Medidas multiparamétricas e cepstrais de professores sem alterações laríngeas: comparação entre os sexos

Abstract

Introduction: The speech-language pathology assessment of voice involves a multidimensional approach to identify vocal alterations, their etiology, and guide rehabilitation according to the individual’s needs. Traditional methods, such as perceptual-auditory evaluation (PAE) and aerodynamic analyses, provide valuable information but have limitations, particularly regarding subjectivity and the requirement for specialized equipment. Traditional acoustic analyses, such as jitter and shimmer, are useful for mild to moderate vocal deviations but present limitations in highly deviant voices due to the aperiodicity of the vocal signal. In contrast, cepstral measures and multiparametric acoustic indices emerge as more reliable and objective tools, being less sensitive to frequency and amplitude perturbations and allowing for connected speech tasks. These indices, such as the Acoustic Voice Quality Index (AVQI), Acoustic Breathiness Index (ABI), and Dysphonia Severity Index (DSI), integrate different acoustic parameters into a single score, facilitating multidimensional vocal assessment. Objective: To analyze the multiparametric acoustic indices (AVQI, ABI, and DSI) and cepstral measures (Cepstral Peak Prominence - CPP e Smoothed Cepstral Peak Prominence - CPPS) of teachers without laryngeal disorders of both genders. Methods: A cross-sectional observational study was conducted with 86 teachers without laryngeal disorders (48 women and 38 men) recruited from the Central Superintendency of Occupational Medical Expertise and Health (SCPMSO). Selection criteria included normal laryngeal exam, excluding smokers, pregnant women, menstruating women, and individuals with conditions that could affect their voice at the time of data collection. Vocal samples were recorded in a silent room with an enviromental noise below 40dB using the Praat software. Tasks included sustained vowels, connected speech, minimal intensity vowel emissions, and ascending glissando. AVQI, ABI, and DSI indices were extracted using specific software scripts, while cepstral measures (CPP and CPPS) were evaluated in sustained vowel and speech tasks. Data were statistically analyzed using Mann-Whitney and T-tests to verify gender differences, with a 95% confidence level. Results: Significant differences between sexes were found in AVQI (p = 0,010) and ABI (p = 0,000), with women presenting higher scores. No significant differences were observed for DSI (p = 0,249). Regarding cepstral measures, men showed higher CPP and CPPs values in sustained vowels (p = 0,032 and p = 0,036), while women demonstrated higher CPP values in connected speech (p = 0,007). For men also obtained higher values in sustained vowels, reflecting greater stability and harmonic structure in male voices. In speech tasks, women showed greater variation due to phonetic and supraglottic characteristics of the female vocal tract. Discussion: The findings reinforce the reliability of multiparametric acoustic indices and cepstral measures for assessing vocal quality, especially in voice professionals like teachers. Gender differences align with the literature, attributing higher breathiness values in women due to the physiological presence of posterior glottic gaps, while men demonstrated greater harmonic structure in sustained voices. Greater variability in speech measures among women may be associated with non-vocal factors such as prosody and articulation. While indices such as AVQI and ABI appear sensitive to gender differences, DSI seems more robust in determining overall vocal quality, regardless of gender. The study highlights the need for normative values adjusted for men and women to improve diagnostic precision. Conclusion: The analyses indicated significant differences in acoustic indices and cepstral measures between genders, even among teachers without laryngeal disorders. These findings underscore the relevance of these measures for objectively and multidimensionally assessing vocal quality, supporting personalized interventions for specific populations, such as teachers. Adopting gender-specific normative parameters could further enhance clinical practice.