The challenge in finding a simple, accurate, reliable, and affordable tool for the objective assessment of excessive daytime sleepiness (EDS)

Oksenberg et al 2025 Physiol. Meas. https://doi.org/10.1088/1361-6579/ae2b4b

Why Measuring Daytime Sleepiness Is So Hard and How Wearable Technology Could Help

Excessive daytime sleepiness (EDS) is far more than feeling a little tired. It’s a physiological state that affects millions of people, impairing attention, slowing reaction time, and increasing the risk of accidents. For patients with sleep disorders like obstructive sleep apnea (OSA), EDS can be the most debilitating symptom, impacting cognitive performance, productivity, and quality of life, but the most dangerous, falling asleep while driving.

Moreover, according to public health organizations, we live in a sleep deprived society, which often has been described as a public health epidemic. Worldwide most adults’ individuals do not get the recommended 7- 9 hours of night sleep. In a 2019 survey in 12 countries found that 62% of adults do not feel they get enough sleep, but the problem is also prevalent in younger populations, and the direct consequence of this behavior is EDS. Studies have found that adults sleeping less that 7 hours are more likely not only to suffer from several chronic health conditions such as depression, anxiety, obesity, high blood pressure, coronary heart disease, increased likelihood for some cancers, diabetes, but also more likely to die early.

But despite decades of research, clinicians and scientists still lack a simple, affordable, and objective tool that can reliably measure sleepiness outside of specialized sleep labs.

A new systematic review from our research group sheds light on why this problem has been so difficult and why advances in wearable photoplethysmography (PPG) and artificial intelligence may offer a pathway forward.

The Problem: Sleepiness Is Not One Thing, It’s Many

Our work highlights a central challenge: EDS is multidimensional.

• Introspective sleepiness: what people think they feel

• Physiological sleepiness: the underlying drive to sleep

• Manifest sleepiness: observable signs like slowed reactions or microsleeps

Subjective short questionnaires such as the Epworth Sleepiness Scale (ESS) are simple and widely used, but they are prone to bias. Many OSA patients underestimate their sleepiness, and subjective reports often fail to reflect physiological risk.

On the other hand, the objective gold-standard tests, the Multiple Sleep Latency Test (MSLT) and Maintenance of Wakefulness Test (MWT) are laborious expensive, time-consuming, and impractical for repeated or real-world use.

That gap leaves clinicians, researchers, and even public-safety organizations without a scalable method to monitor sleepiness where it matters most: during everyday life.

The Stakes Are High: From Clinical Care to Road Safety

The consequences of misjudged sleepiness are profound:

• Between 31 - 80%  of OSA patients report EDS before treatment, and nearly one-third experience residual EDS even after CPAP treatment.

• Drowsiness plays a role in 17.6% of fatal road accidents in the U.S., and fatigued drivers are three times more likely to crash.

• Many occupations — trucking, aviation, shift work, medicine rely on sustained alertness where lapses can be catastrophic.

A practical, validated sleepiness-detection tool could transform patient management, occupational health, and public safety.

PPG: A Simple Wearable Sensor with Untapped Potential

Photoplethysmography (PPG) is already embedded in millions of smartwatches and rings. It measures changes in blood volume using light, enabling signals like:

• Heart rate (HR)

• Pulse rate variability (PRV) / beat-to-beat intervals

• Oxygen saturation (SpO₂)

• Waveform morphology reflecting autonomic nervous system activity

• 
  

Over the last 15 years, researchers have increasingly explored whether PPG can objectively detect sleepiness. To understand the state of the field, our group conducted a rigorous systematic review of 95 studies, narrowing down to 24 high-quality publications focused specifically on using PPG to assess EDS, drowsiness, fatigue, or alertness.

Figure. Systematic analysis of the selected studies.

What the Science Says: Strong Signals, Promising Trends

1. PPG-derived Heart Rate Variability (HRV) metrics can reflect sleepiness

Many studies show associations between sleepiness and changes in autonomic regulation:

• Lower Root Mean Square of Successive Diferences (RMSSD)

• Reduced total power

• Altered LF and HF components

• Decreased baroreflex sensitivity in sleepy OSA patients

These patterns suggest sympathetic activation and reduce cardiac adaptability in sleepy individuals.

2. PPG correlates strongly with ECG

Studies repeatedly demonstrate that pulse-based variability (BRV/PRV) is a reliable proxy for HRV measured from ECG. In some work, the correlation reaches 98%, supporting the use of consumer wearables for physiological monitoring.

3. Raw PPG waveform features matter too

Several papers go beyond HRV, showing that:

• The position of the dicrotic notch

• Pulse morphology

• PPG amplitude fluctuations

can help track fatigue and sleep-wake transitions.

4. AI models perform surprisingly well

Machine-learning systems trained on PPG features achieved:

• 80–97% accuracy for drowsiness detection in lab settings

• Strong correlations with subjective fatigue scales

• Early detection of drowsiness minutes before behavioral lapses

For example, one smartwatch-based system achieved 88–100% early-detection accuracy after filtering out motion artifacts.

5. SpO₂ spectral features may encode sleepiness

Large population studies found that SpO₂ dynamics, not just mean oxygen levels, correlate with subjective EDS severity although classification models still require refinement.

Where Current Studies Fall Short

Despite encouraging results, the research landscape has limitations:

• Most studies are small, often with fewer than 20 participants.

• Many rely on driving simulators rather than real-world conditions.

• Methods are inconsistent: different devices, feature sets, protocols, and ground-truth labels.

• Few studies compare PPG output directly against objective gold-standard MSLT/MWT tests.

• Motion artifacts, device placement, and lighting remain major challenges.

This fragmentation explains why no PPG-based sleepiness detector has yet reached widespread clinical or regulatory validation.

Why We Believe a Breakthrough Is Coming

Our review highlights a convergence of forces:

• Ubiquitous wearable sensors

PPG is now available on nearly every smartwatch and smart ring.

• Advances in AI signal processing

New models can extract robust physiological features even with noise and motion.

• Large datasets coming online

Wearable-collected PPG signals across millions of users open possibilities for population-scale learning.

Together, these advances could enable a simple, reliable, and affordable sleepiness test — something clinicians have needed for decades.

But one crucial step remains.

The Missing Piece: Rigorous Validation Against Gold Standards

Synthesizing the literature, our group concludes that the field must now move toward:

• Controlled studies comparing PPG-derived sleepiness indices to MSLT/MWT

• Testing across diverse populations, sleep disorders, and real-world settings

• Harmonized signal-processing pipelines

• Clear clinical thresholds and normative ranges

• Regulatory-grade reproducibility and robustness

Only through an adequate validation can a PPG-based EDS tool be adopted in clinics, workplaces, transportation systems, and consumer health platforms.

The Road Ahead

The dream of an objective, practical, continuous sleepiness monitor is no longer science fiction. Our systematic review shows that the building blocks — physiology, sensors, signal processing, and AI — are already here. What’s needed now is a coordinated scientific effort to bring these components together and validate them rigorously.

If successful, such a tool could:

• Improve diagnosis and management of several sleep disorders.

• Reduce fatigue-related accidents

• Help shift workers and drivers monitor their alertness

• Enable personalized fatigue-risk prediction

• Transform wearable devices into clinical-grade sleepiness monitors

The science is promising, the technology is ready, and the societal need is urgent.