Journal of neural engineering最新文献

{"title":"Time-frequency embedding with contrastive pre-training allows sub-second seizure detection.","authors":"Helena A Merker, Isabella Dalla Betta, Matthew A Wilson, Francisco J Flores, Emery N Brown","doi":"10.1088/1741-2552/ae98b6","DOIUrl":"10.1088/1741-2552/ae98b6","url":null,"abstract":"<p><p><i>Objective.</i>Rapid and accurate detection of electrographic seizures is critical for both clinical diagnosis and neuroscience research. Although seizure identification is commonly performed in the time domain, analysis in the time-frequency domain provides a more comprehensive representation of seizure characteristics.<i>Approach.</i>In this study, we present a 3D convolutional neural network (CNN) that incorporates a trainable continuous wavelet transform (CWT) layer, enabling adaptive time-frequency feature learning directly from raw EEG. To address common data challenges, we augment the 3D CNN for pre-training with contrastive learning, comparing contrastive predictive coding against bidirectional contrastive learning.<i>Main results.</i>On single-channel and multi-channel data, the standard 3D CNN outperformed both a 2D CNN with pre-computed CWT and a 1D CNN that processes raw signals, achieving95% accuracy down to 0.5-second segments. Contrastive pre-training improved the performance of the 3D CNN in both low-data and class-imbalance scenarios. Further experiments involving band-pass filtering and temporal shuffling revealed that classification is driven primarily by low-frequency patterns and statistical features rather than temporal dependencies. The proposed framework also maintained90% accuracy with moderate noise and downsampling applied to inputs, as well as when cross-subject generalization was evaluated using held-out subjects.<i>Significance.</i>We show that a 3D CNN with a trainable CWT layer and contrastive pre-training enables accurate sub-second seizure detection and effectively mitigates data limitations common in clinical settings. This work demonstrates that time-frequency embedding within CNNs, augmented by self-supervised pre-training, offers a promising path toward architectures for sub-second seizure detection in the presence of practical limitations of real-world scenarios.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148724062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A fused minimum spanning tree framework reveals large-scale cortical network reorganization during dual-task standing in older adults.","authors":"Yi-Ching Chen, Wei-Min Huang, Ing-Shiou Hwang","doi":"10.1088/1741-2552/ae93f7","DOIUrl":"https://doi.org/10.1088/1741-2552/ae93f7","url":null,"abstract":"<p><p><i>Objective.</i>Dual-task standing challenges postural regulation in older adults, but the associated cortical network reorganizations remain incompletely characterized. This study investigated how a visuomotor dual-task modulates postural dynamics and cortical network during unstable standing.<i>Approach.</i>Twenty-six community-dwelling older adults performed foam-surface standing under single-task and visuomotor dual-task conditions. Postural dynamics were quantified using center-of-pressure (COP) measures and stabilogram diffusion analysis (SDA). Cortical network organization was assessed using electroencephalography (EEG) functional connectivity (phase-lag index) and minimum spanning tree (MST) topology. To improve characterization of distributed cortical adaptation, a novel fused MST framework integrating theta, alpha, and beta-band connectivity was evaluated together with conventional band-specific MST.<i>Main results.</i>Dual-task standing significantly increased COP displacement and velocity in both anterior-posterior and medial-lateral directions, accompanied by altered SDA parameters across directional domains, indicating substantial changes in postural control dynamics. EEG analyses revealed task-related cortical network reorganization, with increased theta and alpha band overlap ratios and significant topology changes in the fused MST representation. Compared with band-specific MST representations, the fused MST framework provided an integrated characterization of task-related cortical network adaptation across frequency bands.<i>Significance.</i>Visuomotor dual-task standing induces substantial reorganization of both postural dynamics and cortical network topology in older adults under unstable conditions. These findings suggest that the fused MST framework may provide a useful quantitative approach for assessing cortical adaptation to dual-task balance challenges in older adults.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":"23 5","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148868595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dataset distillation for efficient seizure prediction: tiny-scale data, comparable performance.","authors":"Yuchang Zhao, Xun Chen, Chang Li, Yikai Gao, Ruobing Qian, Aiping Liu","doi":"10.1088/1741-2552/ae9b70","DOIUrl":"10.1088/1741-2552/ae9b70","url":null,"abstract":"<p><p><i>Objective.</i>Deep learning has shown significant potential in electroencephalogram (EEG)-based seizure prediction. However, translating these advances into practical applications faces critical challenges: the prohibitive time costs of reviewing accumulated historical data during frequent model updates, the need for rapid adaptation to diverse model architectures across hardware platforms, and the substantial burden of massive data storage. To this end, we introduce the first dataset distillation study tailored for seizure prediction, which condenses the original dataset into a compact, information-dense synthetic dataset. This tiny-scale proxy allows models to match the performance of those trained on the full-scale dataset.<i>Approach.</i>Methodologically, we formulate the distillation as a min-max optimization task, introducing an EEG-aware alignment objective that jointly preserves latent feature distributions, temporal dynamics in the time-frequency domain, and inter-channel correlation structures. Through signal-compatible amplitude and phase decomposition, we optimize the frequency parameter to maximize the CF discrepancy via a sampling network, while simultaneously tuning the synthetic data to minimize this gap. Additionally, we integrate hard labels to refine soft labels to enhance the generalization of synthetic data.<i>Main results.</i>Extensive validation across 37 public and 35 clinical subjects demonstrate the method's robustness across diverse scenarios and model architectures. With merely 1% of the original data volume, our approach retains roughly 95% of the original performance while reducing training duration by approximately 98%.<i>Significance.</i>This method addresses clinical translation bottlenecks by distilling massive EEG data into compact synthetic sets, offering a promising solution for efficient model maintenance, flexible model switching, and optimized storage.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148803688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MEG resting state alpha and beta band functional connectivity in male children with autism and sensory processing dysfunction.","authors":"Carly Demopoulos, Xuan Jesson, Molly Rae Gerdes, Barbora G Jurigova, Leighton B Hinkley, Kamalini G Ranasinghe, Shivani Desai, Anne Findlay, Srikantan S Nagarajan, Elysa J Marco","doi":"10.1088/1741-2552/ae93fb","DOIUrl":"10.1088/1741-2552/ae93fb","url":null,"abstract":"<p><p><i>Objective.</i>Sensory processing dysfunction (SPD) not only affects most individuals with autism spectrum disorder (ASD), but at least 5% of children without ASD also experience SPD. Our understanding of the relationship between sensory dysfunction and resting state brain activity is still emerging. The objective of this study was to examine group differences and behavioral associations with resting state alpha and beta oscillatory activity in ASD, SPD, and typically developing control (TDC) groups.<i>Approach.</i>This study compared long-range resting state functional connectivity of neural oscillatory behavior in 60 male children aged 8-12 years with (ASD;<i>N</i>= 18), those with (SPD;<i>N</i>= 18) who did not meet ASD criteria, and typically developing control participants (TDC;<i>N</i>= 24) using magnetoencephalography. Functional connectivity analyses were performed in the alpha and beta frequency bands, which are known to be implicated in sensory information processing. Group differences in functional connectivity and associations between sensory abilities and functional connectivity were examined.<i>Main results.</i>Distinct patterns of functional connectivity differences between ASD and SPD groups were found only in the beta band, but not in the alpha band. In both alpha and beta bands, ASD and SPD cohorts differed from the TDC cohort. Distinct patterns of associations between imaginary coherence and performance-based measures of sensory processing and verbal abilities were identified across groups.<i>Significance.</i>These findings demonstrate distinct long-range alpha and beta band phase-lagged neural synchrony alterations in SPD and ASD that are associated with sensory processing abilities in male children. These measures could serve as potential candidate neurophysiological markers for ASD and SPD at the group level, and may provide mechanistic insights relevant to biomarker development.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13522810/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148671680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Distinct tDCS montages act via dissociable lateralization mechanisms to enhance motor function in chronic stroke.","authors":"Yaqian Li, Ying Zhang, Shupei Jiao, Jiewei Lu, Chunfang Wang, Ningbo Yu","doi":"10.1088/1741-2552/ae9aca","DOIUrl":"10.1088/1741-2552/ae9aca","url":null,"abstract":"<p><p><i>Objective</i>. The imbalance in interhemispheric functional connectivity following stroke fundamentally impedes motor recovery. Although transcranial direct current stimulation (tDCS) effectively modulates neuroplasticity, the acute effects of distinct stimulation montages on lateralized brain networks and how these network shifts drive behavioral improvements warrant further investigation.<i>Approach</i>. We employed a single-blind, randomized crossover design involving 28 patients with chronic subcortical stroke. Participants received anodal, cathodal, bilateral, or sham tDCS targeting the primary motor cortex. Resting-state electroencephalography data were acquired immediately before and after each intervention. To quantify network lateralization, we computed dynamic functional connectivity (dFC) using mutual information and constructed an asymmetry index matrix. Key connectivity features were isolated via robust feature selection algorithms, classifying stimulation states and evaluating correlations with acute motor gains on the Jebsen-Taylor Hand Function Test (JTT).<i>Main results</i>. Network dynamics features successfully discriminated pre- versus post-intervention states with high accuracy (AUC: 0.87-0.98). All three active tDCS protocols exhibited montage-specific directional modulation. Crucially, anodal tDCS uniquely reversed contralesional network dominance, facilitating ipsilesional connectivity and shifting the interhemispheric balance toward the affected hemisphere. Conversely, cathodal and bilateral tDCS predominantly induced contralesional network inhibition, whereas sham stimulation exhibited no directional bias. Furthermore, the functional inhibition of the contralesional network (specifically at the P2-P6 connection) induced by bilateral tDCS significantly correlated with acute JTT improvements (<i>r</i>= 0.55,<i>p</i>= 0.04).<i>Significance</i>. This study demonstrates that distinct tDCS montages modulate stroke-induced network imbalances through dissociable lateralization mechanisms. While anodal tDCS drives ipsilesional facilitation to reverse contralesional dominance, bilateral and cathodal montages operate primarily via contralesional inhibition. These findings provide critical mechanistic insights into interhemispheric network dynamics.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148803788","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring muscular and postural correlates of the EEG-derived functional connectome during a quiet-standing stabilometric task.","authors":"Marianna Chianese, Simone Papallo, Giuseppe Prisco, Fabrizio Esposito, Leandro Donisi","doi":"10.1088/1741-2552/ae98b8","DOIUrl":"10.1088/1741-2552/ae98b8","url":null,"abstract":"<p><p><i>Objective.</i>Static postural control is a dynamic feedback process in which the integration of visual and somatosensory inputs within the central nervous system determines the activation of lower kinetic chain muscles, resulting in postural oscillations, to maintain balance during quiet standing. This study explored the association between cortical and muscular activity in relation to postural responses.<i>Approach.</i>Electroencephalography (EEG) is commonly used to assess neural oscillations. Network graph theory (NGT) applied to EEG-derived connectivity outlines the functional connectome. Surface electromyography (sEMG), combined with stabilogram analysis, enables the characterization of muscle activation patterns during postural control. By simultaneously acquiring EEG, sEMG and stabilometric signals, EEG-derived NGT metrics, sEMG-derived time- and frequency-domain features and sway parameters were correlated in twenty-one healthy volunteers (11 males, 29.52 ± 3.70 years) performing static postural control tasks for 30 s, standing upright on a stable surface, with eyes-closed (EC) and eyes-open (EO). Partial Spearman's rank correlation analyses were used to assess multimodal associations between NGT metrics and electromyographic and biomechanical measures.<i>Main Results.</i>During EC sessions, changes in two postural sway measures (jerk, path length) were negatively correlated to changes in the resilience of the EEG connectome, as reconstructed from broadband EEG signals (<i>P</i>< 0.05, corrected for multiple comparisons). During both EC and EO sessions, changes in muscle activation, fatigue-related and postural sway measures were weakly correlated to changes in multiple topological parameters of the EEG connectome, as reconstructed from either broadband (EC) or delta (EO), theta (EO), alpha (EO), beta (EO) and gamma (EC) EEG signals.<i>Significance.</i>This study highlights possible frequency-dependent associations between the global topological organization of the EEG functional connectome and muscular and postural responses during balance maintenance. Albeit preliminary, these exploratory findings may provide a novel perspective for the multimodal investigation of cortico-muscular interactions subserving postural control in health and disease.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148726023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An fMRI-neurofeedback system to train the reward neurocircuitry in ultra-high field MRI: a feasibility study.","authors":"Amir Hossein Dakhili, Ethan Murphy, Saampras Ganesan, Andrew Zalesky, Rebecca Glarin, Hannah Thomson, Anastasia Paloubis, Tudor Sava, Bradford A Moffat, Valentina Lorenzetti, Govinda Poudel, Chao Suo","doi":"10.1088/1741-2552/ae93fa","DOIUrl":"10.1088/1741-2552/ae93fa","url":null,"abstract":"<p><p><i>Objective.</i>This feasibility study provides a technical validation of a functional magnetic resonance imaging-based (fMRI)-neurofeedback system for personalised neuromodulation of reward neurocircuitry in a larger study of cannabis use disorder. We adapted an fMRI-neurofeedback system in 7 Tesla scanner to neuromodulate the top 33% voxels of interest (VOIs) within the anterior cingulate cortex (ACC). We investigated characteristics of the neurofeedback signal and its specificity to VOIs compared to control regions (cerebrospinal-fluid [CSF], precentral gyrus, superior frontal gyrus, and middle temporal gyrus). We also examined whether neurofeedback signal specificity to VOIs changed across 2 neuromodulation runs and 2 conditions, in which participants were instructed to increase (upregulation) or decrease (downregulation) VOIs activity.<i>Approach.</i>The fMRI-neurofeedback system included: MRI scanner; a back-end with Turbo-BrainVoyager software for real-time fMRI data processing; a front-end computer running a custom-developed MATLAB script to calculate/display brain changes during neuromodulation as a thermometer-like bar. The neurofeedback signal was generated by regressing out the percent-signal-change of CSF from that of VOIs and applying weighted sliding window smoothing. To evaluate whether the neurofeedback signal was specific to VOIs, we performed regression analysis using offline preprocessed fMRI data as input and the neurofeedback signal as regressor.<i>Main results.</i>The neurofeedback signal was more specific to VOIs than CSF during both conditions (<i>p</i>< 0.001), and than precentral gyrus during downregulation (<i>p</i>= 0.030). The specificity of the neurofeedback signal to VOIs showed a significant overall increase from Run1 to Run2 (<i>p</i>= 0.025).<i>Significance.</i>The fMRI-neurofeedback system, combined personalised VOIs selection with the higher spatial and temporal resolution afforded by 7 T MRI, which enables more precise spatial localisation, monitoring of brain changes over time, and targeting of functionally defined VOIs of the ACC than is typically achievable at lower field strengths. This system can be adapted to target other regions involved in cognitive processes in normative and clinical populations.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148671693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrated decoding of local and prospective spatial representations for future decision prediction.","authors":"Xueling Wang, Xuan Luo, Yimeng Wang, Xinyi Deng, Chenguang Zheng","doi":"10.1088/1741-2552/ae93f5","DOIUrl":"10.1088/1741-2552/ae93f5","url":null,"abstract":"<p><p><i>Objective.</i>Decoding future spatial decisions from the brain's cognitive map represents a critical step toward cognitive brain-computer interfaces (BCIs) and closed-loop neuromodulation for neurological disorders. However, accurately predicting future paths from hippocampal neural dynamics remains challenging, partly because previous studies have largely overlooked the temporal evolution of early decision-making phases.<i>Approach.</i>We recorded hippocampal CA1 ensemble activity from rats performing a continuous spatial decision task in a modified T-maze. By segmenting the decision process into starting, running, and approaching phases, we examined how local and prospective spatial representations dynamically evolved over time.<i>Main results.</i>Central arm place cells exhibited increasing trajectory dependence as animals approached the choice point, with local theta sequences consistently overrepresenting the actual choice. Concurrently, prospective representations driven by choice arm place cells showed a dynamic transition from preferentially predicting the actual choice during the running phase to representing potential paths more equally near the choice point. Integrating local and prospective features improved decoding performance, reaching 74.4% accuracy for future choice prediction in Test trials and 78.2% accuracy for upcoming trajectory decoding in Sample trials.<i>Significance.</i>These findings demonstrate that incorporating spatiotemporal hippocampal features improves behavioral decoding and provides a framework for developing hippocampus-based BCIs to predict future decisions.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148671648","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Endovascular neuromodulation: feasibility of endovascular stimulation near the cerebellum.","authors":"Weijie Qi, Joel Villalobos, Jingyang Liu, Huakun Xin, Martin J Spencer, Sepehr Kazemi, Stella Ho, Joseph West, David B Grayden, Sam E John","doi":"10.1088/1741-2552/ae998c","DOIUrl":"10.1088/1741-2552/ae998c","url":null,"abstract":"<p><p><i>Objective.</i>Electrical stimulation and neural recording underpin neural prostheses for restoring function and treating neurological disorders, but clinical adoption is limited by the invasiveness of implantation. The Endovascular Neural Interface offers an alternative by accessing intracranial targets through the cerebral vasculature. This work presents the first strength-duration characterization of cortical evoked potentials, elicited by endovascular stimulation, adjacent to the cerebellum.<i>Approach.</i>A polymer-based stent-electrode array was deployed into the left transverse sinus of an ovine model. Biphasic current pulses targeting the cerebellum were delivered via the stent electrodes. Cortical responses were recorded using a subdural electrocorticography grid.<i>Main results.</i>Endovascular stimulation consistently evoked time-locked cortical potentials with early and late components at approximately 40 ms and 100 ms post-stimulation. Electrode functionality and stability were confirmed through impedance monitoring throughout the experiments. Strength-duration analysis revealed rheobase and chronaxie values, providing a quantitative basis for parameter selection and comparison with established intracranial stimulation modalities.<i>Significance.</i>These results demonstrate that endovascular electrodes may access non-superficial brain structures and evoke reproducible cortical responses without open neurosurgery. This work helps establish a foundational framework for endovascular neuromodulation and supports further investigation of its potential for future closed-loop and network-level neuromodulation research.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148763071","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Knowledge distillation for sEMG-based gesture recognition: enhancing wearable HMI systems with lightweight models.","authors":"Fang Qiu, Chenyun Dai, Xiaodong Liu, Xinming Ye","doi":"10.1088/1741-2552/ae93f4","DOIUrl":"10.1088/1741-2552/ae93f4","url":null,"abstract":"<p><p><i>Objective</i>. Gesture recognition using surface electromyography (sEMG) is a promising technology for wearable human-machine interaction systems. However, sEMG-based gesture classification models often suffer from performance degradation in cross-user scenarios due to the variability in individual physiological signals. To address this issue, large-scale deep learning algorithms with a great amount of parameters are employed to enhance model generalization and improve recognition performance across different users. However, the use of deep learning models on wearable devices is challenging due to their limited computational power and memory capacity.<i>Methods</i>. This study addresses this challenge through knowledge distillation, which compresses large teacher models into lightweight student models while maintaining performance. We trained deep teacher models, including DenseNet, InceptionV3, and VggNet, and distilled their parameters into compact convolutional neural network and long short-term memory (LSTM) models. The experiment was performed on the sEMG data of 30 subjects collected from a wrist-band electrode with 32 channels.<i>Main Results</i>. The results demonstrated that student models, particularly LSTM-based ones, achieved classification accuracy close to or even higher than their teacher models, with a highest accuracy improvement. Among the teacher-student combinations, the DenseNet121-LSTM architecture achieved the highest classification accuracy. The relationship between computational complexity floating-point operations (FLOPs) and model performance was also analyzed, showing that the distilled models can effectively approximate high-FLOP models.<i>Significance</i>. This research enhances the feasibility of embedding deep learning models into wearable systems (e.g. wristband or wristwatch), facilitating more responsive and efficient gesture recognition in daily-life applications. The Python implementation of the complete knowledge distillation framework is publicly available at:.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2026-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148671686","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}