You probably know someone who seems to be aging faster than the calendar says, and someone else who seems to be aging slower.
Most of us read those cues from the face and the body. Scientists are now trying to read them from the voice.
An international team has built what they call a speech clock, an automated tool that estimates how old a person sounds from a few minutes of recorded talking.
The study covered 2,928 Spanish-speaking adults across five Latin American countries and was published in Science Advances.
Also read: Forget chicken: These grains are packed with protein
The gap between how old someone sounded and their actual age tracked brain aging, cognitive decline and early signs of Alzheimer's disease.
In other words, the voice carried signals of accelerated aging that regular memory tests had not yet picked up.
What the voice gives away first
The researchers at Trinity College Dublin and partner universities in Argentina, Chile, Colombia, Mexico and Peru did not look for a single feature. Their algorithm weighed hundreds of them.
Seven groups of features drove most of the signal. Timing slowed down, with longer pauses between words. Pitch varied less. Emotional expression in the voice flattened. Vocabulary leaned on more common, everyday words. People said more to convey the same amount of meaning.
Also read: Living near water pollution may raise risk of rare brain disease
And word choice drifted from precise, specific terms toward broader, more concrete ones.
Put together, those shifts added up to a voice that sounded older than the speaker's calendar age.
Why the speech clock lines up with the brain
The team then checked what the speech age was really tracking.
According to the Trinity College Dublin team's press release carried by EurekAlert, a wider speech age gap lined up with brain shrinkage on MRI scans.
Also read: Scientists discover why some cancers may be less common in type 1 diabetes
It also matched three different measures of biological aging taken from DNA samples, and higher blood levels of p-tau217, a protein that rises in Alzheimer's disease.
Healthy adults had the smallest speech age gaps. People with mild cognitive impairment had larger ones.
People with Alzheimer's disease and frontotemporal dementia had the largest.
The gap also tracked how people did on memory and problem-solving tests, including tests that had nothing to do with language.
Also read: Scientists make breakthrough in fight against deadly childhood brain cancer
The National Institute on Aging has flagged earlier research pointing the same way.
On its own subtle changes in speech page, the agency notes that small shifts in how people speak can show up alongside the first biological signs of Alzheimer's in the brain.
What this is not, yet
The speech clock is research. It is not a product you can download, and it should not be used to screen yourself or a relative for dementia today.
The analysis was done on carefully recorded speech samples in Spanish, by trained researchers, and compared against detailed clinical workups.
Also read: Just 30 minutes of daily exercise could make a difference for your liver
The authors also note the study is a snapshot. It captures a difference between groups at one moment in time. It does not yet prove that a widened speech age gap in a healthy person today will become dementia tomorrow. Longer follow-up studies will have to show that.
What the work does suggest is a plausible path to a cheap, non-invasive early warning. A few minutes of recorded speech is something a phone can capture in a primary care waiting room, in a rural clinic with no scanner, or in a telehealth call.
For readers worried about their own memory or a parent's, the practical step has not changed. Persistent memory problems, word-finding trouble or personality shifts are reasons to speak with a doctor, not with an app.
Earlier work has mapped a long window in which Alzheimer's disease may show signs 15 years before diagnosis, which is exactly why tools that can read that window matter.
The research team has said the next step is to test the speech clock in longitudinal studies that follow healthy adults forward in time to see whose voices quietly age ahead of them.
This article is made and published by Ida-Marie Palm Varbæk, who may have used AI in the preparation.