Aging and Linguistic Factors Influence on Speech Audiometry

Abstract

Objective: 1) Assess the influence of word familiarity, phonological neighborhood, and acoustic patterns on speech audiometry variability. 2) Evaluate the interaction between those factors and patients’ age and years of education. Method: Patients (n = 160) underwent otoscopy, tympanometry, pure-tone, and speech audiometry. At an intensity level corresponding to a speech score between 40% and 60%, patients were presented 6 lists of 10 spondees, taken from a standard material used in France. Linguistic characteristics of the spondees and spectral acoustic patterns were characterized. PLS regression analysis was used to determine the factors influencing the percent score recognition for each word (PC). Results: PC varied from 18% to 92%. PC rose significantly with word familiarity ( r = 0.38, P = .005) and 1 kHz frequency band amplitude of each word’s acoustic spectrum ( r = 0.60, P = .001). The percentage variability due to linguistic factors decreased significantly with age (from 30% below 50 years old to less than 10% above 60 years old), and with number of educational years (NEY; from 40% for NEY under 10 down to 18% for NEY above 10, for patients less than 50 years old). PC was significantly lower for spondees with large phonological neighborhood ( F(1, 51) = 12, P = .002), word familiarity being equal) for patients with NEY above 10. Conclusion: When taken independently from hearing loss, variability in PC was explained mostly by linguistic factors, combined with age and years of education, hence the need of carefully considering those factors when conceiving speech audiometry material.