Polysomnography (Sleep Study): What This Test Actually Measures

Why Polysomnography Exists

A sleep study is the only test that can capture what happens across every physiological system while you are unconscious. Questionnaires and wrist-worn trackers estimate sleep duration, but they cannot distinguish between an obstructive apnea, a central apnea, and a hypopnea. PSG can. That distinction drives treatment selection.

The Clinical Gap It Fills

The American Academy of Sleep Medicine (AASM) considers in-lab PSG the reference standard for diagnosing obstructive sleep apnea, central sleep apnea syndromes, narcolepsy with cataplexy, and REM sleep behavior disorder [1]. Home sleep apnea tests (HSATs) record fewer channels. They omit EEG entirely, which means they cannot stage sleep or detect parasomnias. For patients with significant cardiopulmonary disease, neuromuscular conditions, or suspected non-apnea disorders, the AASM clinical practice guideline published in the Journal of Clinical Sleep Medicine recommends attended, in-lab PSG over HSAT [2].

Who Gets Referred for PSG

Referral typically follows a screening pathway. The STOP-BANG questionnaire assigns one point each for snoring, tiredness, observed apneas, blood pressure above 140/90, BMI above 35, age over 50, neck circumference over 16 inches, and male sex. A score of 5 to 8 carries a high probability of moderate-to-severe OSA [3]. Patients scoring in that range, along with those who have unexplained hypersomnia or atypical nocturnal behaviors, move to PSG.

The Channels Recorded During a Sleep Study

Each sensor attached during setup feeds a distinct data stream. The technologist monitors all channels in real time from an adjacent control room. A single night of PSG generates roughly 1,000 pages of raw waveform data.

Electroencephalography (EEG)

Surface electrodes placed on the scalp (typically F4-M1, C4-M1, O2-M1 with contralateral backups) record cortical electrical activity. EEG is the channel that makes sleep staging possible. The AASM scoring manual defines five stages by waveform pattern: Wake (alpha rhythm, 8 to 13 Hz), N1 (theta waves, low-amplitude mixed frequency), N2 (sleep spindles at 11 to 16 Hz, K-complexes), N3 (slow-wave sleep, delta activity ≥ 0.5 to 2 Hz occupying more than 20% of an epoch), and REM (low-amplitude, mixed-frequency activity resembling wakefulness) [4].

Electro-oculography (EOG) and Chin EMG

Two electrodes placed at the outer canthi of each eye detect conjugate eye movements. Rapid eye movements mark REM sleep. Slow, rolling eye movements appear during the transition from wake to N1. Chin electromyography records submental muscle tone. During REM, healthy skeletal muscle tone drops to its lowest point (atonia). Failure to suppress tone during REM is the electrophysiologic hallmark of REM sleep behavior disorder [5].

Respiratory Channels

This cluster of sensors answers the central diagnostic question: is the patient breathing normally? A nasal pressure transducer detects airflow reductions (hypopneas) with greater sensitivity than a thermistor alone, while an oronasal thermistor identifies complete cessation of airflow (apneas). Thoracic and abdominal effort belts (respiratory inductance plethysmography) measure chest wall and diaphragmatic movement. An apnea with continued effort on both belts is obstructive. An apnea with absent effort is central. This distinction changes the treatment plan entirely: CPAP for obstructive events, adaptive servo-ventilation or phrenic nerve stimulation for central events [6].

Pulse Oximetry and ECG

A finger probe continuously samples arterial oxygen saturation (SpO2). The AASM defines a desaturation event as a drop of 3% or more from baseline [4]. The oxygen desaturation index (ODI) counts these drops per hour; it often tracks closely with AHI but can diverge in patients with significant lung disease. A single-lead or modified-lead ECG runs throughout the study, capturing heart rate, rhythm disturbances, and any correlation between arrhythmias and respiratory events. Bradycardia-tachycardia swings during obstructive apneas are common in severe OSA and can be a first signal of cardiovascular risk [7].

Leg EMG

Surface electrodes on the anterior tibialis of each leg record periodic limb movements (PLMs). The AASM defines a PLM as a burst of EMG activity lasting 0.5 to 10 seconds, occurring in a sequence of four or more movements separated by 5 to 90 seconds [4]. A periodic limb movement index (PLMI) of 15 or more per hour in adults, combined with clinical sleep disruption, supports a diagnosis of periodic limb movement disorder.

How PSG Results Are Scored and Reported

After the study ends, a registered polysomnographic technologist scores every 30-second epoch for sleep stage, then marks every respiratory event and limb movement. A board-certified sleep physician reviews and interprets the scored data. The final report includes both numeric indices and narrative interpretation.

The Apnea-Hypopnea Index

AHI is the primary severity metric. It equals the total number of apneas plus hypopneas divided by total sleep time in hours. The AASM severity classification, adopted by the Centers for Medicare and Medicaid Services for CPAP coverage determinations, defines normal as an AHI <5, mild OSA as 5 to 14, moderate as 15 to 29, and severe as ≥30 [1]. A large population study using AASM 2012 scoring criteria (N=1,520) found that 14% of men and 5% of women aged 30 to 70 met criteria for moderate-to-severe OSA (AHI ≥15) [8].

Sleep Architecture Summary

The report breaks total sleep time into the percentage spent in each stage. Healthy adult norms cluster around 2 to 5% N1, 45 to 55% N2, 15 to 25% N3, and 20 to 25% REM [9]. Deviations matter clinically. Patients with severe OSA often show reduced N3 and REM percentages because apneas cause arousals that push the brain back into lighter stages. Treating the apnea with CPAP frequently produces a "REM rebound" on the first treatment night, a sign that previously suppressed deep sleep is being restored.

Sleep Efficiency and Latency

Sleep efficiency equals total sleep time divided by total recording time, expressed as a percentage. Values above 85% are considered normal. Sleep onset latency (time from lights-off to the first epoch of sleep) above 30 minutes may indicate insomnia or a first-night effect. REM latency (time from sleep onset to the first REM period) shorter than 15 minutes raises concern for narcolepsy and should trigger a follow-up multiple sleep latency test (MSLT) [10].

What the Numbers Mean for Hormonal Health

Sleep is not a passive state for the endocrine system. Growth hormone (GH) secretion is tightly coupled to slow-wave sleep (N3). In a study of healthy young men (N=149), Van Cauter et al. Demonstrated that more than 70% of daily GH secretion occurred during the first N3 period of the night [11]. Apneas that fragment N3 can blunt this pulse. Treating OSA with CPAP has been shown to restore the nocturnal GH surge within weeks.

Testosterone and OSA

Testosterone follows a similar circadian pattern. Serum testosterone peaks in the early morning hours, driven by sleep-dependent pulsatile secretion of gonadotropin-releasing hormone (GnRH). A cross-sectional analysis of the Outcomes of Sleep Disorders in Older Men Study (MrOS, N=1,312) found that lower overnight oxygen saturation and more severe OSA were independently associated with lower morning total testosterone levels, even after adjusting for age and BMI [12]. The association was strongest in men with an AHI ≥30.

Cortisol Disruption

Fragmented sleep shifts the cortisol rhythm. Healthy cortisol patterns show a nadir around midnight and a peak around 6 to 8 AM (the cortisol awakening response). Repeated arousals from obstructive events raise nocturnal cortisol levels. Elevated nighttime cortisol has been linked to insulin resistance and visceral adiposity in OSA patients, independent of BMI [13]. PSG identifies the mechanism: the arousal index, measured in arousals per hour, quantifies how often the brain is being yanked out of consolidated sleep. An arousal index above 25 is considered elevated.

Preparing for a Polysomnography Study

Preparation is simple but specific. The goal is to reproduce a typical night of sleep so the data reflects the patient's usual pattern.

Day-of Instructions

Patients should avoid caffeine after noon, skip alcohol entirely on the study day, and avoid napping. Most sleep labs ask patients to arrive between 8 and 9 PM. A technologist applies approximately 22 to 25 sensors using conductive paste and adhesive. The setup takes 30 to 45 minutes. Patients can bring their own pillow.

Medications and Supplements

Certain medications alter sleep architecture and may need to be held or documented. SSRIs suppress REM sleep, which can mask REM-related OSA or obscure a narcolepsy diagnosis. Benzodiazepines reduce N3 percentage. Melatonin at standard doses (0.5 to 5 mg) is generally permitted, but patients should confirm with the ordering physician. Testosterone replacement therapy (TRT) itself does not need to be stopped for PSG, but clinicians should note it on the order because exogenous testosterone may worsen OSA severity in some men [14].

Split-Night Studies

If the AHI exceeds 40 events per hour during the first two hours of a diagnostic PSG and at least two hours of diagnostic recording have been obtained, the AASM allows conversion to a split-night study. The technologist applies a CPAP mask mid-study and titrates pressure during the remaining hours [2]. This approach eliminates the need for a second overnight visit.

Home Sleep Apnea Testing vs. In-Lab PSG

Home sleep apnea tests (Type III or Type IV monitors) are a reasonable alternative for uncomplicated patients with high pretest probability of moderate-to-severe OSA. They record four to seven channels, typically nasal pressure, SpO2, respiratory effort, and sometimes peripheral arterial tonometry.

When HSAT Falls Short

HSAT underestimates AHI because it divides events by recording time, not sleep time (no EEG means no sleep staging). A negative HSAT in a patient with strong clinical suspicion should be followed by in-lab PSG [2]. HSAT also cannot diagnose central sleep apnea, periodic limb movement disorder, narcolepsy, or parasomnias. For patients on opioid therapy, which predisposes to central apneas, or patients with heart failure, in-lab PSG remains the correct first test.

Cost and Access Considerations

In-lab PSG costs between $1,000 and $3,000 before insurance. Medicare covers diagnostic PSG (CPT 95810) when ordered by a physician for suspected sleep-disordered breathing. Most commercial insurers require prior authorization and may mandate HSAT first. The Centers for Medicare and Medicaid Services requires a qualifying PSG or HSAT with AHI ≥5 (with symptoms) or ≥15 (without symptoms) before covering CPAP equipment [15].

After the Study: Interpreting Your Report

The sleep physician issues a final report within one to two weeks. Patients receive a summary that lists total sleep time, sleep efficiency, AHI, ODI, lowest SpO2, arousal index, PLMI, and sleep stage percentages.

Red Flags in PSG Data

Certain findings demand prompt follow-up. A lowest SpO2 below 80% during the study indicates severe desaturation and increases the urgency of treatment. An AHI above 30 combined with excessive daytime sleepiness (Epworth Sleepiness Scale score ≥10) is associated with a 2.5-fold increased risk of motor vehicle accidents according to a prospective cohort study (N=3,201) [16]. A short REM latency (<8 minutes) with two or more sleep-onset REM periods on a subsequent MSLT meets ICSD-3 criteria for narcolepsy type 1, particularly when paired with low cerebrospinal fluid hypocretin-1 levels [10].

Connecting PSG to Treatment

For OSA, treatment selection follows AHI severity. Mild OSA (AHI 5 to 14) may respond to positional therapy, weight loss, or an oral appliance. Moderate-to-severe OSA (AHI ≥15) is typically treated with CPAP, with pressure set based on a titration study or auto-titrating device. The Endocrine Society recommends screening for OSA in men starting testosterone replacement therapy, since TRT may increase AHI by approximately 7 events per hour in susceptible individuals based on data from a randomized controlled trial (N=67) [14]. A PSG before and 3 to 6 months after TRT initiation gives the clearest picture.

Patients with an AHI of 5 to 14 who are starting TRT should have a follow-up PSG or HSAT within 3 to 6 months of therapy initiation, per the Endocrine Society 2018 clinical practice guideline [17].