Melanopic EDI: how to measure circadian light

Melanopic equivalent daylight illuminance (melanopic EDI, or mEDI) is the photopic illuminance of standard D65 daylight producing the same melanopic irradiance as the assessed light. Defined by CIE S 026:2018, it is expressed in lux (lx). Circadian light is light considered for its influence on circadian timing through retinal pathways, especially melanopsin-containing ipRGCs, rather than visual brightness alone. Melanopic EDI characterizes exposure, not a guaranteed biological response. [1, 3]

Measure melanopic EDI using calibrated spectral irradiance or a system validated for melanopic estimation. ActLumus by Condor Instruments estimates Melanopic EDI and Photopic lux with a 10-channel light sensor; its performance was independently evaluated under laboratory and free-living conditions. [16] Interpreting body-worn readings as light at the eye requires placement-specific validation. [15, 27]

By the Condor Research Team

On this page

• Independent sensor validation
• Melanopic EDI versus photopic lux
• Calculation and worked example
• Lighting-design recommendations
• Instrument selection
• Sensor placement
• Circadian entrainment and timing
• Field protocol and quality control
• ActLumus for research
• Frequently asked questions
• References

Independent validation of ActLumus light measurement

Ishihara et al. (2026), published in Sleep, independently compared LYS, ActLumus, GENEActiv and ActiGraph wGT3X-BT with a criterion spectrometer—the study's reference instrument. Here, photopic lx means visually weighted illuminance, distinct from melanopic EDI. [16]

Source: Ishihara et al., 2026, PubMed 41230818. Percent differences are reproduced as reported and apply to the tested devices and conditions; they are not guaranteed error limits or a standalone melanopic EDI accuracy specification.

ActLumus also showed consistently lower interdevice variability. Melanopic illuminance was less accurate on LYS than on ActLumus, particularly under warm indoor lighting. [16] Read the laboratory and free-living validation details.

Melanopic EDI versus photopic lux

Photopic lux and melanopic EDI describe different spectral weightings of the same light. Photopic illuminance uses the visual sensitivity function V(λ); melanopic EDI uses the CIE melanopic action spectrum and a D65 daylight reference. Neither replaces the other when both visual and non-visual exposure are relevant. [1, 2]

Sources: CIE S 026:2018 and its α-opic Toolbox. [1, 5]

Photopic illuminance remains useful for visual tasks, architectural assessment and visual-lighting requirements. Irradiance describes radiant power without photopic weighting, while spectral irradiance preserves its wavelength distribution. Other α-opic EDI metrics address S-cones, M-cones, L-cones and rods individually.

A conventional lux reading does not reveal the underlying spectrum. Daylight, LEDs, displays and decorative sources can produce similar photopic illuminance but different melanopic EDI. The difference is not a defect in the lux unit: it is a mismatch between the measurement and the question.

Photopic V(λ) peaks near 555 nm, whereas the CIE melanopic action spectrum peaks near 490 nm. A photopic-only record cannot uniquely determine melanopic exposure, and there is no universal direction or magnitude of error when photopic lux is used as its substitute. [1, 5]

The five photoreceptor responses and ipRGCs

CIE S 026:2018 defines five α-opic spectral sensitivity functions corresponding to S-cones, M-cones, L-cones, rods and melanopsin-containing intrinsically photosensitive retinal ganglion cells, or ipRGCs. Irradiance weighted by one of these functions is called α-opic irradiance. [1]

Sources: CIE S 026:2018, the CIE Toolbox and Do and Yau (2010). [1, 3, 5]

The pigment maximum and the CIE action-spectrum peak are different quantities. The CIE functions describe sensitivity at the cornea and incorporate specified pre-receptoral filtering; they should not be replaced by the absorption peak of the isolated pigment.

Light supports vision and non-visual responses

ipRGCs are retinal neurons that respond directly to light through melanopsin and also receive input from rods and cones. They contribute to circadian photoentrainment, melatonin suppression and the pupillary light reflex. These interacting pathways make it misleading to describe non-visual responses as completely isolated from the rest of the retina. [3, 6]

Historical observations in 1927 of pupil responses in mice with outer-retinal degeneration helped establish that light detection extends beyond conventional rod and cone vision. Some people with severe rod and cone degeneration can retain light-dependent circadian responses when the relevant inner-retinal pathways remain functional. Retinal disease and ocular transmission nevertheless require individual interpretation. [3, 17]

How to calculate melanopic EDI

Calculate melanopic irradiance first, then express it relative to D65 daylight. A calibrated spectroradiometer can provide the spectral input directly; a validated multichannel instrument can estimate the required weighting from its sensor responses. A photopic reading alone does not contain enough information. [1, 5]

Step 1: apply melanopic spectral weighting

Multiply the spectral irradiance by the CIE melanopic sensitivity function and integrate over wavelength:

E_mel = ∫ E_e,λ(λ) × s_mel(λ) dλ

Here, spectral irradiance E_e,λ is expressed in W/m²/nm, wavelength λ in nm, and the resulting melanopic irradiance E_mel in W/m². The weighting function s_mel is dimensionless. [1, 5]

Step 2: convert to equivalent daylight illuminance

Under CIE S 026:2018:

Melanopic EDI [lx] ≈ 754 [lx per W/m²] × melanopic irradiance [W/m²]

Equivalently, 1 lx of melanopic EDI corresponds to approximately 1.33 mW/m² of melanopic irradiance. The D65 reference constants—melanopic daylight action factor approximately 0.906, dimensionless, and maximum photopic luminous efficacy 683 lm/W—produce this conversion. [1, 5]

D65 represents standard daylight with a correlated color temperature of approximately 6,500 K. For this reference spectrum, melanopic EDI equals photopic illuminance by definition; arbitrary outdoor daylight need not have exactly the D65 spectrum. [1, 5]

MDER: when conversion from photopic lux is possible

The melanopic daylight efficacy ratio (MDER) is melanopic EDI divided by photopic illuminance. It is dimensionless:

Melanopic EDI [lx] = photopic illuminance [lx] × MDER

This relation is useful when the spectrum, or an appropriately validated MDER, is known for the measurement position. D65 has an MDER of 1.0 by definition. Neither correlated color temperature nor an unspecified lamp category supplies a universal MDER. [1, 5]

Legacy melanopic lux is not automatically melanopic EDI

The older equi-energy convention described by Lucas et al. (2014) used a different reference and a conversion of approximately 832.1 melanopic lux per W/m². For values calculated under that convention, multiply legacy melanopic lux by the dimensionless factor 0.9058 to obtain melanopic EDI in lx. [1, 2]

When reviewing older literature, including reports from 2013–2019, check the actual convention rather than inferring it from the publication year or the label melanopic lux.

Worked example: the same photopic lux, different melanopic EDI

The following is an illustrative calculation, not a measurement of a particular office or luminaire. The weighting method and conversion follow CIE S 026:2018. [1, 5]

The melanopic EDI is about one-third below the photopic value. It exceeds the 250 lx daytime melanopic EDI recommendation for healthy adults only if this represents the relevant exposure at the eye; a single reading does not demonstrate that the recommendation is met throughout the day. [9]

The LED ratios are illustrative inputs retained from this guide, not CIE-recommended ratios or fixed properties of those color temperatures. An assumed MDER range of 0.4–0.5 gives 200–250 lx melanopic EDI at 500 photopic lx; the intermediate-white illustration uses a ratio within 0.6–0.7. All these ratios are dimensionless. D65's ratio follows the standard definition. [1, 5]

CIE functions and conversion: [1, 5]. The example spectra and color-temperature labels are not measured luminaire data or a universal CCT-to-MDER conversion.

Melanopic EDI in lighting design

Circadian lighting design evaluates the daily pattern of light at the occupant, not just the brightness or color temperature of the installation. Measuring melanopic EDI alongside photopic illuminance makes daytime, evening and sleeping-environment objectives testable. [9, 18]

Brown et al. (2022) proposed the following indoor-light recommendations for healthy adults: [9]

For seated indoor assessment, the recommendations describe a vertical measurement plane at approximately 1.2 m height, representing eye level. Actual assessment should reflect the occupant's position and viewing direction. [9]

These are expert-consensus exposure recommendations, not CIE-mandated lighting limits, universal biological cutoffs or treatment prescriptions. They do not replace visual-task, glare or safety considerations, and they should not be transferred automatically to infants, patients or shift-working populations.

The broader guidance favors brighter days and darker evenings and nights, while recognizing differences in age, prior light exposure, schedules, health and geography. [18]

Why lighting design needs spectral and occupant-level verification

Color temperature alone is insufficient

Correlated color temperature describes the apparent warmth or coolness of white light, not its complete spectral power distribution. Reflections, transmission, shading, distance and viewing direction further change the exposure reaching an occupant.

A residential-lamp study measured spectral irradiance and calculated melanopic and photopic illuminance. Estimated melatonin-suppression values changed as tunable lamps moved between cool and warm settings. These were source- and setting-dependent estimates, not universal safe or unsafe classifications for home lighting. [20]

For specification and commissioning, ask: What photopic and melanopic exposure reaches the eye, at which position, viewing direction and time? Record daylight conditions, luminaire states, dimming and tuning settings, blinds and curtains—not just the nominal lamp specification.

Office daylight and views

Boubekri et al. (2020) reported 316 equivalent melanopic lux in an optimized daylight-and-views condition, compared with 40.6 equivalent melanopic lux in the comparison environment. Sleep duration was 37 min longer, with higher scores on the reported cognitive simulations, in the optimized condition. [21]

Daylight and views changed together, so the findings do not isolate a melanopic-light effect. The source's equivalent melanopic lux terminology is retained here; these values must not be silently relabeled as melanopic EDI.

Bed location in a stroke unit

Schuerch et al. (2026) found that 250 lx melanopic EDI was reached for approximately 3 h at window beds, compared with 15 ± 30 min at door beds. Bed location changed exposure opportunity; the study does not establish a lighting treatment effect. [22]

Intermediate care unit relocation

A separate study by Schuerch et al. (2026) found lower melanopic EDI in the newly constructed unit, where recommended daytime illuminance levels were never achieved. Delirium incidence did not differ after relocation. The result supports measuring architecture and exposure, but does not show that a lighting change caused a clinical outcome. [24]

Why averages can conceal an inadequate daily pattern

A meta-analysis of 9 office field studies, published by de Vries et al. in 2025, found that none of the offices met the 250 lx melanopic EDI recommendation throughout the entire day. Reported summary statistics gave a different impression: [10]

These counts concern different reported summaries, not interchangeable sets of participants. A daily mean, a median and time above a threshold answer different questions. Retain time-resolved data when evaluating an occupied-day lighting objective.

Choosing a light meter, logger or dosimeter

Instrument names do not guarantee spectral capability. A light logger can be photopic-only or multichannel, and a dosimeter does not automatically provide validated melanopic EDI. Select the measurement system by its documented outputs and performance. [4, 7, 16]

A photopic meter is useful for visual-task or surface measurements; its logging capability varies. A photopic logger preserves exposure timing but not enough spectral information for a unique melanopic conversion.

A calibrated spectroradiometer supports reference measurements, calibration and controlled experiments, but is typically less convenient for continuous free-living monitoring. A multichannel personal logger can support longitudinal exposure research when its calibration, processing and wear protocol are validated. Environmental loggers add room-level context, not necessarily the occupant's personal or retinal exposure.

Match the instrument to the question

• Does an occupied space maintain the daytime recommendation? Measure eye-level melanopic EDI over the occupied day alongside photopic illuminance, using a calibrated spectral reference or validated multichannel logging.
• What exposure did a participant encounter across a week? Collect timestamped personal melanopic and photopic exposure with valid-wear information and a documented placement protocol.
• How much visually weighted light is on a surface now? Use a calibrated lux meter for instantaneous photopic illuminance.
• Can existing photopic records support melanopic conclusions? A defensible source-specific spectrum or conversion model is needed in addition to lux. Without it, retain conclusions as photopic rather than reconstructing unmeasured melanopic EDI.

Characteristics to evaluate before enrollment

• Spectral response: enough validated spectral information for the intended weighting—not simply the presence of RGB channels.
• Dynamic range: check low-light performance near 1 photopic lx and saturation near 100,000 photopic lx if those conditions are relevant. A detection range is not an accuracy guarantee throughout that range. [4, 16]
• Angular response: assess how closely a planar illuminance sensor follows the cosine law. Cosine correction does not mean angle-independent measurement.
• Linearity and thermal behavior: examine performance across expected intensity and temperature conditions.
• Calibration and traceability: document the reference instrument, calibration status and comparability between units.
• Placement and wear detection: validate optical geometry separately from whether the instrument is being worn.
• Sampling, storage and battery: confirm that the chosen configuration supports the planned field duration.

A validated multichannel dosimeter can characterize personal spectral exposure in real-world research, within the capabilities established for its sensor and protocol.

Sensor validation in laboratory and daily life

In the Ishihara et al. (2026) comparison, all tested devices showed linear responses between 50 and 20,000 photopic lx, but their detection ranges, differences from criterion and interdevice variability differed. Interdevice variability describes differences between units of a device model, not the validity of its wear position. [16]

Correlation does not make sensors interchangeable

The free-living arm used simultaneous wear with randomized device positioning at the non-dominant wrist. Reported average photopic illuminance over 24 h differed substantially: [16]

The measurements correlated at r = 0.91, a dimensionless correlation coefficient, despite their different absolute levels. The largest discrepancies occurred below 100 photopic lx. Correlation describes shared variation; it does not establish agreement. [16]

A separate benchmark reinforces device-specific validation

Kore et al. (2026) assessed spectral, spatial, photometric, melanopic and thermal performance in a different set of dosimeters. [4]

The Actiwatch finding prompted caution for scientific use in that evaluation. This comparison did not test ActLumus, and its rankings apply only within the tested set and conditions. [4]

Neither benchmark establishes a universal ranking across all devices, spectra, placements or endpoints. Before enrollment, compare the selected measurement system with an appropriate calibrated reference under study-relevant lighting. Record its model, calibration status, sampling settings and derived-metric method.

Where to place a personal light sensor

A body-worn light sensor measures exposure at the device, not automatically at the eye. An ocular light proxy is an estimate intended to approximate eye-level exposure; the term does not establish equivalence. Define placement as anatomical location, body side, height, sensing-surface orientation and attachment method. [15, 27]

Placement error has distinct components:

• Translational displacement: distance between the sensor and the eye-level target.
• Rotational displacement: difference between sensor orientation and viewing direction.
• Body self-occlusion: obstruction of the light field by the participant's body.

Posture, clothing, head movements, body morphology and the position of windows or luminaires change these relationships.

Wrist placement is practical for repeated wear and integration with actigraphy, but sleeves and frequent orientation changes affect interpretation. Lower-chest placement can be discreet and stable when individually selected. Near-eye placement reduces spatial separation from the target, but field of view, glasses, attachment stability and adherence remain relevant. [15, 27, 30]

Chest placement requires geometric validation

In simulated indoor illumination, lower-chest measurements differed from eye-level measurements by −26.2% to +63.8%, the smallest deviations among the chest locations tested. Under overhead lighting, estimated corneal exposure from chest measurements showed deviations exceeding 25%. [15]

A hybrid measurement-simulation study found that 0% to 46.4% of chest area met its illustrative placement limits, depending on participant and posture. The authors recommended individual selection of a chest position that minimizes displacement factors for typical postures. Neither study supplies a universal chest-to-eye correction factor. [27]

Acceptability is different from validity

An online survey of 145 participants, published in 2025, compared perceived wearability and appeal. [30]

These were acceptability ratings, not measurements of long-term adherence or optical accuracy.

Eye-level measurement still needs a protocol

A neonatal intensive care study estimated eye-level melanopic EDI while accounting for logger position, incubator location, infant head orientation, curtains and correction factors. Its workflow illustrates geometry-sensitive measurement, not a reason to apply adult exposure recommendations to infants. [25]

An outdoor mannequin-head study used a sensor probe at the eye position to examine directional light and the effects of glasses and lenses. It informs eye-level geometry, but came from vision-science and myopia research rather than circadian-specific dosimeter validation. [28]

For a field example using a different body site, see the 14-day collar-worn study of sleep and light in new parents.

Validate placement for the endpoint

A 2026 preprint comparing eye, chest and wrist measurements reported outcome-dependent bias: timing outcomes were more robust than level and temporal-dynamics outcomes. This preliminary finding should not be generalized to every device or population. [32]

Use a validation subsample wearing the intended placement and an eye-level reference concurrently across relevant indoor and outdoor conditions, activities and postures. The evidence assembled here consists mainly of primary validation and acceptability studies, plus a preprint, rather than a dedicated peer-reviewed placement review. Simulated scenes, limited postures and perceived acceptability do not establish universal field validity.

Circadian entrainment and the phase response curve

Circadian entrainment is the ongoing synchronization of an internal, approximately 24-hour rhythm to a recurring environmental cycle. A phase shift is a change in that rhythm's timing after a stimulus. Light is the dominant synchronizing environmental cue for the human circadian system, but its effects depend on biological timing as well as spectrum, intensity, duration and prior exposure. [3, 33]

Signals from ipRGCs reach the suprachiasmatic nucleus, or SCN, through the retinohypothalamic pathway. The SCN's molecular feedback loops, involving clock genes such as Per and Cry, generate near-daily timing; light helps align an endogenous period often around 24.2 h with the 24 h day. [3, 6]

The phase response curve (PRC) plots the direction and magnitude of a phase shift against the biological time of exposure. Dim-light melatonin onset, or DLMO, provides a phase reference. Light during the biological evening and early night generally delays the clock; light after the core body temperature minimum generally advances it. The temperature minimum is often approximately 7–8 h after DLMO, but this is not a universal scheduling rule. [12–14]

The illustrated biological-night interval and melatonin offset around 10.5 h after DLMO are schematic. The controlled experiments below used different stimuli and melatonin assessments to establish circadian phase; their results should remain attached to their original protocols.

Khalsa et al. (2003): a prolonged white-light pulse

In 21 participants, a 6.7 h pulse of approximately 10,000 photopic lx white light produced a Type 1 PRC with a fitted peak-to-trough amplitude of 5.02 h. The curve showed delays before and advances after the temperature minimum, without a prolonged daytime dead zone. [12]

St Hilaire et al. (2012): a shorter white-light pulse

In 18 participants, a 1 h pulse of approximately 8,000 photopic lx white light produced a Type 1 PRC with a fitted amplitude of 2.20 h. This was roughly 40% of the longer white-light PRC amplitude from approximately 15% of its duration, illustrating a nonlinear duration-response relationship. [13]

Rüger et al. (2013): short-wavelength light

In 18 participants, a 6.5 h pulse of 480 nm light at 11.8 µW/cm², approximately 11 photopic lx, produced a maximum delay of −2.6 h and maximum advance of +1.3 h. The response was approximately 75% of that in the reference 10,000 lx white-light PRC. [14]

Together, these experiments demonstrate nonlinear duration-response relationships and the importance of spectral weighting. They do not establish a universal exchange rate between blue and white light.

Exposure involved controlled gaze or full-field conditions; the short-wavelength experiment used dilated pupils. These results cannot be transferred directly to free-living exposure, nor treated as guaranteed shifts or universal upper bounds.

Interpreting light timing without turning a schematic into a prescription

Evening and morning light can shift the illustrated biological night in opposite directions. Interpretation depends on biological phase, not just the clock time printed on a schedule.

The example uses DLMO around 21:00, sleep from 23:00 to 07:00, and a temperature minimum approximately 7.7 h after DLMO. The +2.2 h delay and −1.4 h advance are illustrative horizontal shifts, not measured study outcomes. Positive here means later clock time; in Figure 3, positive means a phase advance.

• Earlier-timing objectives: light after the individual's temperature minimum may support an advance, including in eastward-adjustment research. Local clock time alone is insufficient.
• Later-timing objectives: light in the biological evening may support a delay, including in westward adjustment or selected shift-work protocols. The intended work and sleep schedule matters.
• Evening room and screen exposure: may suppress melatonin or delay timing under some conditions, but is not the sole explanation for social jet lag.
• Daytime lighting: often produces smaller phase shifts than biological-night exposure, but should not be treated as having no circadian relevance.

Sources: laboratory PRCs and reviews of non-visual light responses. [12–14, 18, 33, 36]

An acute increase in alertness is not the same as entrainment. A study investigating phase should retain exposure timing relative to circadian markers, rather than relying only on a daily light total or assumed morning and evening clock-time windows.

Field protocol for comparable light-exposure data

A defensible protocol specifies the exposure quantity, measurement geometry and quality-control rules before collection. Expert field-study recommendations and the MeLiDos metrology initiative emphasize comparability, not simply longer recordings. [8, 10, 11]

1. Define the question and endpoint. Decide whether the target is room-level lighting, personal exposure, an ocular proxy, sleep–wake behavior or circadian phase. Entrainment studies may need independent phase markers such as DLMO or cortisol timing; actigraphy alone does not establish circadian phase.
2. Prespecify metrics and analysis windows. Retain photopic illuminance alongside melanopic EDI. Define clock-time and sleep-relative windows, time above or below thresholds, exposure timing and regularity. Thresholds are protocol parameters, not universal biological boundaries.
3. Document the instrument. Record model, unit identity, firmware or software version when available, spectral channels, derived-metric method, calibration status, reference instrument, detection limits and sampling interval. Check performance under study-relevant spectra, including dim and warm indoor conditions.
4. Specify placement. Record anatomical site, body side, attachment, height and sensing-surface orientation. Pilot overhead, side, window and outdoor illumination during seated, standing, walking and typical head or body movements. State rules for sleeves, accessories and other obstructions.
5. Capture day-to-day variation. Field-study guidance recommends at least 1 week, including weekend days, with epochs of 1 min or finer where appropriate to the endpoint and device. Report season, geographical location, time zone and timestamp handling. [10]
6. Record environmental and behavioral context. For design verification, document occupant viewpoint, daylight, luminaire state, control settings and curtains. For personal monitoring, record charging, removal, nighttime placement and relevant activities with a sleep–wake or exposure diary.
7. Prespecify quality-control rules. Distinguish darkness from non-wear, covered sensors and nighttime storage. Define minimum valid wear, missing-data handling, saturation, implausible values, orientation changes and replacement-device procedures. Preserve the original data and the quality-control decisions. [11]
8. Report distributions and individual time series. Provide valid-wear denominators, uncertainty and individual exposure patterns, not only group means. A duration above a threshold is not equivalent to a daily mean, and between-person variation should remain visible.

Use the light-data quality-control guide for auxiliary-data and cleaning considerations. Processing software can support these decisions, but cannot remove an unvalidated spectral response or placement bias.

Combining light measurements with wear-status data

Movement, orientation, temperature, contact or proximity information and diaries can help distinguish valid exposure from device removal or obstruction. A wrist-specific off-wrist signal should not automatically be treated as a validated adherence measure for chest or glasses mounting.

Patterson Gentile et al. (2026) developed a chest-worn light-logger model using movement, orientation, light and time of day to distinguish wear, non-wear and nighttime placement. [26]

When applied to 20 participants with migraine, the model identified the following adherence to a criterion of at least 80% appropriate use: [26]

These are study-specific results, not universal wear thresholds or guaranteed performance in a new population. The same work identified different cut points for photopic illuminance and melanopic EDI when distinguishing indoor from outdoor conditions; a threshold for one metric should not be transferred automatically to the other. [26]

Wear-status modeling is particularly important when low readings could mean genuine darkness, a covered sensor or a device left on a table. Report the classification method and its validation separately from the optical sensor's calibration.

Integrated light and actigraphy studies: applications and evidence limits

Personal light dosimetry provides ecological exposure data; actigraphy adds behavioral timing and rest–activity information. Together, they support research on when light occurs relative to sleep and activity, without making the optical record a direct measure of sleep stages or circadian phase.

Youth with migraine: exposure patterns

Patterson Gentile et al. (2025) used ActLumus in a 1-week exploratory study. The daytime melanopic recommendation was achieved during 15.1% of the measured day, while evening and nighttime exposure generally remained within the study's limits. Delayed light exposure was strongly associated with monthly headache days. [37]

The association does not show that light timing caused migraine or that altering exposure will improve symptoms.

Adolescents: weekdays and weekends

Emilly Francianne Lamego da Silva et al. (2026) used 7 days of wrist actigraphy with integrated light sensing to characterize time-resolved exposure and sleep-related measures across weekdays and weekends. This demonstrates longitudinal feasibility, not equivalence with eye-level exposure. [31]

Real-world sleep architecture

Akgun et al. (2026) combined personal melanopic light logging, including ActLumus and Spectrawear, with sleep measurements in naturalistic settings. Exposure associations do not make a light logger a standalone sleep-staging instrument. [23]

Adults with late schedules

Facer-Childs et al. (2019) used actigraphy and circadian markers in a randomized, multicomponent behavioral intervention that included earlier light exposure. The study assessed an approximately 2 h advance, but the contribution of light cannot be isolated from the other intervention components. [38]

Circadian Rhythm for Sleep: feasibility research

Seo et al. (2026) used wearable and light-sensor data in a single-arm feasibility study of coaching on stable wake time, morning light and daytime activity. The study reported favorable feasibility and improvements in subjective insomnia measures; without a control group, treatment-effect conclusions remain preliminary. [41]

Sunlight, artificial light at night and sleep stages

Montanari et al. (2026) examined light exposure alongside sleep stages in a 7-day sensor-based observational study. Observational exposure records do not prove causality. [42]

Keep exposure measures and outcome measures distinct

Research on daylight spectrum and intensity reduction has assessed melatonin, cortisol, alpha-amylase, sleep parameters and sleepiness as separate outcomes. These are not interchangeable with a melanopic EDI reading. [34]

Reviews of chronopsychiatry discuss light exposure, activity, heart rate and skin temperature as potential circadian biomarkers, while emphasizing that personalized applications are still developing. This literature does not imply that every signal is available from a single instrument. [39]

For delayed sleep–wake phase disorder, light sensing and circadian modeling are promising assessment and monitoring tools, but the condition is heterogeneous. Interpret exposure alongside behavioral timing, relevant phase markers and clinical context rather than as a standalone diagnostic result. [40]

Analysis tools do not replace measurement validation

The open-source pyLight Python module supports personalized light-exposure analysis, including threshold and timing metrics from compatible logger and actimeter recordings. [29]

The LightLogR R package includes dedicated ActLumus import support. Regardless of software, document the input metric, conversion assumptions, wear masks and processing version. Neither package by itself resolves unknown placement or sensor errors.

What melanopic EDI can and cannot tell you

Melanopic EDI characterizes an exposure, not a diagnosis, hormone concentration or guaranteed sleep outcome. Biological responses also depend on timing, duration, prior light history, pupil and ocular factors, circadian phase, age, schedules and health. [1, 17–19, 33]

A systematic review and meta-analysis found wavelength-dependent acute non-visual effects and concluded that melanopic illuminance may not capture every alerting effect of light. A melanopic metric is therefore valuable without being a complete model of all non-visual responses. [19]

Source-level spectral data may differ from exposure at the eye after reflection, transmission, shading and changes in viewing direction. Wrist and chest measurements add displacement and occlusion effects. A single measurement can miss the daily pattern, and accurate device-level sensing does not establish an accurate ocular proxy.

Light exposure, activity, sleep timing, symptoms and mental health can influence one another. Observational associations remain vulnerable to confounding, reverse causation and incomplete adherence; small samples, short recordings and selected populations limit generalization. Reviews of light and mental health emphasize both the potential of personal sensing and these causal-inference challenges. [35, 42]

For clinical and research teams, use light data as a documented measurement and monitoring input. Evaluate intervention effects with suitable controls and independent circadian, sleep or clinical outcome measures—not from the exposure record alone.

Measure melanopic EDI with ActLumus

ActLumus is a clinical and scientific research actigraph from Condor Instruments with a 10-channel light sensor for Melanopic EDI and Photopic lux estimation. According to Condor Instruments, no competing actigraph offers this 10-channel Melanopic EDI light-sensor combination. View the official ActLumus documentation.

This exclusivity concerns the combined actigraph and 10-channel sensor feature, not the melanopic EDI metric itself. Independent validation results are reported separately above.

For a study team, the measurement workflow includes:

• Multichannel light sensing to retain a melanopic perspective alongside photopic exposure.
• Activity, temperature and dedicated off-wrist sensing to add behavioral and wear-status context.
• Bluetooth connectivity for device communication.
• ActStudio for visualization, extraction and export within the Condor Instruments analysis ecosystem.
• Condor Cloud, an optional additional platform for remote, real-time data monitoring, plus API integration with third-party systems.

The independent results support evaluating ActLumus for the tested optical conditions. They do not remove the need to validate placement, sampling settings or analysis for your endpoint.

Explore ActLumus or discuss your measurement protocol with Condor Instruments. Bring your intended metric, exposure range, body position, recording duration and data workflow.

Frequently asked questions

What is melanopic EDI?

Melanopic equivalent daylight illuminance is the photopic illuminance of D65 daylight that would produce the same melanopic irradiance as the assessed light. CIE S 026:2018 defines the metric, which is expressed in lux and characterizes exposure rather than an individual's response. [1]

What is circadian light?

Circadian light is light evaluated for its influence on circadian timing through retinal pathways, especially melanopsin-containing ipRGCs, rather than visual brightness alone. Melanopic EDI provides a standardized measure of melanopsin-weighted exposure, but does not by itself predict an individual's circadian response. [1, 3]

What is the difference between lux and melanopic lux?

Photopic lux uses the visual sensitivity function V(λ), whereas melanopic quantities use melanopsin-related spectral weighting. The label melanopic lux can refer to an older convention, so specify whether a value is legacy melanopic lux or CIE melanopic EDI. [1, 2]

What is the difference between melanopic lux and melanopic EDI?

The legacy equi-energy convention and the CIE D65 convention use different reference spectra and scales. For values calculated under the Lucas et al. convention, multiplying by the dimensionless factor 0.9058 gives melanopic EDI in lx; verify the method before converting. [1, 2]

How is melanopic EDI calculated?

Weight spectral irradiance by the CIE melanopic action spectrum to obtain melanopic irradiance, then convert to its D65 equivalent. The practical factor is approximately 754 lx per W/m²; a validated multichannel sensor can estimate the necessary weighting, but photopic lux alone cannot. [1, 5]

What is CIE S 026?

CIE S 026:2018 is the metrology standard defining the five α-opic sensitivity functions and related quantities for ipRGC-influenced responses to light. It defines measurement methods, not clinical treatment targets or mandatory exposure limits. [1]

What are ipRGCs?

Intrinsically photosensitive retinal ganglion cells are retinal neurons that detect light through melanopsin and receive rod and cone input. They contribute to circadian entrainment, melatonin responses and pupil regulation. [3, 6]

Why isn't a regular light meter enough for circadian research?

A conventional photopic meter does not uniquely determine melanopic exposure because different spectra can produce the same photopic illuminance. Report photopic lux alongside melanopic EDI when both visual and non-visual exposure matter, using a system validated for the required metric. [1, 16]

Is photopic lux enough for circadian lighting design?

Not by itself: photopic lux describes visual illuminance, while melanopic EDI characterizes melanopsin-weighted exposure. A design concerned with visual and non-visual effects should consider both, without treating either value alone as a complete prediction of alertness, sleep or circadian response. [1, 19]

What is a light dosimeter?

A light dosimeter records exposure over time, often on a participant. Its ability to provide spectral or melanopic information depends on the sensor and calibration; the name alone does not guarantee melanopic EDI capability. [4, 7]

Where should melanopic light be measured?

Measure where the target exposure occurs and document position, height, orientation and timing. An ocular-exposure question calls for eye-level measurement or a placement-specific validated proxy, while a room-level assessment also needs daylight and luminaire-state records. [15, 25, 27]

Can a wearable sensor measure the light reaching the eyes?

A body-worn sensor generally provides an indirect estimate based on its location and orientation, not a direct measurement of retinal exposure. Eye-level measurements or validated ocular proxies still require attention to gaze, head position, optical transmission and obstructions. [15, 27]

Are wearable light sensors interchangeable?

No: independent studies found differences in detection range, photopic and melanopic accuracy, interdevice variability and free-living measurements. Validation must be matched to the model, spectrum, intensity range and intended endpoint. [4, 16]

Is a wrist-worn light sensor a valid measure of light at the eye?

It can provide useful personal exposure data, but should not automatically be treated as eye-equivalent. Wrist orientation, clothing, lighting geometry and the selected endpoint all affect validity. [15, 16, 32]

Is chest placement better than wrist placement?

Neither position is universally better. Chest placement may support acceptability and stability, while wrist placement integrates naturally with actigraphy; both require validation for the intended measurement. [15, 27, 30]

Can a correction factor convert chest or wrist data into eye-level exposure?

A correction may be defensible for a defined device, placement, population and lighting context. The available evidence does not support one universal correction across postures and environments. [15, 27]

What should a sensor-placement protocol report?

Report anatomical site, body side, orientation, attachment method, height, expected wear conditions, sensor model, sampling procedure, non-wear rules and derived metrics. Also document changes in posture, clothing or placement that could alter the relationship to the target exposure. [11, 27]

What is the difference between circadian entrainment and a phase shift?

Entrainment is ongoing synchronization to a recurring environmental cycle. A phase shift is a change in timing following a stimulus; its direction depends on biological phase and exposure conditions. [12–14, 33]

What is the phase response curve (PRC)?

A PRC describes how a stimulus shifts circadian timing according to when it occurs in the biological cycle. Light generally delays phase in the biological evening and advances it after the temperature minimum, but the curve is not a universal clock-time prescription. [12–14]

Why are wear and non-wear models important?

Removed, covered or misplaced sensors can distort exposure summaries, especially when low readings are mistaken for genuine darkness. Models using movement, orientation, light and time can help, but their performance needs evaluation in the target dataset. [11, 26]

Should wearable light data be combined with actigraphy?

Often, yes, when the question concerns relationships among exposure timing, sleep–wake behavior and activity. Actigraphy adds behavioral context but does not validate the optical measurement or replace an independent circadian phase marker. [31, 38]

Can melanopic EDI predict melatonin suppression or sleep outcomes?

It characterizes an important component of the stimulus, but does not by itself predict an individual's outcome. Timing, duration, prior exposure, ocular factors and circadian phase also matter, and melanopic weighting may not capture every acute alerting effect. [19, 33]

Can wearable light measurements prove that light caused a sleep or health outcome?

No: improved exposure measurement does not eliminate confounding, reverse causation or adherence problems. Controlled interventions and appropriate outcome measures are needed to evaluate causality. [35, 41, 42]

References

1. CIE (2018). CIE S 026:2018—CIE System for Metrology of Optical Radiation for ipRGC-Influenced Responses to Light. DOI: 10.25039/S026.2018.
2. Lucas RJ, Peirson SN, Berson DM, et al. (2014). Measuring and using light in the melanopsin age. Trends in Neurosciences 37(1):1–9. DOI · PubMed 24287308 · Europe PMC.
3. Do MTH, Yau KW (2010). Intrinsically photosensitive retinal ganglion cells. Physiological Reviews 90(4):1547–1581. DOI.
4. Kore R, Boyer V, Reichenberger D, Mundinger J, Chang A, Durmus D (2026). Evaluation of wearable light dosimeters for circadian lighting: spectral, spatial, photometric, melanopic, and thermal performance. npj Biological Timing and Sleep 3:33. DOI · PubMed 42477415 · Europe PMC.
5. CIE (2020). CIE S 026 α-opic Toolbox, v1.049a. DOI: 10.25039/S026.2018.TB.
6. Do MTH (2019). Melanopsin and the intrinsically photosensitive retinal ganglion cells: biophysics to behavior. Neuron 104(2):205–226. DOI.
7. Stampfli JR, Schrader B, di Battista C, et al. (2023). The Light-Dosimeter: a new device to help advance research on the non-visual responses to light. Lighting Research & Technology. DOI.
8. Spitschan M, Zauner J, Nilsson Tengelin M, et al. (2024). Illuminating the future of wearable light metrology: overview of the MeLiDos Project. Measurement 235:114909. DOI.
9. Brown TM, Brainard GC, Cajochen C, et al. (2022). Recommendations for daytime, evening, and nighttime indoor light exposure to best support physiology, sleep, and wakefulness in healthy adults. PLoS Biology 20(3):e3001571. DOI.
10. de Vries SW, Gkaintatzi-Masouti M, van Duijnhoven J, Mardaljevic J, Aarts MPJ (2025). Recommendations for light-dosimetry field studies based on a meta-analysis of personal light levels of office workers. Lighting Research & Technology 57(1):47–70. DOI.
11. Zauner J, Stefani O, Bocanegra G, et al. (2026). Auxiliary data, quality assurance and quality control for wearable light loggers and optical radiation dosimeters. npj Biological Timing and Sleep 3:11. DOI.
12. Khalsa SBS, Jewett ME, Cajochen C, Czeisler CA (2003). A phase response curve to single bright light pulses in human subjects. Journal of Physiology 549(3):945–952. DOI.
13. St Hilaire MA, Gooley JJ, Khalsa SBS, et al. (2012). Human phase response curve to a 1 h pulse of bright white light. Journal of Physiology 590(13):3035–3045. DOI.
14. Rüger M, St Hilaire MA, Brainard GC, et al. (2013). Human phase response curve to a single 6.5 h pulse of short-wavelength light. Journal of Physiology 591(1):353–363. DOI.
15. de Vries SW, Mardaljevic J, van Duijnhoven J (2026). Impact of wear position on dosimeter performance: measurement validity under simulated indoor illumination. npj Biological Timing and Sleep 3:19. DOI · PubMed 42086978 · Europe PMC.
16. Ishihara A, Brychta RJ, LaMunion SR, et al. (2026). Performance of wearable light sensors for measuring photopic and melanopic illuminance under laboratory and free-living conditions. Sleep 49(2):zsaf358. DOI · PubMed 41230818 · Europe PMC.
17. Feigl B, Zele AJ (2014). Melanopsin-expressing intrinsically photosensitive retinal ganglion cells in retinal disease. PubMed 24879087 · Europe PMC.
18. Piper M, Broszio K, Denneman J, Gordijn M, et al. (2026). Bright days, dark nights: translating science on the non-visual effects of light into public health guidance. PubMed 42482769 · Europe PMC.
19. Cheshmeh Noor M, Revell V, Mehdizadeh Saradj F, Yazdanfar SA (2023). The impact of wavelength on acute non-visual responses to light: A systematic review and meta-analysis. PubMed 37364848 · Europe PMC.
20. Terán E, Yee-Rendon CM, Sosa-Arámbula HJ, De La Herrán-Arita AK, et al. (2026). Home lighting, blue-light filtering, and their effects on melatonin suppression. PubMed 41565717 · Europe PMC.
21. Boubekri M, Lee J, MacNaughton P, Woo M, et al. (2020). The Impact of Optimized Daylight and Views on the Sleep Duration and Cognitive Performance of Office Workers. PubMed 32384634 · Europe PMC.
22. Schuerch N, Bassetti CLA, Arnold M, Vorster APA (2026). Too noisy, too dark: sleep environment at a typical stroke unit. PubMed 41979096 · Europe PMC.
23. Akgun SG, Gemici B, Roddis C, Bickerstaff L, et al. (2026). Light exposure and sleep architecture in real-world settings. PubMed 42432199 · Europe PMC · Nature Portfolio.
24. Schuerch N, Christ S, Bassetti CLA, Vorster APA (2026). Impact of Intermediate Care Unit Architecture on Noise, Light, and Delirium Incidence. PubMed 42085108 · Europe PMC.
25. Bos R, Belgers S, Dudink J, de Kort Y, et al. (2026). Towards implementation of continuous biologically relevant light measurements at the bedside in the Neonatal Intensive Care Unit. PubMed 42557341 · Europe PMC.
26. Carlyn Patterson Gentile, Adah Thomas, Ryan Shah, Blanca Marquez De Prado, et al. (2026). Evaluating chest-worn light logger adherence: development and application of a wear/non-wear model. PubMed 42248933 · Europe PMC.
27. de Vries SW, Mardaljevic J, van Duijnhoven J (2026). Impact of wear position on dosimeter performance: a hybrid measurement-simulation approach to quantify in-situ factors. PubMed 42086926 · Europe PMC.
28. Jiang X, Kondo S, Otsuka N, Kaneda D, et al. (2025). An analysis of light that reaches the eye surface in an outdoor environment. PubMed 41038838 · Europe PMC.
29. Hammad G, Wulff K, Skene DJ, Münch M, et al. (2024). Open-Source Python Module for the Analysis of Personalized Light Exposure Data from Wearable Light Loggers and Dosimeters. PubMed 39021508 · Europe PMC.
30. Zauner J, Biller AM, Spitschan M (2025). Impact of light logger and dosimeter placement on wearability and appeal in real-life settings. PubMed 41081246 · Europe PMC.
31. Emilly Francianne Lamego da Silva, Guilherme Martins, Francimara Diniz Ribeiro, Leonardo Martins Guimaraes Rossi, et al. (2026). Light Exposure Rhythms and Sleep Organization in Adolescents: Temporal Differences Between Weekdays and Weekends in an Actigraphic Study. PubMed 42029562 · Europe PMC.
32. Zauner J, de Vries S, Didikoglu A, van Duijnhoven J, et al. (2026). Sensor placement causes outcome-dependent bias in ambulatory light-exposure estimates. Preprint. DOI: 10.64898/2026.07.28.741277.
33. Brainard GC, Rollag MD, Hanifin JP (1997). Photic regulation of melatonin in humans: ocular and neural signal transduction. PubMed 9406028 · Europe PMC.
34. Kováčová K, Dzirbíková Z, Grácová L, Turlík M, et al. (2025). Daylight spectrum and intensity reduction and its impact on melatonin, cortisol, alpha-amylase, sleep parameters and sleepiness in humans. PubMed 40916933 · Europe PMC.
35. Spitschan M, Zauner J (2026). Light Exposure as a Modifiable Determinant of Mental Health. PubMed 42047992 · Europe PMC.
36. Dubey SA, Elango N, Coelho LB, Chaudhry M (2026). Evening Light-Emitting Diode Screen Exposure, Melanopsin-Mediated Circadian Disruption, and Optical Filtration for Sleep Hygiene: A Narrative Review in the Indian Context. PubMed 42524168 · Europe PMC.
37. Carlyn Patterson Gentile, Ryan Shah, Blanca Marquez De Prado, Nichelle Raj, et al. (2025). Daily light exposure habits of youth with migraine: a prospective exploratory study. PubMed 41312089 · Europe PMC · Nature Portfolio.
38. Facer-Childs ER, Middleton B, Skene DJ, Bagshaw AP (2019). Resetting the late timing of 'night owls' has a positive impact on mental health and performance. PubMed 31202686 · Europe PMC.
39. Suseela Devi AB, Lyman K, McClung CA (2026). The Role of Chronopsychiatry in Precision Medicine: Circadian Dysregulation as Diagnostic Marker and Therapeutic Target. PubMed 41725981 · Europe PMC.
40. Micic G, Reynolds AC, Phillips AJK, Zee P, et al. (2026). Understanding delayed sleep-wake phase disorder (DSWPD): Mechanisms, comorbidities, and evolving approaches to diagnosis and treatment. PubMed 41679168 · Europe PMC.
41. Seo M, Park S, Jeong J, Nam Y, et al. (2026). The Circadian Rhythm for Sleep digital therapeutic for insomnia: Conceptual background and single-arm feasibility study. PubMed 41800163 · Europe PMC.
42. Montanari A, Wang LM, Birenboim A, Chaix B (2026). The Impact of Sunlight and Artificial Light at Night on Sleep Stages: Evidence From a 7-Day Sensor-Based Observational Study. PubMed 42247676 · Europe PMC.

Plan your circadian light study

For sleep laboratories, chronobiology teams and clinics in the United States and Europe, Condor Instruments can discuss ActLumus, melanopic EDI measurement and integration with actigraphy around the requirements of your protocol.

Contact Condor Instruments · Explore ActLumus

Related research resources

• Actigraphy: measurement and interpretation
• Personal circadian light exposure in real-world research
• Light-logger data quality control
• Sleep and light ecologies in parents and infants

Condor Instruments designs research-grade actigraphs and spectral light dosimeters. Condor, ActLumus, ActTrust and ActStudio are trademarks of Condor Instruments. Independent study findings are distinguished from manufacturer specifications throughout this guide.