Developing Facial and Speech Digital Tools for Clinical Assessments of Patients with Huntington's Disease: An ExploratoryStudy

Facial dyskinesia and voice alterations are typical characteristics of Huntington’s disease, (HD). These two clinical features assessed using digital venues have been shown correlated with the progression of HD independently. However, no studies have investigated the synergistic effects of these two clinical features in a digital way, and their relationships with brain function remained unclear. In this study, we employed a standardized Python-based protocol to extract 17 facial featuresin four facial parts and 42 speech featuresinto five main clusters. HD patients exhibited distinctive facial and speech abnormalities compared to HC, including reduced blink frequency, prolonged blink intervals, and increased facial dyskinesia, as well as loudness, phonation instability, and temporal measures. Correlation analyses revealed that variability in upper facial movements and speech instability were not only associated with motor assessments, but also with cognitive function. Moreover, both facial and speechfeatures in HD correlated with brain structuraland functional changes via neuroimaging analysis. Collectively, our results suggest that the digital facial and speech features represent sensitive and accessible markers with neurobiological evidence for clinical assessment of HD