![[Adobe Stock]](https://sp-ao.shortpixel.ai/client/to_webp,q_glossy,ret_img,w_750,h_375/https://www.psypost.org/wp-content/uploads/2025/11/old-man-talking-750x375.jpg)
A recent study reveals that the rhythm and flow of our natural speech may offer insights into our cognitive health. Researchers found that characteristics like the frequency of pauses and filler words are associated with executive functions, the brain’s high-level management skills, suggesting that analyzing how we talk could one day become a simple method for monitoring brain function. The findings were published in the Journal of Speech Language and Hearing Research.
Executive functions are a set of mental processes that help us plan, focus attention, remember instructions, and juggle multiple tasks. These abilities are fundamental to everyday life, from navigating a conversation to managing a complex project.
While it is understood that these functions can decline during normal aging, a more rapid decline can signal the onset of neurodegenerative conditions. Tracking these changes over time, however, presents a challenge because traditional cognitive tests can be affected by practice effects, where a person’s score improves simply from familiarity with the test. This can mask a genuine cognitive decline.
Seeking a more natural and repeatable way to assess these abilities, a team of researchers from the University of Toronto, Baycrest Hospital, and York University turned to speech. Spoken language is an everyday activity that is unlikely to be altered by repeated assessments.
Advanced automated tools can now analyze hundreds of features in a short speech sample, offering a detailed look at a person’s linguistic and cognitive state. While prior work has shown that speech analysis can help identify dementia, less has been known about its connection to the subtle cognitive variations seen in healthy adults across their lifespan.
The investigation was conducted in two parts. The first study involved 67 healthy older adults between the ages of 65 and 75. Each participant provided two 60-second audio recordings describing complex pictures. They also completed a series of established tests designed to measure various aspects of executive function, such as working memory, inhibition, and mental flexibility.
The researchers used a statistical technique to group the results of these cognitive tests into two main underlying components: one related to inhibition and working memory, and another related to verbal fluency and mental shifting.
To analyze the speech recordings, the team extracted more than 700 distinct linguistic features. To make this data manageable, they grouped these features into eight broad categories, or composites, based on language domains known to change in pathological aging. These composites included measures of syntactic complexity, vocabulary richness, and word-finding difficulty. The word-finding composite specifically captured elements like speech rate, the number and length of pauses, and the use of filler words like “um” and “uh.”
When the researchers examined the relationship between the speech composites and the executive function components, a clear pattern emerged. Of the eight speech categories, only word-finding difficulty was strongly associated with performance on the executive function tests.
Individuals who exhibited more signs of word-finding difficulty in their speech, such as more frequent hesitations, tended to have lower scores on both of the executive function components. This suggested that disfluencies in natural speech could be an indicator of broader cognitive control abilities in older adults.
The second study aimed to see if this connection between speech and executive function extended to a wider population. This time, the researchers recruited 174 healthy adults, with ages ranging from 18 to 90. Participants completed the same picture-description task, along with a different but related set of executive function tests. Again, the cognitive test results were distilled into two main components that were very similar to those identified in the first study.
For the speech analysis in this broader group, the researchers employed a different, data-driven strategy. Instead of grouping speech features based on prior knowledge of dementia, they used a large training dataset of speech samples from over 700 individuals to identify the most prominent patterns of variation within ten different speech domains. This process yielded a set of “speech domain scores” for each participant, including scores for domains like timing, syntax, and coherence.
The results of this second study reinforced the initial findings. The strongest associations were found within the timing domain of speech. Specifically, scores reflecting the frequency and duration of pauses were significantly related to the executive function component for verbal fluency and shifting. People who paused more, whether using short filled pauses or longer silent ones, tended to perform less well on tasks requiring mental flexibility and fluent word generation.
This result confirmed that speech timing is linked to executive abilities not just in older adults, but across the entire adult lifespan. Weaker associations were also noted between executive function and the coherence of speech, or how well a person stays on topic.
Jed A. Meltzer, a senior scientist at Baycrest’s Rotman Research Institute and the study’s senior author, explains the significance of these findings. “The message is clear: speech timing is more than just a matter of style, it’s a sensitive indicator of brain health,” he states. The research highlights that natural speech can provide information about processing speed, a sensitive measure of cognitive health, in a way that is more reflective of real-world behavior than traditional, timed laboratory tasks.
The researchers note some limitations and avenues for future work. The fact that different speech analysis strategies were most effective for each of the two studies suggests that the best approach may depend on the population being studied. For older adults, focusing on features known to degrade in dementia was effective, while for a wider age range, a data-driven approach that captures normal variation worked better. The current study also provides a snapshot in time.
Future research should follow individuals over several years to observe how changes in their speech patterns might predict subsequent cognitive changes. According to Meltzer, this could lead to powerful new tools for monitoring brain health. “This research sets the stage for exciting opportunities to develop tools that could help track cognitive changes in clinics or even at home,” he says. “Early detection is critical for any cure or intervention, as dementia involves progressive degeneration of the brain that may be slowed.” By combining natural speech analysis with other measures, researchers hope to make the early detection of cognitive decline more precise and accessible.
The study, “Natural Speech Analysis Can Reveal Individual Differences in Executive Function Across the Adult Lifespan,” was authored by Hsi T. Wei, Dana Kulzhabayeva, Lella Erceg, Mira Kates Rose, Kiah A. Spencer, Jessica Robin, Ellen Bialystok, and Jed A. Meltzer.
