Voice Assessment in Patients with Amyotrophic Lateral Sclerosis: An Exploratory Study on Associations with Bulbar and Respiratory Function

Abstract

Background: Speech production is a possible way to monitor bulbar and respiratory functions in patients with amyotrophic lateral sclerosis (ALS). Moreover, the emergence of smartphone-based data collection offers a promising approach to reduce frequent hospital visits and enhance patient outcomes. Here, we studied the relationship between bulbar and respiratory functions with voice characteristics of ALS patients, alongside a speech therapist's evaluation, at the convenience of using a simple smartphone.

Methods: For voice assessment, we considered a speech therapist's standardized tool-consensus auditory-perceptual evaluation of voice (CAPE-V); and an acoustic analysis toolbox. The bulbar sub-score of the revised ALS functional rating scale (ALSFRS-R) was used, and pulmonary function measurements included forced vital capacity (FVC%), maximum expiratory pressure (MEP%), and maximum inspiratory pressure (MIP%). Correlation coefficients and both linear and logistic regression models were applied.

Results: A total of 27 ALS patients (12 males; 61 years mean age; 28 months median disease duration) were included. Patients with significant bulbar dysfunction revealed greater CAPE-V scores in overall severity, roughness, strain, pitch, and loudness. They also presented slower speaking rates, longer pauses, and higher jitter values in acoustic analysis (all p < 0.05). The CAPE-V's overall severity and sub-scores for pitch and loudness demonstrated significant correlations with MIP% and MEP% (all p < 0.05). In contrast, acoustic metrics (speaking rate, absolute energy, shimmer, and harmonic-to-noise ratio) significantly correlated with FVC% (all p < 0.05).

Conclusions: The results provide supporting evidence for the use of smartphone-based recordings in ALS patients for CAPE-V and acoustic analysis as reliable correlates of bulbar and respiratory function.

Keywords: ALS; acoustic analysis; digital health; personalized medicine.

Figure 1

An example of voice sound wave analysis, encompassing the reading of phrase C (a,b), from the CAPE-V scale, and the sustainable phonation of the vowel /a/ (c,d); (a,c) were recorded from a single patient in a more advanced disease state (ALSFRS-R total score of 35), while (b,d) depict a patient in a less advanced disease state (ALSFRS-R total score of 46). Notably, even though only through visual observation, discernible distinctions between the two tasks emerge, being particularly more evident during the reading of phrase C. The sentence is presented in both Portuguese (the original language) and English to enhance readability.

Figure 2

Differences in the normalized instrumental-based voice sound features, extracted from phrase C, between the group patients with (white) vs. without (gray) bulbar dysfunction. In general, patients with bulbar impairments experienced more pronounced effects on their speech, characterized by reduced speaking rates and extended durations of pauses. Correlation is significant at the 0.05 level *, 0.001 level **, and <0.001 level ***.

Figure 3

Differences in the normalized CAPE-V scores measures, extracted from phrase c, between the group patients with (white) vs. without (gray) bulbar dysfunction. Correlation is significant at the 0.05 level *, 0.001 level **, and <0.001 level ***.

Figure 4

Representation of the normalized jitter, a feature gauging frequency variability, extracted from sustained phonation of vowel /a/, contrasting with the group of ALS patients with bulbar dysfunction (white) vs. without (gray) bulbar dysfunction. Correlation is significant at the 0.05 level *.