IEEE Transactions on Neural Systems and Rehabilitation Engineering最新文献

{"title":"(Un)supervised (Co)adaptation via Incremental Learning for Myoelectric Control: Motivation, Review, and Future Directions","authors":"Evan Campbell;Fabio Egle;Marius OßWald;Ulysse Côté-Allard;Patrick M. Pilarski;Nicolò Boccardo;Roberto Meattini;Ivan Vujaklija;Levi Hargrove;Michele Canepa;Ethan Eddy;Alessandro Del Vecchio;Claudio Castellini;Erik Scheme","doi":"10.1109/TNSRE.2025.3602397","DOIUrl":"10.1109/TNSRE.2025.3602397","url":null,"abstract":"This paper presents a narrative review of incremental learning methods for myoelectric control, outlining both the historical trajectory and potential of adaptive prosthetic systems. Traditional myoelectric control has evolved from direct control techniques to advanced pattern recognition, yet persistent challenges such as signal non-stationarities and, consequently, the need for frequent recalibration remain. Incremental learning may enable a paradigm shift by continuously updating control models based on real-time, user-in-the-loop data, thereby addressing user-specific variations, environmental changes, and challenges from screen-guided-training based calibration. A central contribution of the paper is its taxonomy of incremental learning strategies, which divides the field into four categories: dedicated on-demand recalibration, unsupervised incremental learning, predictor-dependent incremental learning, and environment-dependent incremental learning. The methodology, strengths, and limitations of each category are discussed, providing a clear framework for evaluating current research and guiding future innovations. Further, this work establishes three settings for incremental learning: domain-incremental, task-incremental, and class-incremental continual learning. In addition, the paper highlights emerging trends such as transfer learning, domain adaptation, and self-supervised regression. It also emphasizes the potential of physiologically-inspired algorithms, novel end-effector designs to enhance prosthetic performance, and human-device co-adaptation. Finally, this paper discusses open challenges for incremental learning like attribution of signal changes to noise vs. behaviours, model complexity vs. data requirements, and user vs. model adaptation. Collectively, these insights pave the way for next-generation myoelectric systems that are more robust, intuitive, and adaptable to the dynamic needs and behaviours of users.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3565-3582"},"PeriodicalIF":5.2,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11137388","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144952630","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}

{"title":"Learning Frame-Level Classifiers for Video-Based Real-Time Assessment of Stroke Rehabilitation Exercises From Weakly Annotated Datasets","authors":"Ana Rita Cóias;Min Hun Lee;Alexandre Bernardino;Asim Smailagic;Mariana Mateus;David Fernandes;Sofia Trapola","doi":"10.1109/TNSRE.2025.3602548","DOIUrl":"https://doi.org/10.1109/TNSRE.2025.3602548","url":null,"abstract":"Autonomous rehabilitation support solutions, such as virtual coaches, should provide real-time feedback to improve motor function and maintain patient engagement. However, fully annotated dataset collection for real-time exercise assessment is time-consuming and costly, posing a barrier to evaluating proposed methods. In this work, we present a novel framework that learns a frame-level classifier using weakly annotated videos for real-time assessment of compensatory motions in stroke rehabilitation exercises by generating pseudo-labels at a frame level. We consider three approaches: 1) a baseline approach that uses a source dataset to train a frame-level classifier, 2) a transfer learning approach that uses target dataset video-level labels and parameters learned from a source dataset with frame-level labels, and 3) a semi-supervised approach that leverages a target dataset video-level labels and a small set of frame-level labels. We intend to generalize to a weakly labeled target dataset with new exercises and patients. To validate the approach, we use two datasets annotated on compensatory motions: TULE, an existing video and frame-level labeled dataset of 15 post-stroke patients and three exercises, and SERE, a new dataset of 20 post-stroke patients and five exercises, created by the authors, with video-level labels and a small amount of frame-level labels. We show that a frame-level classifier trained on TULE does not generalize well on SERE (<inline-formula> <tex-math>${f}_{{1}} = {72}.{87}%$ </tex-math></inline-formula>), but our semi-supervised and transfer learning approaches achieve, respectively, <inline-formula> <tex-math>${f}_{{1}} = {78}.{93}%$ </tex-math></inline-formula> and <inline-formula> <tex-math>${f}_{{1}} = {80}.{47}%$ </tex-math></inline-formula>. Generating pseudo-labels leads to better frame-level classification results for the target dataset than training a classifier with the source dataset (baseline). Thus, the proposed approach can simplify the customization of virtual coaches to new patients and exercises with low data annotation efforts.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3334-3345"},"PeriodicalIF":5.2,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11141498","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144914272","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}

{"title":"Design and Preliminary Evaluation of a Gait Control Strategy for Hip-Knee-Ankle-Foot Prostheses With Motorized Hip Joint","authors":"Farshad Golshan;Edward Lemaire;Hossein Gholizadeh;David Nielen;Natalie Baddour","doi":"10.1109/TNSRE.2025.3602715","DOIUrl":"10.1109/TNSRE.2025.3602715","url":null,"abstract":"Hip disarticulation (HD) amputees face mobility challenges due to the loss of hip, knee, and ankle joints. Current hip-knee-ankle-foot (HKAF) prostheses are entirely passive and require excessive compensatory movements to operate, leading to fatigue and long-term complications. Seeking to address these limitations, this study developed a HD user-centric, walking speed adaptable control strategy paired with a hip-motorized HKAF to emulate gait characteristics of transfemoral amputees. A prototype “Power Hip” was instrumented with internal sensors (IMUs, load cells, joint encoders) to create a prosthetic unit that could be worn without the need for external sensors. A hierarchical gait control strategy was developed to utilize these sensors to calculate the desired hip states and actuate the joint. To evaluate capabilities of the control strategy, an HD amputee participant was recruited to undergo training with Power Hip. Once training was complete, motion captured kinematics and onboard sensor data were analyzed across slow, self-paced, and fast walking speed trials. The Power Hip enabled walking speeds of 0.69–1.01 m/s, with stride parameters aligning with transfemoral amputee outcome measures. Hip extension velocities (60.2–104.9°/s) matched transfemoral kinematics, though swing-phase knee flexion magnitude and velocity were reduced compared to transfemoral benchmarks. The prototype demonstrated a 52° hip range of motion, surpassing conventional passive hip joints, and adapted to speed changes automatically. This research paves the way for advanced prosthetic solutions to improve quality of life for people with hip-level amputations.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3432-3442"},"PeriodicalIF":5.2,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11141500","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144952305","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}

{"title":"Low-Intensity Transcranial Ultrasound Stimulation Modulates the Excitability of Motor Cortical Neural Activity by Stimulating the Cerebellum","authors":"Huifang Yang;Zhe Zhao;Renhao Xu;Jiamin Pei;Jiaqing Yan;Xiangjian Zhang;Hanna Lu;Yi Yuan","doi":"10.1109/TNSRE.2025.3601110","DOIUrl":"https://doi.org/10.1109/TNSRE.2025.3601110","url":null,"abstract":"Functional connectivity between the cerebellum and the motor cortex is critical for motor function. Low-intensity transcranial ultrasound stimulation (TUS) has been proved to directly activate cerebellar neural activity. However, whether and how TUS of the cerebellum modulates neural activity in the motor cortex is unclear. In the study, TUS with different parameters was used to stimulate the mouse cerebellar region while local field potentials (LFP) in the cerebellum and the motor cortex were simultaneously recorded, as well as neuronal calcium ion activity. We discovered that: 1) TUS can increase the amplitude and power intensity of LFP in the cerebellum and motor cortex, which linearly increasing with the increase of ultrasound intensity, and nonlinearly increasing with the increase of the duty cycle; 2) TUS in the cerebellum significantly increased the firing time and peak of neurons of the motor cortex, which also increased linearly with the increase of ultrasound intensity and nonlinearly with the increase of duty cycle; 3) The cross-correlation information of the LFP between the cerebellum and motor cortex was enhanced, which depends on ultrasound intensity and duty cycle; 4) TUS in the cerebellum for 7 days improved the motor ability of ischemic stroke mice. Taken together, the above findings indicate that TUS of the cerebellum can modulate the excitability of neural activity in the motor cortex, which is important in modulating the symptoms of motor cortex-related diseases. We also describe the ameliorative effect of TUS of the cerebellum on movement disorders caused by cerebral ischemia.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3314-3322"},"PeriodicalIF":5.2,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11134687","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144914236","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}

{"title":"Artifact Removal in Electrocorticography Devices With Cardiac Contamination","authors":"K. A. Liyanage;P. E. Yoo;D. B. Grayden;N. L. Opie;T. J. Oxley","doi":"10.1109/TNSRE.2025.3601445","DOIUrl":"https://doi.org/10.1109/TNSRE.2025.3601445","url":null,"abstract":"Electrocorticography (ECoG) devices with electronics housed near the chest are susceptible to artifacts of a differing nature to electroencephalography (EEG) and standard ECoG. Using data obtained via an endovascular neural interface, we compared different artifact removal techniques in an offline setting with the aim of improving the quality and usefulness of clinically acquired data. Three different methods of filtration were applied and assessed: Common Average Referencing (CAR), Independent Component Analysis (ICA) with automated ECG channel selection, and Template-Based Removal (TBR). The automated ECG channel selection method was compared to manual selection. Methods were compared using signal-to-artifact root-mean-squared (RMS) values. The automated ECG source channel selection had high concordance with manual selection. All filtration methods decreased post-artifact RMS amplitudes and improved signal-to-artifact ratios. ICA took the most time to compute but had the most improved signal-to-artifact ratio. In regions with no ECG artifact, TBR preserved the underlying electrocorticography data better than the other methods. ICA with an automated method of ECG channel selection is the preferred method out of the three tested to remove ECG artifact while preserving the underlying signal. We establish methods that can be used to improve neural data of electrocorticography devices susceptible to cardiac contamination to facilitate translation as brain-computer interfaces.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3400-3408"},"PeriodicalIF":5.2,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11133705","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144929185","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}

{"title":"Intelligent Adolescent Idiopathic Scoliosis Detection Method Based on Three-Dimensional Measurements","authors":"Jiahao Li;Chengao Gao;Bohan Liu;Haoyu Wang;Qiong Wang;Shibo Li;Ying Hu;Jianwei Zhang;Xiaozhi Qi;Yu Zhao","doi":"10.1109/TNSRE.2025.3601121","DOIUrl":"https://doi.org/10.1109/TNSRE.2025.3601121","url":null,"abstract":"Adolescent Idiopathic Scoliosis (AIS) is a prevalent latent disorder among adolescents. Early detection and screening are crucial for preventing the progression of deformities. Current assessment methods for spinal curvature rely on X-ray imaging to measure the Cobb angle and detect scoliosis. However, radiation exposure limits its use for widespread screening. Existing tools are time-consuming and rely on physician expertise. This study introduces a novel method based on depth information and high-order curvature calculations to assess asymmetries in the human back, aiming for a harmless and accurate rapid evaluation of AIS. The method utilizes asymmetries in three-dimensional curvatures derived from back point clouds to assess the degree of deformity caused by scoliosis. It achieves this by quantifying the asymmetry index between left and right curvatures, then using these indices to predict the Cobb angle and reconstruct the spinal midline from the spinous processes. Experimental results on patients with varying degrees of Cobb angles show that the spinal midline on X-ray images is highly aligned with the reconstructed spinal midline, and the asymmetry index is significantly positively correlated with the Cobb angle, with a Pearson correlation coefficient of 0.892. Additionally, a screening system built upon this method achieved an accuracy rate of over 0.95 in school screenings involving more than 2,000 subjects, underscoring its potential for large-scale clinical applications.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3259-3270"},"PeriodicalIF":5.2,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11133482","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904632","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}

{"title":"The Efficiency of Manual Editing of High-Density Surface Electromyogram Decomposition Depends on the Recorded Muscle and Contraction Level but Less on the Operator’s Experience","authors":"Nina Murks;Jakob Škarabot;Matej Kramberger;Matjaž Divjak;Gašper Sedej;Tamara Valenčič;Christopher D. Connelly;Haydn Thomason;Aleš Holobar","doi":"10.1109/TNSRE.2025.3600755","DOIUrl":"https://doi.org/10.1109/TNSRE.2025.3600755","url":null,"abstract":"We investigated the agreement and accuracy of manual editing of the high-density electromyogram (hdEMG) decomposition results by seven human operators with various experience levels. All operators edited the same automatically decomposed experimental hdEMG from the first dorsal interosseous (FDI), tibialis anterior (TA), vastus lateralis (VL), and biceps brachii (BB) muscles, and synthetic hdEMG from soleus (SO) and BB muscles at 10%, 30%, 50% and 70% of maximum voluntary contraction. On average, operators kept <inline-formula> <tex-math>$13.7~pm ~7.4$ </tex-math></inline-formula> motor units (MUs) after editing and demonstrated relatively large disagreement in the calculated MU pulse trains (normalized root mean square difference) but relatively high agreement in the identified MU discharges. Inter-operator agreement positively correlated with the initial MU Pulse-to-Noise Ratio used as a quality measure of automatic MU identification, and negatively correlated with the muscle contraction level. Operators agreed more on the results of the simulated than experimental hdEMG. Among the experimental muscles tested, the greatest agreement was demonstrated for VL and the lowest for BB. We obtained similar results when comparing editing to the results of the most experienced operator and to ground truth in simulated cases: the greatest precision and sensitivity were demonstrated for VL, and the lowest for BB. The level of the operator’s experience had a significant impact on the editing of synthetic hdEMG and the detection of the first MU discharge, but not on the rate of agreement or editing time of experimental hdEMG.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3367-3376"},"PeriodicalIF":5.2,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11131216","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144918234","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}

{"title":"Uncovering Low-Dimensional Manifolds of Neural Dynamics for Motor-Imagery Based Stroke Rehabilitation: An EEG-Based Brain–Computer Interface Study","authors":"Tao Liu;Ziwei Wang;Sadia Shakil;Raymond Kai-Yu Tong","doi":"10.1109/TNSRE.2025.3600824","DOIUrl":"https://doi.org/10.1109/TNSRE.2025.3600824","url":null,"abstract":"Stroke rehabilitation aims to repair neural circuits and dynamics through the remapping of neuronal functions. However, there is currently a gap in understanding the alteration of neural population dynamics-the fundamental computational unit driving functions-under clinical settings. In this study, we introduced a novel method to identify stable low-dimensional structures of neural population dynamics in stroke patients during motor tasks. Using whole-brain EEG recordings from chronic stroke patients performing motor imagery (MI) tasks before and after brain-computer interface (BCI) training, as well as a public EEG dataset of acute stroke patients performing MI tasks, we projected EEG signals from sensor space to voxel space via source localization (eLORETA), simulating neural population activity in regions of interest. By applying dimensionality reduction, we successfully obtained low-dimensional neural manifolds to represent neural population dynamics. Our analysis revealed three key findings: (1) For right-handed patients, task-related low-dimensional dynamics in the related brain regions remain stable across subjects, with their features holding potential as biomarkers for stroke rehabilitation; (2) BCI training promotes global and sustained restoration of neural population dynamics; (3) EEG theta-band oscillations show strong correlation with these dynamics, highlighting their macroscopic nature. This study proposes a new, simple, and powerful tool for comprehension and validation of stroke rehabilitation mechanisms confirming the effectiveness of BCI training in restoring neural dynamics.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3281-3292"},"PeriodicalIF":5.2,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11131289","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904631","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}

{"title":"Optimized Coil Design for Enhanced Electric Field Induction in Peripheral Nerve Stimulation","authors":"Jaeu Park;Kyeong Jae Lee;Pritish Nagwade;Jinwoong Jeong;Jeong Hoan Park;Hongsoo Choi;Sohee Kim;Sanghoon Lee","doi":"10.1109/TNSRE.2025.3599634","DOIUrl":"10.1109/TNSRE.2025.3599634","url":null,"abstract":"Peripheral nerve electrical stimulation is widely used for the treatment of neuropathic pain and neural regeneration. However, it often induces adverse biological reactions and unintended activation of surrounding neural tissues. As an alternative, peripheral nerve magnetic stimulation offers a promising, less invasive approach that enables targeted nerve stimulation without direct tissue contact. Despite its potential, it is constrained by the bulkiness of coils and excessive heat generation due to the high currents required. To address these limitations, we conducted a study on coil design optimized for peripheral nerve modulation. Our approach, supported by simulations and animal experiments, focused on optimizing coil geometry to maximize the induced electric field gradient. Among various designs, a four-leaf rhombus-shaped coil demonstrated the highest gradient at the center of the interface. In rat sciatic nerve experiments, this coil, driven by a rectangular pulse with a <inline-formula> <tex-math>$200~mu $ </tex-math></inline-formula>s rise time and 25 V amplitude, successfully elicited compound muscle action potentials in both the tibial anterior and gastrocnemius muscles. This study presents design guidelines for peripheral nerve stimulation (PNS) coils based on magnetic stimulation as an alternative to conventional electrical stimulation. The proposed approach may serve as a foundation for the development of advanced, miniaturized, and energy-efficient neural stimulation coils.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3225-3236"},"PeriodicalIF":5.2,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11127116","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144872965","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}

{"title":"A Real-Time Artifact Removal System for Closed-Loop Deep-Brain Stimulation","authors":"Chenghao Xing;Xi Cheng;Hao Feng;Bin Wu;Yingnan Nie;Chunguang Chu;Xin Zhang;Qiyu Niu;Jia Xiu;Bowen Geng;Liang Chen;Shouyan Wang","doi":"10.1109/TNSRE.2025.3597916","DOIUrl":"10.1109/TNSRE.2025.3597916","url":null,"abstract":"This paper presents a novel real-time signal processing method for removing local field potential (LFP) artifacts during deep-brain stimulation (DBS). Real-time artifact removal is essential for closed-loop DBS systems, as they rely on real-time, artifact-free LFPs to provide stimulation feedback. Building on previous stimulation-sampling synchronization methods, this work introduces a dynamic template subtraction method that achieves precise and efficient real-time removal of stimulation artifacts. By leveraging stimulation-sampling synchronization, the method enables real-time template alignment and artifact removal through subtraction. It can operate even at low sampling rates, requiring a minimum of twice the stimulation frequency. The artifact templates are dynamically updated to adapt to changes in stimulation artifacts, ensuring robust and accurate performance over time. The method was evaluated through simulations and in vitro and in vivo experiments. Simulation tests validated its theoretical feasibility, while it successfully removed stimulation artifacts in vitro, relative errors in the power spectral density between the recovered and reference LFPs in the examined frequency band (1—150 Hz) were 0.64%, 0.31%, 0.58%, and 0.73% under stimulation at 20, 60, 90, and 130 Hz, respectively. In vivo, the method successfully recorded artifact-free LFPs in real time and supported beta-triggered closed-loop DBS. In an additional in vivo evaluation using a commercial medical device, the method recorded artifact-free LFPs with a sampling rate of 260 Hz (twice the stimulation frequency of 130 Hz). The proposed artifact removal method provides important technical support for realizing lightweight closed-loop DBS systems.","PeriodicalId":13419,"journal":{"name":"IEEE Transactions on Neural Systems and Rehabilitation Engineering","volume":"33 ","pages":"3237-3245"},"PeriodicalIF":5.2,"publicationDate":"2025-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11126183","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144859079","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}