Verbal Reaction Time as an Indicator of Sleepiness in Older Adults

Thu 10th Jul, 2025

A recent study conducted by researchers at the University of California, Los Angeles, has demonstrated that Verbal Reaction Time (VRT)--the duration it takes for an individual to respond verbally--can serve as a significant marker for identifying sleepiness in older adults. This research, published in the journal Sleep Science and Practice, highlights the potential for VRT to passively assess excessive sleepiness, particularly in older individuals who are on sedative medications.

Sleepiness poses considerable safety risks in everyday life, yet it often goes unreported, especially among the elderly. Excessive sleepiness has been linked to various hazards including motor vehicle accidents, cognitive decline, and falls, particularly among older individuals prescribed benzodiazepine receptor agonists (BZRAs). Traditional methods of evaluating sleepiness can be intrusive or impractical for real-world application. The findings from this study offer a scalable solution for detecting sleepiness, which could aid in identifying individuals at risk before they encounter accidents or deteriorating health.

The study focused on adults aged 55 and older who had a history of insomnia and were using BZRAs, and participants were sourced from a clinical trial aimed at deprescribing these medications. During the research, participants engaged in cognitive assessments through a mobile application that recorded their verbal responses. The research team measured the VRT, which is defined as the interval from the initiation of recording to the first spoken word, and compared this data with the participants' self-reported levels of sleepiness.

Utilizing advanced analytical tools, researchers explored the correlation between the speed of verbal responses and the participants' perceived sleepiness. Furthermore, they evaluated whether a computer model could accurately predict levels of sleepiness based on voice data. The model successfully forecasted participants' self-reported sleepiness through their voice recordings. Those who took longer to respond verbally after a prompt indicated higher levels of sleepiness. The computer model demonstrated a strong predictive capability, achieving an F1-score of 0.80 ± 0.08, which reflects the model's effectiveness in balancing accuracy and consistency.

Additionally, the voice analysis method reliably distinguished between speaking and silence with an accuracy of 92.5%. These results underscore the potential of voice timing as a practical tool for monitoring sleepiness, especially outside clinical environments. The lead researcher noted that the speed at which an individual begins to speak can provide valuable insights into their level of alertness, suggesting the feasibility of utilizing voice as a passive, scalable method for assessing sleepiness during daily activities.

Future research aims to expand this methodology to larger and more diverse populations, with an eye towards integrating the approach into everyday technologies such as smartphones and telehealth platforms. There is also potential for investigating how voice-based markers may monitor the effects of medications or uncover early signs of cognitive decline.


More Quick Read Articles »