Voice Analysis Could Reveal Biological Age and Cognitive Decline, Study Suggests

Subtle changes in an individual's voice may offer insights into their chronological age and how well they are aging, according to a new study by an international team of researchers. The study developed a "speech clock" using machine learning to analyze numerous vocal characteristics, which could potentially serve as an accessible tool for assessing biological age and identifying cognitive decline, including conditions like Alzheimer's disease.

The research, published in the journal Science Advances, involved nearly 3,000 adults aged 18 to 88 from five countries. The machine learning tool examined hundreds of features in participants' voices, such as speech rate, pitch, vocabulary usage, and emotional inflection. By analyzing these elements, the tool generated a "speech age" estimate for each participant.

The gap between this estimated "speech age" and a person's actual chronological age was then used to infer their biological age, indicating how rapidly their body was aging. Researchers found that individuals whose "speech age" was older than their chronological age were more likely to exhibit signs of cognitive decline.

"Our voice appears to contain much more information about aging than we previously recognized," said Dr. Agustin Ibanez, a senior study author and professor at Trinity College Dublin's Global Brain Health Institute. "It captures both the passage of chronological time and signals coming from cognition, the brain, systemic biology and even our accumulated social environment."

Older voices in the study tended to be slower, flatter in pitch, and lacked emotional variation, while younger voices were faster and more concise. These vocal changes can occur naturally with age as vocal folds lose elasticity and lung capacity decreases, typically starting in one's 50s.

The study included 1,504 healthy participants and 1,069 individuals with Alzheimer's disease, along with others diagnosed with different forms of dementia or cognitive impairment. Participants were Spanish speakers from Argentina, Chile, Mexico, Peru, and Colombia. Healthy participants averaged 61 years old, while those with Alzheimer's averaged 71.

The machine learning model identified six key speech elements: timing, pitch, emotion, vocabulary, correctness, and the number of words needed to convey an idea. The study found that people whose speech patterns indicated an older age relative to their chronological age also tended to have poorer cognitive skills and memory.

Furthermore, when researchers compared the "speech age gap" to brain scans, they observed that larger gaps were associated with a higher likelihood of dementia. For individuals diagnosed with Alzheimer's disease, a greater speech age gap correlated with elevated levels of p-tau217, a blood biomarker for the disease. Alzheimer's and dementia can affect speech by causing difficulties with word-finding, articulation, and swallowing.

The "speech clock" offers a potential alternative to more invasive and costly methods for assessing biological age, such as blood samples and MRI scans. The researchers suggest this accessible approach could be particularly beneficial in underserved regions. However, they emphasize that further research is necessary before the tool can be widely implemented.

"The broader finding is nevertheless striking in that a person’s voice may provide a remarkably compact readout of multiple dimensions of aging," Ibanez stated. "From chronological age to brain aging... information traditionally obtained through very different and often expensive measurements appears to converge, at least partly, in the way we speak."