{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T05:42:04Z","timestamp":1775108524406,"version":"3.50.1"},"reference-count":26,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T00:00:00Z","timestamp":1775088000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Physiol."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Diaphragmatic dysfunction and acute cognitive stress\/delirium are serious complications of mechanical ventilation that prolong intensive care unit (ICU) stay and are associated with increased mortality. Although accumulating evidence suggests a potential lung\u2013brain axis linking impaired respiratory muscle function to adverse neurocognitive trajectories, early risk stratification remains challenging because clinically relevant signals are multimodal, temporally dynamic, and often partially observed. We aimed to develop an interpretable multimodal deep learning model to predict diaphragmatic dysfunction and high cognitive stress\/delirium in mechanically ventilated patients and to identify shared predictors that may reflect lung\u2013brain crosstalk.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We conducted a multicenter retrospective study including 25,751 mechanically ventilated ICU patients. A multimodal long short-term memory (LSTM) network was trained using continuous clinical time-series variables (vital signs, ventilator parameters, and medication dosing trajectories) and diaphragm ultrasound videos from a sub-cohort (n = 4,783). Discrimination was evaluated using the area under the receiver operating characteristic curve (AUC) and the area under the precision\u2013recall curve (AUPRC). Post-hoc SHapley Additive exPlanations (SHAP) were applied to quantify feature contributions and explore cross-modal interactions.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>In the independent test set, the multimodal model outperformed both clinical-only (AUC = 0.811) and video-only (AUC = 0.749) baselines for predicting diaphragmatic dysfunction, achieving an AUC of 0.902 (95% CI, 0.88\u20130.92) and an AUPRC of 0.594 (both P &amp;lt; 0.001 vs. baselines). For high cognitive stress\/delirium prediction, the model achieved an AUC of 0.792. Calibration was acceptable, with close agreement between predicted and observed risks (Brier score = 0.12). SHAP analyses indicated that ultrasound-derived diaphragm thickening fraction (DTF) and cumulative neuromuscular blockade exposure were among the strongest predictors of diaphragmatic dysfunction. Notably, diaphragm excursion and heart rate variability emerged as shared influential predictors for cognitive stress\/delirium, consistent with the lung\u2013brain crosstalk hypothesis in mechanically ventilated patients.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This study presents a multimodal deep learning framework for early identification of mechanically ventilated patients at elevated risk of diaphragmatic dysfunction and high cognitive stress\/delirium. Integrating diaphragm ultrasound with longitudinal bedside physiologic and treatment data improves prediction beyond single-modality models, while post hoc explainability highlights candidate shared predictors relevant to a lung\u2013brain axis. These findings suggest that integrated monitoring of respiratory mechanics and sedation-related physiology may support more proactive ventilator management and neuroprotective ICU care.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fphys.2026.1765898","type":"journal-article","created":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T04:16:11Z","timestamp":1775103371000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Explainable deep-learning models to predict diaphragmatic dysfunction and cognitive stress in ICU patients under mechanical ventilation"],"prefix":"10.3389","volume":"17","author":[{"given":"Yonghua","family":"Wang","sequence":"first","affiliation":[{"name":"Respiratory Intensive Care Unit, The First Affiliated Hospital of Zhengzhou University","place":["Zhengzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuling","family":"Bai","sequence":"additional","affiliation":[{"name":"Respiratory Intensive Care Unit, The First Affiliated Hospital of Zhengzhou University","place":["Zhengzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ge","family":"Jin","sequence":"additional","affiliation":[{"name":"The First Affiliated Hospital of Zhengzhou University","place":["Zhengzhou, China"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2026,4,2]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"e61527","DOI":"10.7759\/cureus.61527","article-title":"Patient outcomes and management strategies for intensive care unit (ICU)-associated delirium: a literature review","volume":"16","author":"Abdelbaky","year":"2024","journal-title":"Cureus"},{"key":"B2","doi-asserted-by":"publisher","first-page":"101203","DOI":"10.1016\/j.aucc.2025.101203","article-title":"Harnessing machine learning for predicting successful weaning from mechanical ventilation: a systematic review","volume":"38","author":"Ahmed","year":"2025","journal-title":"Aust. Crit. Care"},{"key":"B3","doi-asserted-by":"publisher","first-page":"e103581","DOI":"10.1136\/bmjopen-2025-103581","article-title":"Incidence and predictors of delirium in patients admitted to intensive care unit at comprehensive specialised hospitals in amhara region, northwest Ethiopia: a multicentre prospective observational study","volume":"15","author":"Asmare","year":"2025","journal-title":"BMJ Open"},{"key":"B4","article-title":"Predicting risk of delirium from ambient noise and light information in the ICU","author":"Bandyopadhyay","year":"2023"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1099","DOI":"10.1111\/aas.14119","article-title":"Sedative polypharmacy mediates the effect of mechanical ventilation on delirium in critically ill COVID-19 patients: a retrospective cohort study","volume":"66","author":"Bose","year":"2022","journal-title":"Acta Anaesthesiologica Scand."},{"key":"B6","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1186\/2110-5820-2-49","article-title":"Delirium in the ICU: an overview","volume":"2","author":"Cavallazzi","year":"2012","journal-title":"Ann. Intensive Care"},{"key":"B7","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2503.06059","article-title":"MANDARIN: mixture-Of-experts framework for dynamic delirium and coma prediction in ICU patients: development and validation of an acute brain dysfunction prediction model","author":"Contreras","year":"2025","journal-title":"arXiv Preprint arXiv:2503.06059"},{"key":"B8","doi-asserted-by":"publisher","first-page":"22","DOI":"10.1186\/s12883-025-04025-7","article-title":"Delirium at the intensive care unit and long-term survival: a retrospective study","volume":"25","author":"De Trizio","year":"2025","journal-title":"BMC Neurology"},{"key":"B9","doi-asserted-by":"publisher","first-page":"1015","DOI":"10.1007\/s10877-025-01317-8","article-title":"Effectiveness of diaphragmatic ultrasound as a predictor of successful weaning from mechanical ventilation","volume":"39","author":"Elfeky","year":"2025","journal-title":"J. Clinical Monitoring Computing"},{"key":"B10","doi-asserted-by":"publisher","first-page":"e70097","DOI":"10.1111\/nicc.70097","article-title":"Sleep quality and disruptive factors in intensive care units: a comparison between mechanically ventilated and spontaneously breathing patients","volume":"30","author":"Erbay Dalli","year":"2025","journal-title":"Nurs. Critical Care"},{"key":"B11","doi-asserted-by":"publisher","first-page":"356","DOI":"10.1186\/s12912-025-03013-z","article-title":"Efficacy and safety of care bundles in the treatment of delirium in ICU patients: a meta-analysis and systemic review","volume":"24","author":"Fang","year":"2025","journal-title":"BMC Nursing"},{"key":"B12","doi-asserted-by":"publisher","first-page":"27552","DOI":"10.1038\/s41598-025-12264-4","article-title":"Developing and validating machine learning models to predict next-day extubation","volume":"15","author":"Fenske","year":"2025","journal-title":"Sci. Rep."},{"key":"B13","doi-asserted-by":"publisher","first-page":"2263","DOI":"10.3390\/diagnostics14202263","article-title":"Predictive value of diaphragm and lung ultrasonography for weaning failure in critically ill patients with acute respiratory failure due to COVID-19 Pneumonia","volume":"14","author":"Fonseca","year":"2024","journal-title":"Diagnostics"},{"key":"B14","doi-asserted-by":"publisher","first-page":"223","DOI":"10.1002\/jum.16363","article-title":"Ultrasonography to access diaphragm dysfunction and predict the success of mechanical ventilation weaning in critical care: a narrative review","volume":"43","author":"Marques","year":"2024","journal-title":"J. Ultrasound Med."},{"key":"B15","doi-asserted-by":"publisher","first-page":"7830","DOI":"10.62347\/FDPE7053","article-title":"Value of diaphragmatic ultrasound parameters in assessing weaning outcomes and survival in ventilator-dependent intensive care unit patients","volume":"16","author":"Pan","year":"2024","journal-title":"Am. J. Transl. Res."},{"key":"B16","doi-asserted-by":"publisher","first-page":"200","DOI":"10.4266\/acc.2022.01060","article-title":"Evaluating diaphragmatic dysfunction and predicting non-invasive ventilation failure in acute exacerbation of chronic obstructive pulmonary disease in India","volume":"38","author":"Patel","year":"2023","journal-title":"Acute Crit. Care"},{"key":"B17","doi-asserted-by":"publisher","first-page":"e44827","DOI":"10.7759\/cureus.44827","article-title":"Prevalence and risk factors of delirium in patients admitted to intensive care units: a multicentric cross-sectional study","volume":"15","author":"Sadaf","year":"2023","journal-title":"Cureus"},{"key":"B18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.63096\/medtigo3092238","article-title":"Delirium in hospitalized and ICU patients: systematic review of pharmacologic and non-pharmacologic interventions","volume":"2","author":"Sali","year":"2025","journal-title":"Medtigo J. Emerg. Med."},{"key":"B19","doi-asserted-by":"publisher","first-page":"e89771","DOI":"10.7759\/cureus.89771","article-title":"Impact of the ICU No-Visitation policy on delirium in long-term mechanically ventilated patients: a single-center retrospective observational study","volume":"17","author":"Sato","year":"2025","journal-title":"Cureus"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1441","DOI":"10.1016\/j.bja.2025.02.024","article-title":"Weaning failure from mechanical ventilation: a scoping review of the utility of ultrasonography in the weaning process","volume":"135","author":"Sep\u00falveda","year":"2025","journal-title":"Br. Journal Anaesthesia"},{"key":"B21","doi-asserted-by":"publisher","first-page":"e72482","DOI":"10.2196\/72482","article-title":"AI model based on diaphragm ultrasound to improve the predictive performance of invasive mechanical ventilation weaning: prospective cohort study","volume":"9","author":"Song","year":"2025","journal-title":"JMIR Formative Research"},{"key":"B22","first-page":"25","article-title":"Delirium in critically ill patients: incidence, risk factors and outcomes","volume":"22","author":"Vyveganathan","year":"2019","journal-title":"Crit. Care Shock"},{"key":"B23","article-title":"Transformer representation learning is necessary for dynamic multi-modal physiological data on small-cohort patients","author":"Wang","year":"2025"},{"key":"B24","doi-asserted-by":"publisher","first-page":"e094232","DOI":"10.1136\/bmjopen-2024-094232","article-title":"Trajectory and risk factors of cognitive level in ICU delirium patients after transfer out of the ICU: a protocol for a prospective cohort study","volume":"15","author":"Yang","year":"2025","journal-title":"BMJ Open"},{"key":"B25","doi-asserted-by":"publisher","first-page":"7697","DOI":"10.21037\/qims-2025-167","article-title":"Prediction of weaning outcomes from mechanical ventilation in critically ill patients based on the combination of ultrasound parameters of the heart, lung, and diaphragm: a prospective observational cohort study","volume":"15","author":"Zhou","year":"2025","journal-title":"Quantitative Imaging Med. Surg."},{"key":"B26","doi-asserted-by":"publisher","first-page":"e70136","DOI":"10.1111\/crj.70136","article-title":"Real-time predictive analysis of ICU ventilator weaning failure: a prospective validation study","volume":"19","author":"Zhou","year":"2025","journal-title":"Clinical Respiratory Journal"}],"container-title":["Frontiers in Physiology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fphys.2026.1765898\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T04:16:15Z","timestamp":1775103375000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fphys.2026.1765898\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,2]]},"references-count":26,"alternative-id":["10.3389\/fphys.2026.1765898"],"URL":"https:\/\/doi.org\/10.3389\/fphys.2026.1765898","relation":{},"ISSN":["1664-042X"],"issn-type":[{"value":"1664-042X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,2]]},"article-number":"1765898"}}