Category: Huntington's Disease
Objective: To investigate the use of digital technology to assess speech in people with Huntington disease (HD).
Background: Dysarthria is common in Huntington disease and changes in speech may be detected before the manifest of overt motor symptoms (1,2). Thus, there is potential for using speech biomarkers as an objective indicator of HD symptom onset, which may be useful in disease-modifying therapies and clinical trials targeting individuals with premanifest and early manifest HD.
Method: 53 participants were included in the analysis (mean age=43.3±13.2 years). 9 participants were classified as prodromal-HD, 22 as manifest HD (UHDRS score=42.0±15.1), and 22 controls. Participants were asked to perform three speech assessment tests at their natural pace and voice: reading of the Rainbow Passage, counting from 1 to 50, and counting backwards from 50 by 3. Data were obtained using a voice recorder with a microphone clipped to the participant collar. Data were sampled at 44.1 kHz. Dysarthria was assessed as part of the UHDRS (average dysarthria score of HD (1.32±0.65), prodromal-HD (0.02±0.15), controls (0.0±0.0)). Acoustic and clinical metrics such as pitch, pitch variation, loudness, loudness variation, and power were extracted for analysis. Univariate analysis was used to compare across groups while adjusting for age, sex and education level. Significance was set at alpha=0.05 and Cohen’s coefficient (d) was used to measure effect size.
Results: Participants with HD exhibited significantly lower pitch and pitch variation, loudness, speech rate and backwards counting rate, but higher loudness variation and counting error than prodromal-HD patients (d ranges from 0.53-1.57) and controls (d=0.15-2.09), while the largest effect was observed in loudness variation (41% increase, p<0.001, d=1.57 and 58% increase, p<0.001, d=2.09, respectively). Prodromal-HD participants exhibited significantly lower power than controls (p=0.011, d=0.93). The extracted features were able to distinguish between groups with different UHDRS dysarthria scores of 0 (n=32), 1 (n=14), and ≥2 (n=7) (p<0.050, d=0.93-2.85).
Conclusion: Digital technology can automatically assess speech performance in people with HD and can differentiate based on disease status and UHDRS dysarthria score. Development of objective and sensitive metrics to assess speech changes early in disease could be beneficial for novel therapies and clinical trials.
References: [1] Vogel et al. Speech acoustic markers of early stage and prodromal Huntington’s disease: a marker of disease onset? Neuropsychologia 2012;50:3273–3278. [2] Harrington et al. Cognitive domains that predict time to diagnosis in prodromal Huntington disease. J Neurol Neurosurg Psychiatry 2012;83:612–619.
To cite this abstract in AMA style:
J. Adams, E. Dorsey, M. Coffey, M. Pawlik, C. Tarolli, R. Schneider, B. Najafi, H. Zhou, A. Vaziri, H. Nguyen. Automatic and objective speech analysis in Huntington disease [abstract]. Mov Disord. 2021; 36 (suppl 1). https://www.mdsabstracts.org/abstract/automatic-and-objective-speech-analysis-in-huntington-disease/. Accessed December 10, 2024.« Back to MDS Virtual Congress 2021
MDS Abstracts - https://www.mdsabstracts.org/abstract/automatic-and-objective-speech-analysis-in-huntington-disease/