F022 Indentifying key features for predicting clinical status in Huntington’s disease patients using raw speech samples: a parsiminious approach
Year: 2024
Type: conference-paper
Abstract: Background Traditional assessment of HD patients involves time-consuming and complex tests. Speech has emerged as a potential biomarker for HD assessment, with recent studies suggesting its use for predicting clinical scores [Riad et al., 2022]. However, these approaches require costly extraction of hand-crafted features, mostly derived from human annotations, necessitating a more parsimonious approach. Aims This study aims to identify key features relevant for predicting the clinical status of patients with Huntington’s Disease (HD) using machine learning techniques applied to raw voice recordings. Methods We replicated the study by Riad et al. to validate results across the same cohort and a larger cohort. The larger cohort comes from the BioHD (NTC01412125) and RepairHD (NCT03119246) studies, comprising 107 speakers (86 with HD, 21 prodomals; mean age = 52.26 ± 11, Total Functional Capacity (TFC) = 11.1 ± 2.2). On this cohort, we investigate parsimonious modeling using the LASSO technique to identify the most relevant features for prediction. Results Preliminary findings indicate the reproducibility of Riad et al.’s results on both the same and larger cohort. The LASSO regression shows that a few features, primarily demographic features and speech rate, can maintain good prediction performance. Conclusions This study identifies key features to predict the clinical states of HD gene carriers and demonstrates the potential for automating predictions. Previous studies relied heavily on human annotations of the speech task [Riad et al., 2022], highlighting the significance of our approach in reducing the need for extensive manual input. References Riad R, Lunven M, Titeux H, Cao XN, Hamet Bagnou J, Lemoine L, Montillot J, Sliwinski A, Youssov K, Cleret de Langavant L, Dupoux E, Bachoud-Lévi AC. Predicting clinical scores in Huntington’s disease: a lightweight speech test. J Neurol. 2022 Sep;269(9):5008–5021. doi: 10.1007/s00415-022-11148-1. Epub 2022 May 14. PMID: 35567614; PMCID: PMC9363375.
Language: English
Topic: Voice and Speech Disorders
Subfield: Physiology
Field: Medicine
Domain: Health Sciences
Open Access status: gold
OpenAlex ID: W4402373975