Walk into any electronics store and you’ll find dozens of devices that promise to track your sleep. Smartwatches, rings, headbands, under-mattress sensors — the consumer sleep tracking market has exploded. Meanwhile, polysomnography (PSG) remains the clinical gold standard, unchanged in its fundamental approach for over 50 years.

So where does consumer technology end and clinical medicine begin? The answer matters, because millions of Australians are making decisions about their sleep health based on data from devices that were never designed as diagnostic instruments.

What Polysomnography Actually Measures

PSG is comprehensive. It simultaneously records brain electrical activity (EEG), eye movements (EOG), muscle activity (EMG), heart rhythm (ECG), respiratory airflow, respiratory effort, blood oxygen saturation, body position, and limb movements. A trained technician monitors the study in real time.

This multi-channel approach allows clinicians to definitively diagnose sleep disorders — obstructive sleep apnea, central sleep apnea, periodic limb movement disorder, REM sleep behaviour disorder, narcolepsy, and others. It provides the AHI (apnoea-hypopnoea index), sleep architecture breakdown, and detailed event-level data.

PSG’s limitation isn’t accuracy — it’s access. In Australia, lab-based sleep studies require referral, often involve waiting months, cost between $500 and $1,500 (partially covered by Medicare for in-lab studies), and require the patient to sleep in an unfamiliar environment. That last point is clinically relevant. First-night effect — sleeping poorly in the lab because it’s not your bed — is well-documented and can skew results.

What Wearables Actually Measure

Most consumer wearables rely on accelerometry (movement detection) and photoplethysmography (optical heart rate monitoring). Some newer devices add peripheral blood oxygen estimation. A few research-grade wearables include single-channel EEG.

From these signals, algorithms estimate sleep stages, sleep duration, and sometimes respiratory disturbance metrics. The key word is “estimate.” These devices infer sleep states from indirect signals, whereas PSG directly measures the physiological activity that defines those states.

What the Evidence Shows

The research is mixed, but a pattern has emerged. A 2025 systematic review in Sleep Medicine examined 42 validation studies of consumer wearables against PSG. The findings:

• Total sleep time: Most devices were reasonably accurate, typically within 20 to 30 minutes of PSG-measured values. Good enough for general tracking, not precise enough for clinical decision-making.
• Sleep onset latency: Poorly estimated by most wearables. Devices tend to classify quiet wakefulness as light sleep, overestimating how quickly you fall asleep.
• Sleep staging: N1 and N2 discrimination remains poor without EEG. REM detection using heart rate variability is moderate. Deep sleep (N3) estimation varies widely between devices.
• Respiratory events: Wrist-based SpO2 sensors can detect significant desaturation events but miss many hypopnoeas. Sensitivity for moderate-to-severe OSA ranges from 60 to 85 percent depending on the device and threshold used.

The takeaway: wearables are useful for longitudinal tracking and general pattern recognition. They’re not diagnostic tools.

Where Each Fits Clinically

The clinical question isn’t “which is better?” — it’s “what’s the right tool for the situation?”

PSG remains essential for diagnosing complex sleep disorders, titrating treatment, and evaluating cases where the clinical picture is unclear. You can’t diagnose narcolepsy, REM sleep behaviour disorder, or periodic limb movement disorder with a smartwatch.

Wearables are becoming useful for screening and monitoring. A patient whose ring consistently shows fragmented sleep with oxygen dips has a reason to pursue formal evaluation. A CPAP patient can track sleep quality trends between clinic visits. A shift worker can document their sleep-wake patterns for their clinician.

Some sleep physicians are beginning to incorporate wearable data into clinical assessments — not as standalone evidence, but as context. Having two weeks of sleep tracking data at an initial consultation gives a far richer picture than the patient’s recollection alone.

Home sleep apnea testing (HSAT) sits between these two categories. These are medical-grade devices — prescribed, validated, and Medicare-rebatable — that test for obstructive sleep apnea in the patient’s own bed. They don’t measure EEG or sleep stages, but they capture respiratory parameters with clinical-grade accuracy. For uncomplicated suspected OSA in adults, HSAT has become the first-line investigation in many Australian sleep services.

The Convergence Ahead

The gap between consumer and clinical devices is narrowing. Companies are pursuing TGA clearance for wearable-derived sleep metrics. AI development work is helping device manufacturers build more accurate algorithms that can extract clinical-grade insights from consumer-grade sensors.

Within a few years, we’ll likely see wearables cleared for OSA screening — not diagnosis, but screening that triggers referral. This could be particularly valuable in Australia, where an estimated 80 percent of moderate-to-severe OSA remains undiagnosed, largely because people don’t know they should be tested.

For now, the practical advice is straightforward. Use your wearable to understand your sleep patterns and spot trends. Don’t self-diagnose based on its readings. And if it’s consistently showing you something concerning — fragmented sleep, low oxygen, unusually short sleep duration — bring that data to your GP. It’s a starting point for a conversation, not a conclusion.