Summary: Researchers report that everyday speech timing — including pauses, filler words, and subtle pacing patterns — closely reflects executive function, a core set of cognitive abilities that support memory, planning, and flexible thinking. By applying AI-driven analysis to natural speech recordings, the team showed that these linguistic timing features can predict performance on standard cognitive tests independently of age, sex, or education.
Because conversational speech is easy to gather, unaffected by repeated testing practice, and can be recorded unobtrusively, it offers a scalable approach for monitoring early brain changes associated with dementia risk. These findings position natural speech as a promising tool for early detection and long-term tracking of cognitive decline.
Key Facts
- Speech as a biomarker: Timing patterns in everyday speech reliably predict executive function across the adult lifespan.
- AI-driven analysis: Machine-learning methods identified hundreds of subtle speech features — including pauses and fillers — that relate to cognitive health.
- Potential for early detection: Natural speech could enable frequent, low-burden monitoring for people at increased dementia risk.
Source: Baycrest
New research from Baycrest, the University of Toronto, and York University suggests that the way people speak in everyday conversation carries measurable signals about brain health.
The study focused on timing features of natural speech — for example, the length and placement of pauses, the use of filler words such as “uh” and “um,” and moments of word-finding difficulty. These features were consistently associated with executive function, the set of mental processes that support organizing, switching between tasks, holding information in mind, and problem-solving.
This investigation is among the first to link natural conversational speech directly to core cognitive abilities rather than to overt clinical dementia. It builds on earlier work suggesting that overall speaking rate relates to preserved cognitive performance in older adults. The current study broadens that perspective by examining fine-grained timing features and their relation to executive skills across adulthood.
“Speech timing is more than a personal style — it is a sensitive indicator of cognitive function,” says Dr. Jed Meltzer, Senior Scientist at Baycrest’s Rotman Research Institute and senior author of the paper titled “Natural Speech Analysis Can Reveal Individual Differences in Executive Function Across the Adult Lifespan.”
In the experiments, participants described complex pictures aloud while also completing standard tests of executive function. The research team used automated methods and machine learning to extract and evaluate hundreds of subtle speech features from the recorded descriptions. Those features, particularly measures of disfluency and pause patterns, reliably predicted cognitive-test performance even after accounting for differences in age, gender, and education.
Executive functions typically decline with age and often show early deterioration in dementia, but traditional cognitive testing can be time-consuming and vulnerable to practice effects — improvements driven by repeated exposure to the same tests. Natural speech offers an advantage because it is a routine behavior that can be sampled frequently, without artificial testing conditions, and it reflects processing speed and linguistic access in a real-world context.
Because speech can be recorded repeatedly and analyzed automatically, it could be used for ongoing surveillance to detect individuals who are declining faster than expected and who may be at higher risk of developing dementia. Such monitoring could be performed in clinical settings or remotely, providing a continuous and ecologically valid measure of cognitive integrity.
Dr. Meltzer notes that early detection matters: “Tracking subtle cognitive changes sooner creates opportunities to intervene earlier, which is essential if treatments or preventive strategies are to be effective at slowing neurodegenerative processes.”
The authors emphasize the importance of longitudinal research that follows the same individuals over time. Longitudinal data will help distinguish normal age-related changes from the earliest signs of disease. The team also suggests that combining natural speech metrics with other assessments — clinical, imaging, or fluid biomarkers — could improve the accuracy and accessibility of early detection.
Funding: This research received support from the Mitacs Accelerate program and the Natural Sciences and Engineering Research Council of Canada (NSERC).
Key Questions Answered:
A: Subtle timing features in everyday speech, such as pause duration and filler use, closely track executive function — a central indicator of cognitive integrity.
A: Speech can be measured frequently and naturally, without the pressure of formal testing, which may reveal decline earlier than traditional assessments affected by practice effects.
A: Automated analysis identified pauses, fillers, and timing patterns that reliably predicted executive-function test scores across adults of different ages.
About this speech and cognitive decline research news
Author: Natasha Nacevski-Laird
Source: Baycrest
Contact: Natasha Nacevski-Laird – Baycrest
Image: The image is credited to Neuroscience News
Original Research: Closed access.
“Natural Speech Analysis Can Reveal Individual Differences in Executive Function Across the Adult Lifespan” by Jed Meltzer et al. Journal of Speech, Language, and Hearing Research
Abstract
Natural Speech Analysis Can Reveal Individual Differences in Executive Function Across the Adult Lifespan
Purpose:
Automated analysis of naturalistic speech has proven useful for detecting cognitive decline in clinical dementia, but it has been less frequently applied to study the more gradual, ordinary cognitive changes that occur with normal aging. Because executive function (EF) tends to decline across adulthood and is difficult to measure longitudinally due to practice effects, speech-based measures represent an attractive alternative. This study investigated links between EF and characteristics of everyday speech.
Method:
Researchers collected two audio picture descriptions from participants in two experiments alongside standard EF assessments. Study 1 included 67 healthy older adults aged 65–75 years, and Study 2 included 174 healthy adults aged 18–90 years. The team computed composite language scores by aggregating speech features that have been reported to change in pathological aging. Principal components capturing common patterns among speech features were derived from a large training dataset to generate speech-domain scores. The analysis examined relationships between language composites or speech principal components and EF while controlling for age, gender, and education.
Results:
In Study 1, measures of word-finding difficulty, reflected in speech disfluencies, showed significant associations with executive function among older adults. Study 2 extended this finding across the adult lifespan, demonstrating that disfluency measures account for individual differences in EF not only in those over 65 but also in younger and middle-aged adults. Other speech features, such as information units and coherence, showed weaker associations with EF and with Montreal Cognitive Assessment scores; these associations did not remain significant after correction for multiple comparisons.
Conclusion:
The results indicate that word-finding ability evident in natural speech relates to general executive function across adulthood. Natural speech analysis therefore offers a convenient, sensitive, and ecologically valid method for assessing cognitive ability and for developing scalable approaches to detect and monitor cognitive decline.