Novel 'speech clock' can estimate a person's chronological age

A large study shows that machine-learning clocks based on speech can estimate chronological age and may also offer a window into brain aging, biological aging, cognitive health, cumulative burden, and dementia. 

The new study, just published in leading international journal Science Advances, outlines that researchers have developed a "speech clock" that can estimate a person's chronological age from hundreds of acoustic and linguistic characteristics of their speech. The difference between a person's actual age and their speech-predicted age (called the speech age gap) was also associated with multiple independent markers of biological aging, brain health, cognition, social adversity, and dementia.

The study analysed 2,928 Spanish-speaking participants from Argentina, Chile, Colombia, Mexico, and Peru, including healthy adults and people with mild cognitive impairment, Alzheimer's disease, and different forms of frontotemporal dementia. Rather than looking at a single property of the voice, the researchers used machine learning to analyse hundreds of features capturing how people speak and what they say: incorporating speech rate and pauses, pitch, emotional content, vocabulary, semantic precision, and the amount and organisation of verbal output. Machine-learning models combined these features to estimate chronological age and generate an individual speech age gap.

People whose speech appears older than expected for their chronological age also showed signs of accelerated aging across several biological and clinical systems. The speech age gap was associated with brain age measured using structural and functional neuroimaging. It was also related to epigenetic aging, measured through three independent DNA-methylation clocks, that estimate how biologically "old" the body appears based on age-related chemical changes in DNA. 

Additionally, greater speech-age acceleration was associated with poorer global cognition, executive function, functional abilities, and several forms of memory. Importantly, these relationships were not restricted to language tests: speech age was also related to performance on non-linguistic cognitive measures.

The speech clock also differentiated healthy individuals from people with dementia. Healthy participants showed the lowest speech age gaps, while progressively larger gaps were observed across Alzheimer's disease and forms of frontotemporal dementia. The complete speech-age measure discriminated clinical groups better than individual acoustic or linguistic features considered separately. In Alzheimer's disease, it was associated with higher levels of plasma p-tau217, one of the most important blood biomarkers of Alzheimer's pathology. The same speech-derived measure also tracked cognitive and clinical functioning.

Speech also carried a social signal. Among healthy individuals and people with Alzheimer's disease and other dementia, accelerated speech aging was associated with a more adverse social exposome: a combination of lifelong factors such as education, financial conditions, food insecurity, healthcare access, and early-life experiences.

"Our voice appears to contain much more information about aging than we previously recognised," said Agustin Ibanez, Professor in Brain Health at the Global Brain Health Institute and School of Medicine, Trinity College Dublin and senior author of the study. 

"It captures both the passage of chronological time and signals coming from cognition, the brain, systemic biology, and even our accumulated social environment. This raises the possibility that something as simple and accessible as speech clocks, maybe combined with biomarkers, could eventually complement much more expensive measures of aging."

What is the potential impact of this research?

The potential implications are substantial. Many current measures of biological aging require MRI scanners, blood samples, molecular assays, or specialised clinical assessments. Speech, in contrast, can be recorded remotely, repeatedly, non-invasively, and at very low cost. This could be particularly important in countries and communities where advanced diagnostic technologies are difficult to access. 

Because the study was conducted across five Latin American countries (a region historically underrepresented in dementia research) it also provides evidence that sophisticated biomarkers of aging do not necessarily need to depend exclusively on expensive technologies developed in high-resource settings.

The researchers emphasise, however, that the speech clock is not yet a diagnostic test for dementia. The study was primarily cross-sectional, meaning that it cannot establish whether an older-appearing speech profile predicts who will subsequently develop cognitive decline or dementia. Longitudinal studies, validation in additional languages and cultures, and testing in more naturalistic speech settings will be required before clinical implementation.

"The broader finding is nevertheless striking in that a person's voice may provide a remarkably compact readout of multiple dimensions of aging," added Prof. Ibanez. 

"From chronological age to brain aging, epigenetic aging, cognition, Alzheimer's-related pathology, social exposures, and dementia phenotypes, information traditionally obtained through very different and often expensive measurements appears to converge, at least partly, in the way we speak. 

"If confirmed longitudinally and across populations, speech could ultimately become one of the most scalable tools for monitoring healthy and accelerated aging-potentially transforming an everyday human behaviour into a window onto the biology of aging."