EUVIMEDEUVIMED entdecken
Open Access

Physiological predictors of respiratory motor plasticity: A machine-learning reappraisal of phrenic motor facilitation

Lauren McColgan, Thomas V Evans, Robert N Marku, Raphael R Perim, Guang Xu, Isabela P Amorim, Morgan Urdaneta, Gordon S Mitchell

Respiratory Physiology & Neurobiology · 2026-07 · 343 · 104574

DOI: 10.1016/j.resp.2026.104574 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Respiratory Physiology & Neurobiology, 2026-07. Verfasst von Lauren McColgan und 7 weiteren Autor:innen. Beteiligte Einrichtung: Tom & Julie Wood College of Osteopathic Medicine, Department of Physiology & Pharmacology, Marian University, Indianapolis, IN, USA. Hauptthemen laut MEDLINE: Phrenic Nerve, Hypoxia, Long-Term Potentiation, Neuronal Plasticity. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NHLBI NIH HHS, National Heart Lung and Blood Institute.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Acute intermittent hypoxia (AIH)–induced phrenic long-term facilitation (pLTF) is a well-established form of respiratory motor plasticity and a subtype of phrenic motor facilitation (pMF), which also includes pharmacologically induced plasticity. Understanding the mechanisms of pLTF and pMF represents opportunities for therapeutic targets for spinal cord injury, amyotrophic lateral sclerosis and other neurological disorders. Here, we performed a secondary analysis of previously generated experimental datasets to reassess physiological predictors of pLTF and pMF. Given the limitations of expressing pMF as a percent, and to systematically evaluate complex and potentially nonlinear relationships among physiological variables, we employed a supervised machine learning approach using Gradient Boosted Decision Trees and Shapley Additive Explanations (SHAP) analysis to identify and rank both established and novel determinants of respiratory plasticity via percent-based (%) and absolute (Δ) changes in pLTF and pMF. Gradient Boosted Decision Trees models were trained to predict Δ and % outcomes using baseline physiological variables and AIH- or drug-evoked responses as input features, and model interpretability was achieved using SHAP to quantify the contribution of each predictor. The pooled datasets included experiments in which pLTF was induced via AIH (n=75) and pMF via pharmacologic intervention (n=39) using established experimental protocols. All animals underwent standardized measurement of phrenic nerve activity, respiratory frequency, arterial blood pressure, and evoked hypoxic responses, which were incorporated into the predictive models. Hypoxia-induced phrenic nerve response (ΔHypoxicPNA) was the primary predictor of ΔpLTF, while ΔMaximalPNA exerted a lesser influence in our analysis. Similarly, %pLTF was most strongly influenced by %HypoxicPNA but was also influenced by baseline phrenic nerve activity (BL PNA) and the hypoxia-evoked blood pressure response (ΔHypoxicBP). Notably, calculating percent change in pLTF versus absolute change in pLTF presents certain limitations. Each animal has a unique level of baseline phrenic nerve activity, so calculation via percent change can conceal or misrepresent the overall magnitude of respiratory motor plasticity. For ΔpMF, BL PNA was the strongest positive correlate, while body mass, which strongly correlates with age in laboratory rats, was inversely correlated to ΔpMF. Rat body mass was strongly inversely correlated with %pMF, while BL PNA and BL respiratory frequency (RF) exerted some influence. Similarities in determinant variables for pMF and pLTF indicate the two phenomena exhibit overlapping mechanisms of action, while the differing variables in each measure of respiratory motor plasticity demonstrate the unique sensitivities of pLTF and pMF. This study provides a comprehensive evaluation of respiratory motor plasticity, discovering new determinants while validating previous research regarding pLTF. Better understanding its mechanism could further research into new therapies for increasing respiratory drive.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Phrenic NerveHypoxiaLong-Term PotentiationNeuronal PlasticityMachine LearningRespirationAnimalsPredictive Learning ModelsRats

Stichwörter der Autor:innen

Hypoxia-evoked phrenic responseMachine learningNeurorehabilitationRespiratory motor plasticity

Autor:innen und Institutionen

  • Tom & Julie Wood College of Osteopathic Medicine, Department of Physiology & Pharmacology, Marian University, Indianapolis, IN, USA.
  • Tom & Julie Wood College of Osteopathic Medicine, Department of Physiology & Pharmacology, Marian University, Indianapolis, IN, USA. Electronic address: rperim@marian.edu.
  • Tom & Julie Wood College of Osteopathic Medicine, Department of Biostatistics & Public Health, Marian University, Indianapolis, IN, USA.
  • Department of Chemical Engineering, Polytechnic School, University of São Paulo, São Paulo, SP, Brazil.
  • Healthmap Solutions, Tampa, FL, USA.
  • Center for Breathing Research and Therapeutics, Department of Physical Therapy and McKnight Brain Institute, University of Florida, Gainesville, FL, USA.

Förderung

  • NHLBI NIH HHS
  • National Heart Lung and Blood Institute

Zitieren

APA

McColgan, L., Evans, T. V., Marku, R. N., Perim, R. R., Xu, G., Amorim, I. P., Urdaneta, M., & Mitchell, G. S. (2026). Physiological predictors of respiratory motor plasticity: A machine-learning reappraisal of phrenic motor facilitation. Respiratory Physiology & Neurobiology, 343, 104574. https://doi.org/10.1016/j.resp.2026.104574

Vancouver

McColgan L, Evans TV, Marku RN, Perim RR, Xu G, Amorim IP, et al. Physiological predictors of respiratory motor plasticity: A machine-learning reappraisal of phrenic motor facilitation. Respiratory Physiology & Neurobiology. 2026;343:104574. doi:10.1016/j.resp.2026.104574

BibTeX
@article{McColgan2026,
  author = {McColgan, Lauren and Evans, Thomas V and Marku, Robert N and Perim, Raphael R and Xu, Guang and Amorim, Isabela P and Urdaneta, Morgan and Mitchell, Gordon S},
  title = {Physiological predictors of respiratory motor plasticity: A machine-learning reappraisal of phrenic motor facilitation},
  journal = {Respiratory Physiology & Neurobiology},
  year = {2026},
  volume = {343},
  pages = {104574},
  doi = {10.1016/j.resp.2026.104574}
}
Kostenlos starten →