Human Voice and Cognition

summary

• This presentation explores the use of human voice analysis as a non-invasive biomarker for early detection and monitoring of Alzheimer's Disease (AD). The presentation reviews current research and discusses how voice analysis can reveal valuable insights into the progression of AD. The two targeted research questions are 1) to validate the human voice as an AD digital marker and 2) to show the human voice as a culturally and educationally unbiased biomarker. FHS collection of voice data is outlined, both in-clinic and remotely via smartphones.
• Research is summarized regarding acoustic features of speech. These are language-independent sound properties of speech like pitch, without linguistic features related to content and structure. Associations between the acoustic features and various clinical indicators and biomarkers are discussed, including neuropsychological test performance, mild cognitive impairment (MCI), MRI measures, and plasma AD biomarkers. The studies show relationships, some significant, between acoustic features and these indicators. In addition, a study comparing the American FHS cohort and the Malaysian AGELESS cohort assesses whether voice features can serve as culturally and educationally unbiased biomarkers. The acoustic features are similar across both groups. These findings together suggest that voice analysis could be a universally applicable tool for AD detection.