Researchers Link Frequent Speech Pauses to Possible Early Dementia Warning Signs

The brain may reveal its earliest warning signs not through scans or blood tests, but through everyday conversation.

Researchers from Baycrest, University of Toronto, and York University have found that subtle speech patterns — including pauses, filler words like “um” and “uh,” and difficulty finding the right word — may provide important clues about declining cognitive health.

The study suggests that ordinary speech could become a simple and noninvasive tool for detecting early changes linked to memory problems and dementia risk.

Scientists focused on executive function, a group of mental abilities that includes planning, attention, memory, processing speed, and flexible thinking. These functions naturally weaken with age and are often affected early in neurodegenerative diseases such as dementia.

For the research, participants were shown detailed images and asked to describe them naturally. They also completed standard cognitive tests used to measure executive function. Researchers then used artificial intelligence to analyze the speech recordings in detail.

The AI system identified hundreds of subtle speech features, including the frequency and duration of pauses, timing patterns, speaking speed, and the use of filler words. These speech markers strongly predicted how well participants performed on cognitive tests, even after accounting for age, education, and sex.

Researchers say speech analysis could eventually offer major advantages over traditional cognitive testing. Standard tests can be time-consuming and often become less reliable when repeated because people improve through familiarity. In contrast, speech occurs naturally every day and could potentially be monitored continuously and unobtrusively at home or in clinics.

Although more long-term studies are still needed, scientists say combining speech analysis with other health measures could improve early detection of dementia and create more accessible ways to monitor brain health over time.

REFERENCE: Hsi T. Wei, Dana Kulzhabayeva, Lella Erceg, Mira Kates Rose, Kiah A. Spencer, Jessica Robin, Ellen Bialystok, Jed A. Meltzer. Natural Speech Analysis Can Reveal Individual Differences in Executive Function Across the Adult Lifespan. Journal of Speech, Language, and Hearing Research, 2025; 68 (12): 5708 DOI: 10.1044/2025_JSLHR-24-00268