Journal of neural engineering最新文献

{"title":"Assessing age-related proprioceptive changes through active and passive tasks: implications for stroke assessment.","authors":"Erick Carranza, Sreten Franovic, Amy Boos, Elvira Pirondini","doi":"10.1088/1741-2552/adc6bc","DOIUrl":"https://doi.org/10.1088/1741-2552/adc6bc","url":null,"abstract":"<p><strong>Objective: </strong>Voluntary control of motor actions requires precise regulation of proprioceptive and somatosensory functions. While aging is known to impair sensory processing, its effect on proprioception remains unclear. Previous studies report conflicting findings on whether passive proprioception (i.e., during externally driven movements) declines with age, and research on age-related changes in active proprioception (i.e., during voluntary movements) remains limited, particularly in the upper limb. Understanding these changes is critical for identifying and preventing impairments that may affect movement performance and mobility, particularly in neurological conditions such as stroke or Parkinson's disease.</p><p><strong>Approach: </strong>We refined a robotic protocol to assess upper-limb active proprioception and validated its robustness and reliability over multiple sessions. Using this protocol, we compared the performance between young and elderly neurologically healthy adults during both active and passive proprioceptive tasks.</p><p><strong>Main results: </strong>Elderly participants exhibited a significant decline in accuracy when sensing limb position in both active and passive proprioceptive tasks, whereas their precision remained unchanged. These findings indicate that aging primarily affects proprioceptive accuracy rather than variability in position sense.</p><p><strong>Significance: </strong>Our findings contribute to the ongoing debate on age-related proprioceptive decline and highlight the importance of distinguishing between active and passive proprioception. Furthermore, our validated robotic protocol provides a reliable tool for assessing proprioception, with potential applications in studying neurological conditions in clinical settings.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143744663","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":"Vestibular implant stimulation: pulse amplitude modulation versus combined pulse rate and amplitude modulation.","authors":"Stan C J van Boxel, Bernd L Vermorken, Benjamin Volpe, Nils Guinand, Angélica Perez-Fornos, Elke M J Devocht, Raymond van de Berg","doi":"10.1088/1741-2552/adc33a","DOIUrl":"10.1088/1741-2552/adc33a","url":null,"abstract":"<p><p><i>Objective</i>. The vestibular implant is a potential treatment approach for bilateral vestibulopathy patients. To restore gaze stabilization, the implant should elicit vestibulo-ocular reflexes (VORs) over a wide range of eye velocities. Different stimulation strategies to achieve this goal were previously described. Vestibular information can be encoded by modulating stimulation amplitude, rate, or a combination of both. In this study, combined rate and amplitude modulation was compared with amplitude modulation, to evaluate their potential for vestibular implant stimulation.<i>Approach</i>. Nine subjects with a vestibulo-cochlear implant participated in this study. Three stimulation strategies were tested. The combined rate and amplitude modulation setting (baseline rate 50%) was compared with amplitude modulation (baseline rate 50%, and baseline rate equal to the maximum rate). The resulting VOR was evaluated.<i>Main results</i>. Combining rate and amplitude modulation, or using amplitude modulation with a baseline equal to the maximum rate, both significantly increased peak eye velocities (PEVs). Misalignment increased with higher PEVs and higher pulse rate. No significant differences were found in PEVs and misalignment, between both stimulation strategies. Amplitude modulation with a baseline rate at 50%, demonstrated the lowest PEVs.<i>Significance</i>. Combining rate and amplitude modulation, or amplitude modulation with a baseline equal to the maximum rate, can both be considered for future vestibular implant fitting.ClinicalTrials.gov Identifier: NCT04918745.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143672144","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":"Fascicle-selective kilohertz-frequency neural conduction block with longitudinal intrafascicular electrodes.","authors":"Louis Regnacq, Anil K Thota, Arianna Ortega Sanabria, Laura McPherson, Sylvie Renaud, Olivier Romain, Yannick Bornat, James J Abbas, Ranu Jung, Florian Kölbl","doi":"10.1088/1741-2552/adc62a","DOIUrl":"https://doi.org/10.1088/1741-2552/adc62a","url":null,"abstract":"<p><strong>Objective: </strong>Electrical stimulation of peripheral nerves is used to treat a variety of disorders and conditions. While conventional biphasic pulse stimulation typically induces neural activity in fibres, kilohertz (kHz) continuous stimulation can block neural conduction, offering a promising alternative to drug-based therapies for alleviating abnormal neural activity. This study explores strategies to enhance the selectivity and control of high-frequency neural conduction block using intrafascicular electrodes.</p><p><strong>Methods: </strong>In vivo experiments were conducted in a rodent model to assess the effects of kilohertz stimulation delivered via longitudinal intrafascicular electrodes on motor axons within the tibial and common peroneal fascicles of the sciatic nerve.</p><p><strong>Main results: </strong>We demonstrated that a progressive and selective block of neural conduction is achievable with longitudinal intrafascicular electrodes. We showed that the amount of neural conduction block can be tuned by adjusting the amplitude and frequency of kilohertz stimulation. Additionally, we achieved interfascicular selectivity with intrafascicular electrodes, with this selectivity being modulated by the kilohertz stimulation frequency. We also observed a small amount of onset response spillover, which could be minimized by increasing the blocking stimulus frequency. Muscle fatigue was quantified during kHz continuous stimulation and compared to control scenarios, revealing that the muscle was able to recover from fatigue during the block, confirming a true block of motor neurons.</p><p><strong>Significance: </strong>Our findings show that kilohertz stimulation using longitudinal intrafascicular electrodes can be precisely controlled to achieve selective conduction block. By leveraging existing knowledge from conventional stimulation techniques, this approach allows for the development of stimulation protocols that effectively block abnormal neural patterns with reduced side effects.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143733821","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":"The NERVE-ML (neural engineering reproducibility and validity essentials for machine learning) checklist: ensuring machine learning advances neural engineering<sup />.","authors":"David E Carlson, Ricardo Chavarriaga, Yiling Liu, Fabien Lotte, Bao-Liang Lu","doi":"10.1088/1741-2552/adbfbd","DOIUrl":"10.1088/1741-2552/adbfbd","url":null,"abstract":"<p><p><i>Objective.</i>Machine learning's (MLs) ability to capture intricate patterns makes it vital in neural engineering research. With its increasing use, ensuring the validity and reproducibility of ML methods is critical. Unfortunately, this has not always been the case in practice, as there have been recent retractions across various scientific fields due to the misuse of ML methods and validation procedures. To address these concerns, we propose the first version of the neural engineering reproducibility and validity essentials for ML (NERVE-ML) checklist, a framework designed to promote the transparent, reproducible, and valid application of ML in neural engineering.<i>Approach.</i>We highlight some of the unique challenges of model validation in neural engineering, including the difficulties from limited subject numbers, repeated or non-independent samples, and high subject heterogeneity. Through detailed case studies, we demonstrate how different validation approaches can lead to divergent scientific conclusions, highlighting the importance of selecting appropriate procedures guided by the NERVE-ML checklist. Effectively addressing these challenges and properly scoping scientific conclusions will ensure that ML contributes to, rather than hinders, progress in neural engineering.<i>Main results.</i>Our case studies demonstrate that improper validation approaches can result in flawed studies or overclaimed scientific conclusions, complicating the scientific discourse. The NERVE-ML checklist effectively addresses these concerns by providing guidelines to ensure that ML approaches in neural engineering are reproducible and lead to valid scientific conclusions.<i>Significance.</i>By effectively addressing these challenges and properly scoping scientific conclusions guided by the NERVE-ML checklist, we aim to help pave the way for a future where ML reliably enhances the quality and impact of neural engineering research.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11948487/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143618027","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":"Classification of schizophrenia based on RAnet-ET: Resnet based attention network for eye-tracking.","authors":"Ruochen Dang, Ying Wang, Feiyu Zhu, Xiaoyi Wang, Jingping Zhao, Ping Shao, Bing Lang, Yuqi Wang, Zhibin Pan, BingLiang Hu, Renrong Wu, Quan Wang","doi":"10.1088/1741-2552/adc5a5","DOIUrl":"https://doi.org/10.1088/1741-2552/adc5a5","url":null,"abstract":"<p><strong>Objective: </strong>There is a notable need of quantifiable and objective methods for the classification of schizophrenia. Patients with schizophrenia exhibit atypical eye movements compared with healthy individuals. To address this need, we have developed a classification model based on eye-tracking data to assist physicians in the intelligent auxiliary diagnosis of schizophrenia.</p><p><strong>Approach: </strong>This study employed three eye-tracking experiments-Picture-Free Viewing, Smooth Pursuit Tracking, and Fixation Stability-to collect eye-tracking data from patients with schizophrenia and healthy controls. The eye-tracking data of 292 participants (133 healthy controls and 159 patients with schizophrenia) were recorded. Utilizing eye-tracking data in picture-free viewing, we introduce a Resnet-based Attention Network for Eye-Tracking (RAnet-ET) integrated with the attention mechanism. RAnet-ET was trained by employing multiple loss functions to classify patients with schizophrenia and healthy controls. Furthermore, we proposed a classifier for handling multimodal features that combines specific features extracted from the well-trained RAnet-ET, 100 eye-tracking variables extracted from three eye-tracking experiments, and 19 MATRICS Consensus Cognitive Battery scores.</p><p><strong>Main results: </strong>The RAnet-ET achieved good performance in classifying schizophrenia, yielding an accuracy of 89.04%, a specificity of 90.56%, and an F1 score of 87.87%. The classification results based on multimodal features demonstrated improved performance, achieving 96.37% accuracy, 96.87% sensitivity, 95.87% specificity, and 96.37% F1 score.</p><p><strong>Significance: </strong>By integrating attention mechanisms, we designed RAnet-ET, which achieved good performance in classifying schizophrenia from free-viewing eye-tracking data. The synergistic combination of specific features extracted from the well-trained RAnet-ET, MCCB scores, and eye-tracking variables achieved exceptional classification performance, distinguishing individuals with schizophrenia from healthy controls. This study underscores the potential of our approach as a pivotal asset for the diagnosis of schizophrenia.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143733820","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":"LGFormer: Integrating local and global representations for EEG decoding.","authors":"Wenjie Yang, Xingfu Wang, Wenxia Qi, Wei Wang","doi":"10.1088/1741-2552/adc5a3","DOIUrl":"https://doi.org/10.1088/1741-2552/adc5a3","url":null,"abstract":"<p><strong>Objective: </strong>Electroencephalography (EEG) decoding is challenging because of its temporal variability and low signal-to-noise ratio, which complicate the extraction of meaningful information from signals. Although convolutional neural networks (CNNs) effectively extract local features from EEG signals, they are constrained by restricted receptive fields. In contrast, transformers excel at capturing global dependencies through self-attention mechanisms but often require extensive training data and computational resources, which limits their efficiency on EEG datasets with limited samples.</p><p><strong>Approach: </strong>In this paper, we propose LGFormer, a hybrid network designed to efficiently learn both local and global representations for EEG decoding. LGFormer employs a deep attention module to extract global information from EEG signals, dynamically adjusting the focus of CNNs. Subsequently, LGFormer incorporates a local-enhanced transformer, combining the strengths of CNNs and transformers to achieve multiscale perception from local to global. Despite integrating multiple advanced techniques, LGFormer maintains a lightweight design and training efficiency.</p><p><strong>Main results: </strong>LGFormer achieves state-of-the-art performance within 200 training epochs across four public datasets, including motor imagery, cognitive workload, and error-related negativity decoding tasks. Additionally, we propose a novel spatial and temporal attention visualization method, revealing that LGFormer captures discriminative spatial and temporal features, enhancing model interpretability and providing insights into its decision-making process.</p><p><strong>Significance: </strong>In summary, LGFormer demonstrates superior performance while maintaining high training efficiency across different tasks, highlighting its potential as a versatile and practical model for EEG decoding.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143733822","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 wearable brain-computer interface to play an endless runner game by self-paced motor imagery.","authors":"Pasquale Arpaia, Antonio Esposito, Enza Galasso, Fortuna Galdieri, Angela Natalizio","doi":"10.1088/1741-2552/adc205","DOIUrl":"10.1088/1741-2552/adc205","url":null,"abstract":"<p><p><i>Objective.</i>A wearable brain-computer interface is proposed and validated experimentally in relation to the real-time control of an endless runner game by self-paced motor imagery(MI).<i>Approach.</i>Electroencephalographic signals were recorded via eight wet electrodes. The processing pipeline involved a filter-bank common spatial pattern approach and the combination of three binary classifiers exploiting linear discriminant analysis. This enabled the discrimination between imagining left-hand, right-hand, and no movement. Each mental task corresponded to an avatar horizontal motion within the game. Twenty-three healthy subjects participated to the experiments and their data are made publicly available. A custom metric was proposed to assess avatar control performance during the gaming phase. The game consisted of two levels, and after each, participants completed a questionnaire to self-assess their engagement and gaming experience.<i>Main results.</i>The mean classification accuracies resulted 73%, 73%, and 67% for left-rest, right-rest, and left-right discrimination, respectively. In the gaming phase, subjects with higher accuracies for left-rest and right-rest pair exhibited higher performance in terms of the custom metric. Correlation of the offline and real-time performance was investigated. The left-right MI did not correlate to the gaming phase performance due to the poor mean accuracy of the calibration. Finally, the engagement questionnaires revealed that level 1 and level 2 were not perceived as frustrating, despite the increasing difficulty.<i>Significance.</i>The work contributes to the development of wearable and self-paced interfaces for real-time control. These enhance user experience by guaranteeing a more natural interaction with respect to synchronous neural interfaces. Moving beyond benchmark datasets, the work paves the way to future applications on mobile devices for everyday use.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143659527","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":"NeuroNella: Automatic identification of neural activity from multielectrode arrays with blind source separation.","authors":"Carina Marconi Germer, Dario Farina, Stuart N Baker, Alessandro Del Vecchio","doi":"10.1088/1741-2552/adc5a4","DOIUrl":"https://doi.org/10.1088/1741-2552/adc5a4","url":null,"abstract":"<p><strong>Objective: </strong>The identification of individual neuronal activity from multielectrode arrays poses significant challenges, including handling data from numerous electrodes, resolving overlapping action potentials and tracking activity across long recordings. This study introduces NeuroNella, an automated algorithm developed to address these challenges.</p><p><strong>Approach: </strong>NeuroNella employs blind source separation to leverage the sparsity of action potentials in multichannel recordings. It was validated using three datasets, including two publicly available ones: (1) in vitro recordings (252 channels) of retinal ganglion cells from mice with simultaneous ground-truth loose patch data to assess accuracy; (2) a Neuropixel recording from an awake mouse, comprising 374 channels spanning different brain areas, to demonstrate scalability with dense multielectrode configurations in in vivo recordings; and (3) data (32 channels) recorded from the medullary reticular formation in a terminally anaesthetised macaque, to showcase decomposition over long periods of time.</p><p><strong>Main results: </strong>The algorithm exhibited an error rate of less than 1% compared to ground-truth data. It reliably identified individual neurons, detected neuronal activity across a wide amplitude range, and tolerated minor probe shifts, maintaining robustness in prolonged experimental sessions.</p><p><strong>Significance: </strong>NeuroNella provides an automated and efficient method for neuronal activity identification. Its adaptability to diverse dataset, species, and recording configurations underscores its potential to advance studies of neuronal dynamics and facilitate real-time neuronal decoding systems.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143733823","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":"How are opposite neurofeedback tasks represented at cortical and corticospinal tract levels?","authors":"Ioana Susnoschi Luca, Aleksandra Vuckovic","doi":"10.1088/1741-2552/adbcdb","DOIUrl":"10.1088/1741-2552/adbcdb","url":null,"abstract":"<p><p><i>Objective.</i>The study objective was to characterise indices of learning and patterns of connectivity in two neurofeedback (NF) paradigms that modulate mu oscillations in opposite directions, and the relationship with change in excitability of the corticospinal tract (CST).<i>Approach.</i>Forty-three healthy volunteers participated in 3 NF sessions for upregulation (<i>N</i> = 24) or downregulation (<i>N</i> = 19) of individual alpha (IA) power at central location Cz. Brain signatures from multichannel electroencephalogram (EEG) were analysed, including oscillatory (power, spindles), non-oscillatory components (Hurst exponent), and effective connectivity directed transfer function (DTF) of participants who were successful at enhancing or suppressing IA power at Cz. CST excitability was studied through leg motor-evoked potential, tested before and after the last NF session. We assessed whether participants modulated widespread alpha or central mu rhythm through the use of current source density derivation (CSD), and related the change in activity in mu and upper half of mu band, to CST excitability change.<i>Main results.</i>In the last session, IA/mu power suppression was achieved by 79% of participants, while 63% enhanced IA. CSD-EEG revealed that mu power was upregulated through an increase in the incidence rate of bursts of alpha band activity, while downregulation involved changes in oscillation amplitude and temporal patterns. Neuromodulation also influenced frequencies adjacent to the targeted band, indicating the use of common mental strategies within groups. DTF analysis showed, for both groups, significant connectivity between structures commonly associated with motor imagery tasks, known to modulate the excitability of the motor cortex, although most connections did not remain significant after correcting for multiple comparisons. CST excitability modulation was related to the absolute amplitude of upper mu modulation, rather than the modulation direction.<i>Significance.</i>The upregulation and downregulation of IA/mu power during NF, with respect to baseline were achieved via distinct mechanisms involving oscillatory and non-oscillatory EEG features. Mu enhancement and suppression post-NF and during the last NF block with respect to the baseline, respectively corresponded to opposite trends in motor-evoked potential changes post-NF. The ability of NF to modulate CST excitability could be a valuable rehabilitation tool for central nervous system disorders (stroke, spinal cord injury), where increased excitability and neural plasticity are desired. This work may inform future neuromodulation protocols, and may improve NF training effectiveness by rewarding certain EEG signatures.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143569208","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":"Select for better learning: identifying high-quality training data for a multimodal cyclic transformer.","authors":"Jingwei Zhang, Zhaoyi Liu, Christos Chatzichristos, Sam Michiels, Wim Van Paesschen, Danny Hughes, Maarten De Vos","doi":"10.1088/1741-2552/adbec0","DOIUrl":"10.1088/1741-2552/adbec0","url":null,"abstract":"<p><p><i>Objective</i>. Tonic-clonic seizures (TCSs), which present a significant risk for sudden unexpected death in epilepsy, require accurate detection to enable effective long-term monitoring. Previous studies have demonstrated the advantages of multimodal seizure detection systems in reliably detecting TCSs over extended periods. However, the effectiveness of these data-driven systems depends heavily on the availability of reliable training data.<i>Approach</i>. To address this need, we propose an innovative data selection method designed to identify high-quality training samples. Our approach evaluates sample quality based on learning difficulty, classifying samples with lower learning difficulty as higher quality. We then introduce a confidence-based method to quantify the proportion of high-quality samples within the dataset.<i>Main results</i>. Experimental results show that our method improves the performance of a state-of-the-art TCS detection model by 11%.<i>Significance</i>. Using this data selection method, we develop a training pipeline that enhances the training process of multimodal seizure detection models.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143598636","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}