Audio Analysis of Acoustic and Linguistic Features in Huntington’s disease (Audio-HD)

The diagnosis of Huntington’s disease (HD) is based primarily on the motor exam
component of the Unified Huntington’s Disease Rating Scale (UHDRSTM). The UHDRSTM has a
single item to grade dysarthria from 0-4 (normal to anarthria), however automated speech feature
detection could be a more sensitive biomarker of disease onset or progression.
Objectives: Using a data-driven exploratory approach, we sought to identify the features of
speech that differentiated HD from controls using software developed by Canary Speech.

Methods: HD participants (n=13) and healthy controls (HC, n=12) were matched for sex, age,
and education. Participants completed an eight minute tablet-based protocol (Audio-HD) that
included open-ended questions, passage reading, narrative prompt, picture description, and audio
recording of the Stroop Color and Word Test (SCWT).

Results: The HD cohort’s average UHDRSTM total motor score was 16.8 (SD=16.15) and
expanded CAG repeat length 44.3 (SD=3.10). Welch’s t-test (95% CI) identified >1000 features
of speech that differentiated HD from controls. Features acquired during SCWT and passage
reading prompts identified the highest number of significant differences between cohorts.
Among many others, speech dynamics (P < 0.026), duration per word (P < 0.010), words per
second (P < 0.029), bandwidth (P < 0.012), and contrast (P <0.012) were significant.

Conclusions: The Audio-HD protocol using Canary Speech is a sensitive assessment for
detecting speech changes in HD. Identified features will be integrated into learning models to
generate speech biomarkers with the highest sensitivity for early changes in HD.

Keywords: Huntington’s disease, Dysarthria, Speech, Language, Biomarkers

Luis A. Sierra, Clementina J. Ullman, Karen Hildebrand, Julia S. Dierker, Magdaline Mwangi, Henry O’Connell, Samuel A. Frank, Simon Laganiere. “Audio Analysis of Acoustic and Linguistic Features in Huntington’s disease (Audio-HD),” Poster at The 29th Annual Meeting of the Huntington Study Group, Tampa, Florida, Nov 3-5, 2022.