Analysis of Grunting Sound in Infants for Predicting the Severity of Respiratory Distress Syndrome


Satar M., Cengizler Ç., Özdemir M., YAPICIOĞLU YILDIZDAŞ H.

Journal of Voice, 2024 (SCI-Expanded) identifier

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2024
  • Doi Numarası: 10.1016/j.jvoice.2024.07.023
  • Dergi Adı: Journal of Voice
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Periodicals Index Online, CINAHL, Communication Abstracts, Linguistics & Language Behavior Abstracts, MEDLINE, Music Index, Music Periodicals Database, RILM Abstracts of Music Literature
  • Anahtar Kelimeler: Computer-assisted diagnosis, Respiratory distress syndrome, Spectral analysis
  • Çukurova Üniversitesi Adresli: Evet

Özet

Objective: Vocalizations from infants, particularly sounds associated with respiratory distress, are fundamental for observational scoring of respiratory tract issues. Listening to these infant sounds is a prevalent technique for decision-making in newborn intensive care units. Expiratory grunting, indicative of the severity and presence of potential conditions, is valuable, however, this evaluative method is subjective and prone to error. This study investigates the potential of computer-aided analysis to offer an objective scale for assessing the severity of respiratory tract problems, utilizing digital recordings of grunting sounds. Methods: The original data set is formed with a total of 189 grunting sound segments collected from 38 infants. Multiple evaluation approaches were performed to reveal the relation between spectral characteristics of the recordings and the severity or existence of respiratory distress. Results: Three spectral features were evaluated as prominently related to hospital stay duration and respiratory distress. The harmonic ratio of the recordings was graded as the most-related spectral feature that would characterize the severity. Conclusions: The potential of an innovative and objective grading approach is first investigated for replacing the human ear with a computer-aided evaluation system. The results are promising and the detected relation between expert ear-based scoring and harmonic ratio suggests that the spectral character of the grunting sounds would reflect the nature of respiratory conditions. Moreover, this study underlines those spectral features of digital grunting recordings that would be functional for automated prediction and decision-making.