In Focus:
New Light on Healthy Buildings
Insights from architecture and human physiology could guide the design of buildings in which the lighting – natural and artificial – could help people stay healthy, feel good, and work better. Siobhan Rockcastle and Manuel Spitschan are lighting the way.
How well did evolution prepare us to spend so much of our lives inside buildings, largely under artificial lighting? The answer could depend on the mental, cultural, and technical adaptability that evolution conferred on us. The research of the TUM-IAS Focus Groups Human-Centric Building Performance and Chronobiology & Health is shedding new light on the matter. The principal investigators are Hans Fischer Fellow Siobhan Rockcastle, a professor of architecture at the University of Oregon, and Manuel Spitschan, a Rudolf Mößbauer Tenure Track professor at TUM with a joint appointment as research group leader in the Max Planck Institute for Biol ogical Cybernetics. Science journalist Patrick Regan interviewed Rockcastle and Spitschan by videoconference in December 2025.
(Screenshots): Patrick Regan
Q: Looking through some of your published papers, one thing that stands out for me is that – separately and in collaboration – you’ve been studying things that seem intuitive and even “commonsensical” but turn out to be both complex and very tricky to measure and model. Is that a fair assessment?
Rockcastle: I think it is. Understanding the importance of light exposure in buildings is very intuitive, that it would have positive benefits and also some risks. To me it seems obvious, and in a way my work over the last decade has been to model and demonstrate that. And yet I think the physiological aspects are quite complex, so maybe Manuel has a different take on it.
Spitschan: When you brought intuition into it, Patrick, it reminded me of a question I’ve been thinking about: We all have an intuitive understanding of light, personally, when it’s too bright, when it’s time to turn the lights on or off, but I don’t know whether we always make good decisions about light exposure. Do we have the right cognitive machinery to select light exposure conditions that support our health, well-being, physiology, and behavior? And that brings in this interesting tension of how much of it can be changed with behavior, how much of it can be changed with architecture, or for example with different glazing technologies. And of course the way we measure it is very non-intuitive. For example, who would think we can learn something about how light impacts us by taking saliva measurements?
Rockcastle: I am also struck by the ways in which people can talk about their experiences and memories with light in buildings, or buildings that impacted them positively, spaces that made them feel good or that they’d like to return to or spend more time in. But then when I teach entry-level design studios to young architecture students, despite that connection they can make in their own experiences, they don’t intuitively make those connections in their own architectural designs. It’s challenging. You not only have to explain the importance of it, but you also need to help them understand how their design decisions impact how deep light penetrates spaces, and how variable it is over time. Those things aren’t necessarily intuitive.
Q: I was a little shocked when I started peeling back the layers, the kinds of questions that you’re studying and trying to model. We want to live in buildings that make us feel better, make us healthier; we want or should want to manage our exposure to light in a way that is evolutionarily sensible while everything around us may be pushing in another direction. And here you are asking fundamental questions about sensation, psychology, and behavior and on the other hand trying to make better experiences for people in their workplaces and elsewhere in their life. So how you would describe the balance between research on more or less fundamental questions and practical application to real-world solutions?
image: Verena Müller
Spitschan: Basically all of the research I did early on in my career was very mechanistic. In the classic tradition of human psychophysics and sensory biology, I would isolate a specific component of light, for example, and then use that to tickle out some physiological effect that otherwise, under real-world conditions, would not be visible or accessible, or would be masked by other things. We have multiple photoreceptors in the eye, and they have different wavelength preferences. Usually under real-world conditions they’re all activated in some way. There are techniques we’ve developed to study some of these photoreceptors in isolation and drive only those and see how the pupil responds, or how melatonin suppression responds. And ultimately what you gather is a body of work that shows biological capacity and maps out what is possible in this biological system we’re studying, which is, broadly, nonvisual effects of light.
This type of research is relevant for biology, but it’s not directly applicable to people living in spaces where they’re exposed to light. I thought, wouldn’t it be great to try to take this mechanistic knowledge and apply it in real-world conditions, and do that in a framework where you have both biology, with its mechanisms and controls, and this messy real world.
Rockcastle: I came to the sciences from a different path. My education began in the arts and applied sciences, in architectural design and building technology. If it didn’t impact the art and construction of buildings, nobody really wanted to talk to me about it. I started to do experiments in my PhD as a bridge toward modeling something that can be generalized to a broader population, and then can be applied through tools that architects can use to predict and evaluate. I came to it by necessity: I believed there was a way that we could model the emotional effects of light, and that there would be some generalizable thread about the effect of contrast on emotional responses, and I realized: I’m going to have to do experiments to test that hypothesis and create a metric.
That was my inroad into more controlled experimental settings and laboratory experiments. I have experience doing experimental research, and I’ve learned to collaborate with people who can fill in the gaps where my knowledge ends. I’ve become more and more interested in understanding, in a way similar to what Manuel described, what we can find when we put people in a setting and expose them to controlled stimuli. People don’t generally experience the world that way, and they’re also dynamic. They move through spaces throughout the day. I became very interested in both the dynamics of movement and the dynamics of different systems in combination.
I had just been looking at the visual aspects, then became interested in effects of light on circadian rhythms through colleagues I was working with, and that formed the framework for OCULIGHT dynamics, which is a consulting company that we founded in 2018. We started to realize you can’t look at just one element of lighting performance in isolation. You can’t just look at the emotional response or the visual effects of light, you have to take into account comfort as an angle and the health potential and circadian stimulus effect of daylight. So in a way that multi-step platform became the foundation of my more recent research.
In one study we did together – Manuel’s group and the lab at TUM and my own – we started to look at how light can have an effect on thermal physiology as well as cognition. Looking at the thermal side was new for me, unlocking many aspects of the complexity of human biological responses to light. We’re still not even integrating the visual, the aesthetic, as a component. And yet the farther I get in my career, the more I want to focus on a pragmatic approach through simple recommendations that designers can understand and use to improve their buildings. At the same time, I understand that performance and experience are complex, and I’m interested in taking a deep dive into the subjects that I don’t understand.
Q: What roles have your TUM Host, Thomas Auer, and your doctoral candidate, Bilge Kobas, played in the Focus Group collaboration?
Rockcastle: Actually this entire collaboration came about because Bilge Kobas, my PhD candidate, and Thomas Auer reached out to me to explain what the Fellowship was and see if I was interested in working with them. I had a relationship that spans now almost a decade with Thomas, crossing paths in different academic venues, and we’d always had very interesting conversations about applied research. Bilge was fundamental in supporting the early collaboration and bringing me into the lab. The lab that we use, the SenseLab, is part of their chair located in the Munich campus of TUM. Bilge did a lot of the early experimental design for our experiment and has been managing and helping to support the experiments, as well as doing the legwork between a lot of the computational analysis, working with Manuel’s team on that. And then when I made my six-month stay there, I was hosted by the chair in the Munich campus and got the experience of working with students and doctoral candidates who were doing a lot of different projects. Their role has been fundamental on both the facilitation side and the experimental side.
Q: You’ve designed experiments that can give you meaningful measurements of people’s responses to artificial stimuli in realistic settings. Are there findings or new research questions that stand out?
Rockcastle: We have a journal article that has been in review for around a year on the findings from our first experiment. We’ve conducted two experiments through our collaboration, and while we have data sets from the first experiment that we’re still working on, we have quite a bit of new data from the second as well. Some of the things we found that were particularly interesting: We exposed people to a range of different lighting conditions, from very dim to somewhat dim to very bright levels that could be found in a typical indoor environment. We exposed participants to these conditions in hours leading up to sleep, and then measured a range of different cognitive and physiological variables – for example looking at hormone samples, core body temperature, and skin temperature over time as well as sleepiness and various other perceived aspects that we collected through questionnaires. We have reported some of these findings in the article that is under review. What we found is that there is an inverted U-shaped response to many of the variables we measured. In previous research, there’s been this idea that with increasingly bright light exposure, you might have responses that are going to follow a linear trend, and we found a degradation of several of the variables under the brightest lighting conditions. That was surprising.
Spitschan: What has been published before said, essentially: the brighter the conditions, the more of an effect – what you usually find in biology: The more you put in, the more you get out, at least within a certain range. And here I think it’s interesting there’s an optimum of light exposure, which brings us back to the question: Are we making optimal choices about our lighting exposure? If we just maximize something, it’s not going to yield the result we might expect. I don’t think we yet understand why this is the case. It is probably some tradeoff with another function that we don’t understand yet – such as, if it’s too bright, we’re probably distracted by glare or something.
Rockcastle: Compensating in some other way.
Spitschan: Another study we have just finished analyzing looks at a theoretically motivated design that has practical implications for measuring these effects. Usually people come to the lab, we do studies in the evening, and when you do these studies you are aware that you should space them apart by about a week, which we call the “washout period.” So you might wash out the previous effect, and the people come again to lab. No one has ever empirically demonstrated that you need this. So we manufactured a sequence of evenings where people came to the lab and were asked, does the previous evening’s experience influence my response to light on the current evening. We’re just finalizing this, but it has implications both for how we think about the time window that we need to look at these effects, but also for experimental design. Because if we find that we need this washout period, it might be a week, it might be shorter, it might be longer, and it will fundamentally change how we collect data. And now we have evidence to base it on. But it also has a devastating consequence for how we assess and measure light: You cannot really take the past out of the picture either. To predict how they’ll respond today, you would have to think: What did this person experience yesterday or the day before? How did they respond? The more you look, the more the complexity increases. It’s important to do this kind of foundational work. It has implications both for how we do the science and how we think about practical implications.
“The farther I get in my career, the more I want to focus on simple recommendations that designers can understand and use to improve their buildings."
– Siobhan Rockcastle
image: Verena Müller
Spitschan: And even how it gets applied to a population. Think of night shift workers who are working three days with an altered day-night schedule and then go back to their real life. Perhaps there’s a delay, some intervention that’s happening in that abnormal light exposure. This has been an opportunity for me to learn more about the physiological side of this. We saw differentiation in core body temperature within an hour, let’s say, of the measured time sequence, between the really dim lighting conditions and some of the brighter lighting conditions. We saw this perpetually low core body temperature that happened with the dimmest lighting conditions and stretched through the night, through the sleep cycle of our participants.
What we expose people to in buildings affects them while they’re in the building, but it also has an effect on them when they go home and persists through the rest of their daily life, regardless of what they do later. And that has revealed to me the importance of what we do as designers in the built environment. You know, design has really important consequences, and I think we’re only just starting to understand these. Think about schools and workplaces, where you don’t get to opt in to the lighting conditions: They’re set for you, and the fact that those could be creating effects for folks habitually over time, for large periods of their life, is really interesting to me.
Q: Are there other findings you’d like to highlight?
Rockcastle: Yes, on the cognition side. People made more errors on some of the cognitive performance tests under really bright light conditions in the evening. So against common wisdom to say that if you’re working the night shift, let’s just blast people with very bright light to keep them alert. We saw a degradation in performance under these bright conditions. That is an interesting finding of note, to say that we can’t stress the biological system of humans beyond their window of neural adaptation without having negative consequences. So more is not better at all times of the day. That’s one that I would stress.
Spitschan: There’s another finding that I think is very deep in the analyses, and it comes out when we look at each individual data study – for melatonin suppression – and that is that people are very different in how they respond to light. There have been studies on this before. There are big differences in how a group of people will respond to light.
Rockcastle: We had somebody who was almost non-responsive.
Q: Responsive to what exactly?
Rockcastle: Let’s say you have a normal curve of melatonin production as well as other functions in the evening. Then if you expose them to light in brighter and brighter quantities, you’re going to suppress that onset curve more. This has been shown in the literature. What we found is there are some people who don’t have the same curve as others, or who are not suppressing that curve as much with light exposure, and the curve is rising at different rates between participants.
Spitschan: Exactly. You look at the sensitivity in terms of how much their melatonin that is naturally produced by their body will be suppressed by light. If you show the same light to this group of people, they have different levels of how much it is suppressed. You can map this out at different intensities, and measure the dose-respone curve. But because there is so much variability, it’s really hard to make statements about how light exposure in the evening can affect someone. Instead of saying this is how much suppression you’ll get, you have to say this is the range of suppression you’ll get.
So again, we’re moving away from making things simple, recognizing that biology is very messy – but can be understood systematically. Looking forward, one of the questions is how we can communicate that uncertainty, that variability. It’s part of biology, ultimately. It’s clear that we’ll never be finished building models of how humans respond to light, and how light is integrated.
Rockcastle: And it’s kind of against the whole idea of a model to begin with. That’s why the model is challenging, because the model is looking at, let’s say, the confidence interval, and there are people who are falling outside of that. That is really interesting. It doesn’t make our job any easier, but it’s endlessly interesting.
Q: Is there a community, among architects, builders, lighting designers, and their clients, that is receptive to this kind of research? Or have you encountered any resistance? Or simply disinterest?
Rockcastle: There are some architects who don’t want to be getting advice from researchers on how to design buildings, or who don’t appreciate that science is providing some qualitative or quantitive layer of performance in addition to what they see as the value system for defining good design. That part of the field definitely exists.
But I have to say that most designers and engineers are receptive to this because it is intuitive in a way. And they also want evidence to support things that they know to be true, such as that daylight is a positive asset to good design in buildings. If they’re trying to get a client to see the value of integrating more daylight, for example – which may mean a slightly higher budget – we can come in and say it’s not only because you save energy, but it’s also because people are going to feel better and perhaps even be healthier in that building. That’s an argument that pulls heartstrings in a way that energy doesn’t. I’ve never met anyone who dismissed it as unimportant. When they’re told that daylight boosts health, they can understand that. We don’t want occupants of a building to be unhealthy. Of course daylight is important. I’ve found people to be very receptive.
“Once you put a person into a space, you cannot take human biology out of the equation.”
– Manuel Spitschan
Q: Manuel, do you have much contact with architects and the like?
Spitschan: I sometimes do. I’m currently the Speaker of the Daylight Academy, a group of scientists, architects and practitioners focused on understanding daylight and its effects, and optimizing it. These encounters highlight issues that Siobhan and I have tried to cover, to map out, in our research: How can we bring together these disciplines that have different vocabularies, different ways of thinking, and learn from each other, to bridge that gap?
It boils down to this: There’s an understanding that once you put a person into a space, you cannot take human biology out of the equation. This has various consequences, including that we should assess light exposure using the right dose-response functions, the right spectral sensitivities, and this is the key lesson that I think is reaching practice.
Also, there’s a lot of information that should flow back to biomedical science but doesn’t. Understanding how such information flows would be a very interesting meta-scientific task. It’s tedious work, because you have to figure out, well, how do you think about a problem, how do I think about a problem, what are the commonalities, what are the differences, in vocabulary or in systems of thinking. As a neuroscientist or a biologist, one could say this is my playground and I’m going to study this effect, and I want to capture the mechanisms. But for me that’s not enough. The mechanisms are situated in a world, where we can make decisions about design, about windows, and ultimately about our behavior.
Rockcastle: I love something you just said. When you put a human in a space, you can’t take biology out of the equation. When we study humans, we’re always studying them in a space, unless they’re out in the trees with electrodes on their heads. You also can’t take the space out of the equation. They’re intrinsically connected, but they’re studied in such different silos, with people who speak in such different vocabularies, it has been challenging in the past to bridge that gap.
I also want to add that there’s also the risk of people studying these sometimes singular, isolated mechanisms in the lab and then publishing a model, and then industry taking that and developing products around that model, but without the context of how they’re applying it in situations with broader sets of variables. That can also be harmful. So I’m constantly telling my students to be wary of products that are labeled as “healthy lighting” without understanding that, sure, this might have an alerting effect, but people don’t like it, it’s uncomfortable, it produces glare, and it can actually be harmful if turned on at night.
Models should not be applied to the real world without that feedback loop back into the lab, as Manuel was saying, where you have to get evidence now in the real world when these models are applied. Are they working? Are they causing more harm? I had a colleague once who was talking about building science and this ideal of doing no harm, like the Hippocratic oath of doctors – we want to create spaces that do no harm. We can’t do that without the feedback loop back to biologists and neuroscientists to understand whether our applied design recommendations are resulting in healthy conditions for people.
Q: What would you like to do next?
Rockcastle: I’m really interested in looking into the persistence of some of these effects into the night and how long they last. Our second experiment was designed to answer some of these questions on the variability among participants, as well as understanding what happens when you’re exposed to different sequences of light in different patterns of exposure. I think we also really need to understand how this translates to real-world conditions. That’s what I would want to do next.
Spitschan: First we found these interesting foundational effects, but then we also found this variability.
I think the priority for future research would be to understand how diverse populations respond to light, and what modifiers of light sensitivity there are, behavioral and cultural as well as genetic or biological. I think this would be a fascinating research program, because then ultimately we can work toward understanding how variable people are in their response to light, and how they select light, in the real world.
Rockcastle: And following on that, how recommendations should vary in different parts of the world, for different populations. These are among the main follow-up research questions we’d like to be able to pursue.