Actigraphy vs Polysomnography: A Guide for Sleep Specialists

Actigraphy and polysomnography: the essential distinction

Actigraphy is an objective sleep-monitoring method based primarily on movement. A wrist actigraph records activity, and specialized algorithms use patterns of activity and rest to estimate sleep and wake. Measures can include estimated total sleep time, sleep efficiency, wake after sleep onset, awakenings, sleep timing, and rest–activity rhythms. Because the device can be worn outside the laboratory, actigraphy is useful for observing sleep in the participant’s natural environment across multiple days.

Polysomnography is a multi-signal sleep study. In the validation protocol described by Miller and colleagues, PSG included electroencephalography, electrooculography, and submental electromyography, with records manually scored in 30-second epochs according to standard criteria. PSG therefore provides a physiological reference for identifying sleep and wake and distinguishing stages such as light sleep, slow-wave sleep, and rapid eye movement sleep.

The central difference is that actigraphy estimates sleep indirectly from movement, whereas PSG evaluates sleep using physiological signals. Actigraphy is generally more ecological and scalable for repeated measurements; PSG provides richer physiological information during the recorded period. These strengths are complementary rather than interchangeable. ^{[1], [2], [3]}

Definitions for clinical and research teams

Actigrafia

a method that uses an activity monitor, commonly worn on the wrist, to record movement and estimate sleep–wake behavior through an algorithm. Its outputs depend on the device, placement, algorithm, and scoring settings used, which should be reported with the results.

polissonografia

a multi-channel sleep recording that includes physiological signals used to identify sleep and wake and to score sleep stages. PSG is commonly used as the reference standard — the gold standard of sleep studies — when validating actigraphy and other sleep-monitoring technologies.

Sleep-wake estimation

the algorithmic classification of time periods as likely sleep or wake from actigraphy or another indirect signal. It is an estimate, not a direct physiological observation.

Sleep staging

the classification of sleep into physiological stages, including non-rapid eye movement categories and rapid eye movement sleep. Conventional wrist actigraphy does not directly measure these stages because it is based primarily on movement.

^{[1], [2], [3]}

What actigraphy measures well

Actigraphy is most valuable when the research question concerns sleep timing, habitual sleep duration, sleep continuity, day–night patterns, or circadian rest–activity organization over an extended period. Martin and Hakim described applications including documenting sleep patterns before a multiple sleep latency test, evaluating circadian rhythm sleep disorders, assessing treatment outcomes, and supporting home monitoring of sleep-disordered breathing.

Longitudinal measurement is a major advantage. A single PSG night provides detailed physiological information but represents a limited observation window. Actigraphy can capture weekday–weekend differences, irregular schedules, naps, and changes in rest–activity organization that may be missed during a single laboratory night. In a comparison of insufficient sleep syndrome and narcolepsy type 1, actigraphy highlighted weekday sleep restriction with weekend compensation in insufficient sleep syndrome and fragmented but prolonged nighttime sleep with more frequent daytime naps in narcolepsy type 1.

Actigraphy can also support sleep-fragmentation analyses. In a 7-day actigraphy and PSG analysis of 1,908 participants, an actigraphy-derived sleep fragmentation index was strongly correlated with actigraphy-derived sleep efficiency and wake after sleep onset. Its correlations with PSG-derived wake after sleep onset, apnea–hypopnea index, and arousal index were more modest. This supports actigraphy fragmentation metrics as markers of sleep continuity, not direct substitutes for PSG architecture or arousal measures. ^{[1], [4], [5]}

What polysomnography measures that actigraphy cannot

PSG provides physiological information that movement-based actigraphy does not. In the cited validation study, PSG was manually scored in 30-second epochs using electroencephalography, electrooculography, and electromyography. These signals support identification of sleep stages and measurement of sleep architecture, including light sleep, slow-wave sleep, and REM sleep.

This distinction matters when the question depends on sleep-stage transitions, sleep-onset REM periods, arousals, or physiological events that may occur without substantial movement. Studies of hypersomnolence used PSG together with the multiple sleep latency test to evaluate diagnostic features, while actigraphy contributed information about sleep continuity and longer-term sleep–wake organization. A movement-derived estimate can complement such testing, but it does not replace PSG physiology.

PSG is also an appropriate reference method when a new actigraphy device, algorithm, or sleep tracker is being validated. The cited standardized analytical framework recommends discrepancy analysis, Bland–Altman plots, and epoch-by-epoch analysis when comparing sleep-tracking technologies with PSG or another reference method. This is a methodological framework, not a formal clinical guideline; agreement should be evaluated for the specific outcomes and population of interest. The comparisons presented in this article are drawn from validation and comparative studies, not from society-level practice parameters, so clinical indication criteria for individual patients should be checked against formal guidance such as the American Academy of Sleep Medicine clinical practice guideline. ^{[2], [3], [4], [6], [7], [8]}

Actigraphy versus PSG: comparison table

The following comparison summarizes practical differences between actigraphy and polysomnography. Actual performance depends on the device, algorithm, placement, scoring settings, population, and study protocol. ^{[1], [3]}

When to use actigraphy, PSG, or both

Use actigraphy when the primary objective is to characterize sleep–wake behavior over several days or weeks in an ecological setting. Typical questions include whether sleep differs between weekdays and weekends, whether naps occur, whether treatment changes estimated sleep duration or rest–activity rhythms, and whether habitual schedules are documented before a laboratory sleep test.

Actigraphy is practical when repeated PSG is not feasible because of cost, participant burden, laboratory capacity, or the need to measure sleep at home. Its lower burden supports longitudinal intervention studies, circadian research, clinical cohorts, and population-based protocols. In hypersomnolence research, it may help quantify habitual night sleep, daytime naps, and circadian patterns, but algorithm-dependent discrepancies have been reported against 24-hour or 32-hour PSG, particularly for daytime sleep and total sleep time.

Choose PSG when the study requires direct physiological sleep staging, detailed sleep architecture, sleep-onset REM periods, arousal-related measures, or physiological confirmation. PSG is also important when respiratory or movement-related indices are study endpoints, including PSG-derived apnea–hypopnea index or periodic limb movement measures. The method requires multiple sensors, technical expertise, and scoring resources, and is less suitable as the sole tool for long-term home monitoring.

A combined protocol is useful when researchers need both habitual behavior and physiological detail. Actigraphy before or after PSG can document habitual sleep timing, sleep opportunity, circadian patterns, and longitudinal changes, while PSG provides stage and physiological data during a defined recording period. Simultaneous recordings are particularly valuable for validation studies using epoch-by-epoch measures, sensitivity, specificity, accuracy, discrepancy analysis, and Bland–Altman agreement.

A scoping review of sleep outcomes in ADHD randomized controlled trials found that actigraphy and sleep diaries were commonly used, while PSG was used less often; it also noted heterogeneity in sleep variables and technical algorithms across studies. This reinforces the need to select the method according to the endpoint and to document the measurement protocol clearly. ^{[1], [3], [5], [6], [10], [11]}

Why actigraphy can disagree with PSG

The main limitation of actigraphy is the indirect nature of sleep–wake estimation. Quiet wakefulness can be classified as sleep, while movement during sleep can be classified as wake. This commonly produces high sensitivity for detecting sleep but lower specificity for detecting wake.

In hospitalized patients with traumatic injuries, actigraphy algorithms achieved sensitivity of at least 92% for sleep detection and accuracy above 85%, while specificity varied from 48% to 60%. In another validation study involving six healthy adults across 54 sleep records, research-grade actigraphy had 98% sensitivity for sleep, 60% specificity for wake, and 89% overall agreement with PSG. The same study found that actigraphy overestimated total sleep time by 37.6 minutes and underestimated wake by 37.6 minutes.

Performance may be less reliable in people with insomnia, prolonged quiet wakefulness, unusual schedules, daytime naps, hypersomnolence, neurological conditions, or impaired movement; insomnia populations have been the subject of dedicated actigraphy-versus-polysomnography comparisons for this reason, and in one wearable comparison specificity for wake was poor in both insomnia patients and good sleepers. In suspected idiopathic hypersomnia, conventional algorithms mainly overestimated sleep, while a recurrent neural-network approach performed better than the commercial algorithm tested. In a 24-hour comparison involving narcolepsy and idiopathic hypersomnia, actigraphy underestimated nighttime total sleep time in narcolepsy and overestimated daytime total sleep time in both groups.

Algorithm settings are part of the measurement method. Device model, epoch length, wake threshold, scoring algorithm, bed and wake-time handling, missing data, and participant adherence should be documented. Results from one actigraph or algorithm should not automatically be generalized to another. ^{[2], [3], [6], [9], [10], [12], [13]}

Limitations of polysomnography

PSG is resource-intensive and requires technical acquisition and scoring expertise. It can be expensive and time consuming, and its setup requirements make it impractical in some field settings.

PSG also has limited suitability for extended monitoring when the objective is to observe home routines, weekday–weekend variation, or long-term treatment response. A defined recording period provides high-resolution physiological data but may not capture the full range of day-to-day variability.

These limitations do not reduce the importance of PSG as a reference method. They clarify its role: PSG is strongest for physiological detail and reference measurement, while actigraphy is strongest for ecological and longitudinal sleep–wake observation. ^{[1], [2], [3]}

A practical decision framework for sleep studies

Start with the question rather than the device. If the question concerns habitual timing, rest–activity rhythms, or multi-day variability, actigraphy is often the better first-line monitoring method. If it requires sleep staging, stage-specific events, or physiological confirmation, PSG is appropriate.

Next, define the population. Validation results from healthy adults may not apply to patients with insomnia, hypersomnolence, trauma, neurological disease, or atypical movement patterns. Agreement can vary by population and time of day. Daytime sleep and prolonged bedrest may require specific algorithms and should not automatically be interpreted using nighttime settings.

Finally, define the analysis plan. Record device placement, sampling and epoch settings, algorithm version, scoring thresholds, synchronization procedures, non-wear rules, missing-data handling, and any diary procedures. For method-comparison studies, report both agreement and bias. High sleep sensitivity does not mean that wake is detected with equal accuracy, and correlation for total sleep time does not establish agreement for sleep onset latency, wake after sleep onset, or sleep stages.

For additional background, see Condor Instruments’ actigraphy guide, our overview of algorithm validation in actigraphy, and our discussion of sleep fragmentation metrics. Teams replacing discontinued or unavailable devices can also consult the Actiwatch replacement guide. If your protocol combines actigraphy, light measurement, or multi-week monitoring, our team can help align the method with your study endpoints. ^{[3], [5], [6], [10]}

Frequently Asked Question

Referências

1. Wrist actigraphy.. Jennifer L Martin, Alex D Hakim. Chest. 2011. PMID 21652563. Read source
2. A Validation Study of a Commercial Wearable Device to Automatically Detect and Estimate Sleep.. Dean J Miller, Gregory D Roach, Michele Lastella, Aaron T Scanlan, et al. Biosensors. 2021. PMID 34201016. Read source
3. A standardized framework for testing the performance of sleep-tracking technology: step-by-step guidelines and open-source code.. Luca Menghini, Nicola Cellini, Aimee Goldstone, Fiona C Baker, et al. Sleep. 2021. PMID 32882005. Read source
4. Clinical and instrumental comparison between insufficient sleep syndrome and narcolepsy type 1.. Mauro Manconi, Anna Castelnovo, Silvia Miano, Marco De Pieri. Sleep Medicine. 2026. PMID 41411879. Read source
5. Actigraphy-derived sleep fragmentation index: convergent validity and associations with clinical outcomes.. Dana Saleh, Suzanne M Bertisch, Michelle Reid, Andrew Lim, et al. Journal of Clinical Sleep Medicine. 2025. PMID 40471077. Read source
6. The actigraphic evaluation of daytime sleep in central disorders of hypersomnolence: comparison with polysomnography.. Francesco Biscarini, Stefano Vandi, Caterina Riccio, Linda Raggini, et al. Sleep. 2024. PMID 39154204. Read source
7. Diagnostic value of actigraphy in hypersomnolence disorders.. Eva Wiberg Torstensen, Line Pickering, Birgitte Rahbek Kornum, Benedikte Wanscher, et al. Sleep Medicine. 2021. PMID 34265481. Read source
8. Use of Actigraphy for the Evaluation of Sleep Disorders and Circadian Rhythm Sleep-Wake Disorders: An American Academy of Sleep Medicine Clinical Practice Guideline.. Michael T Smith, Christina S McCrae, Joseph Cheung, Jennifer L Martin, et al. Journal of Clinical Sleep Medicine. 2018. PMID 29991437. Read source
9. Validity of actigraphy for nighttime sleep monitoring in hospitalized patients with traumatic injuries.. Julien Lauzier Bigué, Catherine Duclos, Marie Dumont, Jean Paquet, et al. Journal of Clinical Sleep Medicine. 2020. PMID 31992412. Read source
10. Actigraphy against 32-hour polysomnography in patients with suspected idiopathic hypersomnia.. Tugdual Adam, Jérôme Tanty, Lucie Barateau, Yves Dauvilliers. Journal of Sleep Research. 2025. PMID 39979124. Read source
11. Sleep as an outcome measure in ADHD randomized controlled trials: A scoping review.. Scout McWilliams, Ted Zhou, Sylvia Stockler, Dean Elbe, et al. Sleep medicine reviews. 2022. PMID 35313258. Read source
12. Validity of a commercial wearable sleep tracker in adult insomnia disorder patients and good sleepers.. Seung-Gul Kang, Jae Myeong Kang, Kwang-Pil Ko, Seon-Cheol Park, et al. Journal of psychosomatic research. 2017. PMID 28606497. Read source
13. A comparison of actigraphy and polysomnography in older adults treated for chronic primary insomnia.. B Sivertsen, S Omvik, OE Havik, S Pallesen, et al. Sleep. 2006. PMID 17068990. Read source