Healthcare Applications Research Articles (Page 1)

Atrial fibrillation (AF) is a prevalent arrhythmia with significant health risks, often underdiagnosed due to limitations in traditional screening methods. This study investigates the effectiveness of an AI-based electronic stethoscope for AF screening, comparing it to other portable devices. A retrospective study was conducted using 496 cardiac sound recordings from patients with and without AF. The recordings were divided into derivation and validation datasets. An AI model, combining ResNet34 and a 12-layer Vision Transformer (ViT), was developed and trained on the derivation dataset. The model's performance was evaluated using sensitivity, specificity, accuracy, positive and negative predictive values, and the area under the receiver operating characteristic (ROC) curve (AUC). Additionally, a non-consecutive day twice cardiac sound collection was performed on 74 samples to assess the model's consistency. The AI model achieved high performance metrics in both derivation and validation datasets. In the derivation dataset, sensitivity was 0.95 (95% CI, 0.90-0.97), specificity was 0.90 (95% CI, 0.83-0.94), accuracy was 0.92 (95% CI, 0.90-0.96), positive predictive value was 0.92 (95% CI, 0.87-0.96), and negative predictive value was 0.93 (95% CI, 0.86-0.96). In the validation dataset, sensitivity was 0.94 (95% CI, 0.88-0.98), specificity was 0.91 (95% CI, 0.83-0.96), accuracy was 0.93 (95% CI, 0.89-0.96), positive predictive value was 0.93 (95% CI, 0.86-0.97), and negative predictive value was 0.93 (95% CI, 0.85-0.97). The AUC for the derivation dataset was 0.92 (95% CI, 0.89-0.96), and for the validation dataset, it was 0.93 (95% CI, 0.88-0.97). The non-consecutive day cardiac sound collection resulted in a Cohen's Kappa value of 0.74, indicating good consistency in the model's judgments. The AI-based electronic stethoscope shows promise as a reliable and accessible tool for AF screening, with potential applications in primary healthcare and general population screening.

To provide an up-to-date review of existing and current literature in the field of radiological and nuclear disasters to support the needs of research applications for health care and public health preparedness and response. A systematic literature search using 4 databases to identify articles which included a multitude of topics relevant to preparedness for nuclear and radiological disasters. One hundred articles that met inclusion criteria were summarized into 7 themes addressing medical and health care preparedness for nuclear and radiological events. The review generated evidence supporting and defining various measures health care and government entities can take to improve nuclear and radiological disaster readiness and responsiveness in health systems. Strengthening preventive measures and policies, prehospital and hospital mechanisms, training and education, regional collaboration, communication, and infrastructure support were the main gaps identified. An overarching concern regarding the inadequacies of the modern health care system's radiological disaster preparedness was a clear-cut conclusion from the literature. The major challenges and proposed solutions for public safety to the growing threat of radiologic disasters were identified.

The primary objective of deep learning is to have good performance on a large dataset. However, when the model lacks sufficient data, it becomes a challenge to achieve high accuracy in predicting these unfamiliar classes. In fact, the real-world dataset often introduces new classes, and some types of data are difficult to collect or simulate, such as medical images. A subset of machine learning is meta learning, or "learning-to-learn", which can tackle these problems. In this paper, a few-shot classification model is proposed to classify three types of brain cancer: Glioma brain cancer, Meningioma brain cancer, and brain Tumor cancer. To achieve this, we employ an episodic meta-training paradigm that integrates the model-agnostic meta-learning (MAML) framework with a prototypical network (ProtoNet) to train the model. In detail, ProtoNet focuses on learning a metric space by computing distances to class prototypes of each class, while MAML concentrates on finding the optimal initialization parameters for the model to enable the model to learn quickly on a few labeled samples. In addition, we compute and report the average accuracy for the baseline and our methods to assess the quality of the prediction confidence. Simulation results indicate that our proposed approach substantially surpasses the performance of the baseline ResNet18 model, achieving an average accuracy improvement from 46.33% to 92.08% across different few-shot settings. These findings highlight the potential of combining metric-based and optimization-based meta-learning techniques to improve diagnostic support in healthcare applications.

Microphones are sensors common to a variety of the Internet of Things (IoT) and healthcare applications. Many examples have proved that microphones can be useful in detecting, e.g., abnormal breathing rates. There are already applications that serve this purpose, such as respiratory acoustic monitoring and ResApp. Breath signals have been studied using a range of technologies and sensors, including the most common: radar, accelerometer, and wearables. The majority of these sensors are attached to the body of a monitored person. However, the emergence of COVID-19 has drawn particular attention to the importance of using non-contact technologies for monitoring breath signals and other vital signs. This paper presents a comprehensive review of microphone-based non-contact vital sign monitoring, including the methodologies and concepts, while identifying new research gaps and opportunities for future studies.

Objectives This narrative review evaluates the role of artificial intelligence (AI) in healthcare, summarizing its historical evolution, current applications across medical and surgical specialties, and implications for allied health professions and biomedical research. Methods We conducted a structured literature search in Ovid MEDLINE (2018–2025) using terms related to AI, machine learning, deep learning, large language models, generative AI, and healthcare applications. Priority was given to peer-reviewed articles providing novel insights, multidisciplinary perspectives, and coverage of underrepresented domains. Key findings AI is increasingly applied to diagnostics, surgical navigation, risk prediction, and personalized medicine. It also holds promise in allied health, drug discovery, genomics, and clinical trial optimization. However, adoption remains limited by challenges including bias, interpretability, legal frameworks, and uneven global access. Contributions This review highlights underexplored areas such as generative AI and allied health professions, providing an integrated multidisciplinary perspective. Conclusions With careful regulation, clinician-led design, and global equity considerations, AI can augment healthcare delivery and research. Future work must focus on robust validation, responsible implementation, and expanding education in digital medicine.

Artificial Intelligence (AI)-based prediction models of treatment response promise to revolutionize psychiatric care by enabling personalized treatment, but very few have been thoroughly tested in different samples or compared to current clinical standards. Here we present models predicting antipsychotic response and assess their clinical utility in a robust methodological framework. Machine learning models were trained and cross-validated on clinical and sociodemographic data from 594 individuals with established schizophrenia (NCT00014001) and 323 individuals with first episode psychosis (NCT03510325). Models predicted four measures of antipsychotic response at 3 months after baseline. Clinical utility was assessed using decision curve and calibration curve analyses. Model performance was tested in a reduced feature space and across sex, ethnicity, antipsychotic, and symptom change subgroups to investigate model fairness. Models predicting total symptom severity (r = 0.4-0.68) and symptomatic remission (BAC = 62.4%-69%) performed well in both samples and externally validated successfully in the opposing cohort (r = 0.4-0.5, BAC = 63.5%-65.7%). Performance remained significant when the models were reduced to 8-9 key variables (r = 0.53 for total symptom severity, BAC = 65.3% for symptomatic remission). Models predicting symptomatic remission had a net benefit across risk thresholds of 0.5-0.9 and were moderately well-calibrated (ECE = 0.16-0.18). Model performance different across sex, ethnicity and medication subgroups. We present a robust framework for training and assessing the clinical utility of prediction models in psychiatry. Our models generalize across different psychosis populations and show promising calibration and net benefit. However, performance disparities across demographic and treatment subgroups highlight the need for more diverse clinical samples to ensure equitable prediction.

Data for healthcare applications are typically customized for specific purposes but are often difficult to access due to high costs and privacy concerns. Rather than prepare separate datasets for individual applications, we propose a novel approach: building a general-purpose generative model applicable to virtually any type of healthcare application. This generative model encompasses a broad range of human attributes, including age, sex, anthropometric measurements, blood components, physical performance metrics, and numerous healthcare-related questionnaire responses. To achieve this goal, we integrated the results of multiple clinical studies into a unified training dataset and developed a generative model to replicate its characteristics. The model can estimate missing attribute values from known attribute values and generate synthetic datasets for various applications. Our analysis confirmed that the model captures key statistical properties of the training dataset, including univariate distributions and bivariate relationships. We demonstrate the model’s practical utility through multiple real-world applications, illustrating its potential impact on predictive, preventive, and personalized medicine.

Data for healthcare applications are typically customized for specific purposes but are often difficult to access due to high costs and privacy concerns. Rather than prepare separate datasets for individual applications, we propose a novel approach: building a general-purpose generative model applicable to virtually any type of healthcare application. This generative model encompasses a broad range of human attributes, including age, sex, anthropometric measurements, blood components, physical performance metrics, and numerous healthcare-related questionnaire responses. To achieve this goal, we integrated the results of multiple clinical studies into a unified training dataset and developed a generative model to replicate its characteristics. The model can estimate missing attribute values from known attribute values and generate synthetic datasets for various applications. Our analysis confirmed that the model captures key statistical properties of the training dataset, including univariate distributions and bivariate relationships. We demonstrate the model's practical utility through multiple real-world applications, illustrating its potential impact on predictive, preventive, and personalized medicine.

Introduction: The metaverse, a convergence of virtual, augmented, and physical realities, is revolutionizing healthcare delivery, education, and patient engagement. Its backbone technologies include virtual reality (VR), augmented reality (AR), extended reality (XR), artificial intelligence (AI), blockchain, and the internet of things (IoT). Methods: A qualitative synthesis was conducted using peer-reviewed literature retrieved from electronic databases, including PubMed, Scopus, IEEE Xplore, Web of Science, and Google Scholar. The search was restricted to studies published between 2010 and 2023, focusing on metaverse applications in healthcare, such as surgery, education, diagnostics, and telemedicine. Results: Findings revealed the metaverse’s integration into various domains,vsuch as XRassisted surgeries (e.g., Johns Hopkins' AR spine surgery), immersive VR-based rehabilitation, AR-enhanced diagnostics, and AI-driven simulations. Platforms, like Tetra Signum and WHO’s XR training programs, have demonstrated clinical efficacy. Holographic modeling and digital twins have been found to be increasingly used in surgical planning and remote consultation. Discussion: The metaverse has been found to foster real-time, multimodal interaction among clinicians and patients. However, issues, such as data privacy, interoperability, access disparities, and legal ambiguity, challenge full-scale adoption. Ethical implementation and infrastructure upgrades are crucial for equitable integration. Conclusion: Metaverse technologies are transforming traditional medicine into a proactive, personalized, and data-driven system. By embedding immersive experiences into clinical and educational workflows, they promise enhanced outcomes and democratized access to healthcare knowledge. Strategic policies and ethical safeguards are essential to unlock their full potential.

This study investigates the synthesis of silk fibroin nanoparticles (SFNPs) and silver nanoparticles (AgNPs) and their application to cotton textiles to enhance functional properties for potential biomedical use. The nanoparticles were synthesized using chemical reduction and nanoprecipitation methods, and their formation and stability were confirmed through UV–Vis spectroscopy, X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), and scanning electron microscopy with energy-dispersive spectroscopy (SEM–EDS). Cotton fabrics were subsequently modified with SFNPs, AgNPs, and a combined SF-AgNPs formulation. Characterization confirmed the successful deposition and interaction of nanoparticles with cellulose fibers. The treated textiles demonstrated improved antibacterial activity against Staphylococcus aureus and Escherichia coli, along with enhanced antioxidant performance as evidenced by DPPH radical scavenging assays. Notably, the combined SF-AgNPs treatment exhibited synergistic effects, providing stronger antimicrobial durability and higher antioxidant capacity compared to single-nanoparticle treatments. These findings highlight the potential of SFNPs and AgNPs as effective nanomaterials for producing multifunctional, bioactive cotton textiles with promising applications in healthcare and biomedical fields.

Sterilization of medical devices is a critical process to ensure patient safety. However, traditional steam autoclaves may be unsuitable for heat-sensitive materials. In this study, we evaluated an innovative cold sterilization system based on the controlled generation of free radicals with reducing properties. The system has already been validated and marketed following the completion of numerous microbiological tests in compliance with UNI EN standards (13727, 13624, 17126, 14476, 14348). A quantitative suspension test was conducted under controlled conditions, comparing the microbial reduction achieved with the cold system to that obtained with a standard autoclave cycle. The system demonstrated bactericidal efficacy exceeding 6 log10, comparable to that of the autoclave cycle. The results suggest that the free radical system represents a safe, rapid, and effective alternative for the sterilization of heat-sensitive materials, with potential applications in both healthcare and industrial settings.

The tremendous evolvement in AI and ML has started to revolutionize clinical biomedicine by enhancing diagnostic precision and decision-making. This paper proposes a multimodal machine learning framework for image- based clinical diagnosis, integrating multiple biomedical data sources such as medical imaging, clinical records, and laboratory parameters. Image features are extracted using deep learning architectures like Convolutional Neural Networks while clinical and numerical data are processed through natural language processing and statistical models. Fusion of these heterogeneous modalities helps develop a better understanding of patient health conditions. Experimental evaluations demonstrate that multimodal integration significantly improves diagnostic performance compared to unimodal approaches, providing higher accuracy and robustness. The proposed model hence points to the potential of AI-driven multimodal systems for assisting clinicians in early disease detection, reducing diagnostic errors, and supporting precision medicine applications in modern healthcare.

Artificial Intelligence (AI) is changing the way businesses approach data analysis and decision-making in a big way. This paper explores how AI is being applied across sectors, particularly in healthcare, and examines the benefits and challenges that accompany it. One key takeaway is that AI enhances operational efficiency and improves forecasting accuracy, making it easier for industries such as finance, retail, and healthcare to deliver personalised services. In healthcare, AI plays a crucial role in predicting patient outcomes, guiding clinical decisions, and optimising resource allocation. These advancements can lead to significantly better patient care. However, there are hurdles to overcome. Issues such as data governance, security, algorithmic bias, and organisational resistance can slow the effective implementation of AI. Additionally, ethical aspects such as transparency, accountability, and fairness are ultimately vital for ensuring AI is used responsibly. The paper stresses the importance of strong governance frameworks, committed leadership, and a workforce that is ready to adapt. These elements are necessary to ensure that AI deployment meets societal and regulatory standards. Looking ahead, AI enables organisations to transition from merely reacting to data to predicting and shaping outcomes, thereby promoting innovation and enhancing long-term competitiveness. The future of AI in business analytics hinges on responsible usage. It is crucial to balance technological progress with ethical considerations and social responsibility to create fair, sustainable value in our increasingly data-driven world.

Angle-based positioning systems have emerged as critical technologies for precise indoor localization across robotics, healthcare, and industrial automation applications. Ultrawideband (UWB) phase-based angle measurements offers theoretical sub-degree accuracy, but practical implementations suffer from channel inconsistency errors that significantly degrade performance. A dual-layer Bayesian neural network fusion framework (DBNNFF) was presented that effectively addresses these systematic errors through an innovative combination of physical constraints and uncertainty-aware modeling. Experiments were conducted in a microwave anechoic chamber using a customized 5-channel UWB base station and single-channel tags. Data was collected across seven azimuth angles between ± 30°, with 30s cold-start cycles per angle. The DBNNFF framework reduced the angle errors by 94.7% to 0.1036° ± 0.0182°, outperforming many existing algorithms by 25–42.1%. The framework’s dual-network architecture—combining channel correlation model and cold start state distribution estimator—with uncertainty-weighted Bayesian fusion provides well-calibrated confidence intervals and exceptional noise robustness. Experiments conducted in multi-path environments such as office and hallway demonstrated that the DBNNFF algorithm exhibited robust performance, with errors maintained within 0.17°.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-22420-5.

Abstract Airborne particulate matter (PM) poses growing threats to human health and the environment, especially in the context of emerging infectious diseases. Conventional petroleum‐based filters are constrained by an inherent trade‐off between filtration efficiency and air resistance, and additionally suffer from a lack of mechanical durability, thermal protection, and environmental degradability. Here, a multifunctional triboelectric‐assisted aerogel filter is fabricated through a synergistic approach combining cell wall nano‐reconstruction with diatom‐inspired mineralization. This strategy constructs a hierarchical porous framework that regulates charge storage sites, leading to a 71% improvement in charge retention compared with native wood and a significant enhancement in triboelectric output. As a self‐powered filtration platform, the system achieves outstanding removal efficiencies for PM 0.3 (98.75% ± 0.08%), PM 0.5 (99.51% ± 0.21%), and PM 1 (99.98% ± 0.02%) with a low pressure drop of 53 Pa. Meanwhile, the aerogel demonstrates remarkable compressive resilience (18.1 MPa at 60% radial strain) and complete elastic recovery, together with robust thermal insulation and flame retardancy. Notably, the integrated triboelectric signal is analyzed using deep learning algorithms to identify respiratory patterns, thereby enabling real‐time health monitoring and intelligent respiratory management. This work establishes a sustainable, high‐performance material platform for advanced air purification and wearable healthcare applications in extreme environments.

ABSTRACT Wireless body sensor networks (WBSNs) are increasingly used in healthcare for remote monitoring of patients. Although these systems improve access to medical care, they also face serious challenges related to data security and patient authentication. This study proposes a lightweight and secure authentication framework based on a Three‐Tier Secure Message Authentication Code (TTSMAC) protocol. The framework combines three key techniques: Factorized RSA (FRSA) for efficient key generation, Length Pearson Hashing (LPH) for secure token management, and Dual Secret Key Elliptic Curve Cryptography (DSK‐ECC) for protecting stored data. Experimental results showed that the proposed framework reduces encryption/decryption time, lowers key setup overhead, and achieves higher throughput compared with existing methods. Also, the performance evaluations showed substantial improvements in encryption/decryption times and throughput, demonstrating the framework's suitability for resource‐constrained, battery‐powered wearable sensors. Overall, the framework enhances security, maintains patient data privacy, and ensures reliable authentication for WBSN‐based healthcare applications.

Emotion recognition and sentiment analysis are crucial tasks in natural language processing, enabling machines to understand human emotions and opinions. However, the complex, nuanced relationship between emotions and sentiment in conversation poses significant challenges to accurate emotion recognition, as sentiment cues can be easily misinterpreted. Deploying emotion recognition and sentiment analysis tasks on edge devices poses substantial challenges due to computational resource constraints. We present an adaptive multitask learning approach that jointly leverages resource-constrained Mobile Bidirectional Encoder Representations from Transformers (MobileBERT) and Distilled BERT (DistilBERT) models to optimise emotion recognition and sentiment analysis. Our proposed approach utilises prototypical networks to learn effective representations of emotions and sentiment, while a focal weighted loss function effectively mitigates the class imbalance. We adaptively fine-tune the learning process to balance task importance and resource utilisation, resulting in better performance and efficiency. Our experimental results demonstrate the efficacy of our method, achieving the best results on MELD and IEMOCAP benchmark datasets while keeping a compact model size. Despite limited computational demands, our solution demonstrates that emotion and sentiment analysis can deliver performance comparable to resource-intensive large language models (LLMs). Facilitating various applications in human-computer interaction, affective computing, social media, dialogue conversion, and healthcare.

Recently, the integration of smartphone-based platforms into biomedical sensing has provided portable, low-cost, and scalable alternatives to conventional laboratory diagnostics. According to the advances in mobile imaging, embedded sensors, microfluidics, and wireless connectivity, these systems enable the real-time detection and quantitativedetermination of a wide range of biological and chemical targets, including nucleic acids, proteins, cells, and pathogens, directly at the point of care. In this review, we summarize the existing smartphone-integrated biosensing technologies, with a focus on studies published since 2020. We examine various sensing modalities, including optical, e.g., brightfield, fluorescence, dark-field, electrochemical, and hybrid systems, as well as enabling components such as CMOS imagers, illumination sources, and microfluidic chips. We also highlight the integration of artificial intelligence (AI) and deep learning, which enhance diagnostic accuracy through image enhancement, modality translation, and automated quantification. Despite the rapid progress in this field, challenges still exist in terms of hardware variability, standardization, and clinical validation. In the future, the development of modular attachments, such as open-source hardware, and cloud-connected analytics will be akey to scaling smartphone-based diagnostics for decentralized healthcare, environmental monitoring, and global biosensing applications. This review also provides important biomedical applications of smartphone-based biosensing platforms, including the detection of ocular, metabolic, and urological diseasesas well as various cancers, infectious agents, and exosome-derived biomarkers.

This study presents the development and evaluation of MedOne, an AI-powered mobile healthcare application designed to improve healthcare accessibility in Rwanda. MedOne integrates AI-driven diagnostic tools with teleconsultation services, aiming to address critical healthcare challenges in resource-limited settings. The research employs a mixed-methods approach involving 247 participants, including healthcare professionals, end users, and administrators. The system incorporates machine learning algorithms for symptom assessment, natural language processing for multi-language support, and cloud-based architecture for scalability. Findings suggest the system could significantly reduce consultation times by 34%, increase rural healthcare consultations by 67%, and achieve a diagnostic accuracy of 78.5%. The system's design incorporates offline functionality, multi-language support, and cultural adaptation for the Rwandan context.

Current brain-dedicated positron emission tomography (PET) systems (e.g., the High-Resolution Research Tomograph, HRRT) are unable to accurately and precisely measure pharmacologically-specific signals in small brain regions due to insufficient image resolution and sensitivity. The NeuroEXPLORER (NX), a new ultra-high-performance brain-dedicated scanner, promises to address these needs. Seven healthy individuals underwent paired scans on the HRRT and NX with targeted radiopharmaceuticals (18F-FDG; 18F-SynVesT-1; 18F-FPEB; 18F-Flubatine; 11C-PHNO; 18F-FE-PE2I; 11C-DASB). Early (0-10min) and late standard uptake value (SUV) images were visually compared between scanners. The exceptional spatial resolution of the NX can be appreciated in the details of the cortical ribbon and subcortical nuclei (e.g., mediodorsal thalamus) for 18F-FDG (glucose metabolism), 18F-SynVesT-1 (synaptic density), 18F-FPEB (glutamate receptors mGluR5), which have high uptake across gray matter. Tracers with more focal uptake targeting the dopamine system (11C-PHNO, 18F-FE-PE2I) displayed unprecedented anatomical detail (e.g., in the D3 receptor-rich mammillo-thalamic tract and anteroventral thalamus). Similarly, 18F-Flubatine (β2* nicotinic acetylcholine receptors) displayed clear uptake in the brainstem (e.g., inferior olivary nuclei), and 11C-DASB (serotonin transporters) markedly improved delineation of raphe nuclei in the brainstem and multiple cortical areas (e.g., temporal poles, subcallosal area). High-resolution early images of tracer delivery could be obtained from all tracers, especially those with high extraction. We compared measurements of tracer uptake between an ultra-high-performance brain-dedicated PET system (NX) and the previous state-of-the-art system (HRRT) in the same subjects for seven different tracers, demonstrating a substantial gain in image detail, especially for small brain structures. We also discussed the implications of this technology for basic and clinical brain PET research and potential healthcare applications. Not applicable.