Is M-EDI a Reliable Predictor of Circadian Entrainment?

By Bill Chan, President, LiteController Inc.

Melanopic Equivalent Daylight Illuminance (M-EDI) is widely used to assess circadian-effective lighting, but is it truly optimized for human biological responses?

What is M-EDI?

Melanopic Equivalent Daylight Illuminance (M-EDI) is a metric used to evaluate how light impacts our circadian rhythms, based on the sensitivity of Melanopsin, a photopigment in our eyes. It’s a key tool for designing lighting systems that support human health and well-being. M-EDI is specified in CIE S 026 and is also referenced in IES RP-46. At the core of M-EDI is the Melanopic Sensitivity Function (MSF), a spectral weighting function intended to model Melanopsin’s response to light.

Where Does the Melanopic Sensitivity Function (MSF) Come From?

The Melanopic Sensitivity Function (MSF), which M-EDI is based on, originates from in vitro research (Bailes & Lucas, 2013). It was derived using genetically modified HEK293 cells exposed to 1-minute light pulses, measuring Calcium responses as a proxy for Melanopsin activation. MSF peaks at ~490 nm, meaning M-EDI assigns its strongest weight to this range. MSF has not yet been validated through in vivo studies on humans using applicable biomarkers, which raises concerns about the accuracy of its direct translation to human circadian physiology. EML (Equivalent Melanopic Lux) is also based on MSF but evaluated against an equal energy spectrum instead of daylight D65. The conversion factor between M-EDI and EML is 1.104, making them part of the same Melanopic-to-Photopic (M/P) ratio family.

Comparing MSF with Human Melatonin Suppression

Melatonin suppression is a well-established biomarker for circadian disruption, as it reflects the strength of light’s influence on the human biological clock. To validate whether M-EDI accurately predicts circadian entrainment, I compared its spectral contributions with Melatonin suppression data from Rahman et al. (2011). This study measured actual Melatonin levels in humans after 4 hours of exposure to a 5000K CCT fluorescent source.

In the above graph:

The blue curve represents the distribution of α-opic Weighted Spectra, based on M-EDI metrics. The dark yellow bars illustrate Melatonin suppression induced by the same light source. For M-EDI metrics to be a reliable predictor of circadian responses, its weighted spectrum should correlate well with actual human Melatonin suppression. However, as shown in the graph, there is a striking misalignment, raising concerns about its accuracy for human-centric lighting design.

Spectral Contributions

M-EDI predicts 20% of Melanopic response in 380-459 nm, 15% of Melanopic response in 460-480 nm, and 65% of Melanopic response in 481-730 nm. Melatonin suppression (Rahman et al., 2011) shows 29% Melatonin suppression in 380-459 nm, 54% Melatonin suppression in 460-480 nm, and 17% Melatonin suppression in 481-730 nm.

Key Observations

M-EDI is strongest at 490 nm, yet the most potent Melatonin suppression occurs at 460-480 nm. The cumulative irradiance within 460-480 nm makes up only 4% of the total irradiance of the fluorescent light. 460-480 nm accounts for 54% of Melatonin suppression, yet M-EDI assigns it only 15%!

M-EDI predicts the strongest response in 481-730 nm (65%), but Melatonin suppression in this range is only 17%. Further supporting this, Brainard et al. (2001) identified 464 nm blue light as the most effective wavelength for Melatonin suppression in humans—directly within the 460–480 nm range that M-EDI underrepresents. This independent finding underscores the biological significance of this spectral region and strengthens the case for revisiting the assumptions behind the Melanopic Sensitivity Function. I discussed the same issue of the disproportionate relationship between Melatonin suppression and the weighted spectra of EML in my peer-reviewed paper presented at the IES Annual Conference 2020 (IES Annual Conference 2020 Proceedings Pg. 74-82).

Why Does This Matter?

If M-EDI overestimates the impact of longer wavelengths (481-730 nm) and underestimates the impact of 460-480 nm, lighting designs based on M-EDI may not effectively support human circadian health. This could lead to suboptimal lighting in offices and schools, where proper circadian entrainment is critical for well-being and productivity.

Many may assume that LED packages peaking at 480-490 nm optimize circadian entrainment based on M-EDI predictions. However, as shown, this assumption may lead to suboptimal lighting strategies. A more accurate circadian metric would provide LED package manufacturers with better guidance to develop lighting solutions that truly support human biological rhythms.

Why Circadian Lighting Needs More Than Just Melanopsin (M-EDI)

To better understand this, consider an analogy: Melanopsin is like a car’s windscreen, while the SCN (suprachiasmatic nucleus) is the engine that drives the circadian system. Melanopsin is the windscreen of the neuro pathway to the SCN engine. It detects light and sends signals, but it is the SCN that ultimately processes these inputs and regulates circadian rhythms.

Focusing solely on Melanopsin (the windscreen) without considering the SCN’s complex integration of signals (the engine) oversimplifies how light influences circadian rhythms. A car cannot function just with a good windscreen—it needs a working engine. Similarly, circadian entrainment is not just about Melanopsin; it involves the entire SCN network and its dynamic response over time.

Further supporting this, Gooley et al. (2010) demonstrated that green light (~550 nm) suppresses Melatonin at levels comparable to blue light (~460 nm) during the first hour of exposure. However, over longer durations, blue light becomes more effective at Melatonin suppression. This suggests that the spectral sensitivity of Melatonin suppression is time-dependent rather than fixed.

Since Melatonin suppression varies based on light exposure duration, this raises concerns about the validity of a fixed Melanopic Sensitivity Function (MSF) to estimate circadian entrainment efficiency. In my paper, “Green Light Transient Effect Necessitates Replacing the Melanopic Sensitivity Function of Equivalent Melanopic Lux” presented at the IES Annual Conference 2020, I highlighted how Melatonin suppression varies significantly across 2 to 8 hours of exposure. Given that the light source’s irradiance remains constant, this variation suggests that MSF itself may not be static across different exposure times.

Using a fixed MSF curve to estimate the dynamic response of the SCN oversimplifies how light influences circadian rhythms. The SCN is not a static system—it integrates multiple neural and hormonal signals over time to regulate circadian entrainment. Treating MSF as a fixed function for all durations of light exposure ignores this complexity and may lead to inaccurate predictions of circadian responses.

What’s Next? 

The growing field of circadian lighting demands metrics that more accurately reflect human biological responses. While M-EDI has provided a valuable foundation, emerging evidence suggests that refinements may be needed—particularly in spectral weighting and time-dependent sensitivity. As research advances, there is a clear opportunity to improve lighting standards and product development to better reflect the complexity of human circadian physiology.

For lighting professionals, aligning design strategies with evolving biological insights will be essential to creating environments that truly support human health and well-being.

 ABOUT THE AUTHOR

Bill Chan is the President of LiteController Inc., a manufacturer of Circadian Lighting LED fixtures and modules. Bill has also been an Honorary Professor of Electrical Engineering at Tongji University for 32 years. Bill lives and works in Markham, Ontario, Canada.

Image: Bill Chan