EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Using Machine Learning for the Automated Segmentation and Detection of Swallows Obtained by Digital Cervical Auscultation in Preterm Neonates

Stephen So, Timothy Tadj, Belinda Schwerin, Anne B Chang, Seiji Humphries, Thuy T Frakking iD

Dysphagia · 2026-02 · 41 (1) · 275-284

DOI: 10.1007/s00455-025-10879-3 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Dysphagia, online erschienen am 12.09.2025. Verfasst von Stephen So und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Engineering and Built Environment, Griffith University, Parklands Dr, Southport, QLD, 4215, Australia. Hauptthemen laut MEDLINE: Machine Learning, Auscultation, Deglutition, Infant, Premature. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The University of Queensland.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The clinical application of acoustic swallowing sound parameters collected from digital cervical auscultation is limited because of the time-consuming manual segmentation required by trained experts. The automated identification of swallowing sounds in children and adults from swallowing sound audio wavefiles using machine learning have accuracies between 76–95%. No data exists in preterm neonates. To determine if applying automated machine learning using a transfer learning approach could accurately identify and segment swallows from swallowing sounds collected in preterm neonates. Thin fluid swallow sounds were collected from 78 preterm neonates, median birth age 34 weeks gestation (range 25–36 weeks, 52.6% males) across 3 Australian special care nurseries. For the base machine learning model, a deep convolutional neural network (DCNN) pre-trained for audio event classification was used. With raw swallow audio data as input, embedding vectors from the base DCNN were generated and used to train a feedforward neural network to determine the presence of a swallow within an audio segment. The model showed high overall accuracy (94%) in identifying preterm swallows. Better model performance on bottle feeding swallows (Sensitivity, 95%; and specificity, 96%) was seen compared with breastfeeding swallows (sensitivity, 95%, specificity 92%). Interpretation: Our novel study demonstrates the successful use of transfer learning to accurately identify and segment digital swallowing sounds in preterm neonates. Application of this model could support the development of a digital CA app to automatically classify swallow sounds and improve objectivity for CA use in clinical practice within special care nurseries.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningAuscultationDeglutitionInfant, PrematureDeglutition DisordersHumansInfant, NewbornMaleFemaleNeural Networks, ComputerSensitivity and SpecificityAustralia

Stichwörter der Autor:innen

Cervical auscultationDeglutitionMachine learningNeonatePretermSignal processingSwallowing sounds

Autor:innen und Institutionen

  • School of Engineering and Built Environment, Griffith University, Parklands Dr, Southport, QLD, 4215, Australia.
  • Department of Respiratory Medicine, Queensland Children's Hospital, 501 Stanley St,, South Brisbane, QLD, 4101, Australia.
  • Child and Maternal Health Division, Menzies School of Health Research, Charles Darwin University, PO Box 41096, Casuarina, NT, 0811, Australia.
  • Australian Centre for Health Services Innovation, Queensland University of Technology, Level 7, 62 Graham St,, South Brisbane, QLD, 4101, Australia.
  • Research Development Unit, Caboolture Hospital, Metro North Hospital & Health Service, McKean St, Caboolture, QLD, 4510, Australia.
  • Research Development Unit, Caboolture Hospital, Metro North Hospital & Health Service, McKean St, Caboolture, QLD, 4510, Australia. thuy.frakking@health.qld.gov.au.

ORCID-Profile: Thuy T Frakking

Förderung

  • The University of Queensland

Zitieren

APA

So, S., Tadj, T., Schwerin, B., Chang, A. B., Humphries, S., & Frakking, T. T. (2026). Using Machine Learning for the Automated Segmentation and Detection of Swallows Obtained by Digital Cervical Auscultation in Preterm Neonates. Dysphagia, 41(1), 275-284. https://doi.org/10.1007/s00455-025-10879-3

Vancouver

So S, Tadj T, Schwerin B, Chang AB, Humphries S, Frakking TT. Using Machine Learning for the Automated Segmentation and Detection of Swallows Obtained by Digital Cervical Auscultation in Preterm Neonates. Dysphagia. 2026;41(1):275-284. doi:10.1007/s00455-025-10879-3

BibTeX
@article{So2026,
  author = {So, Stephen and Tadj, Timothy and Schwerin, Belinda and Chang, Anne B and Humphries, Seiji and Frakking, Thuy T},
  title = {Using Machine Learning for the Automated Segmentation and Detection of Swallows Obtained by Digital Cervical Auscultation in Preterm Neonates},
  journal = {Dysphagia},
  year = {2026},
  volume = {41},
  number = {1},
  pages = {275--284},
  doi = {10.1007/s00455-025-10879-3}
}
Kostenlos starten →