Language-Independent Acoustic Biomarkers for Quantifying Speech Impairment in Huntington's Disease

Area of Science:

• Neurology
• Speech-Language Pathology
• Acoustics

Background:

• Huntington's disease (HD) significantly impacts voice and speech.
• Objective, language-independent methods are needed for assessing speech impairment in HD.
• Current assessment methods may not be universally applicable across different languages.

Purpose of the Study:

• To identify language-independent acoustic features for quantifying speech dysfunction in HD.
• To assess the utility of these features in English-, Spanish-, and Polish-speaking individuals with HD.
• To explore potential subgroups within the HD population based on speech characteristics.

Main Methods:

• Ninety participants with HD and 83 controls completed speech tasks (vowel, syllable, reading).
• Mobile devices were used for recording speech data.
• Principal Component Analysis (PCA) and unsupervised clustering were applied to identify language-independent features and HD subgroups.

Main Results:

• Forty-six language-independent acoustic features significantly differentiated HD participants from controls.
• PCA revealed four distinct speech clusters within the HD group.
• Speech clusters correlated with Unified Huntington's Disease Rating Scale (UHDRS) scores, disease stage, and dysarthria severity.

Conclusions:

• Acoustic features can objectively quantify speech impairment in HD.
• These features show promise for multilanguage studies assessing disease progression and severity.
• The identified speech clusters may reflect different disease trajectories or subtypes in HD.