AI Speech Clock Can Estimate Biological Age and Detect Dementia Signals

An international research team has developed an artificial intelligence tool called a speech clock to estimate biological aging and detect cognitive impairment through vocal analysis. The model evaluates acoustic and linguistic features from recorded speech to identify disparities between chronological age and biological age, analyzing voice recordings of nearly 3,000 people from five Latin American countries.

Voice Analysis Evaluates Brain Health and Biological Aging

The study, published in Science Advances, analyzed 2,928 Spanish-speaking participants ranging in age from 18 to 88 years, with an average age of 65. The cohort included healthy adults alongside individuals diagnosed with mild cognitive impairment, Alzheimer’s disease, and frontotemporal dementia variants across Argentina, Chile, Colombia, Mexico, and Peru. Research from the Drexel University College of Medicine, published in the Journal of Voice, indicates that specific voice alterations may relate to mental health.

Machine learning models examined more than 700 acoustic and linguistic characteristics derived from brief voice recordings. The tool measures speech rate, pauses, vocal pitch, emotional content, vocabulary, semantic precision, and the organization of verbal production. By synthesizing these metrics, the algorithm estimates a speaker’s chronological age and calculates the speech age gap against their biological age.

Cognitive Decline Correlates With Accelerated Speech Aging

The findings indicate that healthy participants maintained the smallest discrepancies between chronological age and predicted speech age. Conversely, wider gaps emerged among patients with Alzheimer’s disease and various forms of frontotemporal dementia.

An elevated speech age gap also correlated with poorer results in memory tests, the ability to organize tasks, and general cognitive performance. The research also linked accelerated speech aging to adverse lifelong social conditions, including worse economic situation, nutrition, and access to healthcare. Individuals with voice problems showed a 104% higher probability of subsequently receiving a cognitive impairment or dementia diagnosis compared to controls in certain analyses, while hearing loss alone showed a 27% increase and combined voice disorders showed a 29% increase.

Trinity College Dublin Researchers Highlight Diagnostic Potential

Nuestra voz parece contener mucha más información sobre el envejecimiento de lo que sabíamos hasta ahora.

Agustín Ibáñez

Ibáñez noted that many of the techniques currently used to evaluate biological aging require MRI scans, blood samples, molecular analysis, or specialized clinical assessments. In contrast, voice recordings offer a non-invasive, repeatable, and low-cost alternative suited for regions with limited access to advanced diagnostic infrastructure.

Unresolved Questions in Vocal Biomarker Research

The investigative team cautions that the speech clock is not currently a diagnostic test for dementia, nor does it prove that a person whose voice appears older will necessarily develop cognitive problems in the future. Additional research and cross-cultural validation across other languages are required before clinical implementation, and it remains unknown whether treating these voice and hearing issues alters subsequent cognitive health trajectories.