Computers in biology and medicine最新文献

{"title":"A hybrid hierarchical transformer model for ECG classification and age prediction","authors":"Pedro Dutenhefner ,&nbsp;Turi Rezende ,&nbsp;José Geraldo Fernandes ,&nbsp;Diogo Tuler ,&nbsp;Gabriela M.M. Paixão ,&nbsp;Gisele Pappa ,&nbsp;Antønio Ribeiro ,&nbsp;Wagner Meira Jr.","doi":"10.1016/j.compbiomed.2026.111462","DOIUrl":"10.1016/j.compbiomed.2026.111462","url":null,"abstract":"<div><div>Electrocardiograms (ECGs) play a crucial role in cardiovascular healthcare, requiring effective analytical models. ECG analysis is inherently hierarchical, involving multiple temporal scales from individual waveforms to intervals within heartbeats, and finally to the distances between heartbeats. Convolutional Neural Networks (CNNs) have demonstrated strong performance in ECG classification tasks due to their inductive bias toward local connectivity and translation invariance. In other domains, Transformers have emerged as powerful models for capturing long-range dependencies. This paper introduces HiT-NeXt, a hybrid hierarchical model designed to capture both local morphological patterns and global temporal dependencies by combining CNNs with transformer blocks featuring restricted attention windows. The model incorporates ConvNeXt-based convolutional layers to extract local features and perform patch merging, enabling hierarchical representation learning. Transformer blocks are constrained with local attention windows and leverage relative contextual positional encoding to incorporate positional information effectively into embeddings, enhancing robustness to translations in ECG signal patterns. Experimental results demonstrate that HiT-NeXt outperforms state-of-the-art methods on tasks including ECG abnormality classification and cardiological age prediction, achieving superior performance compared to both existing models and cardiologist evaluations.<span><span>²</span></span></div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111462"},"PeriodicalIF":6.3,"publicationDate":"2026-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146028640","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":"Research on breast ultrasound images lesion localization and diagnosis based on knowledge-driven and data-driven methods","authors":"Jianqiang Li ,&nbsp;Lintao Song ,&nbsp;Xiaoling Liu ,&nbsp;Yiming Liu ,&nbsp;Tianbao Ma ,&nbsp;Jun Bai ,&nbsp;Qing Zhao ,&nbsp;Xi Xu","doi":"10.1016/j.compbiomed.2026.111465","DOIUrl":"10.1016/j.compbiomed.2026.111465","url":null,"abstract":"<div><div>Breast cancer poses the most significant threat to women’s health, yet early detection through screening can markedly reduce mortality. Ultrasound imaging, with its affordability, non-invasiveness, and efficacy in dense breast tissue, has emerged as a crucial tool for early screening. Recent advancements in computer vision have spurred the development of computer-aided diagnostic systems that focus on the automated localization and diagnosis of breast lesions. However, challenges such as speckle noise, blurred boundaries, and low contrast in ultrasound images impede accurate lesion detection. This review examines recent studies on breast ultrasound lesion localization and diagnosis, emphasizing model feature construction. It provides an overview of the task, available datasets, and evaluation metrics, and outlines selection criteria through a comprehensive literature analysis. The review categorizes models into three groups: domain knowledge-driven, data-driven, and hybrid approaches. It also discusses current challenges and future directions, aiming to enhance the accuracy of breast lesion localization and diagnosis.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111465"},"PeriodicalIF":6.3,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146017584","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":"Between minds and machines: A neurocognitive comparison of human and chatbot interaction in language learning","authors":"Gamze Turun Ozel ,&nbsp;Semin Kazazoglu ,&nbsp;Burcak Yavuz ,&nbsp;Emir Rusen","doi":"10.1016/j.compbiomed.2026.111499","DOIUrl":"10.1016/j.compbiomed.2026.111499","url":null,"abstract":"<div><div>This study explores neurocognitive differences between human–human interaction (HHI) and human–chatbot interaction (HCI) during English-speaking tasks using EEG analysis. Results showed that HHI elicited significantly greater neural activation, particularly in the left frontal and temporal regions (F3, F7, T3), which are associated with language processing and social cognition. The F3 site exhibited the strongest difference (HHI: 27.09 vs. HCI: 15.5, p &lt; .001, d = −4.02). EEG band analysis revealed higher delta activity during HCI, indicating lower cortical arousal and attentional engagement, while HHI showed greater alpha and beta power (alpha: 6.5 % vs. 2.1 %; beta: 12.1 % vs. 2.4 %), reflecting enhanced cognitive processing and emotional salience. These patterns extended to central, temporal, and parieto-occipital regions, with consistently stronger beta activity in HHI. Findings suggest that natural human interaction elicits deeper and more distributed neural engagement than chatbot communication, offering key insights into the cognitive and emotional dimensions of technology-mediated language use in educational and social contexts.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111499"},"PeriodicalIF":6.3,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146017502","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":"Establishment of threshold of human gut microbes and risk assessment system for colorectal cancer","authors":"Wu Yinhang ,&nbsp;Zhuang Jing ,&nbsp;Qu Zhanbo ,&nbsp;Liu Jiang ,&nbsp;Zhou Qing ,&nbsp;Zhang Qi ,&nbsp;Jin Yin ,&nbsp;Song Jianwen ,&nbsp;Wu Wei ,&nbsp;Han Shuwen","doi":"10.1016/j.compbiomed.2026.111484","DOIUrl":"10.1016/j.compbiomed.2026.111484","url":null,"abstract":"<div><h3>Background</h3><div>Being involved in the occurrence of colorectal cancer (CRC), gut microbes are potential targets for early diagnosis of CRC. Defining the threshold of these characteristic bacteria could provide a basis for the clinical application of microorganisms as novel tumor markers for CRC.</div></div><div><h3>Objective</h3><div>To sort out and define the threshold of related bacteria and the ecological characteristics of gut bacteria.</div></div><div><h3>Methods</h3><div>A total of 8021 fecal samples from healthy people and 497 from CRC patients in the public database were collected to analyse the reference range. CRC-related bacteria and gut microbial characteristics were screened by literature review and analysed. CRC related bacteria and 5–95 % medians of gut microbial characteristics in healthy populations were used as reference value. 16S rRNA Miseq sequencing (175 CRC patients and 175 healthy people) and PacBio sequencing (200 CRC patients and 200 healthy people) were used to detect stool DNA sequence. The community composition of gut microbiota between CRC and healthy subjects was plotted; the species differences were analysed by Lefse analysis. R studio software was used to analyse CRC-related bacteria and gut microbial characteristics.</div></div><div><h3>Results</h3><div>A total of 218 CRC-associated bacteria and 15 gut microbial characteristics, such as enterotypes and Firmicutes/Bacteroidetes ratio, were reviewed and analysed. A 5–95 % threshold for these 218 CRC-associated bacteria and 15 gut microbiome signatures was developed to provide criteria for the normal range of gut bacteria. The CRC evaluation intelligent system software was developed and it could quickly calculate the value of 218 CRC related bacteria and 15 gut microbial characteristics using sequencing data, and assess whether they are within the threshold. And this software has the function of predicting CRC risk. The accuracy of CRC risk assessment ranged from 89.14 % to 91.50 %.</div></div><div><h3>Conclusion</h3><div>We established, for the first time, quantitative thresholds for CRC-associated bacteria and have driven advances in microbial risk prediction for CRC.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111484"},"PeriodicalIF":6.3,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146008911","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":"Generative AI in medicine: A thorough examination of applications, challenges, and future perspectives","authors":"S. Jayasrilakshmi,&nbsp;Ansuman Mahapatra","doi":"10.1016/j.compbiomed.2026.111469","DOIUrl":"10.1016/j.compbiomed.2026.111469","url":null,"abstract":"<div><div>Generative AI, an artificial intelligence, significantly transforms the healthcare sector. Recent breakthroughs in Generative AI include the use of language models and leveraging modern pre-trained Transformer models such as ChatGPT, Bard, LLaMA, DALL-E, and Bing. In medical applications, the advent of Large Language Models (LLMs) is a significant tool for predicting diseases, identifying risk factors, and enhancing diagnostic accuracy by analyzing a massive volume of unevenly distributed medical resources. This study provides a comprehensive review of existing literature on the use of LLMs in healthcare. It elucidates the ‘status quo’ of language models for general readers, healthcare professionals, and researchers. Specifically, this study investigates the capabilities of LLMs, including the transformation of healthcare consultation, enhancement of patient management and treatment, evolution of medical education, optimal resource utilization, and advancement of clinical research. The article organizes the literature based on human organs that will help readers quickly find relevant LLM applications for specific medical fields. The outcome of this survey will help medical professionals, researchers, and the healthcare industry understand the benefits, challenges, observed limitations, future challenges and applications of LLMs in healthcare.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111469"},"PeriodicalIF":6.3,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146008895","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":"Transformer-based feature extraction approach for hematopoietic cancer subtype classification","authors":"Kwang Ho Park ,&nbsp;Younghee Lee ,&nbsp;Wei Ding ,&nbsp;Kwang Sun Ryu ,&nbsp;Keun Ho Ryu","doi":"10.1016/j.compbiomed.2026.111466","DOIUrl":"10.1016/j.compbiomed.2026.111466","url":null,"abstract":"<div><div>Accurate classification of hematopoietic cancer subtypes remains challenging due to the multipotent nature of hematopoietic cells and the absence of definitive genetic markers. To address this, we propose a Transformer-based Autoencoder that captures compact and biologically informative embeddings from gene expression data. Specifically, our method employs multi-head self-attention in the encoder to learn complex nonlinear interactions among genes, with a reconstruction decoder that enforces biological feature retention. We benchmarked our approach against four widely-used feature extraction methods—Principal Component Analysis, Non-negative Matrix Factorization, Autoencoder, and Variational Autoencoder—using transcriptomic data from five hematopoietic cancer subtypes in The Cancer Genome Atlas, totaling 2452 samples. Data were split 60:20:20 into training, validation, and test sets with stratification, and feature-extractor hyperparameters were chosen on the validation set<strong>.</strong> Each method produced 100-dimensional feature vectors, subsequently evaluated using eight multi-class classifiers: Light Gradient Boosting Machine, Extreme Gradient Boosting, Logistic Regression, Random Forest, Decision Tree, Support Vector Machine, and Neural Networks. On the independent test set, the Transformer-based Autoencoder embeddings combined with Light Gradient Boosting Machine achieved F1-score: 0.969, accuracy: 0.986, precision: 0.975, recall: 0.964, specificity: 0.996, G-mean: 0.980, and balanced accuracy: 0.954. For context, we additionally included a supervised tabular Transformer (FT-Transformer) as a reference; while strong, it is not directly comparable to our unsupervised feature extractor. To enhance interpretability and clinical relevance, we applied Shapley Additive exPlanations to identify the twenty most influential genes contributing to subtype discrimination. This analysis revealed key biomarkers related to endoplasmic reticulum function, antigen processing, and ribonucleic acid regulation. These findings demonstrate that transformer-based unsupervised feature extraction substantially improves predictive accuracy and yields valuable biological insights for complex hematologic malignancies. Overall, the study supports attention-driven representation learning for tabular biomedical data and motivates future work in generative/self-supervised representations for gene expression.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111466"},"PeriodicalIF":6.3,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146009068","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":"Helical static-mixer insert for pediatric and neonatal gas blending: RANS-CFD comparison of commercial and in-house monolithic designs","authors":"Shirley Ferraz Crispilho ,&nbsp;Paulo Cesar Duarte Junior ,&nbsp;Martin Poulsen Kessler ,&nbsp;Rudolf Huebner ,&nbsp;Altibano Ortenzi","doi":"10.1016/j.compbiomed.2026.111475","DOIUrl":"10.1016/j.compbiomed.2026.111475","url":null,"abstract":"<div><div>Accurate blending of oxygen and air in pediatric and neonatal respiratory support depends on compact connectors that promote efficient mixing without generating excessive pressure drop or dead volume. In current clinical practice, commercially available T-shaped connectors are often used as passive mixers, but their internal geometry was not originally optimized for this purpose. In this work, the original commercial connector (Geometry A), an in-house modified multi-part connector incorporating a static insert (Geometry B), and a new monolithic helical static-mixer insert (Geometry C) were evaluated under identical flow conditions. Three-dimensional Reynolds-averaged Navier–Stokes computational fluid dynamics simulations were performed to represent oxygen–nitrogen mixing in high-flow nasal cannula circuits, considering realistic flow rates and boundary conditions. For each geometry, mixture quality at the outlet was assessed from the spatial distribution of species mass fraction, hydraulic performance was quantified by the device pressure drop, and residence-time behavior for the helical insert was obtained from a transient scalar-pulse simulation. Geometry B improved outlet homogeneity relative to the commercial connector but required several assembled parts, which complicates handling and sterilization. Geometry C, designed as a monolithic helical static mixer, produced more uniform gas mixing than both previous configurations while maintaining pressure drops within ranges compatible with pediatric and neonatal use. These results indicate that the proposed helical insert has the potential to replace the current multi-part in-house adaptation and to offer a more effective alternative to standard commercial connectors when implemented as a monolithic medical-grade component.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111475"},"PeriodicalIF":6.3,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146008969","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":"Multimodal diagnosis of Parkinson’s disease with an internet-based collaborative agent architecture of medical language models","authors":"Eugenio Peixoto Junior ,&nbsp;Felipe Cordeiro de Sousa ,&nbsp;Junxin Chen ,&nbsp;David Camacho ,&nbsp;Stephen Rathinaraj Benjamin ,&nbsp;Victor Hugo C. de Albuquerque","doi":"10.1016/j.compbiomed.2026.111468","DOIUrl":"10.1016/j.compbiomed.2026.111468","url":null,"abstract":"<div><div>Parkinson’s disease (PD) remains one of the most prevalent neurodegenerative disorders, where delays in diagnosis compromise therapeutic outcomes and increase healthcare costs. Conventional unimodal approaches, based on voice, sensors, or imaging, face critical limitations, including small datasets, lack of reproducibility, and high infrastructure demands. To address these challenges, the proposed multimodal agent-based architecture integrates medical language models, audio signals, and neuroimaging, and is supported by data–machine learning pipelines and an edge–cloud infrastructure. The system leverages ensemble learning, large and vision language models, and Retrieval-Augmented Generation (RAG) to enhance clinical decision support. The transparency of the model was supported by explainability techniques (SHapley Additive exPlanations, permutation importance, partial dependence, and individual conditional expectation), which highlighted the main audio and sensor variables responsible for the predictions. Experimental evaluation confirmed the effectiveness of multimodal fusion. When integrated, the architecture achieved robust performance, with an accuracy of 0.86, an F1-score above 0.88, ROC-AUC greater than 0.93, and both sensitivity and specificity above 0.89. Calibration and hypothesis tests were validated by a low Brier score of 0.205 and an Expected Calibration Error of 0.151, while Decision Curve Analysis confirmed clinical relevance by minimizing false negatives, critical for early screening, and reducing redundant interventions. Multimodal fusion produced accurate, well-calibrated, and interpretable risk estimates for PD screening; larger prospective studies and cost-effectiveness analyses are needed to consolidate clinical applicability.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111468"},"PeriodicalIF":6.3,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146009093","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":"Unveiling the renal therapeutic potential of Nypa fruticans leaves: An integrated experimental and in silico approach","authors":"Farhana Islam ,&nbsp;Mostafa Kamal ,&nbsp;Shoeb Ahmad ,&nbsp;Masum Shahriar ,&nbsp;Fariya Islam Rodru ,&nbsp;Md. Nazmul Hasan ,&nbsp;Md. Nazmul Hasan Zilani ,&nbsp;Md. Ataur Rahman ,&nbsp;Shahad Saif Khandker ,&nbsp;Saquiba Yesmine","doi":"10.1016/j.compbiomed.2026.111482","DOIUrl":"10.1016/j.compbiomed.2026.111482","url":null,"abstract":"<div><div>Chronic kidney disease (CKD) is a progressive, irreversible disorder associated with renal dysfunction, inflammation, and oxidative stress. Given the limitations of current therapies, this study assessed the renal curative effects of <em>Nypa fruticans</em> ethyl acetate leaf extract (EaNFL) in a gentamicin-induced nephrotoxicity rat model. GC‒MS and HPLC analyses identified 23 bioactive compounds in EaNFL, including rosmarinic acid, quercetin, and (−)-epicatechin, which were selected based on ADMET profiling, Lipinski's rule, and DFT analysis. These compounds were further investigated through computational studies against two renal targets: the AT1 receptor (PDB ID: 4YAY) and SGLT2 (PDB ID: 7VSI). Treatment with EaNFL, particularly at 400 mg/kg body weight and in combination therapy, significantly improved renal function and normalized biochemical and hematological parameters, likely due to its potent antioxidant and anti-inflammatory properties. Histopathological data supported these findings, showing reduced tubular necrosis, glomerular damage, and inflammation, especially in the high-dose groups. DFT analysis revealed that rosmarinic acid had the highest HOMO–LUMO energy gap (ΔE = 0.1314 eV), suggesting high chemical reactivity and potential biological compatibility. Molecular docking identified quercetin, rosmarinic acid, and (−)-epicatechin as the top binders, with rosmarinic acid showing the strongest affinity and forming a stable complex, as confirmed by 100 ns MDS. Taken together, the <em>in vivo</em> and <em>in silico</em> results indicate that EaNFL offers renoprotective benefits by targeting the RAAS and glucose transport pathways while also mitigating oxidative stress and inflammation. These findings demonstrate its therapeutic potential and warrant further investigation into its bioactive constituents and potential clinical use in renal treatment.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111482"},"PeriodicalIF":6.3,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146009019","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":"Multi-task non-contact ballistocardiogram-based vital signs monitoring in acupuncture","authors":"Truong Tien Vo ,&nbsp;Quy Phuong Le ,&nbsp;Trong Nhan Nguyen ,&nbsp;Jaeyeop Choi ,&nbsp;Sudip Mondal ,&nbsp;Byeongil Lee ,&nbsp;Junghwan Oh","doi":"10.1016/j.compbiomed.2026.111461","DOIUrl":"10.1016/j.compbiomed.2026.111461","url":null,"abstract":"<div><div>The study introduces an innovative approach for efficient vital signs monitoring in acupuncture by combining multi-channel ballistocardiogram (BCG) signals and multi-task learning, taking advantage of the polyvinylidene fluoride (PVDF) film sensor and deep neural networks. The proposed system utilizes non-contact under-mattress BCG signals and deep learning for heart rate (HR), respiration rate (RR) estimation and lying posture detection. A custom-designed data-logger captures the signal from a BCG sensor located under the patient’s back for data acquisition, and integrates Gated Recurrent Unit (GRU) and Multi-head Self-Attention (MHSA) deep learning mechanisms for efficient HR, RR estimation and posture classification. In experiments with 25 participants, the proposed method achieved 98.7% accuracy for activity recognition and 97.6% for lying posture classification. In HR and RR estimation, the best case of mean absolute error (MAE) for HR achieves 0.77 beats per minute (bpm) in the right lateral posture, while the best value of MAE for RR is 0.43 breaths per minute (brpm) in the seated posture, compared to an FDA-approved device. The results demonstrate the high performance of multi-task learning for vital signs estimation and posture classification with our BCG-based system. This work establishes an innovative and practical pathway for medical assistance tools in non-contact monitoring and management.</div></div>","PeriodicalId":10578,"journal":{"name":"Computers in biology and medicine","volume":"203 ","pages":"Article 111461"},"PeriodicalIF":6.3,"publicationDate":"2026-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145997559","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}