AI Study Links Filler Words and Pauses to Higher Dementia Risk

3 articles · Updated · HuffPost · May 19

Researchers at Baycrest, the University of Toronto and York University found AI analysis of recorded picture-description tasks could predict how participants performed on cognitive tests.
The speech markers included filler words such as “um” and “uh,” pauses and trouble finding words—patterns the study tied to cognitive decline and elevated dementia risk.
Experts not involved in the research said occasional word-finding problems are common in healthy aging, while bigger red flags include repeated questions, getting lost in familiar places and frequently misplacing items.
They also cautioned the findings came from a single time point and may be shaped by cultural or regional speaking styles, limiting how broadly any one person’s speech can be judged.
Doctors said the work is promising for early screening, but emphasized modifiable risks still matter most—especially blood-pressure control, exercise, sleep, heart-healthy diet, sensory care and avoiding smoking or head injury.

Your speech may reveal dementia risk, but can AI distinguish this from the normal memory slips of healthy aging?

AI can now detect dementia risk in your voice. When will this breakthrough be available during a standard doctor's check-up?

Beyond just speech, what other daily digital footprints could AI use to predict our future brain health?