Biomedical Engineering Letters最新文献

{"title":"Impact of proprioceptive deficit on control of joint torques during force matching task in hemispheric stroke survivors.","authors":"Jiwoo Oh, William Zev Rymer, Keonyoung Oh","doi":"10.1007/s13534-025-00538-9","DOIUrl":"10.1007/s13534-025-00538-9","url":null,"abstract":"<p><strong>Purpose: </strong>Proprioceptive loss is related to inaccurate initial reaching movement direction in chronic hemispheric stroke survivors, it potentially also impacts the predictive (feedforward) control of joint torques based on the target and current body joint positions, although this remains to be clarified. Accordingly, we sought to determine whether proprioceptive loss after stroke impairs predictive joint torque control when no visual feedback is provided.</p><p><strong>Methods: </strong>While blindfolded, 12 chronic hemispheric stroke survivors performed a force direction matching task. Each participant's hand was passively moved by a haptic robot to multiple locations on a plane, where they were asked to exert forces in four cardinal directions. The severity of proprioceptive loss was assessed via the finger-to-nose, passive movement detection, and position reproduction tests.</p><p><strong>Results: </strong>We found that the participants with severe proprioceptive deficits on the aforementioned standard tests showed a significantly larger directional error in the generated forces than participants with less impaired or intact proprioception. Specifically, the patients with severely impaired proprioception generated forces in only two directions, corresponding to horizontal shoulder flexion and extension, rather than the four distinct cardinal directions. No significant correlation was found between other parameters (the tactile deficit and the maximum voluntary pushing force) and the directional error in the force direction matching test.</p><p><strong>Conclusions: </strong>These results imply that the quality of proprioception significantly affects the accuracy of predictive control when generating joint torques, potentially due to inaccurate perception of the spatial representation of body joints.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 3","pages":"693-704"},"PeriodicalIF":3.4,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13129131/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147822170","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multimodal vision-language models in chest x-ray analysis: a study of generalization, supervision, and robustness.","authors":"Batoul Aljaddouh, D Malathi, Feisal Alaswad","doi":"10.1007/s13534-025-00535-y","DOIUrl":"10.1007/s13534-025-00535-y","url":null,"abstract":"<p><p>Multimodal vision-language models (VLMs) are increasingly applied to medical imaging, yet systematic evaluations comparing them with unimodal models across datasets, supervision regimes, and clinical domains remain scarce. Prior studies often focus on a single dataset, specific pathologies, or one supervision setting, leaving unclear how these models generalize under realistic variability. We conduct a systematic evaluation of six leading unimodal and multimodal models for chest X-ray (CXR) classification using four widely adopted datasets: MIMIC-CXR, CheXpert, NIH-14, and PadChest. We assess model behavior in both zero-shot (ZS) and fine-tuned (FT) configurations, with a focus on generalization across pathologies, datasets, and linguistic domains. Our findings show that pretrained multimodal models such as CheXzero and CXR-LLaVA perform strongly in zero-shot scenarios, especially on out-of-distribution data, reflecting their capacity for semantic generalization. However, their performance tends to decline after fine-tuning in cross-lingual or noisy-label contexts, indicating susceptibility to overfitting. In contrast, unimodal models gain substantially from supervised fine-tuning, especially on in-domain data. Limitations include evaluation on seven shared pathologies, CXR imaging only, and use of publicly available pretrained models, which may restrict generalization to other clinical tasks. These findings highlight key trade-offs between generalization, robustness, and adaptability, and suggest promise in hybrid training strategies that integrate multimodal priors with targeted domain supervision.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 2","pages":"517-537"},"PeriodicalIF":2.8,"publicationDate":"2025-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13013966/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147522233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction: Pneumatic driven pulsatile ECMO in vitro evaluation with oxygen tanks.","authors":"Changyoung Yoo, Seongmin Kang, Seong-Wook Choi","doi":"10.1007/s13534-025-00532-1","DOIUrl":"https://doi.org/10.1007/s13534-025-00532-1","url":null,"abstract":"<p><p>[This corrects the article DOI: 10.1007/s13534-023-00295-7.].</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 1","pages":"215"},"PeriodicalIF":2.8,"publicationDate":"2025-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12824037/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146054449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Physiological swelling imaging in human calf under stocking compression by sparse Bayesian learning implemented into electrical impedance tomography (SBL-EIT).","authors":"Kota Asano, Ryoma Ogawa, Prima Asmara Sejati, Shinsuke Akita, Masahiro Takei","doi":"10.1007/s13534-025-00524-1","DOIUrl":"10.1007/s13534-025-00524-1","url":null,"abstract":"<p><p>Physiological swelling in human calf has been imaged under stocking compression by the sparse Bayesian learning implemented into electrical impedance tomography (SBL-EIT) to evaluate the in situ treatment effect of various compression pressures. SBL-EIT reconstructs conductivity distribution [Formula: see text] to image excessive extracellular fluid in subcutaneous adipose tissue (SAT), indicating the susceptibilities to physiological swelling due to various compression pressures. The SBL-EIT was applied to the imaging of eight-subject calves during prolonged standing under three types of net compression pressures <i>P</i> ^net measured by pressure sensor - strong pressure: <i>P</i> ^{net, Strong} = 11.9 [Formula: see text] 2.0 mmHg, weak pressure: <i>P</i> ^{net, Weak} = 4.47 [Formula: see text] 3.1 mmHg, and control pressure: <i>P</i> ^{net, Control} = 0.00 [Formula: see text] 0.0 mmHg, respectively. From the experimental results, the spatial-mean conductivity ⟨<b>σ</b>⟩^α2 with two pre-processing steps to eliminate undesirable effects, i.e., the difference in skin condition and effect of wearing stockings itself, is the highest in the case of stocking with control pressure, followed by weak and strong pressures across all subjects. Moreover, the ⟨<b>σ</b>⟩^α<i>2</i> has a strong positive correlation with the conventional inversed impedance 1/<i>z</i> ^BIA by bioelectrical impedance analysis (BIA) (a correlation coefficient 0.528 < <i>R</i> < 0.990; <i>n</i> = 19 and <i>p</i> < 0.05), which is mainly increased during the prolonged standing. Moreover, various susceptibilities to physiological swelling are investigated based on the increase in [Formula: see text] for each subject, which is associated with subject external factors such as postural changes and circumference and internal factors like SAT.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 1","pages":"189-201"},"PeriodicalIF":2.8,"publicationDate":"2025-11-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12824078/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146046921","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An open-source compact stimulator for various transcranial and transcutaneous electrical stimulation in preclinical research.","authors":"Daniel H Mena Fonseca, Hannah Noh, Dong Hyeon Lee, Yong-Il Shin, Byung-Cheul Shin, Joonsoo Jeong","doi":"10.1007/s13534-025-00530-3","DOIUrl":"10.1007/s13534-025-00530-3","url":null,"abstract":"<p><p>While various transcranial electrical stimulation (tES) and transcutaneous electrical nerve stimulation (TENS) devices differ in target area, pulse parameters, and electrode types, they share the fundamental requirement of generating controlled current pulses. However, existing current stimulators are often expensive and limited in their ability to customize a wide range of pulse parameters, especially for preclinical research involving diverse stimulation protocols. Developing a custom stimulator may not be feasible depending on the technical expertise of the researchers. This study proposes an open-source, compact, and easy-to-make current stimulator built using only commercial off-the-shelf components. The design supports a wide range of stimulation parameters without requiring specialized knowledge in electronics. The prototype stimulator has a form factor of 33 mm × 43 mm × 20 mm and is compatible with general electrode types. Pulse parameters can be simply adjusted by using a knob or tuning resistors and capacitors, enabling simple customization by researchers for varying transcranial and transcutaneous stimulations. The wide tunability in current intensity (2 µA to 10 mA), pulse rate (0.1 Hz to 100 kHz), and pulse width (10 µs to DC) is demonstrated, followed by statistical analysis of output waveforms to demonstrate the accuracy and stability of pulse parameters using both high- and low-tolerances components. The results indicate an average accuracy exceeding 96.4% and a coefficient of variance less than 0.57%, suggesting the reliable and robust generation of stimulation waveforms for various tES and TENS applications in preclinical research.</p><p><strong>Supplementary information: </strong>The online version contains supplementary material available at 10.1007/s13534-025-00530-3.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 2","pages":"473-482"},"PeriodicalIF":2.8,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13013780/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147522065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Transcutaneous interference spinal cord stimulation: leadfield-based pareto optimization of electrode montages for improved focality.","authors":"Mariko Teragiwa, Leonel E Medina, Alonso Carvajal, Kanata Yatsuda, Wenwei Yu, Jose Gomez-Tames","doi":"10.1007/s13534-025-00531-2","DOIUrl":"10.1007/s13534-025-00531-2","url":null,"abstract":"<p><p>This study investigates the feasibility of transcutaneous interferential spinal cord stimulation (tISCS), a novel non-invasive neuromodulation method, using temporal interference to enhance focality and comfort in spinal cord stimulation. The central research question is whether tISCS can achieve targeted activation of spinal cord circuits while reducing unwanted stimulation of skin and muscle tissues, which are common limitations of conventional transcutaneous spinal cord stimulation (tSCS). A finite element model of the lower thorax was developed to simulate electric field distributions for various skin electrode montages. To address the computational bottleneck associated with high-resolution modeling and montage optimization, we implemented a leadfield-based Pareto optimization strategy to identify the electrode configuration that maximizes the electric field in the spinal cord and minimizes it in off-target tissues. tISCS montages were compared with tSCS montages in terms of focality and stimulation efficiency. Optimized tISCS configurations significantly reduced electric field intensity in the skin by over 20-fold compared to tSCS. The ratio of spinal cord to skin electric fields increased by at least tenfold, indicating enhanced focality. The injection current efficiency in tISCS can be leveraged to increase spinal cord electric fields by at least fivefold while keeping skin exposure below the levels observed with tSCS. tISCS enables deeper and more selective spinal cord stimulation compared to tSCS, with substantially reduced off-target effects. In conclusion, this is the first computational demonstration of tISCS feasibility, and leadfield-guided Pareto optimization enables efficient montage selection, providing a foundation for future experimental applications.</p><p><strong>Supplementary information: </strong>The online version contains supplementary material available at 10.1007/s13534-025-00531-2.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 2","pages":"483-492"},"PeriodicalIF":3.4,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13013866/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147522221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Advancing neurological disease treatment: a computational approach for fibroblast growth factor detection.","authors":"Farman Ali, Amal Babour, Othman Asiry, Wajdi Alghamdi, Atef Masmoudi, Naif Waheb Rajkhan","doi":"10.1007/s13534-025-00520-5","DOIUrl":"https://doi.org/10.1007/s13534-025-00520-5","url":null,"abstract":"<p><p>Fibroblast Growth Factor plays a crucial role in neurological health, contributing to neuron protection, injury recovery, and angiogenesis. It is also significantly involved in the onset and progression of neurodegenerative disorders such as Huntington's, Alzheimer's, Parkinson's disease, and stroke, making FGF a vital target for therapeutic interventions. Despite its importance, no computational tool has been developed to predict FGF proteins. In this study, we present the first novel deep learning-based computational approach designed for the prediction of FGF proteins. We constructed two novel, high-quality datasets curated from the UniProt database for training and evaluation. Sequences were transformed into numerical representations using three complementary feature encoding methods including Dipeptide Composition, Dipeptide Deviation from Expected Mean, and Grouped Amino Acid Composition. These features capture both local and global sequence information. Multiple deep learning models were explored, including Convolutional Neural Network, Bidirectional Long Short-Term Memory, Generative Adversarial Network, and Gated Recurrent Unit. Among these, our proposed Convolutional Neural Network-based model outperformed all others, achieving an accuracy of 83.50%, sensitivity of 84.30%, specificity of 82.67%, F1 score of 83.42%, and a Matthews Correlation Coefficient of 0.671. The proposed approach has the potential to advance therapeutic discovery by enabling accurate identification of Fibroblast Growth Factor and improving our understanding of their role in neurological health and disease.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 1","pages":"167-176"},"PeriodicalIF":2.8,"publicationDate":"2025-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12824060/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146054447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adversarially regularized transformer with channel-wise noise for robust hand gesture recognition using surface electromyography.","authors":"Doheun Cha, Dong-Gyu Lee, Sangtae Ahn","doi":"10.1007/s13534-025-00529-w","DOIUrl":"10.1007/s13534-025-00529-w","url":null,"abstract":"<p><p>Surface electromyography (sEMG) is a non-invasive method that captures electrical signals through electrodes placed on the skin surface, allowing hand gesture recognition by interpreting muscle activity patterns. Despite its advantages, sEMG signals face challenges from non-stationarity and environmental noise, which complicate reliable classification. To overcome these challenges, we propose a novel network that integrates adversarial training with noise perturbation, improving both robustness and generalization in sEMG-based hand gesture recognition. Our approach achieves state-of-the-art results on benchmark datasets consisting of NinaPro DB1, DB2, and DB4, testing practical applications in prosthetic control and human-computer interaction. In aggregated evaluations, we achieve performance improvements of 3.52%p (DB1), 1.47%p (DB2), and 5.47%p (DB4) compared to EMGHandNet. In subject-wise evaluations, we observe performance gains of 1.73%p (DB1) and 5.15%p (DB4), while maintaining stable performance on DB2 despite its larger and more variable subject pool. Adversarial training also markedly enhances robustness against attacks: under FGSM and PGD attacks, in subject-wise evaluations, performance gains of 1.73%p (DB1) and 5.15%p (DB4) were observed, while maintaining stability in DB2 despite its larger subject pool and outliers. Adversarial training significantly improved robustness against attacks: under fast gradient sign method (FGSM) and projected gradient descent (PGD) attacks ([Formula: see text]=0.01), our model with adversarial training retained 32.92% (FGSM) and 23.88% (PGD) accuracy in DB1, 12.05% (FGSM) and 5.97% (PGD) in DB2, and 18.06% (FGSM) and 12.41% (PGD) in DB4, consistently outperforming models without adversarial training. These results demonstrate that our approach enhances both classification accuracy and adversarial robustness, addressing critical limitations in real-world applications such as prosthetic controls.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 2","pages":"463-472"},"PeriodicalIF":2.8,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13013967/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147522041","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The method for quantified analysis and pattern visualization for eye blinking using high-frame-rate video.","authors":"Woon-Hee Lee, Jongmo Seo, Jeong-Min Hwang","doi":"10.1007/s13534-025-00522-3","DOIUrl":"10.1007/s13534-025-00522-3","url":null,"abstract":"<p><p>This study proposes a visualization and analysis method for eye blinking pattern using high-frame-rate videos. The high-frame-rate video clips for visualization are taken using a camera without additional equipment. The partial video clips of eye blinking except for eyelid flutters and microsleeps are extracted from the entire video clip. The changes in shapes and positions of the upper eyelid during the eye blinking sequences are evaluated, and each eye blinking is visualized as a single image. The various parameters regarding eye blinking are calculated to analyze blinking patterns. The single eye blinking sequence is divided into phases to analyze and classify eye blinking patterns in more detail. In this experiment conducted on 80 volunteers, the proposed method was able to quantitatively analyze eyelid movements, and various parameters related to eye blinking were calculated. Additionally, different types of eye blinking patterns were visualized as graph images, and incomplete eye blinking and consecutive eye blinking were defined and detected. The proposed method can overcome the spatial and situational limitations of conventional bio-signal analysis methods, as it allows non-contact measurement in ordinary environments. In addition, since quantitative eye blink data obtained from high-frame-rate video contain more information than data obtained from bio-signals, it is expected that analysis methods using videos can be easily applied to a wider range of fields.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"15 6","pages":"1097-1107"},"PeriodicalIF":3.4,"publicationDate":"2025-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12638546/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145588469","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An experimental study on estimating sleep period time using accelerometer and electrodermal activity in ring-shaped wearable devices.","authors":"Eunju Lee, Jaewook Shin, Kwang Ik Yang, Jaegeol Cho","doi":"10.1007/s13534-025-00521-4","DOIUrl":"https://doi.org/10.1007/s13534-025-00521-4","url":null,"abstract":"<p><p>There has been growing interest in sleep tracking technologies utilizing ring-shaped wearable devices. This study aimed to develop a method for accurately estimating sleep duration, onset, and offset using such a device. Conventional wrist-worn accelerometer-based devices often show limited accuracy, particularly during periods of low movement. Likewise, algorithms relying solely on electrodermal activity (EDA) signals struggle to detect frequent wake episodes due to their low temporal resolution. To address these limitations, we developed a ring-shaped wearable device and a set of algorithms that integrate both accelerometer and EDA signals. The performance of the proposed algorithms was evaluated through a clinical study involving 25 participants, the majority of whom had sleep disorders. Results showed that combining these complementary signals enabled accurate detection of sleep onset and offset within 10 min and maintained high accuracy even in conditions such as minimal movement or frequent wake episodes. These findings suggest that multimodal sensing may offer a promising direction for enhancing the reliability of sleep monitoring in real-world settings.</p>","PeriodicalId":46898,"journal":{"name":"Biomedical Engineering Letters","volume":"16 1","pages":"177-188"},"PeriodicalIF":2.8,"publicationDate":"2025-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12824073/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146054380","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}