Longitudinal changes in speech features in Huntington’s disease in a multilingual population [Conference Abstract]

Constantin, Valentina Georgea, Fahed, Vitoria, Doheny, Emer, Filan, Ruth, Collazo, Carla, Piotrowska, Joanna, Mann, Elliot, Morgan-Jones, Philippa ORCID: https://orcid.org/0000-0003-2240-691X, Mills, Laura, Drew, Cheney ORCID: https://orcid.org/0000-0002-4397-6252, Rosser, Anne ORCID: https://orcid.org/0000-0002-4716-4753, Cousins, Rebecca, Witkowski, Grzegorz, Cubo, Esther, Busse, Monica and Lowery, Madeleine 2026. Longitudinal changes in speech features in Huntington’s disease in a multilingual population [Conference Abstract]. Journal of Neurology, Neurosurgery and Psychiatry 97 (Supp 2) , A88. 10.1136/jnnp-2026-ehdn.197

Full text not available from this repository.

Abstract

Background: Automated measures of speech impairment have been shown to be associated with motor, cognitive and functional decline in Huntington’s disease (HD). This offers the potential to use speech-based metrics to measure disease progression. However, it is not clear whether such metrics are sensitive to changes over time, or to language for use in multi-site clinical trials. Aims: Here we examined longitudinal changes in temporal speech features during a reading task and a sequential motion rate task in three languages. Methods: Speech was recorded from 43 participants with HD (17 English, 19 Spanish, 7 Polish) using mobile devices at baseline and 12 months later. Bursts of voiced activity and temporal speech features during a passage reading (PR) task and a syllable repetition (SR) task (/pa-ta-ka/) were automatically detected and compared at baseline and follow-up. Statistical analysis was applied using linear mixed models with false discovery rate correction for multiple comparisons. Results: No significant difference was observed between cUHDRS or its individual components at baseline and 12-month follow-up (p>0.05). Significant differences were observed in temporal speech features, including longer speech time (p<0.01) and a higher number of pauses during the PR task, and shorter maximum task duration and lower net speech time during the SR task (p<0.05 for all). Conclusions: The results suggest that automatically detected speech features may be sensitive to longitudinal changes in disease progression that clinical measures do not capture. These findings support further exploration of digital speech measures as an objective clinical tool to monitor Huntington’s ­progression.

Item Type:	Short Communication
Date Type:	Publication
Status:	Published
Schools:	Schools > Engineering Schools > Medicine Research Institutes & Centres > Centre for Trials Research (CNTRR)
Publisher:	BMJ Publishing Group
ISSN:	0022-3050
Last Modified:	26 Oct 2026 13:00
URI:	https://orca.cardiff.ac.uk/id/eprint/190433

Actions (repository staff only)

Edit Item