Acoustic and Aerodynamic Clusters Within Primary Muscle Tension Dysphonia.

Abstract

Purpose: Primary muscle tension dysphonia (pMTD) is a form of vocal hyperfunction with no preexisting tissue trauma to the vocal folds. There are no known structural or neurological causes of pMTD, and there is rarely obvious, confirmatory evidence to reliably diagnose individuals accurately. Furthermore, acoustic and aerodynamic measurements taken during voice assessments vary widely within this population. The purpose of this study was to find subgroups within a sample of pMTD patients based on acoustic and aerodynamic measurements. We use a computational approach to elucidate what has largely been observational in the past.

Method: A retrospective chart review was conducted to collect variables of interest for a sample of 72 pMTD patients seen at the NYU Langone Voice and Swallowing Center from January 1, 2021, to October 1, 2023. An exploratory factor analysis was conducted to find simpler structures in the data. Using factor scores from each patient, a k-means clustering analysis was conducted.

Results: The exploratory factor analysis grouped together variables across patients, which resulted in three principal axes. These three principal axes separately consisted of aperiodicity, fundamental frequency, and aerodynamic measurements. These principal axes explained 44.7% of the total variance. Four clusters of patients were identified across the three principal axes. These were characterized by (a) a high amount of aperiodicity in the voice, (b) lower fundamental frequency values, (c) higher fundamental frequency values, and (d) high aerodynamic values.

Conclusions: The clusters identified in the current study are reliable and moderately separated. Furthermore, these clusters align with previously identified subgroups in related work. The analysis presented here lays the groundwork for additional clustering analyses with new pMTD samples, as well as future work establishing subtype classifications of pMTD.