Physiological measurement最新文献

{"title":"In-silico credibility assessment of a computational physiological model for non-invasive monitoring of respiratory effort in critically ill patients.","authors":"R S P Warnaar, S Geraats, A D Cornet, R G K M Aarts, J Graßhoff, D W Donker, E Oppersma","doi":"10.1088/1361-6579/aea685","DOIUrl":"10.1088/1361-6579/aea685","url":null,"abstract":"<p><p><i>Objective.</i>Monitoring respiratory effort in critically ill patients during assisted mechanical ventilation is essential to individualize ventilatory support and prevent over- or underassistance. A non-invasive method to estimate respiratory muscle pressure () combines respiratory surface electromyography (sEMG) with ventilator pressure-flow waveforms through the integrated equation of motion (iEqM). Implemented within a computational physiological model (CPM), the iEqM links muscle activation to generated pressure. This study investigates the reliability of an iEqM-based CPM under critical care conditions.<i>Approach.</i>CPM performance to estimate respiratory muscle pressure-time product () was evaluated in-silico using simulated patient profiles. Credibility activities included numerical verification, Monte-Carlo-based uncertainty quantification, and sensitivity analysis, exploring variations in respiratory mechanics, effort variability, sEMG signal quality, and patient-ventilator timing. Model outputs were compared with simulated reference values, with an acceptable clinical error margin set at 20%.<i>Main results.</i>Verification confirmed correct model implementation (errors < 0.3%). Input data uncertainty quantification showed limited variability (SD 1.8%). Sensitivity analysis revealed reduced accuracy under lowvariability (< 5.0 cmH₂O), low sEMG signal-to-noise ratios (< 1.4), highmagnitudes (17.5 cmH₂O), and persistent inspiratory efforts during expiration. Calibration using end-expiratory occlusion maneuvers improved accuracy, except at the lowestmagnitude ands.<i>Significance.</i>The iEqM performs reliably when calibrated via end-expiratory occlusion maneuvers. Without calibration, accuracy declined with lower effort variability, poorer sEMG quality, or patient-ventilator asynchrony. These findings emphasize the need for context-aware application, accounting for patient-specific mechanics, signal integrity and ventilator interaction, to ensure credible and reliable monitoring of respiratory effort in critically ill patients.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148945436","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diabetic retinopathy screening using artificial intelligence: a comprehensive performance comparison on datasets from diverse populations and imaging modalities.","authors":"Philipp Schapotschnikow, Vladyslav Dusiak, Murat Osswald, Paolo S Silva, Leo Anthony Celi, Luis Filipe Nakayama","doi":"10.1088/1361-6579/aea2ee","DOIUrl":"10.1088/1361-6579/aea2ee","url":null,"abstract":"<p><p><i>Objective.</i>Diabetic retinopathy (DR) is the leading cause of preventable blindness in adults and poses significant challenges in low- and middle-income regions due to limited access to skilled clinicians and diagnostic facilities. Automated screening solutions using artificial intelligence (AI) have emerged as an efficient alternative, achieving high diagnostic accuracy. However, these solutions are often developed using data from specific populations obtained using relatively expensive high-end devices. This study addresses the potential scope limitations by evaluating the AI-based screening tool<i>retina.help</i>across eight datasets representing diverse populations, imaging modalities, and geographic regions.<i>Approach.</i>The datasets include both public and private sources, with images captured using tabletop and handheld fundus cameras. Key performance metrics for detecting binary referable DR-sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC)-were calculated on an image-by-image basis.<i>Main results.</i><i>Retina.help</i>demonstrated high accuracy on tabletop images, achieving AUROC values of 0.97 on the BRSET and DeepDRiD datasets. Handheld device performance was more variable, with AUROC ranging from 0.88 (Filipino dataset) to 0.99 (Finnish dataset). Sensitivity declined with increased retinal pigmentation, as evidenced by lower values for datasets from Tanzania (62.2%) and Brazil (76.7%) compared to Finland (89.9%). Images from handheld devices often yielded lower sensitivity due to challenges related to low-contrast images. Nonetheless,<i>retina.help</i>generalized well across diverse datasets, showcasing its robustness.<i>Significance.</i>The study highlights the impact of imaging equipment, demographics, and image quality on diagnostic performance. These findings underscore the need for benchmarking AI-based DR screening tools using standardized datasets that encompass diverse populations and imaging conditions. Such evaluations can guide the development of equitable, reliable and robust screening solutions.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148892424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving the performance of sample entropy in ultra-short-term time-series using kernel density estimation.","authors":"Chang Yan, Kaiyue Si, Zhaoyang Cong, Annabella Sihan Dai, Chengyu Liu, Peng Li","doi":"10.1088/1361-6579/aea5d3","DOIUrl":"10.1088/1361-6579/aea5d3","url":null,"abstract":"<p><p>As a widely used nonlinear dynamics metric, sample entropy (SampEn) quantifies the irregularity of time-series. We proposed a reliable approach for estimating SampEn in ultra-short-term time-series based on kernel density estimation, termed kernel SampEn (kSampEn). KSampEn employs a kernel function to generate a smooth estimate of the cumulative distribution function (CDF) of inter-state distances, thereby alleviating the abrupt jumps observed in the empirical CDF used by conventional SampEn. Simulation results demonstrated that for time-series shorter than 30 data points, SampEn generally failed to produce valid estimates, whereas kSampEn remained robust under the tested conditions. Furthermore, SampEn frequently returned invalid estimates when applied to 30 s overnight RR interval segments, whereas kSampEn successfully quantified entropy and revealed higher entropy during non-rapid-eye-movement sleep compared with wakefulness or rapid-eye-movement sleep. In conclusion, kSampEn provides a reliable approach for analyzing cardiovascular signals in the context of sleep research and other applications involving short-term physiological time-series.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148930783","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploratory data analysis of tonometry and PTT/PIR signals for cuffless blood pressure monitoring: dynamic responsiveness versus baseline stability.","authors":"Ye Yang, Qiao Li, Mohamed Almadi, Jacqueline Gong, Ting Xiang, Zijun Liu, Gari D Clifford, Gary Strangman, Yuan-Ting Zhang, Quan Zhang","doi":"10.1088/1361-6579/aeab62","DOIUrl":"https://doi.org/10.1088/1361-6579/aeab62","url":null,"abstract":"<p><strong>Objective: </strong>Sensing modalities for cuffless blood pressure (BP) monitoring remain insufficiently characterized. To address this, we used exploratory data analysis (EDA) to evaluate multimodal physiological signals' associations with reference BP, to gain insight into their BP-related information and characteristics.</p><p><strong>Approach: </strong>Time-synchronized electrocardiogram (ECG), photoplethysmography (PPG), and superficial temporal artery tonometry (STAT) signals were collected from 28 sessions across 9 healthy participants during resting and handgrip-induced BP perturbations, referenced against continuous Finapres BP. Pulse transit time (PTT)/PPG intensity ratio (PIR)-derived and STAT-derived BP estimates were evaluated via participant-level Pearson correlation and &phi;k coefficients, dynamic-responsiveness/baseline-stability measures, and a phase-resolved (stable vs. dynamic) analysis.</p><p><strong>Main results: </strong>PTT/PIR-derived estimates showed weak linear association with reference SBP (r ≈ 0.37-0.41) and weaker association with DBP, with nonlinear association (&phi;k ≈ 0.58-0.62 SBP, 0.38-0.57 DBP) exceeding their linear correlation. STAT-derived estimates showed very weak linear correlation but the strongest nonlinear association (&phi;k ≈ 0.68 SBP, 0.71 DBP) - modest for SBP but more pronounced for DBP. Phase-resolved analysis confirmed STAT tracked BP more closely during the dynamic (handgrip) phase and PTT/PIR during the stable phase, a genuine, phase-specific pattern rather than a pooling artifact; because Finapres can amplify beat-to-beat amplitude, this should be read relative to that reference, not invasive BP.</p><p><strong>Significance: </strong>These complementary properties motivate a fusion strategy using STAT to track rapid BP changes while PTT/PIR anchors the baseline. More broadly, this EDA-derived understanding can guide more data-efficient, interpretable models - properties clinicians require regardless of data volume. Given the modest, healthy-only sample, findings should be regarded as preliminary.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149025288","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Stage-specific EEG spectral signatures of primary insomnia: regional and sex-related differences across NREM and REM sleep.","authors":"Nilantha Premakumara, Chan-Yun Yang, B H Sudantha, Kuo-Wei Chen","doi":"10.1088/1361-6579/aea9ee","DOIUrl":"https://doi.org/10.1088/1361-6579/aea9ee","url":null,"abstract":"<p><p>Sleep architecture is characterized by dynamic transitions between non-rapid eye movement (NREM) and rapid eye movement (REM) stages, each defined by distinct oscillatory signatures. Although quantitative EEG has been widely used to study primary insomnia (PI), reported findings remain inconsistent, partly due to limited consideration of cortical regional differences, sleep-stage specificity, and sex-related variability. This study integrates spectral power analysis with explainable machine learning to provide a comprehensive characterization of neural dynamics in PI. EEG recordings from 100 adults (50 healthy sleepers and 50 PI patients) were analyzed across frontal (F3, F4), central (C3, C4), and occipital (O1, O2) regions. Stage-specific classifiers were developed for NREM and REM using LASSO, random forest, CatBoost and XGBoost models, and Shapley Additive Explanations (SHAP) were employed to quantify the contribution of each spectral and regional feature to model predictions. Models trained on NREM features demonstrated superior discrimination compared with those trained on REM features, suggesting that slow-wave neural activity captures insomnia-related alterations more reliably than REM dynamics. SHAP analysis identified reduced delta power and elevated beta activity most prominently in central regions as the dominant predictors of PI during NREM sleep. In contrast, REM-based models pointed to prefrontal beta1 and occipital alpha as secondary discriminative features. Sex-stratified analyses further revealed systematically higher spectral power in females and distinct SHAP feature-importance patterns between sexes. Collectively, this work establishes an interpretable, stage-specific framework for identifying EEG biomarkers of insomnia and demonstrates the value of explainable artificial intelligence in advancing the neurophysiological understanding of sleep disturbance.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149002377","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Paediatric respiratory sound classification by integrating local feature extraction and global context modeling.","authors":"Yufei Shen, Fan Zhang, Jiakai Wu, Ping Sun, Xiaoling Xu","doi":"10.1088/1361-6579/ae95ad","DOIUrl":"10.1088/1361-6579/ae95ad","url":null,"abstract":"<p><p><i>Background and objective.</i>Automated respiratory sound classification based on deep learning is critical for enhancing diagnostic precision and efficiency in respiratory disease. Research on respiratory sound classification for adults is extensive, whereas work on paediatric respiratory sounds is still scarce. Compared to adults, paediatric respiratory sounds typically contain more high-frequency components, and this inherent difference presents unique challenges for models.<i>Approach.</i>In response to this challenge, we proposed a novel architecture, which integrated the local feature extraction module with the global context model. Specifically, we designed the Mel_Grouper module to serve as a front-end with the purpose of enhancing local pathological representations. The output was fed into the Transformer-based Mel_Encoder to fuse global context.<i>Main results.</i>Experimental results on the SJTU Paediatric Respiratory Sound (SPRSound) dataset demonstrate that our method achieves state-of-the-art performance. Regarding the four subtasks of the SPRSound dataset, it outperforms the previous best results by 3.37%, 3.06%, 6.83% and 5.36%, respectively. To better reflect the real clinical conditions, we further performed experiments on the real-world paediatric respiratory sound dataset.<i>Significance.</i>These results demonstrate the method's significant performance in paediatric respiratory sound classification. The code is available at:.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148679577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Testing the null hypothesis of uncoupling between mean arterial pressure and mean cerebral blood velocity in health and pathology via different surrogate strategies.","authors":"Alberto Porta, Beatrice Cairo, Ilaria Burzo, Beatrice De Maria, Martina Anguissola, Marco Ranucci, Vlasta Bari","doi":"10.1088/1361-6579/ae9f12","DOIUrl":"10.1088/1361-6579/ae9f12","url":null,"abstract":"<p><p><i>Objective</i>. The dynamic cerebral autoregulation (dCA) limits variability of mean cerebral blood velocity (MCBv) despite mean arterial pressure (MAP) variations through a myriad of nonlinear mechanisms. However, dCA operates in association with the flow-to-pressure pathway. This study aims at characterizing the closed loop MCBv-MAP relationship by separating the dependence of MAP on MCBv from that over the reverse pathway using different surrogate approaches.<i>Approach.</i>Analyses were carried out in 26 healthy controls (HCs, age: 43 ± 11 yr, 13 males, 13 females) and 48 severe aortic valve stenosis patients undergoing surgical aortic valve replacement (SAVR, age: 62 ± 14 yr; 35 males, 13 females) before and after surgery. Active standing (STAND) from supine resting (REST) was exploited to challenge dCA. A state space correspondence method based on model-free cross-predictability was exploited to estimate the degree of MCBv dependence on MAP and vice versa and two different types of surrogates destroying nonlinear components or preserving them as much as possible were utilized.<i>Main results.</i>We found that: i) STAND deteriorates dCA in HCs; ii) in SAVR patients STAND does not affect dCA either before or after surgery, but a pre-surgery activation of flow-to-pressure link is present at REST; iii) in HCs preserving nonlinear dynamics when testing the significance of the pressure-to-flow link provides results different from those derived while maintaining exclusively linear dynamics.<i>Significance.</i>Results stressed the pathophysiological relevance of applying a nonlinear directional approach to disentangle closed loop MCBv-MAP relationship and suggest that conclusions might depend on the strategy adopted to generate surrogates.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148831433","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ResoMamba-LLM: multimodal physiological signal stress detection for wearable devices via Mamba and: LLM. reprogramming.","authors":"Yingbin Hong, Dong Jiang, Zhongxiang Lu","doi":"10.1088/1361-6579/ae9b5a","DOIUrl":"10.1088/1361-6579/ae9b5a","url":null,"abstract":"<p><p><i>Objective.</i>The accuracy of psychological stress detection hinges upon the precision with which models capture complex and variable physiological responses. Current research primarily employs deep learning techniques to model multimodal physiological responses, aiming to enhance detection performance by leveraging complementary features across modalities. However, existing deep learning approaches face challenges in modeling long-range dependencies and exhibit suboptimal robustness in cross-modal fusion.<i>Approach.</i>To address this, this paper proposes ResoMamba-LLM, a multimodal stress detection framework integrating Mamba and large language models (LLM). Specifically, ResoMamba-LLM employs dual-channel Mamba encoders to process chest and wrist physiological signals respectively, supplemented by a frequency domain branch to capture periodic rhythms. It further utilizes lightweight reprogramming techniques to map physiological features onto the semantic space of the LLM, thereby stimulating its latent contextual reasoning capabilities.<i>Main results.</i>Achieving accuracies of 95.83% and 81.65% on the WESAD and EmoWear datasets, respectively, experimental results conclusively demonstrate ResoMamba-LLM's robust long-dependency modeling capabilities for multimodal physiological signals and its robust cross-modal fusion abilities.<i>Significance.</i>This provides a novel solution for precise, interpretable psychological stress detection.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148796645","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reliability-aware hierarchical learning for Chagas disease screening from 12-lead ECGs: tackling label uncertainty and class imbalance.","authors":"Hao Wen, Jingsu Kang","doi":"10.1088/1361-6579/ae98e0","DOIUrl":"10.1088/1361-6579/ae98e0","url":null,"abstract":"<p><p><i>Objective.</i>Chagas disease, a neglected tropical disease (NTD) with significant cardiovascular impact, remains underdiagnosed in resource-limited regions. Electrocardiogram (ECG) screening offers a low-cost tool for detecting cardiac involvement, yet algorithm development is challenged by label noise, data scarcity, and the latent nature of infection. This study proposes a robust ECG-based screening framework that explicitly addresses these constraints.<i>Approach.</i>We introduce a<i>Reliability-Aware Hierarchical Learning</i>strategy that calibrates supervision according to data provenance, prioritizing serology-confirmed labels over noisy self-reports. To mitigate data scarcity, we compare a specialized convolutional neural network (CNN) trained from scratch with a transfer learning approach based on a Spatio-Temporal ECG foundation Model (FM). Performance is evaluated across varying data scales, and the representation structure is analyzed to interpret model behavior.<i>Main results.</i>On the official hidden test set of the George B. Moody PhysioNet/Computing in Cardiology Challenge 2025, our approach achieved a Challenge Score of 0.163. We observe that while the specialized CNN performs competitively in data-rich regimes, the FM exhibits superior robustness in extreme low-resource settings. Furthermore, performance reaches a plateau imposed by underlying disease physiology. Bimodal score distributions suggest that models distinguish established cardiomyopathy from indeterminate infection, which remains electrophysiologically indistinguishable from healthy controls.<i>Significance.</i>These findings clarify both the potential and intrinsic limits of ECG-based AI screening for NTD-associated cardiac involvement. Reliability-aware supervision and data-efficient transfer learning provide a practical framework toward scalable and clinically meaningful ECG screening systems in resource-constrained environments.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148725000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RR-interval-based atrial fibrillation detection and burden estimation: cross-dataset validation and calibration-aware probability analysis.","authors":"Mohammad Haekal, Siti Nurul Khotimah, Galih Restu Fardian Suwandi, Freddy Haryanto","doi":"10.1088/1361-6579/ae99aa","DOIUrl":"10.1088/1361-6579/ae99aa","url":null,"abstract":"<p><p><i>Objective.</i>Atrial fibrillation (AF) burden has become an increasingly important endpoint in long-duration rhythm monitoring, but reliable burden estimation requires more than accurate AF detection alone. In particular, when burden is derived by aggregating predicted AF probabilities over time, probability calibration may directly affect burden validity under external dataset shift.<i>Approach.</i>This study developed an interpretable-interval feature model for AF detection and evaluated it using record-wise cross-validation on a development cohort and independent cross-dataset external validation on public Holter electrocardiographic databases. Window-level performance was assessed using the area under the receiver operating characteristic curve (ROC-AUC), area under the precision-recall curve (PR-AUC), Brier score, expected calibration error (ECE), and calibration intercept and calibration slope. Recording-level AF burden was estimated using both probability-based and hard-label aggregation and evaluated using mean absolute error (MAE) and agreement analyses.<i>Main results.</i>The model showed high discrimination in both development and external evaluation, with external ROC-AUC ofand PR-AUC of. However, external calibration deteriorated despite preserved ranking performance, with Brier score of, ECE(15) of, calibration intercept of, and calibration slope of. In the external cohort, probability-based burden estimation preserved strong association with reference burden but showed weaker raw agreement than hard-label aggregation, with MAE ofversus, consistent with systematic probability underprediction. Repeated external recalibration across record-level splits substantially improved probability quality and probability-based burden estimation. Median probability-burden MAE decreased fromwithout recalibration toafter Platt recalibration andafter isotonic recalibration, while median ECE(15) decreased fromtoand, respectively.<i>Significance.</i>These findings indicate that-interval-based AF detection maintained strong ranking performance in the tested external cohort, but probability calibration should be evaluated explicitly when predicted probabilities are aggregated into AF-burden estimates.</p>","PeriodicalId":20047,"journal":{"name":"Physiological measurement","volume":" ","pages":""},"PeriodicalIF":2.5,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148761743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}