Abstract
The mammalian circadian system regulates all biological processes, thereby ensuring optimal function at the appropriate times of day. Animal studies that examine neurobehavioral processes at different times of day, including during the animal's active phase, may provide important new biomedical insights. A logistical problem for the study of nocturnal laboratory rodents is the potential confounding influence of nighttime light exposure, which may cause circadian disruption and alteration of behavior. The historical solution has been to use red light illumination, which is widely believed to be undetected by the rodent visual system. However, some recent studies have questioned this belief. We, therefore, tested the effects of nighttime exposure to commonly used red light conditions on the circadian non-image forming and the image forming visual systems of female and male laboratory rodents. We found that brief dim red light exposure to a range of red light wavelengths produces strong activation of the suprachiasmatic nucleus master clock, rapid suppression of melatonin secretion, and a subsequent phase shift in daily activity onsets. We also found in an operant behavioral task that rats are able to detect long wavelengths of red light, but not near-infrared light. Thus, both the non-image and image forming visual systems of laboratory rodents are responsive to red light conditions that are often used in animal research. The use of red light for laboratory rodent research and animal care should be carefully considered in terms of its possible confounding influences on research objectives.
Significance Statement
There is scientific value in studying nocturnal laboratory rodents during the night, when they are normally awake and active. A challenge for these studies is the potential confounding influence of nighttime light exposure, which may cause circadian disruption and alteration of behavior. Often red light illumination is used for these studies, as it is widely believed to be undetected by the rodent visual system. However, we find through a series of neural, physiological, and behavioral studies that both the circadian non-image forming and image forming visual systems of laboratory rodents are sensitive to commonly used dim red light conditions. Use of red light for laboratory rodent research should be carefully assessed in terms of possible confounding influences on research objectives.
Introduction
The circadian system coordinates the appropriate time of day for operation of all biological processes, thereby ensuring optimal function (Bass and Lazar, 2016; Koronowski and Sassone-Corsi, 2021). The majority of biomedical mammalian animal research (>90%) uses nocturnal laboratory rats and mice (Grimm, 2021). Despite the nighttime active phase of these animals, most research using laboratory rodents is performed during the day (Nelson et al., 2022) when these animals are inactive and sleeping (Mannino et al., 2024). However, studies that have compared various measures in laboratory rodents across their normal active and inactive phases have often found important time of day differences in physiology and behavior (Matveyenko, 2018; Woodruff et al., 2018; Nelson et al., 2021, 2022; Hartsock et al., 2024). Further study of laboratory rodents during their nighttime active phase may reveal new mechanisms that support optimal biological function, and these studies may better generalize to human daytime physiology and neurobehavioral relationships (Peirson and Foster, 2011).
Conducting nighttime studies, however, raises an important logistic issue: how to perform manipulations and measures during the animal's dark phase without introducing circadian disruption because of nighttime light exposure (Fonken and Nelson, 2014)? It is widely believed that researchers can safely perform these dark phase manipulations and measures under dim red light conditions because of an assumption that the visual system of laboratory rodents is insensitive to longer wavelengths of light that fall within the red light range (∼620–750 nm), whereas the human visual system is able to function adequately under these red light conditions (Jennings et al., 1998). This assumption is primarily based on the different complement of photoreceptors found in the retina of laboratory rodents compared with humans. Specifically, rodents lack a third type of cone (L-cone) found in human retinas that is sensitive to longer wavelengths of light (Deegan and Jacobs, 1993; Zhang et al., 2019). Recently, however, one study found surprisingly strong electroretinogram responses of anesthetized rats to flashes of red light (Niklaus et al., 2020). Another study found that rats were able to perform a two-dimensional visual form discrimination task with illumination from red light sources (Nikbakht and Diamond, 2021). Whether the circadian non-image forming visual system is sensitive to red light was not addressed in these studies.
The circadian master clock, located in the suprachiasmatic nucleus of the hypothalamus (SCN), is a primary component of the non-image forming visual system. The SCN receives photic information about daily ambient light levels via direct projections from intrinsically photoreceptive retinal ganglion cells (ipRGCs) in the retina (Gooley et al., 2001; Hattar et al., 2002). The SCN master clock uses this information to align its internal clock phase with the environmental light/dark cycle. When aligned, neuronal activity of the SCN is high during the day and low during the night (Meijer et al., 1998). But the SCN is very sensitive to the “unexpected” presence of light at night, which induces rapid neuronal activation that sets into motion molecular and cellular changes that initiate a phase shift in its function (Shigeyoshi et al., 1997). The SCN also provides daily timing signals that entrain local cellular clocks throughout the body (Mohawk et al., 2012; Koronowski and Sassone-Corsi, 2021). Thus, nighttime activation of the SCN by red light is of concern because of the possible circadian disruption and misalignment it may produce.
In this study we examined the extent to which commonly used red light conditions can affect the non-image and image-forming visual systems of laboratory rodents in such a way that may compromise research study objectives. We found a high sensitivity of the SCN of laboratory rodents to activation by brief dark phase dim red light exposure. This exposure also produced a rapid suppression of nighttime melatonin secretion and a phase shift in daily activity onset. Rats were also able to perform a novel go/no-go operant task by detecting flashes of long wavelengths of red light, but not near-infrared light.
Materials and Methods
Animals
Specific pathogen-free Sprague Dawley rats were obtained from a commercial vendor (Inotiv) and C57BL/6J mice were bred in house. Animals were young adults at time of use, and they were housed within an AAALAC-accredited animal facility at the University of Colorado Boulder. Animals were housed separately by both sex and species in lightproof and sound-attenuated vivarium rooms (2.1 m wide × 1.75 m deep × 2.5 m high, 20–22°C, 48–52% humidity) arranged within a suite of six interior vivarium rooms and an adjacent necropsy workstation. No other animals were housed in the same vivarium room with the experimental animals. Each room permits complete control over light/dark conditions. Animals were acclimated to the vivarium rooms for at least 2 weeks before testing. Rats (two per cage) and mice (three per cage) were housed in translucent polysulfone rat cages (26.7 cm wide × 48.3 cm long × 20.3 cm high) or mouse cages (19 × 29.8 × 12.7 cm) with a stainless steel wirebar grid top and a polysulfone microisolator lid with spunbound polyester filter (Allentown Caging Equipment). The floor of each cage was covered with wood chip bedding material (Teklad Sani-chip; Inotiv). Animals were given ad libitum food pellets (Inotiv Teklad 2918 rodent chow; Inotiv) and reverse osmosis chlorinated water in drinking pouches that were placed on the home cage steel wirebar grid top.
Experimental design
Dark phase home cage light pulse tests
Light pulse tests were administered to animals residing in their home cage and home vivarium room. For each of these tests, cages were arranged side by side on a single shelf of a stainless steel wire rack, ∼110 cm above the ground. With one exception, prior to each light pulse test, animals were maintained on a shifted 12 h light/dark cycle (lights on at 11 P.M., lights off at 11 A.M.). Rats that were given the 658 nm light pulse test were maintained on a standard 12 h light/dark cycle (lights on at 6 A.M., lights off at 6 P.M.). White or red light pulses were administered during the dark phase at zeitgeber time 16 (ZT16; 4 h after lights out). Initiation and termination of a light pulse required an experimenter to momentarily enter a vivarium room and switch on or off the test light source. All lights were turned off in the outer portion of the vivarium suite whenever the door was opened to one of the interior rooms during the dark phase testing, ensuring no extraneous light exposure during the test session. For the control (no-light pulse) condition, the experimenter also momentarily entered the interior room at the appropriate times but did not switch on a light, thereby maintaining complete darkness for animals in that room. In all cases, the white light pulse condition consisted of switching on the standard vivarium installed overhead white light emitting diode (LED) light fixture (array of LEDs that emit ∼3,100 K white light; 54.8 µW/cm2; 130.7 lux within cage; New Star Lighting; Table 1). Animals were killed 30 min after the onset of a light pulse; a time interval that corresponds to the time of peak induction of Per1 and c-Fos mRNA after acute neuronal activation (Kornhauser et al., 1990; Pace et al., 2009; Roszkowski et al., 2016). Animals were rapidly removed from the vivarium room and killed at the nearby necropsy station within 1 min. Rats were killed by guillotine decapitation. Mice were briefly anesthetized (∼45 s) in a small chamber with isoflurane and then decapitated with sharp scissors. Brains were removed and snap frozen in a bath of isopentane chilled to −25 to −35°C with dry ice. Trunk blood was collected for plasma melatonin and corticosterone hormone measures. Brains and plasma were stored in a −70°C freezer until time of processing. The necropsy procedures were performed under the same vivarium-installed overhead red LED light condition described below for the 634 nm light pulse test.
Light pulse stimuli parameters
634 nm light pulse, time duration test
Forty-two male (n = 6) and 36 female (n = 5–6) rats were separated into seven different experimental groups. A control group received no light pulse as described above. Three groups received either a 1, 5, or 15 min exposure to the overhead vivarium installed white LED light, as described above. The other three groups received either a 1, 5, or 15 min exposure to red LED light. The red LED light was emitted from an array of narrowband red LEDs (peak 634 nm; average intensity inside the cage, 2.4 µW/cm2; estimated 0.004 lux rodent melanopic EDI; and 0.028 lux rodent rhodopic EDI) installed in the same overhead light fixtures as the white LED light array (New Star Lighting; Table 1, Fig. 1A). Male and female rats were tested as separate cohorts, but under identical conditions.
Spectrographs of experimental red and near-infrared light stimuli. Relative wavelength intensity profiles are displayed for red and near-infrared light stimuli used for nighttime light pulse experiments (panel A) and for the go/no-go nosepoke operant behavioral tests (panel B). Individual light stimuli are color coded with corresponding labels and peak height wavelength depicted on each graph.
658 nm light pulse
Twelve male rats (n = 4) and nine male and nine female C57BL/6J mice (three male and three female mice per group) were separated into three experimental groups. One group received no light pulse, another group received a 15 min white light pulse (see above), and the other group received a 15 min red LED light pulse. The red LED light was emitted from two commercial red LED headlamps (peak 658 nm; average intensity inside the cage, 19.9 µW/cm2; estimated 0.006 lux rodent melanopic EDI; and 0.038 lux rodent rhodopic EDI; Zebralight) positioned at the same level and facing the animal cages at a distance of 75 cm (Table 1, Fig. 1A).
734 nm light pulse
Male and female rats were separated into three experimental groups (six male and six female per group). One group received no light pulse, one group received a 15 min white light pulse (see above), and one group received a 15 min red LED light pulse. The red LED light was emitted from two lamps fabricated in-house into a configuration that could be used as a headlamp consisting of six LEDs in a 1.5 cm circular diameter array (peak 734 nm; average intensity inside the cage, 51.2 µW/cm2; estimated 0.0001 lux rodent melanopic EDI; and 0.0006 lux rodent rhodopic EDI) positioned at the same level and facing the animal cages at a distance of 38 cm (Table 1, Fig. 1A). Even at this high intensity and close proximity to the cage, this long wavelength red light did not permit adequate visual ability that would be sufficient for a valid in cage animal health check or experimental procedure.
935 nm infrared light pulse test in commercial behavioral test chambers
Twelve male rats were separated into two experimental groups (n = 6). One group received no light exposure, and the other group received 45 min of infrared light exposure. Testing occurred within commercial fear conditioning chambers (30.5 × 24.1 × 21 cm internal dimensions) housed inside lighttight boxes (59.7 × 31.8 × 71.1 cm internal dimensions; MED-VFC-OPTO-USB-R, Med Associates). The boxes included a near-infrared light source installed in the ceiling of the box (NIR-100L2, Med Associates) which we measured to emit narrow band infrared light (peak 935 nm; average intensity, 107.5 µW/cm2 inside the test chamber; Table 1, Fig. 1A). Testing occurred at ZT16, and rats were transferred from their home cage to the test room in light-tight opaque plastic tubs with lids. Rats were placed into and removed from the test chambers while all lights were turned off in the test room.
Activity onset phase shift test
Six male and six female rats were housed separately by sex in our vivarium rooms. A stainless steel running wheel (33 cm diameter) was placed in each cage that was integrated with a stainless steel wirebar grid top which held food pellets and a water bottle (STARR Life Sciences). Cages were organized on two shelves (∼66 and 110 cm above the ground) of a stainless steel wire rack, positioned with the long side of the home cage tub facing out away from the wall. There was a light intensity difference between the cages located on the upper and lower shelves for the overhead white light (upper shelf average, 64.6 µW/cm2; lower shelf average, 39.2 µW/cm2) and overhead 634 nm red light stimuli (upper shelf average, 2.45 µW/cm2; lower shelf average, 1.54 µW/cm2). Running wheel revolutions were continuously recorded via VitalView Activity Data Acquisition Software (STARR Life Sciences). Rats were acclimated to the running wheels for a minimum of 1 month with a 12 h light/dark cycle. Rats were then placed in constant darkness for 14 d on two separate occasions, with a 2 week intervening period when they were re-entrained to a 12 h light/dark cycle. On one occasion they received a 15 min white light pulse at the estimated circadian time 16 (CT16; 4 h after average activity onset) on the fourth night and a 15 min 634 nm light pulse at CT16 on the ninth night. On the other occasion, they received a 15 min mock (no-light) light pulse at CT16 on the fourth night and a 15 min 734 nm light pulse at CT16 on the ninth night. The light pulse procedures were the same as those described above for the dark phase home cage light pulse tests. Daily cage checks and weekly cage changes were performed briefly with use of red head lamps during the middle of the estimated subjective day (CT5–7), a time of circadian day corresponding to the “silent period” of the activity phase-shift light response curve (Nelson, 2005). Phase shifts were determined by two methods. First, we used a pre- and postlight pulse regression method to determine the difference in predicted activity onset for the day after the light pulse based on the 4 d of activity onset prior to the light pulse and the 4 d of activity onset after the light pulse (Daan and Pittendrigh, 1976). Second, we computed the difference between the average daily activity onsets for the 4 d prior to the light pulse and the 4 d after the light pulse. Both methods produced similar statistical results.
Operant go/no-go nosepoke response to light onset detection
A go/no-go operant nosepoke behavioral response task was designed to assess the ability of rats to accurately detect a brief flash of a light stimulus. For the task, rats had to learn to hold their nose in a nosepoke port until a flash of light occurred. If they then rapidly removed their nose from the nosepoke port, they would receive a sucrose pellet. Rats, therefore, initiated a test trial by placing and holding their nose in a nosepoke port. On most trials, a brief flash of light would occur at a random interval of time after the nosepoke onset. Six male rats were trained and tested in a custom computer controlled nosepoke behavior box for 30 min daily sessions during the rat's inactive phase (12 h light/dark, training/testing between ZT4-ZT9). Nosepoke boxes were rectangular chambers (26.7 × 25.7 × 27.9 cm) constructed of Plexiglas walls, with a stainless-steel top and stainless-steel perforated floor, with an underlying removable tray for collection of fecal boli and urine. Located on one wall of these chambers was a nosepoke port situated 1 cm beneath a white glass diffuser window (Edmund Optics #37975) to ensure that all rats viewed similar light stimuli intensity while engaging in a nosepoke trial. For all stimuli, light was scattered evenly over the diffuser windowpane (1.5 cm diameter). A food port was located on the opposite wall. The nosepoke and food ports were outfitted with an infrared light source and detectors that allowed for detection of rat nose entry into the port. A servo-actuated carousel was mounted on the external side of the diffuser window which rotated different LED light sources (white LED; red narrow band LEDs 660, 700,730, 850 nm) and red monochromatic semiconductor laser diodes (635, 670, 690 nm; RPMC Lasers) into alignment with the diffuser window for the appropriate trial condition. Lasers are an excellent light source for this study because they generate a single wavelength, but LEDs were also used to probe sections of the spectrum for which lasers are unavailable (Table 2, Fig. 1B). An infrared camera and near-infrared light source were mounted in the center of the top of the chamber for observing and recording behavior. The test chambers were housed inside of a lighttight, sound-attenuated outer shell (61 × 45.7 × 45.7 cm) made of medium-density fiberboard. The outer shell also contained a computer-controlled sugar pellet dispenser, small electronic interface box (Arduino), and air fan.
Nose poke light stimuli parameters
The general sequence of a nosepoke trial began when a rat placed their nose in the nosepoke port. After a variable hold time, a light stimulus would flash on the diffuser windowpane above the nosepoke port. If the rat then withdrew their nose from the port within a short period of time after exposure to the light stimulus, referred to as the “hit window,” the rat was rewarded with a sucrose pellet, resulting in a “hit” or true positive. If the rat remained in the nosepoke until after the hit window had ended, the trial was counted as a “miss” or false negative. Certain trials (10% of all trials) were designated as “catch trials” in which no light stimulus was shown, to measure false alarm rate. If, by chance, a rat removed their nose within the randomly set hit window of a catch trial, this was scored as a “false alarm” or false positive. If a rat remained in the nosepoke port until after the hit window had passed, awaiting a detectable light exposure, this would result in a “correct rejection” or true negative.
For training and testing, rats were food restricted (90–95% of baseline weight) to increase the salience of the sucrose reward. Using the white light stimulus only, rats underwent initial “nosepoke shaping” followed by a series of “hold training” stages, where nosepoke behavior was made more precise and robust through incremental decreases of the duration of the hit window (0.75–0.6 s), incremental increases in the hold time requirement (0.1–20 s), and a reduction in light stimulus duration (100 ms). After several weeks of behavioral shaping and training, rats could reliably and accurately respond to exposure to the white light stimulus. Thus, rats learned to hold their nose in the nosepoke port for up to 20 s on each trial and to withdraw their nosepoke within 600 msec after the onset of a 100 msec flash of white light.
We then proceeded to the test stages. In each test stage, a single red light stimulus was tested against the white light stimulus. Animals were tested on each stage for ∼10 d. In all test stages the hit window was fixed at 0.6 s, the hold time requirement was set to 2–20 s (pseudorandomly varied for each trial), and the light stimulus duration was fixed at 100 ms. We tested our red light stimuli in the following order: 635 nm laser, 660 nm LED, 700 nm LED, 670 nm laser, 730 nm LED, 685 nm laser, 850 nm LED. For red light, each animal's hit rate and sensitivity were averaged across every training session within a given stage. For white light, each animal's hit rate and sensitivity were averaged across all testing stages. Data were analyzed in two ways, first as the average hit rate for a given test light stimulus and second with a signal detection method that factored in the average performance on catch trials. For the signal detection method, the parameter D-prime was computed that consists of the z score of the false alarm rate subtracted from the z score of the hit rate for each light stimulus.
Double-label fluorescent in situ hybridization and quantification
Coronal brain sections (12 µm) at the rostral-caudal level of the SCN were cut on a crytostat (Leica model 1850; Leica Biosystems), thaw-mounted onto electrostatic charged Colorfrost plus microscope slides (series of six slides with six sections per slide), and stored at −70°C. Brain sections were processed for Per1 mRNA and c-Fos mRNA double-label fluorescent in situ according to a previously published procedure (Ravenel et al., 2024). Brain sections were hybridized overnight at 55°C in hybridization buffer containing both a digoxigenin-11-UTP (Roche #11 209 256 910) labeled riboprobe against Per1 mRNA and a flourescein-12-UTP (Roche #11 427 857 910) labeled riboprobe against c-Fos mRNA. The riboprobe for Per1 mRNA was generated from a 574 bp cDNA template that corresponds to nuclear transcript 974–1547 (GenBank accession no. NM_001034125; Girotti et al., 2009). The riboprobe for c-Fos mRNA was generated from a 680 bp cDNA template that corresponds to nuclear transcript 596–1171 (Accession # X06769.1; courtesy of Dr. T. Curran, St Jude Children's Hospital). Fluorescent labeling of the hybridized Per1 and c-Fos probes was carried out using the TSA-Plus Cyanine 3/Fluorescein System (Perkin-Elmer #NEL753). Slides were coverslipped with ProLong Gold Antifade Mountant with DAPI (Invitrogen #P36931).
In the SCN, Per1 and c-Fos genes are rapidly induced in response to nighttime light exposure in nocturnal rodents (Kornhauser et al., 1990; Masana et al., 1996; Shigeyoshi et al., 1997). The average number of Per1 mRNA and c-Fos mRNA-positive cells within the SCN of tissue sections from each brain was determined as an indirect measure of SCN neuronal activation in response to a nighttime light pulse condition. Tiled three-channel images of regions of tissue sections were collected with 10× objective using high efficiency red, green, and blue fluorescent filter sets to image Per1 mRNA, c-Fos mRNA, and DAPI, respectively (Zeiss Axio Imager M1 epifluorescent microscope, Axiocam 305 monochrome camera, and ZEN software; Carl Zeiss Microscopy). Exposure times for each channel were held constant for all image acquisition. Cell counting was performed using QuPath open-source software for bioimage analysis (Scientific Reports. https://doi.org/10.1038/s41598-017-17204-5). A hand-drawn region of interest was outlined on captured images corresponding to the rat SCN (approximately −0.92 to −1.40 mm relative to bregma; Paxinos and Watson, 2007) or the mouse SCN (approximately −0.46 to −0.82 mm; Paxinos and Franklin, 2001), using proximal anatomical landmarks of the optic chiasm and third ventricle. Positive cells for each channel were counted using the QuPath particle detection feature while holding constant detection parameters for each image. Cell counts were performed separately for each hemisphere on 2–3 sections per brain, yielding 4–6 separate counts per brain. Cell counts were averaged for each brain and expressed as number of cells per region of interest area (cells/mm2).
Melatonin and corticosterone plasma hormone measures
Commercial enzyme immunoassay kits were used to measure plasma levels of melatonin (cat# MEL31-K01; Eagle Biosciences) and corticosterone (catalog #K014; Arbor Assays). For the corticosterone assay, plasma samples were first diluted 1:50 in assay buffer and incubated at 65°C for 60 min in order to denature corticosteroid binding globulin. Samples were run in duplicate, with all samples from the same experiment run in at least one assay, with treatment groups counterbalanced across assay plates. Within and between assay coefficients of variability were 7.8 and 6.5%, respectively, for melatonin, and 5.6 and 17.4%, respectively, for corticosterone.
Light measures
The spectral emission profiles (between 500 and 1,000 nm) of each LED and laser light stimuli used in this study were measured by a broad range spectrometer instrument (either model LR1 V2.1-B, ASEQ Instruments; or model CCS175, Thorlabs; Tables 1, 2; Fig. 1). The average intensities (µW/cm2) of light stimuli were measured with a radiometer (model ILT2400 radiometer with model SED033/F/W detector, International Light Technologies). The spectral power distribution of the white light and red light stimuli used for the light pulse studies was measured with a spectroradiometer (model ILT350, International Light Technologies). Light intensity and spectral power distribution measures were performed under experimental conditions with the radiometer detector placed inside a representative animal cage. For comparison purposes, we also used a light meter, similar to the ones widely used in laboratory and animal facilities, to measure illuminance (lux; model LT300, Extech Instruments). It should be noted, however, that standard lux meters are not designed to accurately measure light intensity in the red light range (Lucas et al., 2014; Dauchy and Blask, 2023). For example, our lux meter measurements, in contrast to the radiometric measurements, substantially underestimated the relative light intensity of longer wavelengths of red light compared with white light (Table 1). We include measures of illuminance (lux) for comparison purposes as this measure is widely used for light intensity measures in laboratory and animal facilities. In addition, for the light pulse stimuli, we used recently available online calculator tools (Lucas et al., 2024) to estimate the rodent α-opic equivalent daylight illuminance (EDI) for each of the red light relevant rodent retinal photopigments: melanopsin, rhodopsin, and M-cone opsin (Table 1). The α-opic EDI provides a standardized metric for reporting experimental light effective intensities expected to be present at the retinal level for each photoreceptor pigment present in that species (Lucas et al., 2014, 2024; Lucas and Peirson, 2024; McDowell et al., 2024).
Statistical analysis
Statistical analysis of data was performed using the software packages IBM SPSS Statistics 29.0.2.0 (IBM) and GraphPad Prism 10.4.0 (GraphPad Software). Analysis of variance (ANOVA) was used to determine statistical significance for experimental factors in each experiment with more than two comparison groups. For the 634 nm light pulse test, a one-way ANOVA followed by Dunnett’s t test was used to compare each light pulse condition with the no-light pulse control group. This was followed by analysis of each of the light pulse conditions with a 2 (light type) × 3 (light pulse duration) × 2 (sex) ANOVA and with Tukey's post hoc test for individual pairwise group comparisons. For the 658 and 734 nm light pulse experiments, an initial evaluation with a 3 (light pulse condition) × 2 (sex) ANOVA found no effect of sex or sex by light pulse condition interaction. Thus, statistical results presented are for one-way ANOVA (collapsing across sex), with Tukey's post hoc test for individual pairwise group comparisons. For the activity onset phase shift test, an initial 4 (light pulse condition) × 2 (sex) mixed model ANOVA was examined, with light pulse condition as a within groups factor and sex as a between groups factor. Since there was no effect of sex or a sex by light pulse condition interaction, the statistical results presented are for one-way repeated-measures ANOVA (collapsing across sex), with Bonferroni’s correction for post hoc dependent-groups pairwise comparisons. For nosepoke behavior, each stage of testing was treated as an independent test, and data were analyzed by one-way ANOVA, with Tukey's post hoc test for individual pairwise group comparisons. Student's t test was used for the near-infrared light pulse study. For all statistical tests, the alpha level was 0.05. Data graphs depict treatment group bars (mean + SEM), with superimposed individual subject data points.
Results
Brief nighttime exposure (1–15 min) to dim red light (634 nm) activates the SCN and rapidly suppresses melatonin secretion
We first tested whether brief dim nighttime red light exposure to a commercial overhead red LED light fixture would activate the SCN of female and male rats that were otherwise undisturbed within their home cages. These light fixtures were installed in our animal research vivarium rooms during a facility remodel. Some commercial light fixture suppliers advertise that illumination by red LEDs is ideal for use in animal facilities since red light “is invisible to rodents” and therefore will not disrupt activity cycles. These light fixtures contain an array of white LEDs (∼3,100°K white light; 54.8 µW/cm2, 130.7 lux within cage) for standard daytime illumination and a separate array of red LEDs (peak wavelength 633.6 nm; 2.4 µW/cm2, 4.7 lux within cage) which can be independently switched on and off (Table 1, Fig. 1A). The red light intensity level within the cage, as measured with a radiometer (2.4 µW/cm2) or lux meter (4.7 lux), is widely considered to be “dim” red light (Fonken and Nelson, 2014; Dauchy et al., 2015).
Rats were maintained on a 12 h light/dark cycle, and we tested the effect of 1, 5, or 15 min exposure to white or red light stimuli at ZT16 (4 h after lights out). Control rats received no nighttime light pulse. These are light exposure conditions that are likely to occur when experimenters or animal care personnel perform brief tasks in animal or procedure rooms during the animal's dark phase. Neuronal activation of the SCN was assessed by rapid Per1 and c-Fos gene induction present 30 min after the onset of the light pulse. There was minimal Per1 and c-Fos mRNA expression in the SCN of control brains at this time of day, whereas all light pulse conditions produced large increases in Per1 and c-Fos mRNA (Dunnett's multiple-comparisons test, p < 0.05; Fig. 2A–C). We found no significant differences in the response to red light versus white light and no sex differences in response to the light pulses. There was a main effect of light exposure duration (Per1: F(2,50) = 4.04, p = 0.024; c-Fos: F(2,31) = 3.46, p = 0.044), with 15 min of exposure producing a greater response than 1 min of exposure when collapsing across type of light (Tukey’s post hoc test, p < 0.05). There were no significant interactions between light type, exposure duration, or sex.
Brief nighttime exposure (1–15 min) to dim red light (634 nm) activates the SCN and rapidly suppresses melatonin secretion. Nighttime (ZT16) exposure of female and male rats to 1, 5, or 15 min of white light (∼3,100 K, 54.8 µW/cm2) or red light (634 nm, 2.4 µW/cm2) produced a rapid increase in SCN Per1 mRNA (panel A) and c-Fos mRNA (panel B), and a rapid suppression of plasma melatonin (panel D), but no effect on plasma corticosterone (panel E). Brain tissue and blood samples were collected 30 min after light pulse onset. Bars show group mean + SEM, with superimposed data points for individual animals. Letters above the bars denote results of post hoc tests of group differences; groups that are not different from each other (p > 0.05) share the same letter. Panel C shows representative epifluorescent photomicrographs (10× objective) of SCN from rats receiving 5 min of white light exposure, 5 min of red light exposure, or no-light control exposure. Local anatomical landmarks are indicated: 3v, third ventricle; ox, optic chiasm, with white dashed line showing dorsal border. Scale bar, 100 µm.
We also measured plasma melatonin and corticosterone 30 min after the onset of nighttime white light or red light pulse exposure. All light pulse conditions produced rapid and comparable suppression of nighttime plasma melatonin levels compared with the no-light condition (Dunnett's multiple-comparisons test, p < 0.05; Fig. 2D). There was no effect of the light pulses on plasma corticosterone levels in either female or male rats (Fig. 2E), suggesting that the light pulses did not elicit a generalized stress response. As expected, there was a main effect of sex (F(1,63) = 15.1, p = 0.0002), with female rats having higher plasma corticosterone levels during the night than males (Spencer and Deak, 2017).
Brief nighttime exposure (15 min) to a longer wavelength of red light (658 nm) activates the SCN in rats and mice and suppresses melatonin secretion
We next tested whether 15 min nighttime exposure to a commercial-sourced red LED light with an emission wavelength longer than 630 nm could be safely used without activating the SCN of laboratory rats and a widely used strain of inbred mice (C57BL/6J). Red LED headlamps are often used by animal vivarium staff and researchers for brief health checks and experimental procedures, such as drug injections of laboratory rodents during their dark phase. These commercial headlamps typically have red LEDs with a wavelength in the 630 nm range. However, to test the sensitivity of the SCN response to a nighttime pulse (15 min) of a longer wavelength of red light, we found a commercial headlamp with a peak red LED wavelength of 658.3 nm (19.9 µW/cm2, 2.5 lux within cage; Table 1, Fig. 1A). Although this illumination produced a low lux reading (<3 lux) within the home cage, the radiometric intensity (19.9 µW/cm2) was considerably greater than that produced by the overhead 634 nm red LED light fixture (2.4 µW/cm2, 4.7 lux), as is to be expected for a commercial head lamp intended to produce red light emission of a longer wavelength that was bright enough for human use.
Again, we saw that this longer wavelength red light pulse produced a strong activation of the SCN in both rats (Per1: F(2,7) = 12.50, p = 0.005; c-Fos: F(2,7) = 36.32, p < 0.001; red and white light different from no-light condition, Tukey's post hoc test, p < 0.05; Fig. 3A,B) and mice (Per1: F(2,15) = 37.02, p < 0.001; c-Fos: F(2,15) = 13.17, p < 0.001; red and white light different from no-light condition, Tukey's post hoc test, p < 0.05; Fig. 3C,D). In the mouse, but not the rat, the red light pulse produced somewhat less activation of the SCN than the white light pulse (trend for c-Fos, p = 0.098, significant for Per1, p = 0.009, Tukey's post hoc test).
Brief nighttime exposure (15 min) to a longer wavelength of red light (658 nm) activates the SCN in rats and mice and suppresses melatonin secretion. Fifteen min exposure to white light (∼3,100 K, 54.8 µW/cm2) or red light (658 nm, 19.9 µW/cm2) at ZT16 produced a rapid increase in SCN Per1 mRNA and c-Fos mRNA in rats (panels A and B) and mice (panels C and D) and a rapid suppression of plasma melatonin in rats (panel A). Plasma melatonin was not detected in this strain of mouse (C57BL/J6) for any of the light conditions. There was no effect of light pulse on plasma corticosterone (panels A and C). Brain tissue and blood samples were collected 30 min after light pulse onset. Error bars show group mean + SEM, with superimposed data points for individual animals. Letters above the bars denote results of post hoc tests of group differences; groups that are not different from each other (p > 0.05) share the same letter. Representative photomicrographs of SCN from rats (panel B) or mice (panel D) receiving 15 min of white light exposure, 15 min of red light exposure, or no-light control exposure are shown. Local anatomical landmarks are indicated: 3v, third ventricle; ox, optic chiasm, with white dashed line showing dorsal border. Scale bar, 100 µm.
This 658 nm red light nighttime pulse also rapidly suppressed circulating melatonin levels in rats (F(2,9) = 6.22, p = 0.02; red and white light different from no-light condition, Tukey's post hoc test, p < 0.05; Fig. 3A). Many standard laboratory mouse strains, including C57BL/6J mice, are known to be deficient in circulating melatonin levels due to mutations in the gene encoding the enzyme hydroxyindole O-methyltransferase which is necessary for pinealocyte synthesis of melatonin (Kasahara et al., 2010). Consistent with this fact, melatonin levels were undetectable in all of our mouse plasma samples. There was no effect of this light pulse on plasma corticosterone levels in rats or mice.
Brief nighttime exposure (15 min) to far-red light (734 nm) activates the SCN but does not suppress melatonin secretion
To test the possible upper limit of a viable red light wavelength that could be safely used for nighttime procedures with laboratory rodents, we fabricated a far-red LED light source (peak wavelength 733.9 nm; 51.2 µW/cm2, 1.4 lux within cage; Table 1, Fig. 1A). Although this far-red light produced a relatively high radiometric intensity level of light exposure (51.2 µW/cm2 inside the animal cage), it provided very limited useful illumination for the experimenters. Nevertheless, 15 min nighttime exposure to this far-red light was again sufficient to produce strong activation of the SCN of female and male rats (Per1: F(2,29) = 49.82, p < 0.0001; c-Fos: F(2,29) = 35.70, p < 0.0001; red and white light different from no-light condition, Tukey's post hoc test, p < 0.001; Fig. 4A–C). In the case of c-Fos induction, the red light produced less activation than the white light (Tukey's post hoc test, p = 0.001; Fig. 4B). In contrast to the shorter wavelengths of red light that we tested, 15 min exposure to this far-red LED light did not produce a rapid suppression of circulating melatonin levels (Fig. 4D). There was an unexpected effect of light pulse condition on corticosterone levels. When collapsing across sex, red light exposure produced lower corticosterone levels compared with the white light condition (p = 0.015), but not compared with the no-light condition (Fig. 4E). The extent to which brief far-red light exposure has a suppressive effect on corticosterone secretion compared with white light exposure requires further study.
Brief nighttime exposure (15 min) to far-red light (734 nm) activates the SCN but does not suppress melatonin secretion. Nighttime (ZT16) exposure of female and male rats to 15 min of white light (∼3,100 K, 54.8 µW/cm2) or far-red light (734 nm, 51.2 µW/cm2) produced a rapid increase in SCN Per1 mRNA (panel A) and c-Fos mRNA (panel B), but it did not produce a rapid suppression of plasma melatonin (panel D). When collapsing across sex, there was a reduction of plasma corticosterone in response to red light compared with white light (panel E). Brain tissue and blood samples were collected 30 min after light pulse onset. Error bars show group mean + SEM, with superimposed data points for individual animals. Letters above the bars denote results of post hoc tests of group differences; groups that are not different from each other (p > 0.05) share the same letter. Representative photomicrographs of SCN from rats (panel C) receiving 15 min of white light exposure, 15 min of red light exposure, or no-light control exposure are shown. Local anatomical landmarks are indicated: 3v, third ventricle; ox, optic chiasm, with white dashed line showing dorsal border. Scale bar, 100 µm.
Nighttime exposure (45 min) to near-infrared light (935 nm) does not activate the SCN or suppress melatonin secretion
Near-infrared light illumination, while not typically visible to humans, is an effective strategy for providing high resolution infrared video camera images of rodent behavior in the dark. Near-infrared light sources are often provided in commercial lighttight behavioral test chambers or are used for home cage monitoring of rodent behavior in the dark. Given some reports of human perceptual responses and rodent retinal responses to infrared light with wavelengths as long as 1,000 nm (Palczewska et al., 2014), we tested the sensitivity of the SCN to 45 min (typical duration of some extended behavioral training sessions) of nighttime exposure to near-infrared LED light (peak wavelength 935.4 nm; 107.5 µW/cm2 in test chamber; Table 1, Fig. 1A). We found that this infrared light exposure did not induce Per1 mRNA or c-Fos mRNA in the SCN (mean number of positive cells ± SEM, Per1: no light = 5.5 ± 2.8 cells per mm2; infrared light = 10.1 ± 4.4 cells per mm2; c-Fos: no light = 2.9 ± 3.5 cells per mm2; infrared light = 8.7 ± 6.2 cells per mm2, t test, p > 0.05). Furthermore, infrared light exposure did not suppress plasma melatonin levels (mean ± SEM: no light = 170.0 ± 24.6 pg/ml; infrared light = 170.3 ± 51.6 pg/ml, t test, p > 0.05). Thus, near-infrared light illumination can be safely used with infrared cameras for nighttime behavioral procedures or home cage monitoring without producing circadian disruption.
Brief nighttime exposure (15 min) to dim 634 nm red light, but not 734 nm far-red light, causes a subsequent phase shift in activity rhythms
In order to test whether brief exposure to nighttime red light would produce a phase shift in subsequent daily activity onsets, we individually housed female and male rats in cages with running wheels. After stabilization of daily wheel running behavior under 12 h light/dark conditions, rats were switched to constant darkness housing conditions and were then given subsequent 15 min exposure to white LED light, 634 nm red LED light, 734 nm red LED light, or no light (control condition) in a repeated-measures design. Light exposures were presented at the estimated circadian time (CT) of CT16 (4 h after the average free running activity onsets). Both white light and 634 nm red light produced similar phase delays of ∼100 min (p < 0.05, Bonferroni’s correction), whereas 734 nm red light and control exposure (no light) produced no significant shift in activity phase onset (Fig. 5). Because wheel running cages were located on two rows of shelves in each room that differed in light intensity (see Materials and Methods), we also tested whether there was an effect of shelf location on the phase shift response. We found a significant interaction between light condition and shelf (F(3,30) = 4.72, p = 0.008) that was driven by a greater phase shift for rats on the upper shelf when exposed to the 635 nm red light compared with the white light, whereas for rats on the lower shelf, there was no difference in the magnitude of light shift when exposed to the 635 nm red light compared with the white light.
Brief nighttime exposure (15 min) to dim 634 nm red light, but not 734 nm far-red light, causes a subsequent phase shift in activity rhythms. Female and male rats housed in cages with running wheels were placed in constant darkness. In a repeated-measures design, rats were subsequently exposed to 15 min of white light (∼3,100K, 54.8 µW/cm2), red light (634 nm, 2.4 µW/cm2), far-red light (734 nm, 51.2 µW/cm2), or no-light at circadian time 16 (4 h after average activity onset, which was designated as CT12). White light exposure and 630 nm red light exposure but not 734 nm far-red light exposure caused a subsequent phase delay in activity onsets (panel A). Error bars show group mean + SEM, with superimposed data points for individual animals. Letters above the bars denote results of post hoc tests of group differences; groups that are not different from each other (p > 0.05) share the same letter. Panels B and C show representative actograms for one rat that in a first sequence of constant darkness testing, panel B received a white light pulse (white star) on Day 4 and a 634 nm red light pulse (red star) on Day 9, and in a separate sequence of constant darkness testing, panel C received no-light pulse (black star) on Day 4 and a 734 nm far-red light pulse (red star) on Day 9. Each row of the actogram shows 24 h of wheel running activity. Numbers on the left of each actogram indicate day within a test sequence and red dashed lines show the estimated line of best fit for the activity onsets for the 4 d before and 4 d after each light pulse test.
Rats can reliably perform a light detection operant go/no-go nosepoke task in response to long wavelengths of red light, but not to near-infrared light
Our SCN measures indicate that red light extending into the far-red range (734 nm) can activate the non-image forming visual system of laboratory rodents. To test whether long wavelengths of red light can be detected by the rodent image forming visual system, we examined whether male rats could learn to perform a go/no-go operant behavioral task that depended on detection of a brief flash of red light. We examined performance for a wide spectral range of red light stimuli that included LEDs with a relatively narrow bandwidth (660, 700, and 730 nm LEDs), as well as monochromatic lasers (635, 670, and 685 nm lasers; Table 2, Fig. 1B). We also tested nosepoke responses to a near-infrared LED light (850 nm). For this task, rats learned to place their nose in a nosepoke port and hold it there until they detected the onset of a brief flash of white or red light. If they withdrew their nose from the nosepoke within 600 ms after the onset of the 100 ms light flash, they then received a sucrose pellet reward. Over several weeks of initial training with the white light stimulus, rats became very adept at accurately performing this difficult timing task, typically receiving a sucrose pellet (hit rate) on ∼75% of trials (Fig. 6A; see Movie 1 for representative example with white light stimulus). Rats were able to perform nearly as well for each red light stimulus (Fig. 6A; see Movie 2 for a representative example with 730 nm red light stimulus), but they performed poorly with the near-infrared light stimulus of 850 nm (hit rate <10%). Thus, there was a main effect of light condition (F(7,40) = 17.11, p < 0.0001), with the 850 LED condition significantly different from each of the other conditions (Tukey's post hoc test, p < 0.05). Accurate light stimuli detection of the white light stimulus and red light stimuli was confirmed by D-prime signal detection analysis (F(7,40) = 17.9, p < 0.0001; followed by Tukey's post hoc test; Fig. 6B). Given the spectral properties of each of the red light stimuli used in this task (Table 2, Fig. 1B), we can confidently conclude that the image forming visual system of these rodents was able to detect a flash of red light with a wavelength longer than 685 nm and likely at least as long as 715 nm (the lower half-maximum wavelength range of the 734 nm LED stimulus).
Rats can reliably perform a light-detection operant go/no-go nosepoke task in response to long wavelengths of red light, but not to near-infrared light. Rats were trained in a go/no-go nosepoke operant behavior task to hold their nose in a nose-poke port until they detect a 100 ms flash of light. Rats received a sucrose pellet reward if they removed their nose from the nosepoke within 600 ms after the onset of the flash of light. The ability of rats to detect a flash of white light or one of a range of red light wavelengths (635–730 nm) or near-infrared light (850 nm) was tested across separate blocks of trials that each included white light trials interspersed with trials of one of the red light or near-infrared light stimuli. Rats were able to reliably detect white light flashes (Movie 1 for representative example with white light stimulus) and all red light flashes (Movie 2 for a representative example with 730 nm red light stimulus), but not the near-infrared light flash as determined by percent hit rate (panel A) or D-prime signal detection analysis (panel B). Error bars show group mean + SEM; superimposed individual data points represent the average performance for an individual rat across all trials with that specific light stimulus. Letters above the bars denote results of post hoc tests of light stimuli group differences; groups that are not different from each other (p > 0.05) share the same letter.
Representative example of a rat accurately performing go/no-go nosepoke responses to a flash of white light. [View online]
Representative example of a rat accurately performing go/no-go nosepoke response to a flash of 730 nm far-red light. [View online]
Discussion
The use of nocturnal laboratory rodents for biomedical research has led to revolutionary advances in understanding the biological basis of many aspects of human health and disease (Mukherjee et al., 2022; Chang and Grieder, 2024). Because the circadian system dynamically regulates all biological processes (Bass and Lazar, 2016; Koronowski and Sassone-Corsi, 2021; Nelson et al., 2021), it is important to study nocturnal rodents during both their inactive phase (daytime) and during their active phase (nighttime). However, a logistic challenge when studying these nocturnal animals is to ensure that routine experimental and husbandry practices do not introduce potentially confounding influences of circadian disruption that compromise scientific conclusions and limit important new discoveries.
Nighttime white light exposure is an especially problematic source of circadian disruption (Dauchy et al., 2010; Fonken and Nelson, 2014; Emmer et al., 2018). Although there is widespread belief that the visual system of laboratory rodents is insensitive to red light, in contrast to shorter wavelengths of light (Jennings et al., 1998), some recent studies have called into question that assumption (Niklaus et al., 2020; Nikbakht and Diamond, 2021). We therefore tested the circadian effects of nighttime exposure to red light conditions designed to simulate working conditions often used in research and animal husbandry procedures with laboratory rodents.
We began our investigation by studying the impact on laboratory rodents of nighttime short duration dim red light (634 nm) exposure. The source of the red light was a commercial overhead red LED light fixture that was installed in some of our animal facility vivarium and procedure rooms. We were surprised to find that as little as 1 min of dim red light exposure strongly activated the SCN and rapidly suppressed melatonin hormone secretion. We also found that 15 min nighttime exposure to this red light condition was sufficient to produce a subsequent phase delay in activity rhythms.
We then explored whether light sources that emit longer wavelengths of red light could instead be safely used for performing animal procedures without producing an alteration of laboratory rodent circadian function. It appears, however, that this strategy is not a good solution. Even brief exposure (15 min) to a far-red light (734 nm) produced strong activation of the SCN. Although that far-red light stimulation did not produce acute suppression of melatonin secretion or a subsequent daily activity phase shift, it provides very limited functional visibility for humans. For example, in our experience, the illumination is not sufficient for performing a daily health check or administering a drug injection despite having an overall radiometric intensity approximately equivalent to the overhead white lights (Table 1). There are some reports of retinal activation of humans and laboratory rodents to even longer wavelengths of light that extend into the infrared light range (Palczewska et al., 2014; Vinberg et al., 2019). We found, however, that exposure of rats to near-infrared light (935 nm) did not activate the SCN or suppress melatonin secretion, indicating that exposure to near-infrared lights can be safely used without producing circadian disruption.
We are not aware of other studies, besides ours, that have evaluated the effects of similar brief red-light exposure conditions on laboratory rodent SCN activation. However, other studies have found that daily nighttime exposure to dim red light can entrain daily activity rhythms in rats and mice (McCormack and Sontag, 1980; Butler and Silver, 2011). Other studies have also shown that red light exposure parameters similar to ours rapidly suppress melatonin secretion in rats (Honma et al., 1992; Dauchy et al., 2015). Moreover, Dauchy et al. (2015) found that daily exposure of rats to dim red light throughout the dark phase substantially suppresses daily melatonin secretion and alters a number of metabolic and physiological parameters important for health and well-being. The extent to which the altered metabolic and physiological measures observed in the Dauchy et al. study (Dauchy et al., 2015) were secondary to the dramatic reduction in daily melatonin levels is unknown. But if some of those alterations were melatonin dependent, then perhaps the physiological effects of daily nighttime red light exposure may differ for the widely used laboratory mouse strains, such as C57Bl/J6 mice, that lack significant melatonin production (Kasahara et al., 2010). The Dauchy et al. study (Dauchy et al., 2015) also found that daily nighttime red light exposure produces an overall blunting and phase shift in daily corticosterone secretion. In our study, we did not see an acute effect of nighttime white light or red light exposure on corticosterone hormone levels, suggesting that these light pulse experiences were not acutely stressful.
Our SCN measures indicate that red light extending into the far-red range can activate the SCN component of the non-image forming visual system of laboratory rodents. To determine if the rodent's image-forming visual system can also detect red light, we trained rats on an operant behavioral go/no-go task that depended on detection of a flash of light. We found that rats performed well on this task with each red light stimulus tested, but not with a near-infrared LED light. This spectral sensitivity is very similar to the results of a study by Nikbakht and Diamond (2021) which found that laboratory rats were able to perform a two-dimensional visual form discrimination task with illumination from red light LED sources but not a near-infrared light source. Our study demonstrates that rats can not only use red light illumination to perceive objects in their environment (Nikbakht and Diamond, 2021), but they can also behaviorally detect the occurrence of a flash of red light extending into the far-red range. The ability of laboratory rodents to visually perceive red light has implications for its presence during experimental conditions. For example, red light illumination may serve as spatial cues and if uncontrolled from one test to the next may produce variable behavioral results. There is also the concern that red indicator lights on experimental equipment could serve as an unwanted temporal cue that may confound interpretation of experimental results.
The results of our study raise the question of which rodent retinal photoreceptors may be responsible for transducing functional responses to far-red light. Although we did not perform direct retinal measures or manipulations in this study, we can make some inferences based on the known relationship between retinal photoreception physiology and non-image forming visual function (Schmidt et al., 2011). The SCN is directly informed about the presence of ambient light exclusively from the M1 class of melanopsin expressing ipRGCs (Güler et al., 2008; Schmidt et al., 2011). The photopigment melanopsin has peak absorption of short wavelength light (∼480 nm), and this short wavelength sensitivity has contributed to the misconception that the SCN is primarily sensitive to only shorter wavelengths of light, especially those in the blue light range. However, the ipRGCs receive input from rods and cones, and therefore ipRGC neuronal activity integrates information about retinal photic stimulation via all photoreceptor types. Although photoentrainment of the SCN in mice requires the presence of melanopsin ipRGCs (Güler et al., 2008), it does not require melanopsin expression (Altimus et al., 2010). Instead, photoentrainment depends primarily on the presence of rhodopsin in retinal rods (Altimus et al., 2010; Lall et al., 2010). Since rods have a similar spectral sensitivity in both rodents and humans (peak sensitivity ∼500 nm), the non-imaging forming visual system of laboratory rodents may have sensitivity to red light similar to that of humans.
The somewhat surprising result from our study is that brief nighttime exposure to narrow bandwidth red LED light was very effective at activating the SCN. This was observed even for red light in the far-red range (734 nm). Applying the recommended standardized α-opic equivalent daylight illuminance (EDI) metrics to our study (Lucas et al., 2024; Lucas and Peirson, 2024; McDowell et al., 2024), we find that our 734 nm red light stimulus, despite relatively high overall radiometric intensity within the cage, yields very low rodent melanopic and rhodopic lux EDI estimates (Table 1). These low EDI values are due to the long wavelength restricted nature of the spectral power distribution for this light stimulus. The SCN-dependent photoentrainment sensitivity to very low intensities of shorter wavelengths of light in the blue/green range has long been known, and some studies have reported photoentrainment of rats and mice to daily dim red light with wavelength and intensity parameters similar to the 634 and 658 nm red light pulse stimuli used in our study (McCormack and Sontag, 1980; Butler and Silver, 2011). Our results are also consistent with the guidance of Peirson et al. (2018) that recommends using light intensities of <11.8 Log photon flux for longer wavelengths of red light in order to prevent nonvisual system responses. This is an overall light intensity level less than each of our red light pulse stimuli (Table 1).
Concerning the image-forming visual system, perception of shades of red light may be greater in humans by virtue of the presence of L-cones in the retina that are absent in rodents. However, under low illumination conditions, the visual system relies primarily on rods, and so the visual perceptual ability of laboratory rodents under red light conditions may also not differ much from humans.
Concluding statement
We found that both the non-image and image forming visual systems of laboratory rodents are sensitive to long wavelengths of light that extend throughout much of the red light range. Our study reinforces and extends the conclusions of some other papers which caution against the notion that nighttime exposure of laboratory rodents to red light has no physiological or behavioral consequences (Butler and Silver, 2011; Niklaus et al., 2020; Nikbakht and Diamond, 2021). Despite the consequences of red light exposure observed in this study, we believe that there is much value in the study of laboratory rodents during their circadian active phase (Nelson et al., 2021). We recognize that there are many research procedures that require adequate light illumination for proper execution in the presence of the animal subjects during their dark phase.
In general, the retinal response to light is substantially reduced with longer wavelengths of light extending into the red range (Rocha et al., 2016; Emmer et al., 2018; Peirson et al., 2018). Thus, red light illumination may have less of an impact on various physiological and behavioral measures than white light, although this can only be determined empirically for each illumination condition. There is a risk, however, that uninformed use of dim red light conditions produces a false sense of security, such that the duration and intensity of red light exposure is not carefully controlled across treatment conditions and minimized to the same extent that researchers and vivarium staff would vigilantly adhere to for nighttime exposure to white light. Our study supports the utility of increased use of infrared cameras to monitor nighttime homecage activity and experimental behavioral responses of nocturnal rodents as a strategy to minimize unwanted effects of nighttime light exposure on non-image and image forming systems. Although several studies have demonstrated that substituting very low levels of dim red light for total darkness during the “dark” portion of rodent light/dark cycles has physiological, behavioral, and circadian system functional consequences (Bedrosian et al., 2013; Dauchy et al., 2015), it is unknown whether short-term daily nighttime red light exposure of rodents would have similar effects. Several papers provide thoughtful discussion of practical considerations for use of red light at night in laboratory rodent research (González, 2018; Peirson et al., 2018; Lucas and Peirson, 2024). Finally, because our study further illustrates that nighttime red light conditions commonly used in animal research have acute effects on non-image and image forming visual systems, it is important for scientific rigor and reproducibility (Matveyenko, 2018) to fully document in research reports the light parameters that experimental subjects (including homecage controls) are exposed to during their dark phase (Lucas et al., 2024; Lucas and Peirson, 2024; McDowell et al., 2024).
Footnotes
This work was supported by a GLAS grant from the American Association for Laboratory Animal Science (R.L.S and J.D.R.), National Institutes of Health (NIH) grant National Institute of Mental Health and Neurosciences MH115947 (R.L.S), NIH grant MH119734 (A.M.S. and R.L.S), and NIH training grant T32 NHLBI HL149646 fellowship (H.K.S.). We thank Dr. Sara Hashway (Director, Office of Animal Resources), Micah Stoltz (Operations Manager), and all the animal care support staff at the University of Colorado Boulder for their quality assistance with these studies; Dr. Kenneth Wright (University of Colorado Boulder) with assistance in measuring spectral power distributions for our red light stimuli; and Dr. Sondra Bland (University of Colorado Denver) and Dr. Jerry Rudy (University of Colorado Boulder) for helpful suggestions for the manuscript.
The authors declare no competing financial interests.
- Correspondence should be addressed to Robert L. Spencer at Robert.spencer{at}colorado.edu.
