Adrián Castillo-Allendes, Eric J Hunter
Perceived swallowing burden and objective swallowing safety reflect dissociable physiological domains captured by different acoustic measures. High-frequency noise during connected speech emerged as a candidate acoustic correlate of airway invasion severity. These findings are preliminary and require replication in larger prospective studies.
OBJECTIVE: Laryngeal function deteriorates with age, but the relationship between acoustic markers of phonatory quality, patient-reported swallowing burden, and imaging-derived airway invasion severity remains poorly characterized. This exploratory study examined whether distinct acoustic dimensions of laryngeal function are differentially associated with self-reported swallowing symptoms and airway invasion severity in community-dwelling older adults.
METHODS: Thirty-two community-dwelling adults aged 65-86 years completed the Eating Assessment Tool (EAT-10), Voice Handicap Index-10 (VHI-10), and Voice Symptom Scale (VoiSS), and underwent Fiberoptic Endoscopic Evaluation of Swallowing (FEES) with Penetration-Aspiration Scale (PAS) scores. Acoustic measures were extracted from sustained vowel /a/, connected speech (Rainbow Passage), and pitch glide tasks. Spearman correlations between acoustic features and PAS severity scores were the primary analysis. Binary grouping (PAS ≥3, airway invasion risk) and logistic regression were examined as secondary exploratory analyses.
RESULTS: EAT-10 correlated with VHI-10 (ρ = 0.45, P = 0.011) and VoiSS (ρ = 0.61, P < 0.001) but not with airway severity scores (ρ = 0.09, P = 0.616). Patient-reported outcomes did not differentiate PAS risk groups. Shimmer-based amplitude perturbation measures from the pitch glide task were the dominant acoustic correlates of EAT-10 (ρ up to 0.56), whereas frequency perturbation indices showed weaker associations (ρ = 0.38). These burden-related measures were acoustically dissociable from those associated with swallowing safety: high-frequency noise during connected speech (HFNO) was the primary acoustic correlate of airway invasion severity (ρ = 0.41, P = 0.020), and the only variable to discriminate airway invasion risk group (PAS ≥3; P = 0.018, |r| = 0.59). In secondary exploratory analyses, PCA identified an exploratory laryngeal inefficiency dimension (PC1, 44.5% variance; AUC = 0.83).
CONCLUSIONS: Perceived swallowing burden and objective swallowing safety reflect dissociable physiological domains captured by different acoustic measures. High-frequency noise during connected speech emerged as a candidate acoustic correlate of airway invasion severity. These findings are preliminary and require replication in larger prospective studies.