Language age as an indicator of brain health and dementia

Some people enjoy listening. Because of the way they emphasize certain words, which words they use and after which they take a particularly long pause. But whether someone has a beautiful narrative voice seems less interesting than the knowledge that a kind of language age could be calculated from these characteristics.

Researchers have now discovered this. For the study, published in the journal “Science Advances,” researchers evaluated voice recordings from 2,928 Spanish-speaking people from Argentina, Chile, Mexico, Peru and Colombia. 1504 participants were cognitively healthy, 1068 had Alzheimer's dementia, 332 had frontotemporal dementia and 24 had mild cognitive impairment.

The participants had to complete several language tasks: retell a short story, describe a film and name as many animals, vegetables or words with a specific first letter as possible within a minute.

The researchers then analyzed more than 700 features of the recordings, such as speech pace and syllable length, pauses, pitch, word choice, emotional coloring and the amount of information spoken.

Using these characteristics, they trained a computer model to estimate a person's chronological age. This is how a kind of “language clock” was created. They called the difference between the predicted and actual ages the “speech age gap.” If the model estimated that a person was older than they actually were based on their language, the result was a higher language age. However, the estimate is not precise: on average, the model was around nine years off when it came to age.

Healthy participants had the lowest deviations on average. In contrast, people with Alzheimer's or frontotemporal dementia showed a higher “language age”. The deviation was particularly large in forms of frontotemporal dementia in which language itself is severely affected.

Language age was also related to mental performance. The higher the calculated language age, the worse the participants performed on tests of memory, attention and other cognitive abilities. The connections were particularly clear in Alzheimer's and certain forms of frontotemporal dementia; They were strongest in memory performance.

Biological markers were also reflected in language. In Alzheimer's patients, older language age was associated with higher levels of phosphorylated tau-217 in the blood - a biomarker linked to Alzheimer's changes in the brain.

The connection with so-called brain aging clocks was even clearer. These use MRI images to estimate whether a brain appears biologically older or younger than its actual age suggests. People with older language ages were also more likely to have higher estimated brain ages on structural and functional brain scans. The researchers found similar, albeit weaker, connections with epigenetic aging markers in the blood.

In the long term, language could become a simple and inexpensive additional marker for brain health. In contrast to MRI examinations or blood tests, voice recordings can be collected almost anywhere and without medical equipment.

The voice clock is not yet sufficient for a dementia test. The study initially only shows that a higher language age is related to cognitive performance and biological markers. Whether this can also be used to predict who will later develop dementia cannot be determined from the predominantly cross-sectional study. In addition, only Spanish-speaking people from five Latin American countries were examined. Further studies must show whether the process works similarly in other languages ​​and cultures.