Wavelength dependence of acute, light-evoked recruitment of brainstem pain-modulating systems in female rats Authors: Zhigang Shi, Stephanie Radu, Melissa Martenson, Mary Heinricher Published: 2026 ABSTRACT Highlights•Light engages pain-modulating neurons in the rostral ventromedial medulla (RVM).•RVM neurons respond to blue and green, but not red, light.•Blue and green light activated RVM ON-cells and suppressed RVM OFF-cells. Introduction Chronic pain is often associated with complaints of multisensory hypersensitivity, with increased sensitivity to non-somatic inputs including sound, smell, and light. Photophobia in migraine is the most prominent of these increased sensitivities, with the threshold for light-induced discomfort lowered even outside of the headache phase (Drummond, 1986, Vanagaite et al., 1997). However, sensory hypersensitivity is not limited to migraine disorders alone. Individuals living with other chronic pain disorders, such as those with fibromyalgia, frequently report that normal levels of light and sound are aversive (Geisser et al. 2008.; Friend and Bennett, 2011, Martenson et al., 2016). The lack of an obvious neural mechanism for enhanced sensitivity across multiple modalities has been puzzling, often leading to clinical questioning of the “reality” of this experience. There is evidence that sensory hypersensitivity is not simply a general amplification of incoming stimulation and cortical sensory processing (Carrillo-de-la-Peña et al. 2006). Instead, two distinct mechanisms are receiving increasing attention: convergence of non-somatic sensory information with nociceptive transmission pathways at the level of the thalamus, and recruitment of brainstem pain-modulating systems by non-somatic stimuli. These two ideas are not mutually exclusive, and may in fact be complementary. Support for the first mechanism derives from evidence that thalamic neurons that respond to noxious stimulation of the dura can also be activated by light stimuli (Noseda et al., 2010, Noseda et al., 2016). Evidence for the second mechanism comes from the recent finding that light stimuli can alter the activity of identified pain-modulating neurons in the brainstem. This occurs via irradiance detection pathways rather than through the thalamus (Martenson et al., 2016, Hryciw et al., 2023). The notion that non-somatic stimuli can recruit pain-modulating systems to facilitate pain is consistent with the long-standing recognition that these systems are dysregulated in many chronic pain conditions. A review of work on idiopathic pain syndromes including irritable bowel syndrome, temporomandibular disorders, fibromyalgia, and tension-type headache, found a pattern of low pain-inhibitory capacity (Yarnitsky 2010). Certainly, an inability to maintain an appropriate balance between facilitation and inhibition of nociceptive signals could be a pathogenetic factor in the development of chronic pain states (Chen and Heinricher 2019). Understanding how non-somatic sensory systems modulate pain requires a full understanding of the stimulus characteristics that influence pain. Recent work has suggested that different wavelengths of light may have distinct contributions to photophobia in migraine. Human studies utilizing psychophysical assessments in migraine patients with normal eyesight found that green light exacerbated migraine headaches significantly less than white, blue, amber or red light, suggesting that at least some aspects of migraine photophobia could originate in cone-driven retinal pathways. Complementary studies in rats of light-evoked changes in activity of thalamic neurons that responded to stimulation of the dura (and presumably contribute to headache pain), found that green light activates these neurons to a lesser extent than other colors tested (Noseda et al. 2016). Another study looking at the utility of blue wavelength-blocking glasses in children with migraine found that visually provoked beta activity was suppressed by red light and enhanced with blue light in migraine patients, implying varying effects of wavelength on cortical activity (Good et al. 1991). Taken as a whole, these findings suggest that wavelength-selective photophobia involves cone-driven retinal pathways relaying to thalamocortical somatosensory circuits separate from the main visual pathways through visual thalamus and cortex (Noseda et al. 2016). The present study considers the wavelength sensitivity of brainstem pain-modulating neurons. White light is known to activate a significant subset of pain-facilitating neurons and suppress firing of pain-inhibiting neurons in the rostral ventromedial medulla (RVM) of both male and female rats (Martenson et al., 2016, Hryciw et al., 2023). The RVM is the final output node of a critical pain-modulating system (Heinricher, 2016, Chen and Heinricher, 2019). It regulates nociceptive transmission through projections to the dorsal horn of the spinal cord. Activation of pain-facilitating neurons in RVM (referred to as “ON-cells”) and inhibition of pain-inhibiting neurons (referred to as “OFF-cells”) by light could therefore lower the threshold for pain sensation, causing light to be perceived as aversive. Indeed, the threshold for heat-evoked withdrawal of the hindpaw was found to be lower when white-light was delivered to the eye (Martenson et al. 2016). Information about retinal illumination was communicated to RVM not through a thalamic relay, but via the olivary pretectal nucleus, an important relay in the irradiance-detection pathway (Martenson et al. 2016). It is possible that this pathway has a spectral sensitivity distinct from that of retino-thalamic systems implicated in photophobia by Noseda and colleagues (2016). We therefore recorded the activity of identified ON– and OFF-cells in the RVM of lightly anesthetized female rats, and tested their responses to the blue, green, and red components of a white light stimulus sufficient to alter their activity. Methods These studies took advantage of a well-characterized lightly-anesthetized rat model, in which activity of OFF- and ON-cells has been shown to respectively suppress and enhance nociceptive behaviors (Heinricher et al., 1989, Heinricher et al., 1994, Martenson et al., 2009, Martenson et al., 2016). All experimental procedures followed the guidelines of the National Institutes of Health and were approved by the Institutional Animal Care and Use Committee at Oregon Health & Science University. Female Sprague-Dawley rats (150–250 g) were obtained from Charles River, and were acclimatized for at least one week in a room with a 12/12 h light/dark cycle. All rats were pair-housed and provided with food and water ad libitum. Experiments were performed during the light phase. Estrus stage was determined by vaginal smear at the end of each experiment. Animal preparation and anesthesia On the day of the experiment, animals were deeply anesthetized (5% isoflurane in 100% oxygen) and a catheter inserted in the external jugular vein for subsequent infusion of methohexital. They were then transferred to a stereotactic frame, and methohexital given continuously via the jugular catheter. With the animal still at a surgical anesthetic plane, a craniotomy was drilled posterior to lambda and the dura removed for access to the RVM. Heart rate was monitored using electrocardiography. Body temperature was monitored and maintained at 35.5–37.5 °C with a heating pad and a room space heater. Once surgical preparation was completed, atropine and ophthalmic ointment were applied to the eyes to eliminate any differences in the amount of light reaching the retina due to variations in pupillary diameter, between animals or trials. The room was then maintained at low ambient light conditions (< 20 lx). The rate of methohexital infusion was adjusted such that the animal withdrew its hindpaw to a radiant heat stimulus (< 53 °C), without spontaneous movement, whisking or other signs of distress (48.0 ± 0.7 mg/kg/h). Animals were maintained at a stable anesthetic rate for at least 30 min before beginning data collection. Electrophysiological recording in RVM Extracellular single-unit recordings were made with stainless steel microelectrodes (DZAP to 100 kΩ, Frederick Haer Co., Brunswick, ME). Signals were amplified (10,000x), sampled at 32,000 Hz, bandpass filtered (400 Hz to 15 kHz) and digitized. Waveforms were monitored continuously to ensure proper discrimination throughout the experiment. Typically, one neuron was studied per experiment, although occasionally two or more neurons were recorded in the same animal. Heat and light stimuli Heat-evoked paw withdrawal was used to characterize RVM neurons as ON– or OFF-cells: ON-cells become active (if not already active) at the time of noxious-evoked paw withdrawal, OFF-cells cease firing immediately prior to withdrawal, or remain silent if inactive at the time of heat onset. Paw withdrawal latency (PWL) was used as a behavioral measure of thermal nociceptive sensitivity. To evoke a withdrawal, the left hindpaw was placed on a platform overlying a feedback-controlled halogen bulb. Paw surface temperature was monitored using a thermocouple in contact with the paw. Temperature was ramped at approximately 1.4 °C/s from a holding temperature of 35 °C until a paw withdrawal occurred, or to a maximum of 53 °C. The heat ramp was terminated immediately upon detection of paw movement. Paw withdrawals were detected using a mechanoelectric bridge transducer attached to the paw. Heat trials were separated by at least 6 min. White-light stimuli (10 klux, 19.4 mW) were delivered to the left eye using a fiber-optic source (Thermo Oriel Model 66881), which was placed 4 cm from the eye. An Oriel Model 76,995 Electronic Shutter Controller was used to control the exposure duration. Narrow band-pass filters (full-width half-maximum 10 nm, transmission >85%, Edmund Optics, Barrington, NJ) with center wavelengths of 430 nm (blue, 500 lx, 6.4 mW), 532 nm (green, 400 lx, 7.9 mW), and 660 nm (red, 300 lx, 7.6 mW) were used to isolate the respective color components of the white-light stimulus. The stimulus intensity was regularly verified (Fig. 1, left).Fig. 1Experimental setup and protocol. Left: A recording electrode was placed in the RVM to monitor neuronal activity. Light of different wavelengths was delivered to the eye, and noxious heat to the hindpaw. Top right: Heat and light (white or blue/green/red component of the white light stimulus) were delivered in alternation. Lower right: Light was delivered for 10 min, with a noxious heat trial given at 9 min, and again 10 min later after the light was terminated. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Experimental setup and protocol. Left: A recording electrode was placed in the RVM to monitor neuronal activity. Light of different wavelengths was delivered to the eye, and noxious heat to the hindpaw. Top right: Heat and light (white or blue/green/red component of the white light stimulus) were delivered in alternation. Lower right: Light was delivered for 10 min, with a noxious heat trial given at 9 min, and again 10 min later after the light was terminated. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Experimental Protocols Following identification of an ON– or OFF-cell, its responses to white light and heat were assessed using two 30–s white-light trials interleaved with three heat trials at 3-min intervals, a protocol based on previous work using light pulses of 10 s to 1 min duration (Noseda et al., 2010, Martenson et al., 2016, Noseda et al., 2016, Hryciw et al., 2023). Responses to colored light (blue, green, or red; one color per experiment) were then assessed using two 30–s colored-light trials interleaved with two heat trials at 3-min intervals (Fig. 1, right top). This interleaved design provided a consistent pattern of noxious stimulation, and paw withdrawal latency remained stable across trials. Light-evoked responses were assessed by the averages of the two white-light trials and two colored-light trials. To test whether colored light altered nociceptive withdrawal threshold, the same colored light was delivered continuously for 10 min. A heat stimulus was applied to the hindpaw 9 min after light onset and another heat stimulus was applied 9 min after light offset (Fig. 1, right bottom). All three wavelengths were tested in a subset of rats. Four 30-s trials (white, blue, green, and red) were delivered in random order at 3-min intervals at the end of the experiment. Histology Recording sites were marked with an electrolytic lesion at the conclusion of the experiment. Animals were euthanized by methohexital overdose and perfused transcardially with physiological saline. Brains were removed, post-fixed in 10% formalin and subsequently sectioned at 60 µm. Lesion sites were identified and mapped according to the atlas of Paxinos and Watson (2009). The RVM was defined as the nucleus raphe magnus and adjacent reticular formation medial to the lateral edge of the pyramids at the level of the facial nucleus. Data analysis Spontaneous firing was determined from 30-s sampling intervals immediately preceding each stimulus trial. For ON-cells, heat-evoked responses were defined as the total number of spikes occurring within the withdrawal-related burst. For OFF-cells, heat-evoked responses were defined as the duration of the firing pause that began immediately prior to paw withdrawal. A neuron was considered “light-responsive” if it exhibited a change in firing rate of at least 50% during exposure to white-light compared to the firing rate during the 30-s pre-stimulus period. For ON-cells that were silent during the 30 s prior to light onset, a cell was deemed responsive if a minimum of 10 spikes was exhibited during light stimulation. A ceiling of 500% of baseline was conservatively applied to limit the influence of neurons with low spontaneous activity in the analysis. Light-evoked responses were further quantified as the duration of the longest single silent period initiated during light (OFF-cells) or the total spike count triggered by the light stimulus (ON-cells). In cases where the ON-cell was already active before stimulus onset, the evoked response was assigned the total number of spikes during the light stimulus. Statistical analyses were performed using GraphPad Prism 10 (GraphPad Software). Given that cell parameters were not normally distributed, comparisons were conducted using non-parametric tests or a paired t-test or repeated-measures ANOVA on log-transformed data. For the latter, data are presented as geometric means and 95% confidence intervals (CIs). Paw withdrawal latencies were compared using a paired t-test, and these data are presented as mean ± SEM. P < 0.05 was considered statistically significant. Results We recorded 41 OFF-cells and 46 ON-cells that were responsive to white light in 67 rats (2–4 cells were recorded simultaneously in 17 of the 67). The distribution of neurons within the RVM is shown in Fig. 2. The estrous stage was determined in 64 rats (14 in proestrus, 12 in estrus, 18 in metestrus, and 20 in diestrus) and did not significantly affect paw withdrawal latency, RVM neuronal activity, or responses to heat or white light. Given the relatively small number of animals/cells at each stage, estrous-stage comparisons are not presented.Fig. 2Recording sites of OFF- and ON-cells in the RVM. Recorded neurons were distributed between − 1.52 mm and − 2.50 mm relative to the interaural line. Recording sites of OFF- and ON-cells in the RVM. Recorded neurons were distributed between − 1.52 mm and − 2.50 mm relative to the interaural line. Overall, mean paw withdrawal latency was 4.6 ± 0.2 s. For OFF-cells, the spontaneous firing rate in the 30 s prior to light exposure was 12.1 ± 1.4 spikes/s, which decreased to 2.2 ± 0.5 spikes/s (17.2 ± 2.3% of baseline) during white light. ON-cells fired at 1.8 ± 0.9 spikes/s before light exposure, which increased to 12.7 ± 1.9 spikes/s during white light. Spectral sensitivity of RVM neurons To determine the spectral sensitivity of RVM neurons to light, neurons identified as responsive to white light were then tested with the blue, green, or red component of the white light stimulus (Table 1). The majority of these neurons responded to blue and green light, with decreased firing in OFF-cells and increased firing in ON-cells. None of the tested neurons responded to red light.Table 1Percentage of OFF- and ON-cells that responded to the blue, green, or red components of the white-light stimulus. Only neurons that were responsive to white light were tested. Responsiveness was defined as a change in firing rate of at least 50% relative to the 30-s period preceding light onset. Values indicate the percentage of cells responding.BlueGreenRedOFF-cellResponsive64.3%78.6%0.0%number tested141413 ON-cellResponsive81.3%53.3%0.0%number tested161515 Table 1Percentage of OFF- and ON-cells that responded to the blue, green, or red components of the white-light stimulus. Only neurons that were responsive to white light were tested. Responsiveness was defined as a change in firing rate of at least 50% relative to the 30-s period preceding light onset. Values indicate the percentage of cells responding.BlueGreenRedOFF-cellResponsive64.3%78.6%0.0%number tested141413 ON-cellResponsive81.3%53.3%0.0%number tested161515 Percentage of OFF- and ON-cells that responded to the blue, green, or red components of the white-light stimulus. Only neurons that were responsive to white light were tested. Responsiveness was defined as a change in firing rate of at least 50% relative to the 30-s period preceding light onset. Values indicate the percentage of cells responding. RVM neurons responded concordantly to heat, white light, blue light, and green light, as shown in representative recordings in Fig. 3. Irrespective of stimulus modality or wavelength, OFF-cells exhibited a pause or decrease in firing, whereas ON-cells displayed a burst or increase in activity during these stimuli.Fig. 3Responses of representative RVM ON– and OFF-cells to white and colored light. Recordings from three OFF-cells (left, A, C, E) and three ON-cells (right, B, D, F) during heat stimulation and exposure to white, blue, green, and red light. Heat evoked a pause in OFF-cells and a burst in ON-cells (arrows). White, blue, and green light suppressed OFF-cell firing and activated ON-cells, whereas red light evoked no response. Ratemeters show firing in spikes per second (sp/s). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Responses of representative RVM ON– and OFF-cells to white and colored light. Recordings from three OFF-cells (left, A, C, E) and three ON-cells (right, B, D, F) during heat stimulation and exposure to white, blue, green, and red light. Heat evoked a pause in OFF-cells and a burst in ON-cells (arrows). White, blue, and green light suppressed OFF-cell firing and activated ON-cells, whereas red light evoked no response. Ratemeters show firing in spikes per second (sp/s). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Group analysis (Fig. 4) confirmed that both blue and green light reduced OFF-cell firing compared to pre-stimulus baseline (blue: W = − 101, p = 0.004, n = 14; green: W = 105, p = 0.0001, n = 14, Wilcoxon’s signed rank test), while red had no effect (W = 3, p = 0.95, n = 13). Similarly, both blue and green light evoked increased firing in ON-cells (blue: W = 91, p = 0.0002, n = 16; green: W = 51, p = 0.0059, n = 15), while red had no effect (W = − 36, p = 0.12, n = 15).Fig. 4Comparison of RVM OFF- and ON-cell responses to white versus colored light. A,C,E: For OFF-cells, responses are expressed as firing rate during light exposure relative to the 30-s pre-light baseline (100% = no change, dotted line). B,D,F: For ON-cells, responses are expressed as total evoked spikes during the light-evoked burst. The magnitude of the response to blue or green light was less than that to white light. No responses were seen with red light. Data are presented as geometric means and 95% confidence intervals. *p < 0.05 compared to white light. A. t13 = 3.23, p = 0.0065, n = 14; B, t15 = 2.46, p = 0.027, n = 16; C. t13 = 9.32, p < 0.0001, n = 14; D. t14 = 4.06, p = 0.0012, n = 15. E. t12 = 6.88, p 0.0001, n = 13. F. t14 = 5.96. p < 0.0001, n = 15. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Comparison of RVM OFF- and ON-cell responses to white versus colored light. A,C,E: For OFF-cells, responses are expressed as firing rate during light exposure relative to the 30-s pre-light baseline (100% = no change, dotted line). B,D,F: For ON-cells, responses are expressed as total evoked spikes during the light-evoked burst. The magnitude of the response to blue or green light was less than that to white light. No responses were seen with red light. Data are presented as geometric means and 95% confidence intervals. *p < 0.05 compared to white light. A. t13 = 3.23, p = 0.0065, n = 14; B, t15 = 2.46, p = 0.027, n = 16; C. t13 = 9.32, p < 0.0001, n = 14; D. t14 = 4.06, p = 0.0012, n = 15. E. t12 = 6.88, p 0.0001, n = 13. F. t14 = 5.96. p < 0.0001, n = 15. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Moreover, while OFF-cell firing during blue and green light was significantly reduced relative to the 30-s pre-light baseline, the magnitude of the firing rate suppression was less than that observed during white light (Fig. 4A, C). For ON-cells, responses were quantified as the number of spikes during the light-evoked burst of activity. Blue and green light evoked fewer spikes than white light (Fig. 4B, D). Consistent with the representative examples, neither OFF‑cells nor ON‑cells as a group responded to red light (Fig. 4E, F) despite the substantial change in firing evoked by white light for the tested neurons. To further elucidate the ability of light to recruit OFF- and ON-cell populations, we measured the latencies to OFF‑cell suppression and ON‑cell activation during white, blue, and green light exposure (Fig. 5). The latency to OFF‑cell suppression evoked by blue light was not significantly different from that evoked by white light. By contrast, the latency to the ON-cell burst evoked by blue light was longer than that evoked white light. For both OFF- and ON-cells, the responses to green light occurred at a longer latency than during delivery of white light.Fig. 5Latency to OFF-cell suppression and ON-cell activation during blue and green light. OFF-cell response latency during blue light was comparable to that during white light. However, ON-cell response latency was significantly longer during blue light compared with white light. Latency was significantly longer during green light compared to white light for both OFF-cells and ON-cells. Data are presented as geometric means and 95% confidence intervals. *p < 0.05 compared to white light. A. t12 = 1.50, p = 0.17, n = 13; B. t13 = 2.59, p = 0.022, n = 14; C. t12 = 6.69, p < 0.0001, n = 13; D. t12 = 10.24, p < 0.001, n = 13. *p < 0.05 compared to white light. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Latency to OFF-cell suppression and ON-cell activation during blue and green light. OFF-cell response latency during blue light was comparable to that during white light. However, ON-cell response latency was significantly longer during blue light compared with white light. Latency was significantly longer during green light compared to white light for both OFF-cells and ON-cells. Data are presented as geometric means and 95% confidence intervals. *p < 0.05 compared to white light. A. t12 = 1.50, p = 0.17, n = 13; B. t13 = 2.59, p = 0.022, n = 14; C. t12 = 6.69, p < 0.0001, n = 13; D. t12 = 10.24, p < 0.001, n = 13. *p < 0.05 compared to white light. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) To directly compare responses to different colors, particularly blue and green light, we then tested different colors in a subset of OFF-cells (n = 16) and ON-cells (n = 24). Consistent with the results described above, responses to blue and green light were smaller than those evoked by white light. Direct comparison of blue and green light showed no significant difference in OFF‑cell responses (blue: 42.6 ± 10.5% of pre-light baseline, green: 46.4 ± 7.4% of baseline, t15 = 1.23, p = 0.24, n = 16). By contrast, ON‑cells activation by green light (253 ± 69 spikes) was less than by blue light (439 ± 99 spikes, t23 = 2.28, p = 0.032, n = 24). Longer-duration light exposure A substantial subset of RVM OFF- and ON-cells thus responded during exposure to blue and green, but not red, light, with suppression of OFF-cell firing and activation of ON-cells. Given the known antinociceptive influence of OFF-cells and pronociceptive influence of ON-cells, we would therefore predict that exposure to blue or green, but not red, light would lead to behavioral hyperalgesia. Consistent with this prediction, longer exposure (9 min) to blue or green light produced a significant decrease in heat-evoked paw withdrawal latency (Fig. 6). Consistent with the lack of RVM response to red light, exposure to red light had no effect on withdrawal latency.Fig. 6Exposure to blue or green light induces hyperalgesia. Heat-evoked paw withdrawal latency (PWL) was significantly reduced during 10-min exposure to blue (t22 = 2.21, p = 0.038, n = 23) or green (green, t22 = 2.48, p = 0.021, n = 23) light compared with ambient conditions (grey). Red light (t20 = 0.83, p = 0.39, n = 21) had no effect on PWL. Data are presented as mean ± SEM. *p < 0.05 vs ambient conditions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Exposure to blue or green light induces hyperalgesia. Heat-evoked paw withdrawal latency (PWL) was significantly reduced during 10-min exposure to blue (t22 = 2.21, p = 0.038, n = 23) or green (green, t22 = 2.48, p = 0.021, n = 23) light compared with ambient conditions (grey). Red light (t20 = 0.83, p = 0.39, n = 21) had no effect on PWL. Data are presented as mean ± SEM. *p < 0.05 vs ambient conditions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Consistent with these behavioral results, OFF-cell firing was suppressed throughout exposure to blue or green light (Fig. 7). While ON-cells exhibited activation upon light onset, their firing rates returned to pre-stimulus levels before the end of the exposure.Fig. 7Neuronal responses during 10-min colored-light exposure. Top panels show representative recordings of OFF-cells (A) and ON-cells (B) during exposure to blue, green, or red light. Colored light was delivered for 10 min (overlay). Triangles indicate the application of heat to the hindpaw at 9 min after light onset and again at 9 min after light offset. Bottom panels summarize population responses of OFF-cells (C) and ON-cells (D). Blue light (n = 14) and green light (n = 14) significantly suppressed OFF-cell firing during 10-min illumination. ON-cells exhibited activation upon exposure to blue light (n = 16) and green light (n = 15), but returned to levels comparable to baseline (time 0 min) and ambient conditions (time 19 min). Red light had no effect on the firing rate of either OFF-cells (n = 13) or ON-cells (n = 15). Data are plotted as mean ± SEM. Statistical comparisons were performed using repeated-measures one-way ANOVA on log-transformed data followed by Tukey’s post hoc test. * p < 0.05 vs baseline (0 min, 30 sec before light onset), #p < 0.05 vs ambient (19 min).C. Blue: F7, 91 = 11.15, p < 0.001; Green: F7, 91 = 9.6, p < 0.001; Red: F7, 84 = 1.92, p = 0.076. D. Blue: F7, 105 = 8.82, p < 0.001; Green: F7, 98 = 3.48, p = 0.0023; Red: F7, 96 = 1.90, p = 0.078. Sp/s: spikes per second. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Neuronal responses during 10-min colored-light exposure. Top panels show representative recordings of OFF-cells (A) and ON-cells (B) during exposure to blue, green, or red light. Colored light was delivered for 10 min (overlay). Triangles indicate the application of heat to the hindpaw at 9 min after light onset and again at 9 min after light offset. Bottom panels summarize population responses of OFF-cells (C) and ON-cells (D). Blue light (n = 14) and green light (n = 14) significantly suppressed OFF-cell firing during 10-min illumination. ON-cells exhibited activation upon exposure to blue light (n = 16) and green light (n = 15), but returned to levels comparable to baseline (time 0 min) and ambient conditions (time 19 min). Red light had no effect on the firing rate of either OFF-cells (n = 13) or ON-cells (n = 15). Data are plotted as mean ± SEM. Statistical comparisons were performed using repeated-measures one-way ANOVA on log-transformed data followed by Tukey’s post hoc test. * p < 0.05 vs baseline (0 min, 30 sec before light onset), #p < 0.05 vs ambient (19 min).C. Blue: F7, 91 = 11.15, p < 0.001; Green: F7, 91 = 9.6, p < 0.001; Red: F7, 84 = 1.92, p = 0.076. D. Blue: F7, 105 = 8.82, p < 0.001; Green: F7, 98 = 3.48, p = 0.0023; Red: F7, 96 = 1.90, p = 0.078. Sp/s: spikes per second. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) For heat-evoked responses during 10-min light exposure, blue light increased the duration of the OFF-cell pause compared with ambient conditions (ambient: 4.8 ± 1.0 s; blue light: 16.8 ± 6.1 s, t9 = 2.95, p = 0.016, n = 10). However, the total number of spikes in the ON-cell burst was not increased during blue light (t15 = 0.82, p = 0.43, n = 16), and green light had no effect on the heat-evoked responses of either OFF-cells (t9 = 0.53, p = 0.61, n = 10) or ON-cells (t14 = 0.55, p = 0.59, n = 15). Discussion There is a significant literature focused on how the spectral characteristics of the light environment influence the circadian cycle and mood (Berson et al., 2002, Digre and Brennan, 2012). More recently, parallel studies in rodent and human have provided evidence for an interaction between light and pain-transmission pathways at the level of thalamus and cortex (Noseda et al., 2016, Nir et al., 2018). However, integration of information about environmental light levels also converges with somatic input at the level of individual pain-modulating neurons in the RVM. Importantly, this light-related information was shown to reach the RVM not via the thalamus, but through irradiance detection pathways (Martenson et al., 2016, Hryciw et al., 2023). In the present experiments, we considered how different wavelengths of light affect the activity of pain-modulatory systems by recording the responses of identified ON– and OFF- cells in the RVM to blue, green, and red components of a white-light stimulus in our well-established lightly anesthetized rat model. In this preparation, spontaneous or experimental recruitment of OFF-cells results in measurable hypoalgesia, whereas engagement of OFF-cells produces hyperalgesia (Heinricher et al., 1989, Heinricher et al., 1994, Martenson et al., 2009, Martenson et al., 2016). After first confirming that white light increases the activity of ON-cells and decreases that of OFF-cells (Martenson et al., 2016, Hryciw et al., 2023), we determined that the blue and green components of the white-light stimulus evoked similar pronociceptive changes in the activity of these pain-modulating neurons, while the red component did not. We used female rats, not only because of the greater prevalence of migraines and chronic pain syndromes in women compared to men, but because the anatomical organization and functional activation of the PAG-RVM system is known to be sexually dimorphic (Loyd and Murphy, 2006, Greenspan et al., 2007, Hauser et al., 2011). Despite this, the majority of work on pain-modulating systems in both animal models and humans has historically used males. The present analysis thus contributes to a more balanced look at the function of these systems. We first tested the effect of exposure to blue light and found the majority of OFF- and ON-cells responded to blue light. While the magnitude of the response to the blue component of the white stimulus was less than that evoked by the white stimulus of which it is a component, it was nonetheless sufficient to produce a significant thermal hyperalgesia. This pronociceptive effect is consistent with the expected effects of shifting RVM output such that ON-cells are dominant (Heinricher et al., 1989, Heinricher et al., 1994, Martenson et al., 2009, Martenson et al., 2016). It is also consistent with reports of a pronociceptive influence of blue light in human migraine patients (Main et al., 2000, Noseda et al., 2016). By contrast with shorter wavelengths, green light has been reported to be less aversive or even, with multi-day exposures, antinociceptive (Noseda et al., 2016, Ibrahim et al., 2017, Martin et al., 2021a, Martin et al., 2021b, Tang et al., 2022, Wu et al., 2023). Green light was found to exacerbate migraine headache significantly less than exposure to white, blue, amber, or red lights in patients with normal eyesight (Noseda et al. 2016). Migraineurs also reported a higher discomfort threshold when exposed to middle wavelengths compared to shorter and longer wavelengths, even outside of migraine attacks (Main et al., 1997, Main et al., 2000). Indeed, green light has been reported to reduce migraine-headache pain in some patients (Martin et al., 2021a, Martin et al., 2021b, Lipton et al., 2023). These findings raised the possibility that the green component of the white-light stimulus might suppress the firing of ON-cells and fail to inhibit OFF-cells. However, this was not the case: like blue and white light, green light increased the activity of ON-cells and decreased that of OFF-cells. As with blue light, the majority of both OFF- and ON-cells responded to green light, although again, the magnitude of the response was smaller than that with white light. Nonetheless, green light produced measurable behavioral hyperalgesia. This finding accords with those of Noseda and colleagues in migraine (2016). Moreover, although others have reported that green light is antinociceptive or analgesic, this is certainly related to the duration of exposure, since the antinociceptive actions of green light in rodents requires hours to days of exposure (Ibrahim et al., 2017, Martin et al., 2023). Indeed, it is possible that the effects of green light tested here would be reversed with chronic exposure. This latter idea is not inconsistent our observation that ON-cell activation by green light adapted over the course of a 9-minute exposure. Nonetheless, the fact that light engages the RVM via the irradiance detection pathway rather than retino-thalamic circuits suggests that this is a functionally distinct circuit (Martenson et al. 2016). Finally, none of the cells tested here showed a change in activity when challenged with the red component of the white-light stimulus, and exposure to red light had no effect on thermal withdrawal latency in these animals. Our findings contrast with those of Khanna and colleagues, who reported hyperalgesia in rats exposed to low intensity (50 lx) red light-emitting diodes for 8 h daily for 5 days (Khanna et al. 2019). Notably, this hyperalgesia was not evident with acute red-light exposure, and required at least three days to fully develop, again suggesting that duration of exposure may be an important variable. Interestingly, Khanna and colleagues suggested that the hyperalgesic effect of red light was mediated by the RVM. However, they used focal application of bicuculline to test the role of the RVM. This is known to activate all RVM neurons, including OFF-cells, leading to analgesia (Drower and Hammond, 1988, Heinricher et al., 1991, Heinricher and Kaplan, 1991), and would mask any hyperalgesia due to red light whether or not the mechanism involved the RVM. Whether any hyperalgesic effect of red light is mediated by RVM therefore remains to be fully tested. Khanna and ccolleagues’ findings nevertheless highlight the potential and unknown influence of intensity and duration of light exposure in producing either pronociceptive or antinociceptive effects. The spectral response characteristics of RVM ON– and OFF-cells described here are consistent with those of the rodent retina, which possesses ultraviolet and middle-wavelength photoreceptors as well as intrinsically photosensitive ganglion cells, and more generally, those of the irradiance detection system implicated in control of circadian rhythms and other non-image-forming visual functions (Jacobs et al. 2001). Our findings thus accord with previous evidence that the influence of light on activity of RVM pain-modulating neurons is mediated by irradiance detection pathways (Martenson et al. 2016). Indeed, our findings suggest that RVM is engaged by pathways distinct from those influencing pain-transmission neurons in thalamocortical systems (Noseda et al. 2018). ON– and OFF-cells responded to the blue and green components of the white-light stimulus, whereas the red component was not sufficient to engage these pain-modulating neurons. Although rats do in fact see red light, and red light can produce circadian phase shifts, behavioral sensitivity to red light is at least an order of magnitude less than that for shorter wavelengths (Zhang et al., 2017, Niklaus et al., 2020). Although it is possible that a more prolonged or more intense red-light stimulus would have recruited ON– and OFF-cells, the present experiments indicate that the red component of our effective white-light stimulus is insufficient to activate these neurons. In conclusion, there are likely multiple pathways by which light can impact pain, but we focused here on how acute exposure to light of varying wavelengths might modulate the descending pain-modulatory system, a paradigm relevant to chronic-pain-related multisensory hypersensitivity. A substantial subset of RVM OFF- and ON-cells responded to exposure to blue and green, but not red, light. The suppression of OFF-cell firing and activation of ON-cells overall suggested that exposure to blue or green, but not red, light would produce pronociception. This was confirmed by a significant thermal hyperalgesia during exposure to blue and green light. While different wavelengths of light may elicit distinct biological responses, the duration and intensity are likely also important and have the potential to dramatically alter behavioral effects. Whether these effects are pronociceptive or antinociceptive, our results demonstrate that the pain-modulating cells of the RVM can be engaged by light of specific wavelengths. CRediT authorship contribution statement Zhigang Shi: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Conceptualization. Stephanie Radu: Writing – original draft, Investigation, Conceptualization. Melissa Martenson: Writing – review & editing, Methodology, Conceptualization. Mary Heinricher: Writing – review & editing, Funding acquisition, Formal analysis, Conceptualization. Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Mary M Heinricher reports financial support was provided by National Institutes of Health. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.