Computers in biology and medicine最新文献

{"title":"Design of patient-specific mandibular reconstruction plates and a hybrid scaffold","authors":"Sait Emre Dogan ,&nbsp;Cengizhan Ozturk ,&nbsp;Bahattin Koc","doi":"10.1016/j.compbiomed.2024.109380","DOIUrl":"10.1016/j.compbiomed.2024.109380","url":null,"abstract":"<div><h3>Background:</h3><div>Managing segmental mandibular defects remains challenging, requiring a multidisciplinary approach despite the remarkable progress in mandibular reconstruction plates, finite element methods, computer-aided design and manufacturing techniques, and novel surgical procedures. Complex surgeries require a comprehensive approach, as using only reconstruction plates or tissue scaffolds may not be adequate for optimal results. The limitations of the treatment options should be investigated towards a patient-specific trend to provide shorter surgery time, better healing, and lower costs. Integrated hybrid scaffold systems are promising in improving mechanical properties and facilitating healing. By combining different materials and structures, hybrid scaffolds can provide enhanced support and stability to the tissue regeneration process, leading to better patient outcomes. The use of such systems represents a significant advancement in tissue engineering and a wide range of medical procedures.</div></div><div><h3>Materials and Methods:</h3><div>A head and neck computed tomography (CT) data of a patient with odontogenic myxoma was used for creating a three-dimensional (3D) mandible model. Virtual osteotomies were performed to create a segmental defect model, including the angulus mandibulae region. The first mandibular reconstruction plate was designed. Finite elemental analyses (FEA) and topology optimizations were performed to create two different reconstruction plates for different treatment scenarios. The FEA were performed for the resulting two plates to assess their biomechanical performance. To provide osteoconductive and osteoinductive properties a scaffold was designed using the defect area. A biomimetic Tricalcium phosphate-Polycaprolactone (TCP-PCL) hybrid bone scaffold enhanced with Hyaluronic acid dipping was manufactured.</div></div><div><h3>Results:</h3><div>The results of the in-silico analysis indicate that the designed reconstruction plates possess robust biomechanical performance and demonstrate remarkable stability under the most rigorous masticatory activities. Using the Voronoi pattern decreased the mass by %37 without losing endurance. Using reconstruction plates and hybrid scaffolds exhibits promising potential for clinical applications, subject to further in vivo and clinical studies.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109380"},"PeriodicalIF":7.0,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The quality, uniqueness, and causality of NMR-based prediction models for low-density lipoprotein cholesterol subfractions in human blood plasma","authors":"Yongxin Ye ,&nbsp;Bo Markussen ,&nbsp;Søren Balling Engelsen ,&nbsp;Bekzod Khakimov","doi":"10.1016/j.compbiomed.2024.109379","DOIUrl":"10.1016/j.compbiomed.2024.109379","url":null,"abstract":"<div><div>Low-density lipoprotein (LDL) cholesterol (<em>chol</em>) subfractions are risk biomarkers for cardiovascular diseases (CVD). A reference analysis, ultracentrifugation (UC), is laborious and may be replaced with a rapid prediction using proton NMR spectra of human blood plasma. However, the quality and uniqueness of these prediction models of biologically related subfractions remains unknown. This study, using two independent cohorts (n = 277), investigates the inter-correlations between LDL cholesterol in the main fraction and five subfractions, as well as the independence of their NMR-based prediction models. The results reveal that the prediction models utilize both shared and unique spectral information from the NMR spectra to determine concentrations of LDL subfractions. Analysis of variance contributions for prediction and causality assessments demonstrate that the NMR spectra contain unique predictive information for the LDL1<em>chol</em>, LDL2<em>chol</em>, and LDL5<em>chol</em> subfractions. In contrast, the spectral signatures for LDL3<em>chol</em> and LDL4<em>chol</em> are either insufficient or confounded. Our findings indicate that these five CVD biomarkers represent two independent clusters, reflecting their biosynthetic pathways, and confirm the presence of causal relationships between certain LDL <em>chol</em> subfractions. This highlights the importance of employing caution when interpreting the concentrations of specific LDL subfractions as standalone biomarkers for CVD risk.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109379"},"PeriodicalIF":7.0,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142719888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fusing CNNs and attention-mechanisms to improve real-time indoor Human Activity Recognition for classifying home-based physical rehabilitation exercises.","authors":"Moamen Zaher, Amr S Ghoneim, Laila Abdelhamid, Ayman Atia","doi":"10.1016/j.compbiomed.2024.109399","DOIUrl":"https://doi.org/10.1016/j.compbiomed.2024.109399","url":null,"abstract":"<p><p>Physical rehabilitation plays a critical role in enhancing health outcomes globally. However, the shortage of physiotherapists, particularly in developing countries where the ratio is approximately ten physiotherapists per million people, poses a significant challenge to effective rehabilitation services. The existing literature on rehabilitation often falls short in data representation and the employment of diverse modalities, limiting the potential for advanced therapeutic interventions. To address this gap, This study integrates Computer Vision and Human Activity Recognition (HAR) technologies to support home-based rehabilitation. The study mitigates this gap by exploring various modalities and proposing a framework for data representation. We introduce a novel framework that leverages both Continuous Wavelet Transform (CWT) and Mel-Frequency Cepstral Coefficients (MFCC) for skeletal data representation. CWT is particularly valuable for capturing the time-frequency characteristics of dynamic movements involved in rehabilitation exercises, enabling a comprehensive depiction of both temporal and spectral features. This dual capability is crucial for accurately modelling the complex and variable nature of rehabilitation exercises. In our analysis, we evaluate 20 CNN-based models and one Vision Transformer (ViT) model. Additionally, we propose 12 hybrid architectures that combine CNN-based models with ViT in bi-model and tri-model configurations. These models are rigorously tested on the UI-PRMD and KIMORE benchmark datasets using key evaluation metrics, including accuracy, precision, recall, and F1-score, with 5-fold cross-validation. Our evaluation also considers real-time performance, model size, and efficiency on low-power devices, emphasising practical applicability. The proposed fused tri-model architectures outperform both single-architectures and bi-model configurations, demonstrating robust performance across both datasets and making the fused models the preferred choice for rehabilitation tasks. Our proposed hybrid model, DenMobVit, consistently surpasses state-of-the-art methods, achieving accuracy improvements of 2.9% and 1.97% on the UI-PRMD and KIMORE datasets, respectively. These findings highlight the effectiveness of our approach in advancing rehabilitation technologies and bridging the gap in physiotherapy services.</p>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"109399"},"PeriodicalIF":7.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142726599","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using the coefficient of determination to identify injury regions after stroke in pre-clinical FDG-PET images.","authors":"Wuxian He, Hongtu Tang, Jia Li, Xiaoyan Shen, Xuechen Zhang, Chenrui Li, Huafeng Liu, Weichuan Yu","doi":"10.1016/j.compbiomed.2024.109401","DOIUrl":"https://doi.org/10.1016/j.compbiomed.2024.109401","url":null,"abstract":"<p><strong>Background: </strong>In the analysis of brain fluorodeoxyglucose positron emission tomography (FDG-PET) images, intensity normalization is a necessary step to reduce inter-subject variability. However, the choice of the most appropriate normalization method in stroke studies remains unclear, as demonstrated by inconsistent findings in the literature.</p><p><strong>Materials and methods: </strong>Here, we propose a regression- and single-subject-based model for analyzing FDG-PET images without intensity normalization. Two independent data sets were collected before and after middle cerebral artery occlusion (MCAO), with one comprising 120 rats and the other 96 rats. After data preprocessing, voxel intensities in the same region and hemisphere were paired before and after the MCAO scan. A linear regression model was applied to the paired data, and the coefficient of determination R² was calculated to measure the linearity. The R² values between the ipsilateral and contralateral hemispheres were compared, and significant regions were defined as those with reduced linearity. Our method was compared with voxel-wise analysis under different intensity normalization methods and validated using the triphenyl tetrazolium chloride (TTC) staining data.</p><p><strong>Results: </strong>The significant regions identified by the proposed method showed a large degree of overlap with the infarcted regions identified by TTC data, as measured by the Dice similarity coefficient (DSC). The average DSC of the proposed method was 59.7%, whereas the DSCs of the existing approaches ranged from 41.4%∼51.3%. Additional validation using receiver operating characteristic (ROC) demonstrated that the area under the curve (AUC) of the average ROC curves reached 0.84 using the proposed method, whereas existing methods achieved AUCs ranging from 0.77∼0.79. The identified regions were consistent across the two independent data sets, and some findings were corroborated by other publications.</p><p><strong>Conclusions: </strong>The proposed model presents a novel quantitative approach for identifying injury regions post-stroke using FDG-PET images. The calculation does not require intensity normalization and can be applied to individual subjects. The method yields more sensitive results compared to existing identification methods.</p>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"109401"},"PeriodicalIF":7.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142726600","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Detection of retinal diseases using an accelerated reused convolutional network","authors":"Amin Ahmadi Kasani,&nbsp;Hedieh Sajedi","doi":"10.1016/j.compbiomed.2024.109466","DOIUrl":"10.1016/j.compbiomed.2024.109466","url":null,"abstract":"<div><div>Convolutional neural networks are continually evolving; with some efforts aimed at improving accuracy, others at increasing speed, and some at enhancing accessibility. Improving accessibility broadens the application of neural networks across a wider range of tasks, including the detection of eye diseases. Early diagnosis of eye diseases and consulting an ophthalmologist can prevent many vision disorders. Given the importance of this issue, various datasets have been collected from the cornea to facilitate the process of making neural network models. However, most of the methods introduced in the past are computationally complex. In this study, we tried to increase the accessibility of deep neural network models. We did this at the most fundamental level—specifically, by redesigning and optimizing the convolutional layers. By doing so, we created a new general model that incorporates our novel convolutional layer named ArConv layers. Thanks to the efficient performance of this new layer, the model has suitable complexity for use in mobile phones and perform the task of diagnosing the presence of disease with high accuracy. The final model we present contains only 1.3 million parameters. In comparison to the MobileNetV2 model, which has 2.2 million parameters, our model demonstrated better accuracy when trained and evaluated on the RfMiD dataset under identical conditions, achieving an accuracy of 0.9328 versus 0.9266 on the RfMiD test set.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109466"},"PeriodicalIF":7.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modelling of coronary artery stenosis and study of hemodynamic under the influence of magnetic fields","authors":"Chandra Shekhar Maurya,&nbsp;Abhijeet Kumar","doi":"10.1016/j.compbiomed.2024.109464","DOIUrl":"10.1016/j.compbiomed.2024.109464","url":null,"abstract":"<div><div>Investigating magnetic blood flow characteristics through arteries and micron-size channels for clinical therapies in biomedicine is becoming increasingly important with the rise of point-of-care diagnostics devices. A computational fluid dynamics (CFD) investigation is conducted to explore blood flow within a coronary artery affected by an elliptical stenosis near the artery wall under the influence of a magnetic field. The novelty of our study is the integration of Navier-Stokes and Maxwell's equations to calculate body forces on fluid flow, coupled with the application of magnetic fields both longitudinally and vertically, and the use of the Carreau-Yasuda model to analyse non-Newtonian blood rheology. Blood flow is modelled by solving the incompressible continuity and momentum equations, considering laminar and non-Newtonian properties, with the finite element-based solver COMSOL Multiphysics. The CFD model is validated using previously published analytical and computational data. This study investigates the effects of magnetic fields on blood flow through stenotic arteries with 25 %, 35 %, and 50 % stenosis, examining how the magnetic field and its orientation impact variations in velocity profiles, pressure drop, and wall shear stress (WSS). Our results show that magnetic fields can effectively manipulate blood flow, causing acceleration or deceleration depending on field direction. Significant changes in hemodynamics are observed, particularly at 50 % arterial stenosis, highlighting the profound impact of stenosis on flow characteristics. Compared to healthy arteries, the velocity change in stenosed arteries increased by 16.5 %, 29.4 %, and 62.1 % for 25 %, 35 %, and 50 % stenosis, respectively. The findings advance experimental models of blood flow in magnetic fields, highlighting the critical importance of regulating blood velocity and pressure. These insights are particularly valuable for developing drug delivery systems and magnetic-driven blood pumps.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109464"},"PeriodicalIF":7.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel gait quality measure for characterizing pathological gait based on Hidden Markov Models","authors":"Abdelghani Halimi ,&nbsp;Lorenzo Hermez ,&nbsp;Nesma Houmani ,&nbsp;Sonia Garcia-Salicetti ,&nbsp;Omar Galarraga","doi":"10.1016/j.compbiomed.2024.109368","DOIUrl":"10.1016/j.compbiomed.2024.109368","url":null,"abstract":"<div><div>This study addresses the characterization of normal gait and pathological deviations caused by neurological diseases. We focus on the angular knee kinematics in the sagittal plane and we propose to exploit Hidden Markov Models to build a statistical model of normal gait. Such model provides a log-likelihood score that quantifies gait quality. Hence allowing to assess deviations of pathological cycles from normal gait. Our approach allows a refined characterization of motor impairments of three different patients’ groups. In particular, it detects the affected lower limb in Hemiparetic patients. Comparatively to the Gait Variable Score and a Dynamic Time Warping-based metric, our results show that our statistical method is more effective for finely quantifying pathological deviations. Finally, we show the potential use of our methodology to assess therapeutic impacts during gait rehabilitation, which represents a promising avenue for improving patient care.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109368"},"PeriodicalIF":7.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Discrete element method-based simulation system for predicting natural stone evacuation pathways in patients with urolithiasis","authors":"Naoaki Okabayashi ,&nbsp;Katsuki Hirai ,&nbsp;Takayuki Nagata ,&nbsp;Kazuhide Makiyama ,&nbsp;Junichi Matsuzaki ,&nbsp;Kota Fukuda ,&nbsp;Shun Takahashi ,&nbsp;Mitsuru Komeya ,&nbsp;Hiroshi Kimura","doi":"10.1016/j.compbiomed.2024.109454","DOIUrl":"10.1016/j.compbiomed.2024.109454","url":null,"abstract":"<div><div>Urolithiasis is a globally prevalent disease with high incidence and recurrence rates and is often accompanied by severe pain. Its ideal treatment is spontaneous stone passage, avoiding the invasiveness associated with surgery. However, the mechanisms underlying spontaneous stone passage remain unclear. Therefore, in this study, we developed a kidney stone trajectory prediction simulation system using the discrete element method (DEM) to elucidate spontaneous passage mechanisms by analyzing and visualizing stone trajectories within the kidney. We compared this simulation system with experiments using a three-dimensional kidney replica of patients with urolithiasis to optimize critical DEM parameters, including the collision margin <em>ε</em>, friction coefficient <em>C</em>_f<em>,</em> and restitution coefficient <em>C</em>_r. The reliability of these optimized parameters was validated using kidney shapes that differed from those used in the optimization experiments. The simulation system with optimized parameters consistently demonstrated high fidelity to the experimental results, regardless of kidney shape, initial stone position, or stone size. These findings demonstrate the reliability of the simulation system, underscoring its potential contribution to developing new and effective treatments for urolithiasis by improving the accuracy of stone trajectory predictions.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109454"},"PeriodicalIF":7.0,"publicationDate":"2024-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700618","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Disentangled representational learning for anomaly detection in single-lead electrocardiogram signals using variational autoencoder","authors":"Maximilian Kapsecker ,&nbsp;Matthias C. Möller ,&nbsp;Stephan M. Jonas","doi":"10.1016/j.compbiomed.2024.109422","DOIUrl":"10.1016/j.compbiomed.2024.109422","url":null,"abstract":"<div><div>Wearable technology enables the unsupervised recording of electrocardiogram (ECG) signals. Analyzing these high-dimensional ECG data poses challenges regarding statistical approaches and explainability. This work investigates the feasibility of medically explainable anomaly detection through disentangled representational learning of ECGs and personalization to mitigate inter-subject variations. Five open-source ECG datasets were converted into a set of denoised one-second traces of lead I signal, each covering individual features such as wave morphologies and pathologies. A beta total correlation variational autoencoder was optimized on four of these datasets for 68 systematic parameterization variants. The best-performing model revealed disentanglement in the 12-dimensional embedding space, specifically between atrial- and ventricular features. Within the embedding space, a k-nearest neighbor classifier was evaluated on a left-out test set tailored for anomaly detection. The result is a F1 score of 0.94 for the binary prediction of sinus rhythm and the pathological classes: Left bundle branch block, right bundle branch block, myocardial infarction, and AV block (1st degree). The 90.94% accuracy in anomaly detection falls within the range of established detectors (89.38%–99.77%) but offers the advantage of being explainable and largely unsupervised. Model fine-tuning for each of 100 randomly sampled individuals of the Icentia11k dataset mitigated inter-subject variations. The associated F1 score for predicting normal, premature atrial contraction, and premature ventricular contraction from the embedding space was 0.93. The distribution plots of pathologies along the explainable axis were reasonably consistent with medical expertise. The results suggest the presented disentangled variational autoencoder as a robust method for explainable ECG representation.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109422"},"PeriodicalIF":7.0,"publicationDate":"2024-11-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700620","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Towards real-world wearable sleepiness detection: Electrodermal activity data during speech can identify sleep deprivation","authors":"Jihye Moon ,&nbsp;Andrew Peitzsch ,&nbsp;Youngsun Kong ,&nbsp;Pranav Seshadri ,&nbsp;Ki H. Chon","doi":"10.1016/j.compbiomed.2024.109320","DOIUrl":"10.1016/j.compbiomed.2024.109320","url":null,"abstract":"<div><div>Accurate assessment of sleepiness is pivotal in managing the fatigue-associated risks stemming from sleep deprivation. Speech signals are easy to obtain, allowing detection of sleepiness anywhere. Previous machine learning (ML) studies using speech have not been successful in achieving reliable estimation of perceived sleepiness levels, which results in inaccurate sleepiness determination. In this paper, we propose that these challenges primarily stem from the inherent complexities of speech signals with inaccurate labels of sleepiness. Because the physical effects of sleepiness become pronounced after prolonged wakefulness, we conducted a 25-h sleep deprivation study. We collected electrodermal activity (EDA) and speech data from 30 subjects during speech production every 2 h over the 25-hour period, along with various sleepiness level labels—their cognitive impairment scores derived from the psychomotor vigilance test, their self-reported sleepiness scores, and the h awake scores. The data analysis compared EDA recorded during speech versus only the speech data and examined which approach provided better sleepiness level estimation and detection using ML. The ML result is that features derived from only EDA during speech production provided the most accurate sleepiness determination. Specifically, EDA ML models trained using the hours awake scores provided the best sleepiness level estimation, with 0.53 correlation, and better detection of sleepiness (which is related to cognitive performance deterioration), with 0.85 accuracy (0.80 sensitivity), when compared to ML features derived from speech, which obtained 0.40 correlation for sleepiness level estimation and 0.69 accuracy (0.59 sensitivity) for sleepiness detection. Moreover, the EDA data collected during speech production offered the best performance for sleepiness detection compared to EDA collected during other activities, such as visual vigilance (0.68 accuracy and 0.65 sensitivity). Given the potential of EDA data during speech production, this work demonstrates the promise of future wearable devices that could collect EDA data from speech activity, along with speech signals, for more advanced and accurate real-world sleepiness detection.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"184 ","pages":"Article 109320"},"PeriodicalIF":7.0,"publicationDate":"2024-11-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}