Sleep

Hypnodensity sleep metrics from fingertip photoplethysmography are better associated with daytime sleepiness and fatigue than traditional metrics from polysomnography.

TL;DR

PPG-based hypnodensities improve the characterisation of sleep quality and better predict patient-reported outcomes in SDB patients than traditional hypnogram-based sleep metrics from polysomnography.

Key Findings

PPG-derived hypnodensity metrics correlated with self-reported daytime fatigue and sleepiness at levels better or similar to traditional polysomnography-based metrics.

  • Study included 2,280 patients (65% male, mean age 62 years, mean AHI 23.5/h)
  • Comparisons were made between high-frequency (1/s) PPG-derived hypnodensities and manually scored polysomnography hypnogram-based metrics
  • The level of correlation was described as 'better or similar compared to traditional hypnogram-based sleep metrics derived from manually scored polysomnography'
  • Patient-reported outcomes assessed were daytime fatigue and sleepiness

PPG-based hypnodensities for N1 and N3 sleep showed a statistically stronger association with fatigue and sleepiness than manually scored N1 and N3 sleep.

  • The stronger association for N1 and N3 hypnodensity metrics was statistically significant at p < 0.005
  • This comparison was made against the corresponding manually scored N1 and N3 sleep stages from polysomnography
  • The hypnodensity approach provides probabilities for each sleep stage at 1-second resolution rather than 30-second epoch classifications

Multivariable logistic regression revealed significant associations between multiple PPG-based hypnodensity metrics and daytime sleepiness.

  • Significance threshold met at p < 0.01 for multiple PPG-based hypnodensity metrics
  • The model was adjusted for age, sex, body mass index, AHI, and total sleep time below 90% blood oxygen saturation
  • Multiple (not just one or two) PPG-based hypnodensity metrics showed significant associations in this adjusted model

Traditional polysomnography sleep metrics are limited by 30-second epoch classification, which oversimplifies sleep stage continuity and fragmented sleep typical in sleep-disordered breathing.

  • In standard polysomnography hypnograms, each 30-second epoch is classified into one of five sleep stages
  • This approach oversimplifies 'sleep stage continuity and fragmented sleep, typical in sleep-disordered breathing (SDB)'
  • Deep learning-derived hypnodensities provide probabilities for each sleep stage at higher temporal resolution (1/s in this study)

The study population was a large clinical cohort of sleep-disordered breathing patients with moderate average disease severity.

  • Total sample size was 2,280 patients
  • 65% were male with a mean age of 62 years
  • Mean apnoea-hypopnoea index (AHI) was 23.5/h, consistent with moderate sleep-disordered breathing
  • Sleep recordings were nocturnal fingertip photoplethysmography (PPG)

PPG-based hypnodensities were captured from nocturnal fingertip photoplethysmography using a deep learning model at 1-second resolution.

  • Recordings were made using fingertip PPG devices, which are simpler and less burdensome than full polysomnography
  • The deep learning model produced sleep stage probabilities (hypnodensities) at 1/s frequency
  • The authors suggest future use could include tracking sleep quality with fingertip pulse oximeters outside of sleep centres, potentially over multiple nights

What This Means

This research suggests that a simpler approach to measuring sleep quality — using a finger sensor that measures blood flow (photoplethysmography, or PPG) combined with artificial intelligence — can predict how sleepy and fatigued patients feel during the day at least as well as, and sometimes better than, the gold-standard sleep lab test (polysomnography). The key innovation is that instead of labeling every 30-second chunk of sleep as a single stage (like 'light sleep' or 'deep sleep'), the new approach continuously estimates the probability of being in each sleep stage every second. This richer picture of sleep, called a 'hypnodensity,' appears to capture the disrupted, fragmented sleep patterns common in people with sleep-disordered breathing (such as sleep apnea) more accurately. The study found these new metrics were significantly better associated with patient-reported fatigue and sleepiness for the N1 (lightest) and N3 (deepest) sleep stages in particular. The study was conducted in a large group of 2,280 patients with sleep-disordered breathing, most of whom were middle-aged men with moderate severity disease. Even after accounting for important factors like age, sex, body weight, sleep apnea severity, and low oxygen levels during sleep, multiple PPG-derived hypnodensity measures remained significantly linked to daytime sleepiness. This matters because standard sleep lab metrics often fail to explain why some patients feel very sleepy while others with similar test results do not — a longstanding puzzle in sleep medicine. This research suggests that simple, wearable fingertip devices could eventually replace or supplement expensive sleep lab studies for monitoring sleep quality in people with breathing-related sleep disorders. Because PPG sensors are relatively inexpensive and easy to use at home, this approach could allow patients to be monitored over multiple nights in their natural environment, potentially giving doctors a more complete and realistic picture of how well a patient is sleeping and how their treatment is working.

Check Your Own Numbers

Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.

Upload Your Labs

Have a question about this study?

Citation

Rusanen M, Tuyinsenge V, Muller S, Tamisier R, Baillieul S, Myllymaa S, et al.. (2026). Hypnodensity sleep metrics from fingertip photoplethysmography are better associated with daytime sleepiness and fatigue than traditional metrics from polysomnography.. Pulmonology. https://doi.org/10.1080/25310429.2026.2683326