How to Read a Sleep Doctor's Report: PSG Findings Translated

You Got Your Sleep Report. Now What?

Most sleep reports look like a wall of acronyms. AHI. RDI. ODI. PLMI. TST. SE. Then a paragraph at the bottom that may or may not feel reassuring.

If you're reading one right now, you're probably a little anxious. That's normal. A polysomnography report (PSG) is written for clinicians, not for the person whose body it describes. The good news: once you know what each line means, the report stops looking like a stress test.

This post walks through a typical adult sleep report and where the thresholds come from. We'll spend extra time on position-dependent results, because for a lot of people, the supine-versus-side breakdown is the most actionable thing in the whole study.

This is educational, not diagnostic. Your specific report should be reviewed with the sleep specialist who ordered it.

Key Facts

• AHI severity (adults): under 5 normal, 5 to 14 mild, 15 to 29 moderate, 30 or more severe (AASM 2017 guideline)
• Sleep efficiency: 85% or higher is typically considered normal (NCBI primer)
• REM sleep: roughly 20 to 25% of total sleep time in healthy adults (StatPearls)
• Sleep latency: 10 to 20 minutes is the typical normal range (Sleep Foundation)
• PLMI clinical threshold: greater than 15 events per hour in adults (ICSD-3 / narrative review)
• Positional OSA (Cartwright): supine AHI at least double the lateral AHI (Sleep & Breathing)

The Recording Summary

Every PSG report opens with a few timing numbers. They sound boring. They aren't.

Total Recording Time (TRT) is how long the equipment was on you, lights out to lights on. Total Sleep Time (TST) is how much of that time you actually spent asleep. The gap between the two matters.

Sleep Efficiency (SE) is the ratio. TST divided by TRT, expressed as a percent. A normal value sits around 85% or higher for healthy adults, and most labs flag anything below 80% as low ¹. If your SE is 70%, you spent a lot of the night awake in bed, which has its own implications (insomnia, frequent arousals, environmental disruption, anxiety).

Wake After Sleep Onset (WASO) is the total time awake after first falling asleep. High WASO with low SE points toward fragmented sleep, not trouble falling asleep.

Sleep Architecture: N1, N2, N3, REM

Healthy sleep moves through stages in roughly 90 minute cycles. A PSG report breaks down what percentage of your sleep was spent in each stage.

Approximate normal ranges in healthy adults ^{2 3}:

• N1 (lightest sleep): about 2 to 5%
• N2 (light sleep, most of the night): about 45 to 55%
• N3 (deep / slow-wave sleep): about 10 to 25%
• REM (dream sleep): about 20 to 25%

N3 percentage drops with age, so an older adult with 8% N3 isn't automatically abnormal. REM clusters in the second half of the night, so a study cut short can show artificially low REM. If the report shows almost no REM, ask about it, since apnea often gets worse during REM.

Sleep Latency and REM Latency

Sleep Onset Latency (SOL) is how long it took you to fall asleep after lights out. The typical normal range is 10 to 20 minutes ⁴. Falling asleep in under 5 minutes consistently can suggest sleep deprivation. Taking 30+ minutes can suggest insomnia or anxiety, especially if it shows up across multiple nights at home.

REM Latency is how long after falling asleep before you reached your first REM period. Normal is roughly 70 to 110 minutes ². A very short REM latency (under about 15 minutes) is one of the criteria sleep doctors look at when narcolepsy is on the differential, but it can also reflect REM rebound from sleep deprivation.

One night in a lab is one data point. Latencies vary a lot night to night.

Respiratory Events: The Apnea Section

This is the section that gets the most attention, and reasonably so.

Apneas are pauses in airflow lasting at least 10 seconds. Hypopneas are partial reductions in airflow paired with either a drop in oxygen or an arousal. The AASM scoring manual defines the exact criteria, and there's a long-running debate about the "3% versus 4%" desaturation rule that affects how many hypopneas get scored ⁵.

The headline number is the Apnea-Hypopnea Index (AHI), which is the average number of apneas plus hypopneas per hour of sleep. Severity in adults ⁶:

• Under 5/hr: within normal limits
• 5 to 14/hr: mild OSA
• 15 to 29/hr: moderate OSA
• 30/hr or more: severe OSA

You may also see RDI (Respiratory Disturbance Index), which adds RERAs (respiratory effort-related arousals) to the count, and tends to run higher than AHI on the same study.

Oxygen Desaturation Index (ODI) is how often per hour your blood oxygen dropped by a defined amount (usually 3% or 4%) from the prior baseline. ODI tracks AHI fairly closely in most people, and an ODI greater than 10 has reasonable performance for catching moderate-to-severe OSA ⁷.

Lowest SpO2 (O2 nadir) is the single lowest oxygen reading recorded during the study. Healthy adults rarely dip below 90% during sleep. Nadirs in the low 80s or below are usually flagged in the report. % time below 90% (T90) is another oxygenation marker that captures the "burden" of low-oxygen exposure across the whole night.

Position-Dependent Results: The Most Important Section You're Probably Skipping

This is the part of the report that matters most for a lot of people, and it's often the part that gets a one-line mention.

A standard PSG breaks AHI down by body position. You'll typically see four numbers:

• Supine AHI (on your back)
• Non-supine AHI (sometimes split into left, right, and prone)
• Supine time (how much of the night you spent on your back)
• Non-supine time

Why does this matter? Because in the supine position, gravity pulls the soft palate, tongue, and pharyngeal tissue toward the back of the throat, narrowing the airway. The lateral (side) position relieves a lot of that load ⁸.

For a meaningful percentage of people with OSA, the entire problem is supine. Cartwright's classic definition of positional OSA is a supine AHI at least twice the lateral AHI ⁹. Mador's stricter definition adds that the non-supine AHI must drop below 5 ¹⁰. Across populations, prevalence of positional OSA runs around 27% by Mador's criteria, with some estimates as high as 50 to 75% by broader definitions ^{10 8}.

What this means in practice: someone with an overall AHI of 20 (moderate) might have a supine AHI of 45 and a side-sleeping AHI of 4. That person doesn't really have moderate sleep apnea on average. They have severe sleep apnea on their back and basically none on their side. The treatment implications are completely different.

The catch: PSG only sees the positions you happen to sleep in that night. If you spent 20 minutes supine in the lab and 7 hours on your right side, the supine number is based on a tiny sample. Some labs note this as "limited supine data." It's one of the reasons multi-night data at home (which is what Komori is being built to capture) gives you a richer baseline than a single in-lab night.

For more on what positional data unlocks, see our piece on positional sleep apnea and what position data can tell you.

Periodic Limb Movements (PLMI)

If your legs twitched in regular intervals during the night, you'll see a Periodic Limb Movement Index (PLMI), which is the count per hour of sleep.

Severity bands commonly cited ¹¹:

• PLMI under 5: minimal
• 5 to 14.9: mild
• 15 to 29.9: moderate
• 30 or more: severe

The ICSD-3 clinical threshold for periodic limb movement disorder in adults is greater than 15 per hour ¹¹. Most labs also report PLMAI (PLM-arousal index), which counts only the limb movements that woke you up. That number tends to track better with daytime tiredness than PLMI alone, though research on the link is mixed.

Limb movements often co-occur with OSA, and treating the apnea sometimes reduces the limb movement count.

Arousal Index

An arousal is a brief shift toward wakefulness on the EEG that lasts at least 3 seconds ¹¹. You usually don't remember them. The Arousal Index is the count per hour.

A common rough cutoff is that an arousal index over 10/hr is on the elevated side, with values north of 20/hr clearly fragmenting sleep. Arousals get sub-classified by what triggered them: respiratory, limb movement, spontaneous, or snore. The trigger breakdown matters more than the total. A high respiratory arousal index points one direction; a high spontaneous arousal index points another.

Cardiac

Most reports include average heart rate during sleep and lowest and highest heart rate. They'll flag any arrhythmias caught on ECG, including bradycardia, tachycardia, pauses, or atrial fibrillation episodes.

This section is short on most reports because the cardiac findings are either "unremarkable" or specific enough that your doctor will discuss them with you directly. Untreated severe OSA is associated with cardiovascular consequences, which is part of why severity matters even when symptoms feel mild.

Snoring

Snoring shows up as a Snore Index (snores per hour) or as a percentage of sleep time spent snoring. The snore index correlates with OSA severity in groups, but on an individual level the relationship is messy. You can have heroic snoring and a normal AHI, or near-silent breathing and severe apnea ¹².

Snoring data is most useful as context, not as a diagnostic by itself. If your bed partner is the one who pushed you to get the study, the snore index is a sanity check that the lab caught what they've been hearing every night.

For why snoring is sometimes a position problem in disguise, see snoring and sleep position.

The Final Impression

The last section is the clinical interpretation, written by the sleep physician. It's where the numbers get pulled into a story. You'll see things like:

• "Severe OSA, predominantly supine, with significant oxygen desaturation."
• "Mild OSA. PAP titration recommended."
• "Sleep efficiency reduced. Frequent arousals not clearly attributed to respiratory events."
• "Within normal limits."

If the impression doesn't match how you feel, that's a conversation worth having. Sleep medicine is full of people whose numbers look fine and whose lives feel terrible, and vice versa. Talk to your sleep specialist about your specific report. That's the actual next step.

The Big Lookup Table

Use the table as a quick reference, not as self-diagnosis. The specialist sees the whole picture.

Home Sleep Tests vs In-Lab PSG

If your study was a Home Sleep Apnea Test (HSAT) instead of an in-lab PSG, your report will be shorter, and that's by design.

The AASM's clinical guidance is clear: HSAT is appropriate for uncomplicated adults with high pretest probability of moderate to severe OSA. It's not meant for everyone ¹³.

What HSAT typically captures:

• Airflow
• Respiratory effort
• Oxygen saturation
• Heart rate or pulse
• Sometimes body position

What HSAT typically misses:

• EEG, which means no real sleep staging (no N1/N2/N3/REM breakdown)
• Arousals, especially the cortical ones the AASM scoring rule includes
• Limb movements (no EMG)
• True total sleep time (the denominator becomes recording time, which often makes the AHI look lower than it really is)

The AASM points out that sensor dislodgement and signal quality issues during HSAT can cause the "true" AHI to be underestimated, which is why a negative HSAT in a high-suspicion patient often gets followed up with in-lab PSG ¹³.

If your HSAT came back clean but you still snore loudly, gasp at night, or feel exhausted, that's a reason to ask about an in-lab study, not a reason to drop it.

Where Komori Fits

Komori doesn't replace a sleep study. It can't measure airflow, EEG, or EMG. It's a wellness device, not a diagnostic.

What it's designed to do well: track sleep position, movement, and bed-exit events night after night, in your own bed, with no contact and no camera. That gives you something a single PSG night can't, which is a baseline. How much time do you actually spend supine on a typical Tuesday? Has your night-to-night supine fraction shifted since you started side-sleeping with a wedge pillow? Are your wake-ups clustering at certain times?

For people whose PSG flagged positional OSA, that long-running positional data is the most useful complement to the sleep doctor's report we know how to build. For people whose PSG was clean, multi-night context can still help you understand patterns the single lab night didn't catch. We've written about why we don't compress your sleep into a single score and what movement actually says about sleep if you want to go deeper.

The PSG report tells you what happened on one night. Your body has thousands of nights. Both views matter, and they're better together.

If your report raised questions, the right next step is the sleep specialist who ordered it.

---

Footnotes

1. Patil SP. "How to interpret the results of a sleep study." Journal of Community Hospital Internal Medicine Perspectives, 2014. ↩ ↩² ↩³

2. Patel AK, et al. "Physiology, Sleep Stages." StatPearls Publishing, 2024. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

3. Boulos MI, et al. "Reference Data for Polysomnography-Measured and Subjective Sleep in Healthy Adults." Journal of Clinical Sleep Medicine, 2019. ↩

4. Sleep Foundation. "Sleep Latency." Last updated 2024. ↩ ↩²

5. AASM. "Summary of Changes: AASM Manual for the Scoring of Sleep and Associated Events, Version 3." American Academy of Sleep Medicine, 2024. ↩

6. Kapur VK, et al. "Clinical Practice Guideline for Diagnostic Testing for Adult Obstructive Sleep Apnea." Journal of Clinical Sleep Medicine, 2017. ↩ ↩²

7. Chung F, et al. "Oxygen desaturation index from nocturnal oximetry: a sensitive and specific tool to detect sleep-disordered breathing." Journal of Clinical Sleep Medicine / related literature, 2022. ↩ ↩² ↩³

8. Heinzer R, et al. "A review of supine position related obstructive sleep apnea: classification, epidemiology, pathogenesis and treatment." Sleep Medicine Reviews, 2023. ↩ ↩²

9. Frank MH, et al. "Positional OSA part 1: towards a clinical classification system for position-dependent obstructive sleep apnoea." Sleep and Breathing, 2015. ↩ ↩²

10. Mador MJ, et al. Cited in "The undervalued potential of positional therapy in position-dependent snoring and obstructive sleep apnea." Sleep and Breathing, 2013. ↩ ↩²

11. Stefani A, Högl B. "Periodic limb movements during sleep: a narrative review." Journal of Thoracic Disease, 2021. ICSD-3 thresholds referenced. ↩ ↩² ↩³ ↩⁴ ↩⁵

12. Kim J, et al. "Definition of the snoring episode index based on the analyses of snoring parameters and the apnea hypopnea index." Scientific Reports, 2022. ↩ ↩²

13. Rosen IM, et al. "Clinical Use of a Home Sleep Apnea Test: An Updated AASM Position Statement." Journal of Clinical Sleep Medicine, 2018. ↩ ↩²