Voice can reveal the health of the brain and body

The study involved 2,928 Spanish‑speaking adults from five Latin American countries — Argentina, Chile, Colombia, Mexico, and Peru. Among them were both healthy individuals and participants with mild cognitive impairment, Alzheimer’s disease, and various forms of frontotemporal dementia. The researchers did not look for a single “marker of aging” in the voice; instead, they used machine learning to combine multiple parameters: speech rate and pauses, pitch, emotional tone, vocabulary, the precision of formulations, and the volume and structure of statements. Based on these data, the model predicted chronological age and calculated the “speech age gap” — the difference between a person’s real age and the age that “sounds” in their speech.

It turned out that this gap is associated with several independent indicators of aging. People whose speech sounded older than their actual age were more likely to show signs of accelerated aging in various body systems. The difference between speech age and chronological age correlated with brain age, which was assessed using structural and functional neuroimaging, as well as with epigenetic aging — it was measured using three independent “DNA methylation clocks”: they show how old the body appears at the level of chemical changes in DNA.

A more pronounced gap was also associated with worse cognitive performance: overall cognitive function, executive abilities, and various types of memory declined. It is important that the connection was observed not only with speech and language tests but also with tasks unrelated to language—that is, the speech profile reflected broader cognitive processes.

Particularly noticeable differences were found between healthy individuals and patients with dementia. In healthy participants, the gap was smallest, while in people with Alzheimer’s disease and frontotemporal dementia, it progressively increased. Moreover, the comprehensive speech indicator distinguished these groups more accurately than individual acoustic or linguistic parameters taken separately. In the case of Alzheimer’s disease, the speech index also correlated with the level of p‑tau217 in plasma — one of the key biomarkers of Alzheimer’s pathology. In addition, it reflected the degree of cognitive and clinical impairments.

The social dimension also turned out to be unexpected: in healthy individuals and patients with dementia, accelerated “speech aging” was associated with a more unfavorable social background — a combination of factors throughout life, including the level of education, financial situation, food security, access to medical care, and early life experiences. It turns out that speech reflects not only biological processes but also the accumulated social burden.

According to Professor Agustín Ibáñez, the voice contains much more information about aging than was previously believed. It simultaneously reflects the passage of time and signals from the brain, body systems, cognitive functions, and even from the social environment.

One of the main advantages of this approach is the low cost and simplicity of data collection. Unlike many modern methods (MRI, blood tests, molecular tests), speech can be recorded remotely, repeatedly, non‑invasively, and at almost no cost.

“Speech clocks” are not yet a diagnostic test for dementia. The study was cross‑sectional, meaning it does not prove that “older” speech predicts the development of cognitive impairments in the future. To be implemented in clinical practice, long‑term observations are needed, as well as testing in more natural speech conditions, and verification in other languages and cultures.