Journal of neural engineering最新文献

{"title":"Data augmentation using masked principal component representation for deep learning-based SSVEP-BCIs.","authors":"Wenlong Ding, Aiping Liu, Longlong Cheng, Xun Chen","doi":"10.1088/1741-2552/add9d1","DOIUrl":"10.1088/1741-2552/add9d1","url":null,"abstract":"<p><p><i>Objective.</i>Data augmentation has been demonstrated to improve the classification accuracy of deep learning models in steady-state visual evoked potential-based brain-computer interfaces (BCIs), particularly when dealing with limited electroencephalography (EEG) data. However, current data augmentation methods often rely on signal-level manipulations, which may lead to significant distortion of EEG signals. To overcome this limitation, this study proposes a component-level data augmentation method called masked principal component representation (MPCR).<i>Approach.</i>MPCR utilizes a principal component-based reconstruction approach, integrating a random masking strategy applied to principal component representations. Specifically, certain principal components are randomly selected and set to zero, thereby introducing random perturbations in the reconstructed samples. Furthermore, reconstructing samples via linear combinations of the remaining components effectively preserves the primary inherent structure of EEG signals. By expanding the input space covered by training samples, MPCR helps the trained model learn more robust features. To validate the efficacy of MPCR, experiments are performed on two widely utilized public datasets.<i>Main results.</i>Experimental results indicate that MPCR substantially enhances classification accuracy across diverse deep learning models. Additionally, in comparison to various state-of-the-art data augmentation approaches, MPCR demonstrates both greater performance and high compatibility.<i>Significance.</i>This study proposes a simple yet effective component-level data augmentation method, contributing valuable insights for advancing data augmentation methods in EEG-based BCIs.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144087127","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":"Cortical stability and chaos during focal seizures: insights from inference-based modeling.","authors":"Yun Zhao, David B Grayden, Mario Boley, Yueyang Liu, Philippa J Karoly, Mark J Cook, Levin Kuhlmann","doi":"10.1088/1741-2552/add83f","DOIUrl":"10.1088/1741-2552/add83f","url":null,"abstract":"<p><p><i>Objective.</i>Epilepsy affects millions globally, with a significant subset of patients suffering from drug-resistant focal seizures. Understanding the underlying neurodynamics of seizure initiation and propagation is crucial for advancing treatment and diagnostics. In this study, we present a novel, inference-based approach for analyzing the temporal evolution of cortical stability and chaos during focal epileptic seizures.<i>Approach.</i>Utilizing a multi-region neural mass model, we estimate time-varying synaptic connectivity from intracranial electroencephalography (iEEG) data collected from individuals with drug-resistant focal epilepsy.<i>Main results.</i>Our analysis reveals distinct preictal and ictal phases characterized by shifts in cortical stability, heightened chaos in the ictal phase, and highlight the critical role of inter-regional communication in driving chaotic cortical behaviour. We demonstrate that cortical dynamics are consistently destabilized prior to seizure onset, with a transient reduction in instability at seizure onset, followed by a significant increase throughout the seizure.<i>Significance.</i>This work provides new insights into the mechanisms of seizure generation and offers potential biomarkers for predicting seizure events. Our findings pave the way for innovative therapeutic strategies targeting cortical stability and chaos to manage epilepsy.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144012298","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":"Closed-loop electrical block of vagus nerve scales from rodent to porcine cardiac models.","authors":"Shane Bender, David Green, Joseph Hadaya, Sahil Haridas, Christopher Chan, Ronald Challita, Al-Hassan Dajani, Jeffery Ardell, Tina Vrabec","doi":"10.1088/1741-2552/add8be","DOIUrl":"10.1088/1741-2552/add8be","url":null,"abstract":"<p><p><i>Objective</i>. Direct current (DC) electrical block of the vagus nerve has shown the ability to downregulate the parasympathetic input to the heart. Previous investigations used static prescribed values, but the main advantage of electrical nerve block is the ability to modulate the block effect in real time. Here we investigate the potential of real-time, closed loop control of heart rate (HR), and how these control schemes translate across species.<i>Approach.</i>In anesthetized rats and pigs, proximal vagus stimulation was applied as a perturbation to simulate overactive vagal activity, causing a decrease in HR. DC nerve block was applied distally to mitigate this perturbation and raise HR. The block amplitudes applied were normalized to a block threshold (BT), or the amount of current to block the nerve completely in 60 s. Two static levels of 10% and 50% BT were compared to a closed-loop controlled current.<i>Main Results.</i>In both the rat and the pig models, the closed-loop nerve block was able to control the HR to the desired setpoint (SP). Neither of the static values were able to achieve a reliably consistent level of block, with the controlled trials showing a much tighter spread of HR over time. In the pigs, a higher-gain controller was able to reach the SP more quickly. In the rat, the controller reduced both the injected charge and the time to recovery after block. In the pig, the charge was increased, but near-instant recovery times were retained. A closed-loop system is required for precision control of cardiac output.<i>Significance.</i>Both the rat and pig models showed success in closed-loop control of HR. Translating from rat to pig models only required minor changes to the controller, indicating that the system is robust. The ease of this translation effort bodes well for potential future translation to human therapies.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12108926/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144083009","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":"Role of frequency-dependent and capacitive tissue properties in spinal cord stimulation models.","authors":"Niranjan Khadka, Boshuo Wang, Marom Bikson","doi":"10.1088/1741-2552/add76e","DOIUrl":"10.1088/1741-2552/add76e","url":null,"abstract":"<p><p><i>Objective.</i>Spinal cord stimulation (SCS) models simulate the electric fields (<i>E</i>-fields) generated in targeted tissues, which in turn govern physiological and then behavioral outcomes. Notwithstanding increasing sophistication and adoption in therapy optimization, SCS models typically calculate<i>E</i>-fields using quasi-static approximation (QSA). QSA, as implemented in neuromodulation models, neglects the frequency-dependent tissue conductivity (dispersion), as well as propagation, capacitive, and inductive effects on the<i>E</i>-field. The objective of this study is to calculate the impact of frequency-dependent tissue conductivity and permittivity in SCS models, across a broad frequency range.<i>Approach.</i>We solved a high-resolution RADO-SCS finite element model to simulate<i>E</i>-field magnitudes in spinal column tissues under voltage-controlled (VC) and current-controlled (CC) SCS. Varied combinations of epidural space and dura conductivity based on prior SCS modeling studies (under the QSA-method), as well as values from the Gabriel (1996<i>Compilation of the Dielectric Properties of Body Tissues at RF and Microwave Frequencies</i>) dataset for 1 Hz, 1 kHz, 2.5 kHz, 16.66 kHz, and 1 MHz were considered. We assessed the relative contribution of epidural space and dura permittivity on peak<i>E</i>-field magnitude and neural activation, and compared results to the QSA-method models.<i>Main results.</i>Across published SCS models, the conductivities of epidural space (considered either fat or mixed tissues; 0.025-0.25 S m^-1) and dura (0.02-0.6 S m^-1) vary by over an order of magnitude, associated with differences in predicted spinal cord peak<i>E</i>-field magnitudes for VC-SCS (6.55-43.71 V m^-1per V) and CC-SCS (10.94-25.20 V m^-1per mA). These literature variations in conductivity and resulting peak<i>E</i>-field magnitude are greater than from epidural/dura tissue dispersion (1 kHz-1 MHz) based on Gabriel (1996<i>Compilation of the Dielectric Properties of Body Tissues at RF and Microwave Frequencies</i>) database (VC-SCS: 7.26-8.09 V m^-1per V; CC-SCS: 21.14-21.25 V m^-1per mA). Changes in<i>E</i>-field magnitudes were not associated with significant changes in relative spatial profiles of the<i>E</i>-field or activating function. The impact of epidural space/dural permittivity (at 1 kHz) on<i>E</i>-field magnitudes and activating function was minimal (⩽1%) for both SCS modes.<i>Significance.</i>The impact of dispersion/permittivity is significantly less than existing variations in tissue conductivities used across SCS modeling studies. As relative<i>E</i>-field or activating function profiles were not significantly changed by tissue conductivities, any impact of neuronal activation thresholds tracks changes in<i>E</i>-field magnitude. We limited our analysis to a single geometry and epidural/dural properties to isolate the impact of QSA.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144012304","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":"Memory optimized random forest classifier for EEG seizure detection in implantable monitoring and closed-loop neurostimulation devices.","authors":"Sotirios Kalousios, Jonathan Larochelle, Nicolas Zabler, Nico Locker, Peter Woias, Laura Maria Comella, Andreas Schulze-Bonhage, Matthias Dümpelmann","doi":"10.1088/1741-2552/add76f","DOIUrl":"10.1088/1741-2552/add76f","url":null,"abstract":"<p><p><i>Objective</i>. Up to one third of epilepsy patients do not achieve satisfactory seizure control and may benefit from implantable devices for responsive neurostimulation or online seizure monitoring. Beyond energy efficiency, the limited memory capacity in these devices, imposes significant constraints to algorithmic design of seizure detection models. This study aims to evaluate the performance of cross-patient random forest (RF) models optimized for low-power microcontroller applications by assessing various channel integration strategies and measuring their memory requirements.<i>Approach</i>. Fifty patients undergoing electroencephalographic monitoring with 362 seizures were included in the analysis, with approximately one hour of signal for each seizure. One central and four peripheral electrodes over the epileptogenic focus were selected to resemble the layout of a novel neurostimulation device. Fifteen features were extracted from 2 s non-overlapping segments. RF models comprised either 500 or 125 trees, with varying depths. Three early channel integration (EI) strategies were compared with late integration (LI), using three channel fusion methods. A leave-one-patient-out cross-validation approach was used for evaluation, and memory requirements, alongside with inference energy and latency for 8-bit integer and 32-bit floating point models were computed on a microcontroller.<i>Main results.</i>The performance of EI feature sorting and LI were comparable. LI was favored by the 32-bit floating point format and more complex models, with the median channel fusion achieving a median area under the receiver operating characteristic curve score of 0.925. Feature sorting performed best with medium-sized models and was largely unaffected by the 8-bit integer format. Following causal output post-processing, false stimulations per hour were reduced to 5.5 at 100% sensitivity and fell below 3 at∼80% sensitivity.<i>Significance</i>. Our findings suggest that RF models with minimal energy and memory requirements can achieve state-of-the-art performance, making them well-suited for embedded applications in implantable devices. The complex interplay of the investigated factors is critical to performance, and along hardware specifications, should guide algorithmic design.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144002004","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":"Metaplasticity and continual learning: mechanisms subserving brain computer interface proficiency.","authors":"Shuo-Yen Chueh, Yuanxin Chen, Narayan Subramanian, Benjamin Goolsby, Phillip Navarro, Karim Oweiss","doi":"10.1088/1741-2552/add37b","DOIUrl":"10.1088/1741-2552/add37b","url":null,"abstract":"<p><p><i>Objective.</i>Brain computer interfaces (BCIs) require substantial cognitive flexibility to optimize control performance. Remarkably, learning this control is rapid, suggesting it might be mediated by neuroplasticity mechanisms operating on very short time scales. Here, we propose a meta plasticity model of BCI learning and skill consolidation at the single cell and population levels comprised of three elements: (a) behavioral time scale synaptic plasticity (BTSP), (b) intrinsic plasticity (IP) and (c) synaptic scaling (SS) operating at time scales from seconds to minutes to hours and days. Notably, the model is able to explain<i>representational drift</i>-a frequent and widespread phenomenon that adversely affects BCI control and continued use.<i>Approach.</i>We developed an all-optical approach to characterize IP, BTSP and SS with single cell resolution in awake mice using fluorescent two photon (2P) GCaMP7s imaging and optogenetic stimulation of the soma targeted ChRmine_Kv2.1. We further trained mice on a one-dimensional BCI control task and systematically characterized within session (seconds to minutes) learning as well as across sessions (days and weeks) with different neural ensembles.<i>Main results.</i>On the time scale of seconds, substantial BTSP could be induced and was followed by significant IP over minutes. Over the time scale of days and weeks, these changes could predict BCI control proficiency, suggesting that BTSP and IP might be complemented by SS to stabilize and consolidate BCI control.<i>Significance.</i>Our results provide early experimental support for a meta plasticity model of continual BCI learning and skill consolidation. The model predictions may be used to design and calibrate neural decoders with complete autonomy while considering the temporal and spatial scales of plasticity mechanisms. With the power of modern-day machine learning and artificial Intelligence, fully autonomous neural decoding and adaptation in BCIs might be achieved with minimal to no human intervention.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12101542/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144038934","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":"Feasibility of phase-locked transcranial magnetic stimulation of cerebellum for the treatment of essential tremor.","authors":"Xu Zhang, Roeland Hancock, Sabato Santaniello","doi":"10.1088/1741-2552/add76c","DOIUrl":"10.1088/1741-2552/add76c","url":null,"abstract":"<p><p><i>Objective.</i>Cerebellar transcranial magnetic stimulation (TMS) has been proposed to suppress limb tremors in essential tremor (ET), but mixed results have been reported so far, both when pulses are applied repetitively TMS (rTMS) and in bursts. We aim to investigate the cellular effects of TMS on the cerebellum under ET through numerical simulations.<i>Approach.</i>A computational model of the olivo-cerebello-thalamocortical pathways exhibiting the main neural biomarkers of ET (i.e. circuit-wide tremor-locked neural oscillations) was expanded to incorporate the effects of TMS-induced electric field (E-field) on Purkinje cells. TMS pulse amplitude, frequency, and temporal pattern were varied, and the resultant effects on ET biomarkers were assessed. Four levels of cellular response to TMS were considered, ranging from low to high cell recruitment underneath the coil, and three stimulation patterns were tested, i.e. rTMS, irregular TMS (ir-TMS, pulses were arranged according to Sobol sequences with average frequency matching rTMS), and phase-locked TMS (PL-TMS).<i>Main results.</i>rTMS can suppress ET oscillations, but its efficacy depends on tremor frequency and recruitment level, with these factors shaping a narrow range of effective settings. The ratio between tremor and rTMS frequencies also affects the neural response and further narrows the span of viable settings, while ir-TMS is ineffective. PL-TMS is highly effective and robust against changes to cell recruitment level and tremor frequency. Across all scenarios, PL-TMS provides a rapid (i.e. within seconds) suppression of tremor oscillations and, when both PL-TMS and rTMS are effective, the time to tremor suppression decreases by 50% or more in PL-TMS versus rTMS. At the cellular level, PL-TMS operates by disrupting the synchronization along the olivo-cerebellar loop, and the preferred phases map onto the mid-region of the silent period between complex spikes of the Purkinje cells.<i>Significance.</i>Cerebellar PL-TMS can provide robust suppression of ET oscillations while operating within safety boundaries.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144014305","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":"Fast FEM-based electric field calculations for transcranial magnetic stimulation.","authors":"Fang Cao, Kristoffer Hougaard Madsen, Torge Worbs, Oula Puonti, Hartwig Roman Siebner, Arno Schmitgen, Patrik Kunz, Axel Thielscher","doi":"10.1088/1741-2552/add76d","DOIUrl":"10.1088/1741-2552/add76d","url":null,"abstract":"<p><p><i>Objective</i>. To provide a finite-element method (FEM) for rapid, repeated evaluations of the electric field induced by transcranial magnetic stimulation (TMS) in the brain for changing coil positions.<i>Approach</i>. Previously, we introduced a first-order tetrahedral FEM enhanced by super-convergent patch recovery (SPR), striking a good balance between computational efficiency and accuracy (Saturnino<i>et al</i>2019<i>J. Neural Eng.</i><b>16</b>066032). In this study, we refined the method to accommodate repeated simulations with varying TMS coil position. Starting from a fast direct solver, streamlining the pre- and SPR-based post-calculation steps by implementing these steps as parallel sparse matrix multiplications strongly improved the computational efficiency. Additional speedups were achieved through efficient multi-core and GPU acceleration, alongside the optimization of the volume conductor model of the head for TMS.<i>Main Results</i>. For an anatomically detailed head model with ∼4.4 million tetrahedra, the optimized implementation achieves update rates above 1 Hz for electric field calculations in bilateral gray matter, resulting in a 60-fold speedup over the previous method with identical accuracy. An optimized model without neck and with adaptive spatial resolution scaled in dependence to the distance to brain grey matter, resulting in ∼1.9 million tetrahedra, increased update rates up to 10 Hz, with ∼3% numerical error and ∼4% deviation from the standard model. Region-of-interest (ROI) optimized models focused on the left motor, premotor and dorsolateral prefrontal cortices reached update rates over 20 Hz, maintaining a difference of <4% from standard results. Our approach allows efficient switching between coil types and ROI during runtime which greatly enhances the flexibility.<i>Significance</i>. The optimized FEM enhances speed, accuracy and flexibility and benefits various applications. This includes the planning and optimization of coil positions, pre-calculation and training procedures for real-time electric field simulations based on surrogate models as well as targeting and dose control during neuronavigated TMS.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144063674","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":"EEG-based assessment of long-term vigilance and lapses of attention using a user-centered frequency-tagging approach.","authors":"S Ladouce, J J Torre Tresols, K Le Goff, F Dehais","doi":"10.1088/1741-2552/add771","DOIUrl":"10.1088/1741-2552/add771","url":null,"abstract":"<p><p><i>Objective.</i>Sustaining vigilance over extended periods is crucial for many critical operations but remains challenging due to the cognitive resources required. Fatigue and other factors contribute to fluctuations in vigilance, causing attentional focus to drift from task-relevant information. Such lapses of attention, common in prolonged tasks, lead to decreased performance and missed critical information, with potentially serious consequences. Identifying physiological markers that predict inattention is key to developing preventive strategies.<i>Approach.</i>Previous research has established electroencephalography (EEG) responses to periodic visual stimuli, known as steady-state visual evoked potentials (SSVEP), as sensitive markers of attention. In this study, we evaluated a minimally intrusive SSVEP-based approach for tracking vigilance in healthy participants (<i>N</i>= 16) during two sessions of a 45 min sustained visual attention task (Mackworth's clock task). A 14 Hz frequency-tagging flicker was either superimposed on the task or absent.<i>Main results.</i>Results revealed that SSVEP responses were lower prior to lapses of attention, while other spectral EEG markers, such as frontal theta and parietal alpha activity, did not reliably distinguish between detected and missed attention probes. Importantly, the flicker did not affect task performance or participant experience.<i>Significance.</i>This non-intrusive frequency-tagging method provides a continuous measure of vigilance, effectively detecting attention lapses in prolonged tasks. It holds promise for integration into passive brain-computer interfaces, offering a practical solution for real-time vigilance monitoring in high-stakes settings like air traffic control or driving.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144061428","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 CNN-based approach for detecting eye blink episodes in EEG signals.","authors":"Izabela Rejer, Izabela Gago","doi":"10.1088/1741-2552/add49d","DOIUrl":"10.1088/1741-2552/add49d","url":null,"abstract":"<p><p><i>Objective.</i>This study aims to develop and evaluate a convolutional neural network (CNN)-based architecture for detecting eye blink episodes in electroencephalographic (EEG) signals, with a focus on the precise detection of individual events rather than their classification into predefined categories.<i>Approach.</i>The proposed method integrates a CNN-based architecture with a dedicated data augmentation technique that can capture the characteristic time patterns of the blink episodes.<i>Main results.</i>The performance of the proposed approach was validated using EEG data collected from 10 subjects across three experimental setups. The average detection rates reached 96.91% and 97.18% for individual subject tests, and 94.45% for cross-subject evaluation.<i>Significance.</i>The results demonstrate the high effectiveness and strong generalization capabilities of the proposed method, emphasizing its potential applications in improving neural data quality, cognitive state monitoring, and assistive technologies.</p>","PeriodicalId":94096,"journal":{"name":"Journal of neural engineering","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144035239","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}