaging parameters, and further refining this method.
{"title":"End-to-End Mandarin Speech Reconstruction Based on Ultrasound Tongue Images Using Deep Learning","authors":"Fengji Li;Fei Shen;Ding Ma;Jie Zhou;Shaochuan Zhang;Li Wang;Fan Fan;Tao Liu;Xiaohong Chen;Tomoki Toda;Haijun Niu","doi":"10.1109/TNSRE.2024.3520498","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3520498","url":null,"abstract":"The loss of speech function following a laryngectomy usually leads to severe physiological and psychological distress for laryngectomees. In clinical practice, most laryngectomees retain intact upper tract articulatory organs, emphasizing the significance of speech rehabilitation that utilizes articulatory motion information to effectively restore speech. This study proposed a deep learning-based end-to-end method for speech reconstruction using ultrasound tongue images. Initially, ultrasound tongue images and speech data were collected simultaneously with a designed Mandarin corpus. Subsequently, a speech reconstruction model was built based on adversarial neural networks. The model includes a pretrained feature extractor to process ultrasound images, an upsampling block to generate speech, and discriminators to ensure the similarity and fidelity of the reconstructed speech. Finally, both objective and subjective evaluations were conducted for the reconstructed speech. The reconstructed speech demonstrated high intelligibility in both Mandarin phonemes and tones. The character error rate of phonemes in automatic speech recognition was 0.2605, and tone error rate obtained from dictation tests was 0.1784, respectively. Objective results showed high similarity between the reconstructed and ground truth speech. Subjective perception results also indicated an acceptable level of naturalness. The proposed method demonstrates its capability to reconstruct tonal Mandarin speech from ultrasound tongue images. However, future research should concentrate on specific conditions of laryngectomees, aiming to enhance and optimize model performance. This will be achieved by enlarging training datasets, investigating the impact of ultrasound tongue imaging parameters, and further refining this method.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"140-149"},"PeriodicalIF":4.8,"publicationDate":"2024-12-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10810495","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142905886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pressure ulcers (PUs) pose a significant challenge in the care of bedridden patients, to which automated tilt nursing beds have emerged as a promising solution. However, the lack of effective models to elucidate the mechanical responses of deep tissue during assisted repositioning and identify the optimal tilt angle has hindered the implementation of effective automatic assisted repositioning systems for long-term care patients. Therefore, this study developed a novel computational model that integrates the buttocks with a support mattress to simulate automatic assisted repositioning, thereby analyzing deep tissue responses and optimizing tilt angles for effective load offloading. Inverse modeling was used to reconstruct the 3D shape of the buttocks, nodal equivalence techniques were employed to simplify the mesh and accurately represent internal tissue contacts, and soft tissue parameters were optimized using Response Surface Methodology (RSM). Finally, finite element (FE) analysis was conducted to evaluate the biomechanical responses and optimize the repositioning strategies. Model validation demonstrated a deformation error of $6.93~pm ~7.41$ mm (mean ± standard deviation) and interface pressure differences within 22.4%, demonstrating the efficacy and bio-fidelity of the system. Repositioning simulations at angles from 0° to 30° showed a 20% reduction in total soft tissue strain, with peak equivalent stress decreasing by 22.27% at the mattress- to-buttock interface and by 20.43% at the muscle-to-adipose tissue interface. These simulations suggest that a 30° turning angle is beneficial for alleviating pressure concentration, which may inspire the design and optimization of automatic assisted repositioning strategies in rehabilitation practices.
{"title":"Biomechanical Modeling and Evaluation of Buttocks Automatic Assisted Repositioning in Bedridden Patients","authors":"Liqing Yang;Ganjun Song;Di Luo;Haotian Xu;Jiamei Han;Mingzhao Xiao;Dingqun Bai;Shan Tian;Wensheng Hou;Jisi Tang;Xin Zhang;Lin Chen","doi":"10.1109/TNSRE.2024.3520146","DOIUrl":"https://doi.org/10.1109/TNSRE.2024.3520146","url":null,"abstract":"Pressure ulcers (PUs) pose a significant challenge in the care of bedridden patients, to which automated tilt nursing beds have emerged as a promising solution. However, the lack of effective models to elucidate the mechanical responses of deep tissue during assisted repositioning and identify the optimal tilt angle has hindered the implementation of effective automatic assisted repositioning systems for long-term care patients. Therefore, this study developed a novel computational model that integrates the buttocks with a support mattress to simulate automatic assisted repositioning, thereby analyzing deep tissue responses and optimizing tilt angles for effective load offloading. Inverse modeling was used to reconstruct the 3D shape of the buttocks, nodal equivalence techniques were employed to simplify the mesh and accurately represent internal tissue contacts, and soft tissue parameters were optimized using Response Surface Methodology (RSM). Finally, finite element (FE) analysis was conducted to evaluate the biomechanical responses and optimize the repositioning strategies. Model validation demonstrated a deformation error of \u0000<inline-formula> <tex-math>$6.93~pm ~7.41$ </tex-math></inline-formula>\u0000 mm (mean ± standard deviation) and interface pressure differences within 22.4%, demonstrating the efficacy and bio-fidelity of the system. Repositioning simulations at angles from 0° to 30° showed a 20% reduction in total soft tissue strain, with peak equivalent stress decreasing by 22.27% at the mattress- to-buttock interface and by 20.43% at the muscle-to-adipose tissue interface. These simulations suggest that a 30° turning angle is beneficial for alleviating pressure concentration, which may inspire the design and optimization of automatic assisted repositioning strategies in rehabilitation practices.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"175-182"},"PeriodicalIF":4.8,"publicationDate":"2024-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10807345","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142918341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号