A study of nearly 3,000 Spanish-speaking adults in five Latin American countries found an association between speech-estimated age and signs of cognitive and biological aging. The researchers caution that the model cannot yet predict an individual's dementia risk.
Published in Science Advances, the study describes an artificial intelligence model that analyzes voice quality and frequency, pauses between words, semantic coherence and emotional modulation.
The participants included healthy adults and people with mild cognitive impairment, Alzheimer's disease and frontotemporal dementia. Speech that the model rated as older than a person's chronological age was associated with stronger signs of biological aging and less favorable social conditions, including financial hardship and limited access to health care.
Healthy participants had the smallest gap between speech-estimated and chronological age. Larger gaps were seen among people with Alzheimer's disease and some forms of frontotemporal dementia.
The study was mainly cross-sectional, so it cannot show whether an older-sounding speech profile predicts who will later develop cognitive decline or dementia. The researchers said the model is not yet a diagnostic test.