Is Our Framework for Lighting Sufficient?

Part 2: Daylight and Health

Why Is Sunlight Necessary for Health?

Stage I — Color Vision and Circadian Entrainment

"Light is light" — sounds the deep wisdom from many skeptics, when I try to explain to them why we are building Pixun and why spectral properties of light matter for health.

I think what they mean is that light makes things visible, creates colors, and yeah maybe also keeps us awake... but that's it. So why would you put all this effort into using daylight, when LEDs are so simple and cheap?

Even people in circadian lighting share these views, as their concept of "the non-visual effects of light" is actually singular (relying solely on ipRGCs in the retina).

The assumption is that providing good color rendering and ample amounts of light that trigger circadian entrainment should give us all that we need.

Stage II — We Need the Whole Spectrum

But there's a massive body of literature showing that various frequencies of light have deep physiological effects in all tissues of the body.

So let's use LEDs with smoother spectra, extending to the infrared and maybe UV.

And there you have it, we have successfully mimicked sunlight. We can even call it sunlight... or at least that's what many companies do. By the way, a huge number of people have told me that they believed we can create sunlight with electric sources — perhaps a testament to the effectiveness of such marketing.

However, even within the most simplistic framework of light spectrum, we are very far from recreating sunlight. Here are a few obvious reasons:

1. Precise Spectrum

Different frequencies of light can have specific effects. Even a 50 nm difference in the near-infrared can determine whether metabolic facilitation or inhibition occurs — highlighting that the specific frequencies and their proportions are not negligible.

As phosphor technology develops, we achieve higher and higher resolutions in mimicking daylight spectrum, but it is not quite there yet.

Where exactly? — We don't have a target. We don't even have a frequency range where the comparison should be made. Is UV interesting too, or maybe far-infrared?

How similar? — We also don't have any metrics for knowing what threshold should be achieved in similarity.

People have a tendency to pick one mechanism as "the" reason for everything. In circadian lighting, it was ipRGCs. In photobiomodulation, it is metabolic facilitation by near-infrared. So there have been suggestions for using this as the one dimension along which the spectrum should be compared, where blue light would get a "bad" score and red to near-infrared a "good" score.

Such thinking is not only overly simplistic, it also misses the fact that

• even within the near-infrared, some regions inhibit metabolic activity;

• such effects also depend on intensity and dosage (too high and potentially too low levels of irradiation can sometimes reverse the effect);

• we already know about countless further mechanisms and interactions — beyond metabolic effects — to take into account.

Most of all, this is an area where knowledge gaps are certainly gigantic.

Energy Efficiency

Interestingly, the closer we try to mimic daylight, the more expensive it becomes (both regarding the light source and energy use).

We do not yet have a good measure for comparing daylight and electric light in this regard. If we use luminous efficacy, then only the simplest visible aspects of light are captured — so a higher quality (more daylight-mimicking) light source will fare worse than a cheap LED.

Conversely, the relative cost of daylighting applications would appear much lower if they were compared to electric sources not only in terms of lumens (visible light) per dollar, but if the whole spectral output was taken into consideration. This is not a simple question to solve, as who would decide how to weigh different bins of the spectrum?

This issue is just as complex as putting numbers to the long-term health effects of light — perhaps a reason why the topic is still so ignored despite all the glaring evidence.

2. Dynamics

Any comparison between electric lighting and daylight gets further complicated by the intricate changes that daylight goes through, not just on a daily but also on other time scales.

Modern electric sources are controlled digitally. So, what resolution is sufficient to claim that they mimic the gradual changes in natural light? This issue relates to the smoothness of spectral changes in digital vs. analog systems — and it also ties to the problem of flicker.

3. Flicker

The more LEDs we try to mix for a smoother and better adjustable spectral power distribution, the harder it becomes to avoid flicker. While theoretically it is possible, the cost and efficiency burden makes truly flicker-free LED applications not feasible in real life. This is especially true for lamps that run on mains power.

Stage III — Is It Just the Spectrum?

So this is where it gets really interesting. If we could perfectly mimic daylight's spectrum, would we never miss sunshine?

Across the population, there is a correlation between sunlight exposure and health. The assumption of Stage I, but even of Stage II, is that if we used our framework for lighting and used enough electric light indoors to match daylight, this association would diminish. We would live just as happy and healthy if we never went outside, at least when it comes to light physiology.

Is this correct? Has it ever been tested? Can it even be tested? Certainly, there are many confounding variables which would need to be eliminated and that can only be done in animal studies.

Issues with Stage I: Empirical Discrepancies

With the currently mainstream circadian lighting applications, several comparisons between matched daylight and electric light installations are available. 

Based on studies of the action mechanisms of our melanopic response and comparisons to daylight intensity, thresholds have been applied to match the (short-wavelength) intensity of electric lighting so that daytime activations in an office space match that of outdoor exposure.

The result? Many research participants (who were certainly not sensitive — after all, they signed up voluntarily) didn't even complete the study and turned off the intrusive lights.

Focusing only on one factor and neglecting the plethora of others, this approach severely upset the balance within the light spectrum. Since circadian lighting also calls for higher intensities (in an effort to mimic daylight), these imbalances, as well as the negative effects of flicker, are further increased.

Such cold and bright light sources do not only cause glare (which is mainly a spatial inadequacy of the installation), they also increase oxidative stress in the eyes and other tissues, negatively alter the skin's microbiome, reduce color sensitivity, likely hinder proper emmetropization, and disrupt the balance of many other physiological aspects of the spectrum.

Issues with Stage II: No Good Hypothesis

Unlike the concept of melanopic responses, Stage II efforts to mimic sunlight are not well defined. Some tests of it do show, however, that a more complete LED spectrum is better regarding mood, performance, and even eye development outcomes.

Can we test, though, if this approach could ever reach the effectiveness of daylight? Currently, no.

Instead, we can lengthen the list of physical properties we examine, where we will find further and further aspects of measurable differences between daylight and electric light.

Issues with Stage III:

Stage I is a good hypothesis: clearly defined and falsifiable. Its only issue is that it has long been falsified.

Stage II is not a good hypothesis: it's not clearly defined and it is often framed in a way that's not falsifiable. 

Stage III is even more problematic: we have basically nothing to grasp here.

It stems from the realization that daylight is essential for health. If history will repeat itself, the known and measured properties of light will not give full explanation of this (as Stage II would assume). So we are back to the original problem: we simply don't know what light is and how to measure it.

On a pragmatic level though, none of this matters; we can still measure effects on humans. Contrary to popular belief, staying on the descriptive level is no less scientific than pretending to understand the mechanisms through building a model that only works if we don't look at all the evidence...

So I'll stay on the safe side and assume that daylight is strictly necessary — and we don't yet know why.