Introduction
Light has widespread effects on physiology, behaviour and animal welfare, providing vision as well as regulating circadian rhythms, sleep and hormones (Peirson et al. Reference Peirson, Brown, Pothecary, Benson and Fisk2018). However, lighting standards for laboratory animals are typically based on human vision – not the animals who live under it. This challenge was recently addressed in guidelines on the measurement and standardisation of lighting for laboratory mammals (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024). This approach builds upon established methods of light measurement for circadian responses to light in humans, which adjust for the differences in sensitivity between visual and circadian systems (Lucas et al. Reference Lucas, Peirson, Berson, Brown, Cooper, Czeisler, Figueiro, Gamlin, Lockley, O’Hagan, Price, Provencio, Skene and Brainard2014; CIE 2018; Brown et al. Reference Brown, Brainard, Cajochen, Czeisler, Hanifin, Lockley, Lucas, Munch, O’Hagan, Peirson, Price, Roenneberg, Schlangen, Skene, Spitschan, Vetter, Zee and Wright2022). Here, we aim to provide a summary of advice in the form of common questions about the effects of light that may arise from researchers, veterinarians or animal care staff interested in applying these methods to refine animal facility lighting, with a focus on laboratory mice (Mus musculus).
What is light?
Visible light is electromagnetic radiation that can be perceived by the human eye, typically covering wavelengths from 380–780 nm (Sliney Reference Sliney2016). This fundamental description of light highlights a core problem of light measurement – that it is based upon what humans see – not other species. This human-centric approach to light poses a recurring problem. It affects how we define light, how we produce light and how we measure light (Cronin et al. Reference Cronin, Johnsen, Marshall and Warrant2014).
A good way of illustrating this is to consider the white light sources that light our homes, workplaces and animal facilities. White light can be produced in many different ways – such as the incandescent, fluorescent or light emitting diode (LED) bulbs we are all familiar with. All of these involve emitting power across the 380–780 nm range. However, the distribution of this power can be very different between such light sources and is effectively optimised to trick our eye into seeing the light as appearing the same. This is because the human eye is most sensitive to light in the yellow-green part of the spectrum, with a peak around 555 nm, which is described by the photopic sensitivity curve – also called V(λ) (Figure 1). As a result of this sensitivity, these white light sources are optimised to appear bright to the human eye, overlapping with V(λ).
Example of four commonly encountered white light sources (daylight, incandescent light, fluorescent light and a cool white light emitting diode [LED]) and their very different spectral power distribution (300–780 nm). The blue lines show the different spectral power distribution of the light sources (measured in μW/cm2/nm). The red curve on each image shows the photopic sensitivity curve (Vλ) – a measure of human visual sensitivity – that is used to calculate photometric lighting quantities such as lumens and lux.
Figure 1. Long description
Four line graphs arranged horizontally. Each graph shares a common x-axis representing wavelength in n m from 300 to 750 and a y-axis representing spectral power in mu W dot cm super minus 2 dot n m super minus 1. A red bell-shaped curve representing the photopic sensitivity curve V lambda is consistent across all panels, peaking at approximately 555 n m.
* The first panel, Daylight, shows a blue line that rises sharply at 300 n m, remains high and jagged across the visible spectrum, and slightly tapers off toward 750 n m.
* The second panel, Incandescent, shows a blue line starting near zero at 300 n m and exhibiting a steady, linear increase as it moves toward the infrared end of the spectrum.
* The third panel, Fluorescent, shows a blue line with several sharp, narrow peaks at specific wavelengths, most notably near 400, 435, 545, and 610 n m, with very low power between these spikes.
* The fourth panel, White L E D, shows a blue line with a sharp, narrow peak in the blue region at 450 n m, followed by a broader, lower hump peaking around 600 n m.
Light sources optimised for human visual function are often suboptimal for the visual sensitivity of other species. As such, when considering the effects of light on different animals, we need to use a sensitivity function that matches what they see. This involves an appreciation of both visual and non-visual responses to light in animals.
What do mice see?
Mice see the world very differently from humans (Baker Reference Baker2013; Seabrook et al. Reference Seabrook, Burbridge, Crair and Huberman2017). In mammals, light responses are mediated by the eye, and light is detected by specialised light sensitive cells in the retina called rods and cones. These cells can be defined by their peak sensitivity (λmax). Rods detect dim light and are used for night-time vision when light levels are low. Humans and mice have similar rod sensitivity (λmax = 496 nm and 498 nm, respectively) (Bridges Reference Bridges1959). Cones detect brighter light and allow colour vision by comparing the wavelength composition of light. Humans have three types of cone: red (λmax = 558 nm); green (λmax = 530 nm); and blue (λmax = 419 nm) (Dartnall et al. Reference Dartnall, Bowmaker and Mollon1983). The cones are concentrated in a central region of the retina called the fovea, where red and green cones predominate, giving rise to the V(λ) sensitivity peak of 555 nm. By contrast, mice have no fovea and just two types of cone: M cones sensitive to green light (λmax = 508 nm) and S cones sensitive to ultraviolet light (λmax = 360 nm) (Jacobs et al. Reference Jacobs, Neitz and Deegan1991; Sun et al. Reference Sun, Macke and Nathans1997). As a result, the sensitivity of mice is shifted to shorter wavelengths – with reduced sensitivity to long wavelength (red) light and higher sensitivity to short wavelength (violet/ultraviolet) light in comparison to humans (Figure 2). These cones are also found in a gradient across the retina, with more M cones in the upper (dorsal) retina and more S cones in the lower (ventral) retina (Applebury et al. Reference Applebury, Antoch, Baxter, Chun, Falk, Farhangfar, Kage, Krzystolik, Lyass and Robbins2000). In this regard, the mouse retina is more like the human peripheral retina, with more rods and fewer cones (Leamey et al. Reference Leamey, Protti, Dreher, Chalupa and Williams2008). In addition, the visual acuity of mice is poor. By comparison with humans, mice will therefore see the world as a blurry image with increased brightness of violet/ultraviolet colours and reduced sensitivity to reds (Huberman & Niell Reference Huberman and Niell2011; Baker Reference Baker2013) (Figure 3).
Comparison of the retinal visual pigments found in the retina of both (a) humans and (b) mice (Mus musculus).
Figure 2. Long description
The figure consists of two panels, a and b. Both panels share identical axes: the y-axis represents Relative Sensitivity from 0 percent to 100 percent in increments of 25 percent, and the x-axis represents Wavelength in n m from 300 to 800 in increments of 100.
Panel a, labeled humans, shows five bell-shaped curves. From left to right, the peaks are:
- Blue cone (dark blue line) peaking at approximately 420 n m.
- Melanopsin (light blue line) peaking at approximately 480 n m.
- Rod (black line) peaking at approximately 500 n m.
- Green cone (green line) peaking at approximately 530 n m.
- Red cone (red line) peaking at approximately 560 n m.
Panel b, labeled mice, shows four bell-shaped curves. From left to right, the peaks are:
- S-cone (purple line) peaking at approximately 360 n m.
- Melanopsin (light blue line) peaking at approximately 480 n m.
- Rod (black line) peaking at approximately 500 n m.
- M-cone (green line) peaking at approximately 510 n m.
In both panels, all curves except the human blue cone and mouse S-cone show a secondary, much lower sensitivity bump around 350 n m.
A simple illustration of the differences in human vs mouse (Mus musculus) visual function. Showing (A) a photograph of South Park in Oxford, UK and (B) a simple approximation of how the scene in (A) would appear to a mouse, with increased blurring due to the poor visual acuity of the mouse eye, combined with a short wavelength shifted spectral sensitivity, leading to reduced green/red and enhanced violet colours.
What other responses does light affect?
As well as providing vision, the eye also plays an important role in detecting the brightness of the external environment to regulate many other processes. Virtually all organisms possess an internal circadian clock with a period around 24 h, which optimises physiology and behaviour to the varying demands of the day/night cycle. This biological clock means that even in complete darkness, animals still show ~24 h patterns of activity and rest. These circadian rhythms (from the Latin ‘circa diem’ – about a day) are generated by an intracellular genetic mechanism found in cells and tissues throughout the body. In mammals, the master circadian pacemaker is localised within the suprachiasmatic nuclei (SCN) of the hypothalamus, which coordinates peripheral rhythms in organs and tissues throughout the body (Reppert & Weaver Reference Reppert and Weaver2002; Dibner et al. Reference Dibner, Schibler and Albrecht2010).
Light detected by the eye is the primary signal used by the body to align (entrain) the circadian clock to the external environment (Schmidt et al. Reference Schmidt, Do, Dacey, Lucas, Hattar and Matynia2011; Hughes et al. Reference Hughes, Jagannath, Hankins, Foster and Peirson2015; Do Reference Do2019). As such, the daily cycle of light and dark is perhaps the most important environmental cue for regulating physiology and behaviour. Studies on circadian rhythms in mice lacking rods and cones showed that these animals could still entrain to the light/dark cycle, providing early evidence that there may be another type of photoreceptor in the eye (Provencio & Foster Reference Provencio and Foster1995; Yoshimura & Ebihara Reference Yoshimura and Ebihara1996; Freedman et al. Reference Freedman, Lucas, Soni, von Schantz, Munoz, David-Gray and Foster1999). Subsequent studies identified a subset of retinal ganglion cells, the retinal neurons whose axons project to the brain and form the optic nerve, were intrinsically photosensitive and expressed a blue-light sensitive protein called melanopsin (Provencio et al. Reference Provencio, Jiang, De Grip, Hayes and Rollag1998, Reference Provencio, Rodriguez, Jiang, Hayes, Moreira and Rollag2000; Berson et al. Reference Berson, Dunn and Takao2002; Hattar et al. Reference Hattar, Liao, Takao, Berson and Yau2002).
Since the discovery of this new photoreceptor system in the eye, melanopsin-expressing intrinsically photosensitive retinal ganglion cells (variously referred to as ipRGCs, mRGCs or pRGCs) have been implicated in a wide range of different physiological and behavioural responses to light in mice. This includes regulating circadian rhythms, pupil constriction, sleep, mood, development and even pain (Schmidt et al. Reference Schmidt, Do, Dacey, Lucas, Hattar and Matynia2011; Lucas et al. Reference Lucas, Peirson, Berson, Brown, Cooper, Czeisler, Figueiro, Gamlin, Lockley, O’Hagan, Price, Provencio, Skene and Brainard2014; Do Reference Do2019). These cells have also been characterised in non-human primates (Dacey et al. Reference Dacey, Liao, Peterson, Robinson, Smith, Pokorny, Yau and Gamlin2005) and subsequently shown to be involved in regulating circadian and other non-visual responses in humans (Zaidi et al. Reference Zaidi, Hull, Peirson, Wulff, Aeschbach, Gooley, Brainard, Gregory-Evans, Rizzo, Czeisler, Foster, Moseley and Lockley2007; Lucas et al. Reference Lucas, Peirson, Berson, Brown, Cooper, Czeisler, Figueiro, Gamlin, Lockley, O’Hagan, Price, Provencio, Skene and Brainard2014; Mure et al. Reference Mure, Vinberg, Hanneken and Panda2019).
How can I account for visual and circadian responses to light?
Melanopsin ipRGCs have a different spectral response to vision with a peak sensitivity (λmax) around 480 nm. As these cells play a key role in regulating circadian and other non-visual responses to light, in 2014 an expert working group recommended using sensitivity functions based on melanopsin for such responses to light (Lucas et al. Reference Lucas, Peirson, Berson, Brown, Cooper, Czeisler, Figueiro, Gamlin, Lockley, O’Hagan, Price, Provencio, Skene and Brainard2014). However, rods and cones also contribute to these responses, as ipRGCs receive extrinsic light input in addition to their intrinsic photosensitivity. This was addressed by providing a sensitivity function for each of the five photoreceptors in the human eye – rods, red cones (L-cone), green cones (M-cone), blue cones (S-cone) and melanopsin in humans. By analogy to photopic lux, these are also referred to as rod-opic, L-cone opic, M-cone opic, S-cone opic and melanopic lux. These measures are collectively referred to as α-opic equivalent daylight illuminance (EDI), where α refers to the specific photoreceptor of interest.
This recommendation was accepted in a modified version by the International Commission on Illumination (CIE) as α-opic EDI (CIE 2018). The unit of these measurements is lux and the EDI component refers to how α-opic lux values are related to photopic lux. By definition, under daylight conditions, these photoreceptor-specific α-opic lux values are identical to photopic lux. For example, daylight of 2,000 photopic lux will give 2,000 lux for rod-opic, L-cone opic, M-cone opic, S-cone opic and melanopic lux. As such, measurements in a α-opic EDI can be thought of as how much daylight (in photopic lux) would be needed to give that much activation of each photoreceptor. A key advantage of this approach is that for any light source, the amount of light available for each photoreceptor in the eye can be determined, providing a much more accurate description of the light experienced for both visual and circadian responses.
Can I measure light for mice in the same way?
As the retinal photoreceptors of mice are well-characterised, it is possible to calculate similar α-opic EDI values for mice (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024). Based on the mouse retina, this gives four measurements – rod-opic, M-cone opic, S-cone opic and melanopic EDI. Conveniently, as the rods, M-cones and melanopsin have very similar peak sensitivity (λmax), these three values will usually be very similar. By contrast, the UV-sensitive S-cones in the mouse retina usually receive very little input from many artificial light sources optimised for human vision. As colour vision is based upon comparison of cone inputs (in mice, S-cones and M-cones), this means that a light source that looks white to a human is likely to look very different to a mouse, which may have consequences for any behavioural task where colour provides visual cues. This may affect the reproducibility of behavioural tests that rely upon visual cues.
Different light sources of the same intensity in photopic lux will also appear different to mice. As most artificial white light sources are optimised for human vision (peaking at 555 nm), mice will be less sensitive to such lights as they lack a long-wavelength L-cone. How much dimmer depends upon the type of light source and its spectral composition. The problem is amplified where coloured light is measured in photopic lux. For example, a theoretical study comparing the effects of 100 photopic lux of light from LEDs with an emission of 470 nm (blue) light versus 625 nm (red) may be equally bright to the visual system of humans. But to a mouse, the blue LED will be 3,430 times brighter based on melanopic sensitivity. Without considering the differences in sensitivity of the mouse visual and circadian system, such studies can come to erroneous conclusions. As nocturnal rodents, mice are photophobic – which is exploited in many behavioural tests of exploratory behaviour (Crawley Reference Crawley2008). Furthermore, circadian rhythms and acute light exposure regulate adrenal corticosterone levels (Ishida et al. Reference Ishida, Mutoh, Ueyama, Bando, Masubuchi, Nakahara, Tsujimoto and Okamura2005) as well as exerting effects on mood (LeGates et al. Reference LeGates, Fernandez and Hattar2014). Given these physiological and behavioural effects, accurately measuring light for mice has the potential to inform improvements to routine housing and husbandry as well as improving scientific reproducibility.
Finally, an advantage of this system is that it makes simulating daylight for mice simple – all you need to do is make these four α-opic EDI values identical. This offers further opportunities for refinement, as it enables researchers and animal care staff to optimise animal facility lighting to take the animal’s experience into account and provide more naturalistic lighting conditions.
What type of light meter should I use?
Most commonly available light meters measure light in photopic lux. As described above, this is based on human visual sensitivity to light (Vλ) rather than the visual or circadian photoreceptors of mice. There are three possible approaches to measuring light in mouse-relevant units:
Spectrophotometer
A spectrophotometer enables measurement of power at each wavelength across the spectrum, which can be used to calculate α-opic EDI values. This is the most accurate approach and is essential for those working in research on visual or circadian biology where light is the primary stimulus. This is also the only approach that will enable all four mouse α-opic EDI values to be determined, which is important when studying colour vision. This is particularly important where mouse UV-sensitive S-cones may contribute to physiological or behavioural responses. Examples of such spectrophotometers are available elsewhere (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024).
Melanopic EDI meter
Light meters that can measure melanopic EDI are becoming increasingly available due to the interest in circadian lighting for humans. Whilst mouse equivalents are not available, because human and mouse melanopsin ipRGCs have a similar sensitivity, human melanopic EDI will provide a good estimation of mouse melanopic EDI.
Lux measurement with conversion
In many cases a simple approximation may be made, based on measuring light in photopic lux and applying a correction. This depends upon the type of light source used. If the melanopic:photopic (M:P) lux ratio is known, this can be applied to photopic lux measurements. For example, if a warm white LED has a M:P ratio of 0.59, 100 photopic lux will give 59 melanopic lux. Some example white light sources and their M:P ratio are provided in Table 1. A quick comparison of this approach across a range of light intensities compared to a calibrated spectrophotometer was conducted in SNP’s lab. This gave results within 16% of accurately measured melanopic EDI values. This approach obviously cannot replace accurate measurement, but it enables those with limited resources or existing light meters to approximate the animal’s experience. It is also a helpful heuristic to interpret existing published data on the effects of light on mouse physiology and behaviour.
Examples of common laboratory white light sources and melanopic:photopic (M:P) ratios. The light source, a description of the type of lighting and an M:P ratio is provided. The M:P ratio can be used as a conversion factor between photopic lux and melanopic EDI lux
Table 1. Long description
The table consists of three columns: Light source, Description, and M P ratio.
* Daylight: Sun above the horizon with minimal cloud, D 65 series, 6,504 K. M P ratio is 1.00.
* Fluorescent: Modern F 11 fluorescent tube, including compact fluorescent bulbs, C C T 3,000 K. M P ratio is 0.53.
* Incandescent: Tungsten-halogen light bulbs, 2,856 K. Halogen light bulbs give broadly comparable spectrum. M P ratio is 0.44.
* Cool white L E D: White L E D, C C T 4,600 K. M P ratio is 0.65.
* Warm white L E D: White L E D, C C T 3,000 K. M P ratio is 0.59.
A note below the table states that color temperature is given for black body emission for daylight or incandescent light sources. For light sources that do not emit as a black body, correlated color temperature C C T is given.
Note: Colour temperature is given for black body emission for daylight or incandescent light sources. For light sources that do not emit as a black body, correlated colour temperature (CCT) is given.
What practical steps can I take to refine lighting conditions for mice?
Most guidelines for lighting conditions for laboratory mice are based on what the staff need to work and not the experience of the animals (Jennings et al. Reference Jennings, Batchelor, Brain, Dick, Elliott, Francis, Hubrecht, Hurst, Morton, Peters, Raymond, Sales, Sherwin and West1998). By using species-relevant units, such as mouse α-opic EDI, it is possible for the first time to account for the animal’s experience. For routine husbandry, melanopic EDI is the most appropriate unit. This will account for the circadian effects of light which will impact a wide range of physiology and behaviour. Moreover, as mouse melanopsin, rods and M-cones are spectrally similar, this will also account for many aspects of visual sensitivity over a wide range of light intensities.
For general animal husbandry, a stable light/dark cycle should be provided with daytime light > 10 lux melanopic EDI and night-time light < 0.1 lux melanopic EDI. This night-time light threshold of < 0.1 melanopic EDI can also be used as a threshold when using dim red light during the mouse’s dark phase. Critically, light should be considered from the animal’s perspective and therefore measured from within the cage, rather than at benchtop height or in the centre of the room. In addition, mice should be given the opportunity to escape or modulate their light exposure – for example, by providing an enclosed shelter or adequate nesting material to build an enclosed nest. In the future, it may be possible to further optimise animal facility lighting to simulate more natural lighting conditions (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024).
What type of light is optimal?
The short answer here is that we simply do not know as we could not measure this before! However, as all animals evolved under natural lighting conditions, daylight is an ideal reference point. The α-opic lux system simplifies simulating daylight to a mouse, as this just requires making each of the four α-opic EDI values the same. In the future, this provides opportunities of simulating more natural lighting conditions for laboratory mice.
Due to the ultraviolet sensitive S-cones in the mouse retina, it is often difficult to achieve significant S-cone opic activation with artificial light sources designed for human vision. A simple way of achieving this is to supplement standard white LEDs with violet light. Violet LEDs typically have a peak at 405 nm, which is enough to activate mouse S-cones without using potentially harmful ultraviolet light. In the future, alternative types of white LED may be able to accomplish this. White LEDs consist of a blue LED peak (450–470 nm) that stimulates phosphors creating a second longer wavelength peak (530–560 nm). The height of these two peaks affects how they appear to the human eye: If the blue LED peak is larger, these appear cool white, and if the second phosphor peak is larger, these appear as warmer yellow colours. Rather than using a blue LED in these light sources, violet-pumped LEDs are being developed that use a violet LED (405 nm) to activate the longer wavelength phosphors. In the future, such violet-pumped LEDs could be used to provide white light that can simultaneously simulate daylight for both humans and mice (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024).
What are the animal welfare and ethical issues?
There is already emerging evidence that the above recommendations for laboratory animal lighting can exert measurable effects on mouse behaviour. For example, by measuring light and locomotor activity within the home cage, the effects of cage position (and thus light levels) in modifying mouse activity patterns can be better understood. This work also showed that mice still can entrain their circadian activity rhythms even in red cages, which still transmit light > 0.1 lux melanopic EDI (Steel et al. Reference Steel, Tir, Tam, Bussell, Spitschan, Foster and Peirson2022). In studies where mice could avoid light by using a dark nest-box, animals reduced their daily light exposure to less than 1 h per day. And by providing gradual light/dark transitions to replicate twilight, this resulted in increased twilight (crepuscular) light sampling behaviour from nest-box to cage (Steel et al. Reference Steel, Tam, Brown, Foster and Peirson2024). Taking this nest-box paradigm further, simulating daylight and the spectral changes that occur at twilight resulted in changes in the activity onset of mice (Steel et al. Reference Steel, Hankins, Foster and Peirson2026). Together, these studies highlight how measuring light exposure based upon the animal’s experience can provide a greater understanding of mouse behaviour and welfare needs. The increasing accessibility of home cage monitoring, with artificial intelligence, should afford deeper insights into the effects of animal’s environments on their behaviour and welfare (Kizielewicz et al. Reference Kizielewicz, Molk, Camacho, Thuman and Beale2026).
Inappropriate light intensity is cited as a negative, unavoidable physical condition within the Five Domains model of animal welfare (Mellor et al. Reference Mellor, Beausoleil, Littlewood, McLean, McGreevy, Jones and Wilkins2020). Being unable to escape from light is therefore likely to be a significant stressor for the many millions of mice used in research and testing around the world every year. This is a highly significant welfare issue as well as a deep ethical concern due to the effects on reproducibility, which could mean that animals have been used – and wasted – in poor quality science. For example, time-of-day effects have been implicated in translational failures between rodent models and clinical trials in stroke (Esposito et al. Reference Esposito, Li, Park, Sencan, Guo, Shi, Lan, Lee, Hayakawa, Sakadzic, Ji and Lo2020).
A final application of species-relevant light measures could be to increase the standardisation in laboratory mouse studies. However, over-standardisation should be avoided, as this can lead to ‘local truths’ that hold only under very specific conditions, rather than greater reproducibility (Richter et al. Reference Richter, Garner and Wurbel2009). Instead, α-opic EDI measurements can be used to avoid systematic biases in experimental design. For example, by counterbalancing control and experimental groups so that both are exposed to a similar range of light levels, it is then possible to analyse data to determine how light influences different experimental outcomes. This should lead to increased benefits from more translatable research, which represents a more ethical use of animals if a Non-Animal Method is not currently available.
Future opportunities
Understanding the world from the perspective of another species is incredibly challenging. There are dramatic differences between the visual systems of humans and mice, which pose a real challenge for appreciating how laboratory mice see and respond to their environment. However, as described in this article, measuring light for mice based upon human brightness perception is clearly inappropriate. As the understanding of mouse visual and circadian systems has increased, it is now possible to use lighting measures based upon the sensory abilities of mice and other species (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024). This provides new opportunities to assess how light affects both laboratory mouse science and welfare, and this approach should be considered during facility design or light replacement within existing facilities. For example, from a scientific perspective, many behavioural tasks involve visual cues or use exposure to light to study exploratory or avoidance behaviour (Crawley Reference Crawley2008). Accounting for the animal’s experience of light exposure during such tasks should help understand sources of variability in such tests, increasing reproducibility. As described above, such data could be collected as part of ongoing studies by incorporating differing light levels in the experimental design, whilst ensuring that these are comparable between experimental groups.
Our understanding of the lighting preferences of laboratory mice is still very limited. As a nocturnal burrowing rodent, routinely housing mice under 12 h of light per day with a limited ability to avoid this exposure is likely to be detrimental to welfare. However, there is always a balance between monitoring laboratory animal welfare vs disturbing the animal’s normal behaviour. As mice are nocturnal, daily welfare checks and most experiments are conducted in the light during the human working day, when the animal should be sleeping. As such, it has been suggested that lighting cycles could be reversed to study mice in their dark phase when they are normally awake or otherwise altered to reduce the conflict between the needs of humans and mice (McLennan & Taylor-Jeffs Reference McLennan and Taylor-Jeffs2004; Hawkins & Golledge Reference Hawkins and Golledge2018). But studying, or caring for, mice at any time of day require the researchers and animal care staff to be able to see, so red light is typically used, based upon the common misconception that mice cannot see red light. As mice are simply less sensitive to longer wavelength light, if this red light is bright, this will still evoke responses (Peirson et al. Reference Peirson, Brown, Pothecary, Benson and Fisk2018). Based upon thresholds for mouse circadian responses, such light should be < 0.1 melanopic EDI, which can be readily achieved using longer wavelength (680 nm) red LEDs whilst providing sufficient light for human vision (Lucas et al. Reference Lucas, Allen, Brainard, Brown, Dauchy, Didikoglu, Do, Gaskill, Hattar, Hawkins, Hut, McDowell, Nelson, Prins, Schmidt, Takahashi, Verma, Voikar, Wells and Peirson2024). However, it should also be recognised that dim red lighting may be suboptimal for performing detailed welfare checks, for example, post-operative checks. Table 2 summarises some of the action points relevant to research, animal technologists and care staff, journal editors, regulators and Animal Ethics Committees or similar bodies.
Action points, based on current understanding of the mouse’s experience of light. Here we provide a list of common issues in rodent husbandry and research and recommended actions based on recent guidelines on rodent visual and circadian photobiology
Table 2. Long description
The table consists of four main rows under the headers Issue and Recommended actions.
* Row 1: The issue is that mice avoid light and have minimal exposure in the wild. Recommended actions include providing nesting material and boxes for protection, and considering reversing light-dark phases to minimize disturbance during sleep.
* Row 2: The issue is that mice experience light differently from humans. Recommended actions include using light meters to record melanopic E D I from within cages, and setting a stable cycle with daytime light above 10 lux melanopic E D I and nighttime light below 0.1 lux melanopic E D I.
* Row 3: The issue is that mice can see red light but are less sensitive to it. Recommended actions include keeping dim red light below 0.1 melanopic E D I using 680 nanometer red L E Ds, and performing detailed welfare checks during the light phase if brighter light is required.
* Row 4: The issue is that more work is needed to understand light quality impacts. Recommended actions include correlating light quality with behavior as part of husbandry, treating light intensity as a variable in research, and reporting spectrum and timing in publications according to P R E P A R E and A R R I V E guidelines.
1 norecopa.no/prepare
2 arriveguidelines.org
In conclusion, light exerts profound effects on the physiology and behaviour of laboratory mice, but this has often been poorly accounted for. This is because most lighting guidelines and research papers report light in units based upon human sensitivity to light, which ignores the differences between human and mouse sensory physiology. Due to new guidelines for the measurement of light for animals – based on their known visual and circadian responses – it is now possible to measure light as experienced by laboratory mice. This provides many new opportunities for refinement in terms of scientific research, routine husbandry and animal welfare. The future is bright – and now we can really ask how mice may prefer it themselves.
Acknowledgements
We would like to thank the whole expert working group who contributed to developing the methods described herein, specifically: Robert Lucas, Annette Allen, George Brainard, Timothy Brown, Robert Dauchy, Altug Didikoglu, Michael Tri Do, Brianna Gaskill, Samer Hattar, Roelof Hut, Richard McDowell, Randy Nelson, Jan-Bas Prins, Tiffany Schmidt, Joseph Takahashi, Vandana Verma and Sara Wells. The authors would also like to thank James Bussell (University of Oxford) for feedback on the manuscript. The original workshop was supported by awards from UFAW (11-22/23), the Committee International de l-Eclairage, NASA Ames Research Center, and the University of Manchester. SNP is funded by the BBSRC (BB/X002357/1) and Wellcome (227093/Z/23/Z).
Competing interests
None.
Open access
