Predicting State and Trait Vulnerability to Sleep Loss Using Functional Brain Networks

Sleep deprivation is a significant risk to health and safety. In the UK, 8.7 million shift workers are routinely exposed to sleep loss, while one in ten UK adults regularly sleep less than is healthy and safe. Exposure to sleep deprivation has a widespread impact on the brain, increasing the risk of potentially catastrophic errors. Several high-profile disasters have been partly caused by sleep deprivation including the Space Shuttle Challenger disaster and Chernobyl nuclear meltdown. However, sleep deprivation leads to serious injury and fatality every day across the UK, causing an estimated 20% of all motor vehicle crashes and 13% of all workplace injuries. The cost to our economy is estimated at £39 billion, every year. As sleep deprivation remains common in a 24/7 society, our challenge is to develop strategies to keep people safe. The use of technology is a rapidly growing strategy and either (i) continuously monitors individuals to warn of fall asleep events or (ii) identifies those most at risk from the effects of sleep deprivation. In our recent review of technologies however, most were either not validated or were unable to predict vulnerable individuals prior to exposure. New approaches and markers predicting both state and trait vulnerability to sleep loss are therefore urgently needed, and our project will make exciting progress in this respect. Our project aims to develop easy-to-obtain markers that detect (i) when brain failure occurs during sleep deprivation (state vulnerability), and (ii) identify who is most at risk of this occurring prior to exposure (trait vulnerability). By assembling a multidisciplinary team of sleep scientists, applied mathematicians and cognitive neurophysiologists, we will apply advanced mathematical techniques to brain activity (EEG) and combine with gold standard performance outcomes from a one-of-a-kind sleep deprivation dataset. Objective 1 will examine whether novel EEG markers detect real-time performance impairment due to sleep deprivation. Functional brain networks will be derived from EEG across 24-hours of sustained wakefulness, and compared against instability in sustained attention, the gold standard measure of sleep-related performance impairment. Objective 2 will develop a novel EEG marker that accurately classifies individuals vulnerable to sleep loss, using machine learning techniques. Objective 3 will validate our biomarker as a fitness for duty test, that is, determine how well it predicts performance impairment during a simulated night shift. Objective 4 will validate our marker as a fitness to drive test, that is, determine how well it predicts individuals who experience a near-crash event while driving a real car following loss of sleep. Potential application of our project is strong with multiple beneficiaries. We will advance understanding of brain networks and behavioural resilience during sleep deprivation, while discovering a novel EEG biomarker that predicts state and trait vulnerability to sleep loss. EEG is easy-to-collect and cost-effective, making future scale-up realistic. Our project outcomes can be incorporated into future ‘fitness for duty/fitness to drive’ technologies, addressing a major gap in a rapidly growing, economically lucrative market. We will work with technology innovators to seek commercialisation of outcomes. Any future device would benefit individuals and operators in safety critical sectors by identifying and supporting those most vulnerable to the effects of sleep deprivation. Longer-term, this has potential to benefit society-at-large, socially and economically, by saving lives in the workplace, on our roads, and beyond.