Smart Health最新文献

{"title":"Prediction of body physiological parameter changes after experiencing infinity walk: A prospective observational view","authors":"K. Murugavalli ,&nbsp;R. Ramalakshmi ,&nbsp;G. Vishnuvarthanan ,&nbsp;M. Pallikonda Rajasekaran","doi":"10.1016/j.smhl.2025.100609","DOIUrl":"10.1016/j.smhl.2025.100609","url":null,"abstract":"<div><div>To develop a clinical prediction tool for estimating the changes in the body physiological parameters during six months for the people practicing infinity walk. Comparing the parameters of normal walk and infinity walk for recording progression and forecasting conversion in the physiological parameters. All eight subjects pursue a balanced diet as per the instructions delivered by dietitian. Four participants planned to practice the infinity walk, in the open or in a room, going from south to north and vice versa. Another four participants are advised to practice regular walking. All the individuals walked for 30 min in both morning and evening. Eight participants, comprising 4 men and 4 women. The participants, who included healthy individuals, people with diabetes, people with hypertension and those with both diabetes and hypertension, ranged in age from 30 to 60. The primary outcome is changes in weight, blood sugar level, blood pressure, and oxygen saturation when infinity walk is employed as a stimulus. In this prospective observational study, four of the subjects in each group are practicing the infinity walk, while the remaining four are walking normally for an hour each day (morning and evening). Physiological data are collected weekly once by clinician for analysis. Statistical analysis has been performed using linear regression and Wilcoxon Rank sum test in order to predict the changes in the physiological parameters in the future, when the subjects continuing the procedure for a long duration. In order to increase prediction accuracy, the Generalized Regression Neural Network (GRNN) and the Gaussian Process Regression (GPR) model are combined to create the GPR-GRNN. The statistical analysis states that the better changes in the body physiological parameters can be identified on the subject's performing infinity walk than the subjects practicing normal walk. Participants who perform infinite walking over regular walking show progressively better results. The findings of this study laid the groundwork for a reliable prediction of improvements in body physiological parameters over a period of time when infinite walking is practiced. Also, the results can assist in justifying the reason for the suggestion/preference given by the doctors/researchers to patients/common people to select infinity walk as an exercise.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100609"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266258","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"iSPHYNCS: Enabling long-term actigraphy with vital signals in a comparison between Fitbit Inspire 2/HR and MotionWatch 8 for non-parametric circadian rhythm analysis","authors":"Rafael Morand ,&nbsp;Oriella Gnarra ,&nbsp;Julia van der Meer ,&nbsp;Jan D. Warncke ,&nbsp;Annina Helmy ,&nbsp;Livia G. Fregolente ,&nbsp;Elena Wenz ,&nbsp;Kseniia Zub ,&nbsp;Lorenzo Brigato ,&nbsp;Claudio L.A. Bassetti ,&nbsp;Stavroula Mougiakakou ,&nbsp;Markus H. Schmidt","doi":"10.1016/j.smhl.2025.100614","DOIUrl":"10.1016/j.smhl.2025.100614","url":null,"abstract":"<div><div>Actigraphy is a tool to study an individual’s rest-activity rhythm, but its use is limited secondary to both cost and its availability to only specialized medical centers. Consumer-grade activity trackers are widely used by the general population. However, it remains to be determined if these devices provide equivalent information compared to traditional actigraphy and if the additional measurements of vital signals may further facilitate the identification of specific sleep disorders. We propose a new approach to perform long-term actigraphy with consumer-grade activity trackers that provide a broader array of physiological vital signals. We recorded one week of simultaneous actigraphy (MotionWatch 8) and activity tracking (Fitbit Inspire HR/2) in a study population (n <span><math><mo>=</mo></math></span> 40) that included individuals with central disorders of hypersomnolence and healthy controls. We leverage information from the individuals’ heart rate, step count, and calorie consumption measured by Fitbit to calculate non-parametric circadian rhythm analysis. Performing Bland-Altman analyses to assess the agreement between the clinical actigraphy and our proposed methods, we show that calorie consumption of Fitbit offers an acceptable alternative to the actigraphy device, superior to step count. Our results suggest that consumer-grade activity trackers have the potential to expand current medical device-based actigraphy by enabling additional signal detection and longer recording times at a lower cost.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100614"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145226830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Significant clinical factors for mortality prediction in ICU sepsis patients: A machine learning approach","authors":"Camilo Santos ,&nbsp;Judy Ximena Ramos Garzón ,&nbsp;Said D. Pertuz ,&nbsp;Carlos A. Fajardo","doi":"10.1016/j.smhl.2025.100613","DOIUrl":"10.1016/j.smhl.2025.100613","url":null,"abstract":"<div><h3>Background:</h3><div>Sepsis is a leading cause of ICU mortality worldwide, requiring early risk identification for improved outcomes. However, the complexity of sepsis limits the clinical applicability of predictive models. For machine learning (ML) models to be useful in practice, they must be robust performing consistently across different ICU settings and actionable, meaning they are feasible to implement with routinely used clinical factors. This study aims to identify significant clinical predictors to develop actionable models for robust mortality prediction in sepsis patients across diverse clinical settings.</div></div><div><h3>Methodology:</h3><div>This retrospective study analyzed data from 15,100 ICU patients in the MIMIC-IV v3.0 dataset for training and 8,201 patients in the eICU v2.0 dataset for external validation. Eight ML models were trained, and the best-performing model was selected based on the highest AUC. Significant clinical factors were identified using odds ratios and adjusted odds ratios to assess their predictive association.</div></div><div><h3>Results:</h3><div>The best-performing model achieved an AUC of 0.84 (MIMIC-IV) and 0.75 (eICU). Sequential Feature Selection reduced the model to 25 clinical factors without compromising performance, as confirmed by the DeLong’s test (AUC = 0.84, <span><math><mi>p</mi></math></span>-value = 0.507). The study suggests that vital signs and laboratory results are significant predictors, reinforcing their clinical relevance.</div></div><div><h3>Conclusion:</h3><div>This study provides a robust and actionable ML model for prediction of mortality in ICU sepsis patients. Its ability to maintain high performance with significant clinical factors extends to resource-limited environments, offering an important step forward in improving critical care patient outcomes.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100613"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145226831","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"(LiFT) Lightweight Fitness Transformer: A language-vision model for Remote Monitoring of Physical Training","authors":"Alexander Postlmayr,&nbsp;Pamela Cosman,&nbsp;Sujit Dey","doi":"10.1016/j.smhl.2025.100610","DOIUrl":"10.1016/j.smhl.2025.100610","url":null,"abstract":"<div><div>We introduce a fitness tracking system that enables remote monitoring for exercises using only a RGB smartphone camera, making fitness tracking more private, scalable, and cost effective. Although prior work explored automated exercise supervision, existing models are either too limited in exercise variety or too complex for real-world deployment. Prior approaches typically focus on a small set of exercises and fail to generalize across diverse movements. In contrast, we develop a robust, multitask motion analysis model capable of performing exercise detection and repetition counting across hundreds of exercises, a scale far beyond previous methods. We overcome previous data limitations by assembling a large-scale fitness dataset, <em>Olympia</em>, covering more than 1,900 exercises. To our knowledge, our vision-language model is the first that can perform multiple tasks on skeletal fitness data. On <em>Olympia</em>, our model can detect exercises with 76.5% accuracy and count repetitions with 85.3% off-by-one accuracy, using only RGB video. By presenting a single vision-language transformer model for both exercise identification and rep counting, we take a significant step towards democratizing AI-powered fitness tracking.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100610"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145332479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A systematic review on machine learning methods for dementia care: Problems, data and features","authors":"Xiaowen Su ,&nbsp;Tyler Morris ,&nbsp;Skylar Simpson ,&nbsp;Jordis Blackburn ,&nbsp;Luke Macdougall ,&nbsp;Xiaopeng Zhao","doi":"10.1016/j.smhl.2025.100553","DOIUrl":"10.1016/j.smhl.2025.100553","url":null,"abstract":"<div><div>The prevalence of dementia is increasing in many countries thanks to increased longevity across the world. Timely and effective care for dementia has significant societal and economic benefits. In recent years, there have been new explorations of machine learning methods for dementia care issues such as assessment, management, treatment of persons with dementia, which yield promising results while facing great challenges. We have conducted a systematic review of the research on the topic of machine learning methods for dementia care to access the feasibility and the state-of-the-art in the field. The problems solved by machine learning methods and the data and features used to facilitate those methods have been reviewed. The challenges have been discussed to indicate future directions.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100553"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145105416","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"GAN-enhanced hybrid deep learning model for real-time cardiac beat detection in the CardioHelp app with edge computing","authors":"Ucchwas Talukder Utsha,&nbsp;Bashir I. Morshed","doi":"10.1016/j.smhl.2025.100601","DOIUrl":"10.1016/j.smhl.2025.100601","url":null,"abstract":"<div><div>Cardiovascular diseases represent a major global health challenge, often complicated by limitations in early and accurate diagnosis due to the subtle and diverse nature of cardiac arrhythmias. Thorough diagnosis typically requires a review of medical history, physical examination, and specialized tests, including echocardiograms and electrocardiograms (ECG/EKG). Specific challenges include severe class imbalance within ECG datasets, which leads to reduced detection accuracy of rare arrhythmic events, such as Supraventricular Ectopic (S), Ventricular Ectopic (V), and Fusion (F) beats. Moreover, there is a scarcity of systems capable of detecting cardiac abnormalities in real time directly on edge devices, as most current solutions rely on offline analysis or cloud-based processing, which limits their effectiveness for continuous, real-world monitoring. To address these challenges, we propose an advanced edge-computing solution leveraging a Generative Adversarial Network (GAN)-augmented hybrid CNN-LSTM model for accurate real-time ECG beat classification. Our method first employs GAN-generated synthetic beats to effectively mitigate class imbalance. Subsequently, the CNN component extracts robust spatial ECG features, which are passed to the LSTM to capture temporal dependencies, ensuring comprehensive feature learning for precise arrhythmia classification.</div><div>The proposed solution includes a wearable custom data acquisition device interfaced with our CardioHelp smartphone app. This app provides real-time visualization of ECG and respiratory signals, continuous heart and respiratory rate monitoring, and immediate notifications for detected cardiac anomalies. The GAN-augmented hybrid model achieved exceptional results, attaining a test accuracy of 98.44% on the MIT-BIH Arrhythmia Database, with notable F1-scores: 96% for S beats, 98% for V beats, and 90% for F beats. Upon deployment on an edge device, the model maintained high validation accuracy (97.50%), demonstrating robustness under real-world conditions, with F1-scores of 85.1% (S), 89.3% (V), and 76.2% (F). Real-time validation with 10 healthy participants further confirmed its practical usability and reliability for continuous everyday monitoring. Additionally, the model operated efficiently with an average inference time of 4.6 ms per beat, suitable for instantaneous ECG classification on resource-constrained edge devices. The results confirmed the effectiveness of our system, paving the way for further pilot studies involving real-life cardiac patients. By enabling early intervention and remote patient monitoring, our developed system has the potential to significantly improve cardiovascular health outcomes.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100601"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144864264","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"CicloExergame: Player experience evaluation of an exergame for patient telerehabilitation","authors":"Carlos Henrique R. Souza,&nbsp;Saulo S. de Oliveira,&nbsp;Daniela F. Nascimento,&nbsp;Luciana O. Berretta,&nbsp;Sergio T. Carvalho","doi":"10.1016/j.smhl.2025.100611","DOIUrl":"10.1016/j.smhl.2025.100611","url":null,"abstract":"<div><div>Numerous factors are involved in functional rehabilitation, including patient engagement in activities and active participation of physiotherapists. Challenges in accessing therapy have prompted initiatives focusing on telerehabilitation. In this context, therapists’ active involvement is compromised, and exercises are often repetitive and monotonous, reducing patient engagement and motivation, thus impacting session outcomes. Consequently, this study aimed to outline the evaluation process involving patients using an exergame (CicloExergame) designed with a distributed architecture for conducting telerehabilitation sessions incorporating a cycle ergometer. The experiments, focusing on assessing Player Experience, were conducted with two groups: 12 healthy individuals and 10 patients at the Clinical Hospital of the Federal University of Goiás, Brazil. The study consisted of using the game in short sessions and administering questionnaires before and after use. Among the healthy participants, 75% rated the experience as “highly satisfactory,” while the remaining 25% found it “satisfactory.” This satisfaction trend was similarly observed among the patient group (60% “highly satisfactory” and 40% “satisfactory”). These promising results highlight the motivating and engaging potential of the proposed exergame.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100611"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145048868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TUCUXI-Core: An Open Source generic and efficient computing engine for Model-Informed Precision Dosing","authors":"Yann Thoma ,&nbsp;Thierry Buclin","doi":"10.1016/j.smhl.2025.100600","DOIUrl":"10.1016/j.smhl.2025.100600","url":null,"abstract":"<div><div>The absorption, distribution, metabolism, and excretion of drugs vary from one individual to another. As a result, drug concentration profiles in the blood do not evolve in the same way throughout the population. Measuring drug concentrations and individualizing drug dosages therefore makes it possible to respond to the specific needs of patients, so that their circulating exposure to the drug reaches targets that are both effective in achieving treatment objectives and safe regarding potential toxicity.</div><div>In this context, software tools can usefully assist prescribers in the process of adjusting drug dosages based on the therapeutic monitoring of blood concentrations, which is increasingly available in clinical laboratories. Bayesian adaptation based on prior population pharmacokinetic knowledge is the gold standard strategy for this. This article presents the core computing engine of <em>TUCUXI</em>, a software package that aims to support clinicians in this activity through a user-friendly graphical interface. We describe its main features and architecture, providing a general overview of this open-source project.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100600"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144864367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Parkinson’s Disease detection through multimodal data analysis","authors":"Arina Ivanova ,&nbsp;Aleksei Shcherbak ,&nbsp;Sergey Nesteruk ,&nbsp;Ekaterina Bril ,&nbsp;Anna Baldycheva ,&nbsp;Andrey Somov","doi":"10.1016/j.smhl.2025.100618","DOIUrl":"10.1016/j.smhl.2025.100618","url":null,"abstract":"<div><div>Parkinson’s disease (PD) is a slowly progressive neurodegenerative disease which still lacks objective tools for diagnosis. According to recent research results, misdiagnosis of PD may reach up to 25%. In this article, we report on the medical decision support system based on wearable sensors and video cameras with consequent “multimodal” data analysis using Machine Learning (ML) methods. For data collection reasons 169 subjects performed eleven exercises recommended by the neurologists. The proposed smart system is assessed through ML metrics and outperform the state-of-the-art solutions by achieving precision 98.6%, recall 98.1%, and F1-micro 98.3%. This decision support system opens wide vista for its application in hospitals as well as at home settings for controlling the undergoing therapy.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100618"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145424545","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparing robotic and computer vision assessments of unilateral and bilateral reaching in healthy adults","authors":"Desmond Asante ,&nbsp;Shelby Ziccardi ,&nbsp;Stephen J. Guy ,&nbsp;Rachel L. Hawe","doi":"10.1016/j.smhl.2025.100615","DOIUrl":"10.1016/j.smhl.2025.100615","url":null,"abstract":"<div><div>Assessment of reaching is foundational to upper limb neurorehabilitation. Current neurorehabilitation needs have increased the demand for quantitative clinical assessments of bilateral coordination. Robotics and computer vision for motion tracking are two means to provide relevant quantitative metrics but have many differences including the dimensionality of reaching movements (planar versus three-dimensional) and data acquisition. We do not know how consistent measures of bilateral coordination performance are between these different assessments. In this study, we examined how one robotic and one computer vision method can identify differences between symmetrical and asymmetrical reaching, and the correlations in movement time, and hand lag between these two approaches. Thirty healthy young adults completed four reaching games using the Kinarm exoskeleton robot and a custom developed augmented reality assessment using computer vision.</div><div>We found that both approaches were able to detect well-established movement time and hand lag differences between symmetrical and asymmetrical reaching, with the differences between symmetrical and asymmetrical being larger with the computer vision approach. Moderate correlations were found between approaches for unilateral and symmetric reaching in both movement time and hand lags; however, no significant correlations were found between approaches for asymmetric reaching.</div><div>Our results show that reaching task performance differs between robotic and computer vision-based assessment. However, both approaches provide quantitative metrics of unilateral and bilateral reaching that are consistent with prior research. There are benefits and tradeoffs to each approach, and this study informs how clinicians and researchers can consider the methodological differences when determining which assessment method to use.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"38 ","pages":"Article 100615"},"PeriodicalIF":0.0,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145265534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}