Sleep deprivation is already known to significantly impair cognitive performance, slowing reaction times and reducing alertness in ways that can resemble the effects of alcohol intoxication. Despite its impact on public safety, there is currently no quick or reliable biochemical test to measure dangerous fatigue levels in real-world settings such as roads or workplaces.
Researchers publishing in the Journal of Proteome Research have now reported early evidence that chemical changes in saliva could provide a basis for such a diagnostic tool. In a controlled study involving 20 healthy young adult men, scientists observed clear metabolic differences between samples collected after a full night of sleep deprivation and those taken after a normal night of rest.
The findings point toward the possibility of a non-invasive test that could be used to detect extreme fatigue before accidents occur, particularly in settings such as transportation, emergency services and shift-based industries where sleep loss is common.
Drowsy driving alone is believed to contribute to tens of thousands of road crashes annually in the United States, prompting growing interest in technologies that can identify impaired alertness in real time.
“Until now, sleep deprivation has been impossible to measure biochemically, yet it remains one of the greatest burdens of our time,” said Thomas Kraemer, the study’s corresponding author. “This study introduces the first direct biomarkers of sleep loss in saliva under real-world conditions, marking a milestone in forensic investigations.”
To conduct the research, scientists recruited 20 healthy men who typically sleep between seven and nine hours per night. Each participant completed three different sleep conditions in a randomized design: a full night of total sleep deprivation, four nights of restricted sleep (two hours less than usual), and a well-rested condition involving approximately eight hours of sleep.
Saliva samples were collected before and after each condition and analyzed to detect changes in metabolic compounds. The researchers identified 10 key molecular differences that clearly distinguished sleep-deprived samples from well-rested ones.
However, the study found no significant metabolic differences between participants who were well-rested and those who had experienced partial sleep restriction, suggesting that the biological signals of moderate sleep loss may be harder to detect.
Using the data, the team developed a machine-learning model capable of predicting whether a person was sleep-deprived based solely on saliva composition. The system correctly identified sleep-deprived samples with an accuracy rate of 94 per cent.
Researchers noted that some misclassifications may be linked to individual differences in metabolism. In certain cases, participants who had remained awake for 24 hours still showed incomplete recovery markers even after a subsequent full night of sleep, suggesting that biological recovery from sleep loss may vary widely between individuals.
The results suggest what scientists describe as a potential “sleepiness fingerprint” — a distinct pattern of salivary metabolites that could one day be used to detect dangerous fatigue in real time, similar to roadside alcohol testing.
Such a development could have significant implications for road safety and occupational health, offering authorities a potential tool to prevent accidents linked to impaired alertness.
The research team is now preparing a larger international study involving more than 1,000 samples, including shift workers, women, and frequent drivers, to validate and refine the model.
While still in its early stages, the study highlights the growing potential of bio-sensing technologies in addressing sleep-related safety risks, and suggests that saliva testing could one day become a practical tool for identifying when fatigue reaches dangerous levels.