Green light exposure reduces central sensitization in cats with osteoarthritis-associated somatosensory impairment Authors: Aliénor Delsart, Aude Castel, Guillaume Dumas, Frédéric Sauvé, Colombe Otis, Bertrand Lussier, Franck Péron, Marc Hébert, Nicolas Lapointe, Nicolas Quaegebeur, Pierre Grandjean, Maxim Moreau, Johanne Martel-Pelletier, Jean-Pierre Pelletier, Eric Troncy Published: 2026 ABSTRACT SummaryIn humans, chronic pain is associated with multisensory hypersensitivity, which can be modulated by sensorial exposure. Although cats experience chronic pain, their sensorial perception under this condition has never been investigated. Adult neutered healthy (n = 6) or osteoarthritic (n = 32) cats were assessed for functional alteration, central sensitization, and sensorial (olfactory and auditory) tolerance. The effects of specific light wavelengths were tested for neuromodulation via visual exposure. Hypersensitized cats displayed stronger unpleasant odor avoidance and lower sensitivity to repeated repulsive sounds. Odor avoidance correlated with delta and theta band proportions, inverse markers of endogenous inhibitory control. Functional alteration correlated with beta activity, a pain marker. After green light exposure, neuro-sensitization decreased for 80% of the less sensitized cats, and delta and theta band proportions increased further (no longer requiring inhibition), while beta and gamma band proportions decreased. Neuro-sensitization appears central in feline osteoarthritis pain, is associated with sensorial modifications, and is modulable by non-pharmacological approaches.Keywords: osteoarthritis, endogenous control, non-pharmacological, electroencephalogram, cat Introduction Among the feline population, 25.6% of cats tested (aged 0.2–20 years) suffer from osteoarthritis (OA), and its pain management remains a significant challenge.1 OA pain is governed by a dynamic balance between facilitatory and inhibitory endogenous controls (ECs) specific to each individual.2 An imbalance is associated with peripheral and central sensitization, as reported in humans3 and in cats,4 phenotypically assessed using quantitative sensory testing (QST).2 Briefly, peripheral sensitization results from lowered activation threshold and is characterized by allodynia and hyperalgesia.5,6 Persistent input from C-fibers can induce spinal sensitization (e.g., wind-up phenomenon),7 when local neuro-excitation, neuro-inflammation, and inhibition of inhibitory interneurons promote neuroplasticity and central sensitization.5,8 The inhibitory-ECs pass through brainstem structures such as the locus coeruleus (LC), rostral ventromedial medulla (RVM), and periaqueductal gray matter (PAG).2,9 A specific chronic pain electroencephalographic (EEG) signature was reported in human patients as being sensitive to analgesic treatment.10,11 Those with severe OA and pain intensity had higher beta (high frequency) and lower theta (low frequency) power activities at rest,11 whereas the inhibitory-EC intensity was negatively correlated with delta and theta brain oscillations.12,13 Integration of sensory input from the different senses involves several cerebral regions that overlap with those responsible for pain perception,14 suggesting that in pathological situations, multiple sensorial perceptions can be altered and influence each other. Indeed, migraine or fibromyalgia human patients often suffer from central sensitization15,16 and have reported increased sensitivity to non-painful sensory stimulations such as light,17,18,19 odors,20,21 and sounds.22,23 This phenomenon named “central hypersensitivity” denotes an increase in neuro-sensitization to both painful and non-painful sensory stimuli.24 Although tactile and thermal hypersensitivities have been reported in OA cats,4 chronic pain influence on other senses remains poorly studied. Sensory stimulations also have the capacity to modulate pain perception. The latter was shown to be reduced by pleasant smells and exacerbated with neutral or unpleasant olfactory stimuli.25,26 A narrow-band green light-emitting diode (LED; 520–525 nm wavelength) exposure decreased neuro-sensitization in sodium monoiodoacetate (MIA) OA pain rat models and pain perception in human patients with chronic pain.27,28,29,30,31 No such data are available in cats, but some olfactory (e.g., prey odor and catnip),32 auditory (e.g., specialized music),33 and visual (e.g., prey movement)34 environments were suggested as enrichment for clinic stress reduction and for shelter cats’ welfare improvement. Conversely, ultrasounds and citrus smells were described as repellents for this species.35,36 A recent feline EEG feasibility study found that narrow-band blue light perception differed from that of green and red in cats with chronic pain, but exposure duration was insufficient to draw any conclusion on its analgesic effect.37 Considering the neuro-sensitization development in cats affected by OA chronic pain and the lack of studies on sensorial perception, this study aimed to evaluate the interaction between chronic OA pain and sensorial alterations and possible neuromodulation. The hypotheses were as follows: (1) feline sensorial perception will be measurable and will be correlated with pain and neuro-sensitization measurements (QST and EEG), and (2) exposure to varying types of light (for wavelength and color temperature) can modulate neuro-sensitization, with the most pronounced effects observed under narrow-band green light stimulation. Results This study was divided into three phases (Table 1). Phase I aimed to characterize sensorial perception using previously reported repellent stimuli (grapefruit smell35 and ultrasounds36) according to various degrees of neuro-sensitization and functional impairment. Phase II explored neuromodulation following narrow-band green light exposure using QST. Phase III investigated neuromodulation during sensorial stimulation (olfactory and visual) or following exposure to specific wavelengths and color temperatures (green, artificial, and natural light) using EEG spectral analysis. Functional OA impairment was quantified by the validated Montreal instrument for cat arthritis testing for use by veterinarians [MI-CAT(V)] score, categorized into mild (9%–20%), moderate (21%–35%), or severe (>35%) clusters.38,39 Peripheral and central sensitization were confirmed using two validated QSTs: paw withdrawal threshold (PWT; −32% compared to the cutoff value for allodynic, i.e., hypersensitized, cats)40 and response to mechanical temporal summation (RMTS) stimuli (−56% compared to the cutoff value, meaning <14 stimulations).4 Intraclass correlation coefficient (ICC) estimates and their confidence interval at 95% (CI95) were calculated and, when repeatable, data were averaged. Mann-Whitney U test, Wilcoxon signed-rank test, correlation (Spearman rank rhoS, ρS, or Pearson, r) coefficient, spectral analysis, and linear mixed-effects modeling were performed. Table 1Outcomes and distribution of cats in each phasePhaseSenses testedCats (number)Health statusAge (median [min–max])(I)Neuro-sensitization (QST) versus innovant sensorial testingolfactory and auditory stimulations11 (6 females)OA8 [3–14]Neuromodulation versus QST (II)visual stimulations6 (2 females)healthy2 [1–4]24 (17 females)OA11 [6–16]Versus EEG (III)olfactory and visual stimulations11 (6 females)OA8 [3–14]visual stimulations6 (2 females)OA10 [8–14] Outcomes and distribution of cats in each phase Phase I: Neuro-sensitization (QST) versus innovant sensorial testing Central sensitization and functional alteration description Osteoarthritic cats (n = 11) presented moderately repeatable RMTS scores (ICC = 0.738 CI95 [−0.009 to −0.934]; p = 0.030). The mean score of the duplicate sessions (total of four values) was used in the following olfactory and auditory correlation analysis. A wide range of neuro-sensitization was reported, as assessed by RMTS (8.50 stimulations [5.50–30.00]), alongside functional impairments measured by MI-CAT(V) (19.70% [1.50–34.80]). The RMTS and MI-CAT(V) scores were moderately and negatively correlated (ρ = −0.633; p = 0.032). Cats classified within the mild MI-CAT(V) cluster (n = 6; 15.95% [9.10–19.70]) tended to demonstrate lower central sensitization compared to cats in the moderate cluster (n = 3; 24.2% [22.70–34.80]), with a 2-fold higher RMTS score compared to the latter (p = 0.095, Mann-Whitney U test; Figure 1A). The influence of RMTS and MI-CAT(V) on olfactory perception (A) RMTS score according to MI-CAT(V) clusters. Cats mildly affected (n = 6) tended to demonstrate less central sensitization than cats classified in the moderate MI-CAT(V) cluster (n = 3) according to Mann-Whitney U test (p = 0.095). (B) Avoidance of GEO according to RMTS score of osteoarthritic cats (n = 11). A moderate positive Spearman rank correlation (ρS = −0.653; p = 0.026) was observed with cats presenting greater central sensitization to more likely avoid the unpleasant smell. (C) Avoidance of GEO according to MI-CAT(V) total score of osteoarthritic cats (n = 11). A strong positive Spearman rank test correlation was revealed (ρS = 0.913; p < 0.001). Cats with greater functional impairments were more likely to avoid the unpleasant smell. (D) Avoidance of GEO according to MI-CAT(V) clusters. Cats mildly affected (n = 6) were less repulsed by the GEO smell compared to cats in the moderate MI-CAT(V) cluster (n = 3) according to a Mann-Whitney U test (p = 0.024). Data are represented as mean (SD) for RMTS score, MI-CAT(V) total score, and as percentage of avoidance for GEO smell. RMTS, response to mechanical temporal summation; MI-CAT(V), Montreal instrument for cat arthritis testing for use by veterinarians; GEO, grapefruit essential oil. Olfactory perception The grapefruit essential oil (GEO) 0.1% avoidance was moderately and negatively correlated with RMTS score (ρS = −0.653; p = 0.026; Figure 1B) and strongly and positively correlated with MI-CAT(V) score (ρS = 0.913; p < 0.001; Figure 1C). Neuro-sensitized cats (lower RMTS) or with more functional impairments (higher MI-CAT[V]) were more likely to avoid the GEO smell. Cats in the mild MI-CAT(V) cluster were 25% less repulsed by the smell compared to those in the moderate cluster (p = 0.024, Mann-Whitney U test; Figure 1D). Auditory perception Based on the brainstem auditory evoked response (BAER), one cat was considered as an outlier due to technical issues and removed for the subsequent analyses. The number of tolerated repeated ultrasound tended to be repeatable between sessions (ICC = 0.587 CI95 [−1.329 to −0.940]; p = 0.171) probably due to a type II statistical error; the mean of the duplicate sessions was used for the Spearman correlation analysis. The auditory tolerance was strongly and negatively correlated with RMTS scores (ρS = −0.811; p = 0.002; Figure 2A) and positively correlated with MI-CAT(V) scores (ρS = 0.698; p = 0.022; Figure 2B). Neuro-sensitized cats (lower RMTS) or with more functional impairments (higher MI-CAT[V]) were more tolerant to ultrasound. No difference was found between MI-CAT(V) clusters (type II statistical error, n = 2 cats in the moderate cluster). The influence of RMTS and MI-CAT(V) on auditory perception (A) Repeated exposure to ultrasound according to RMTS score. A strong negative Spearman rank test correlation was observed (ρS = −0.811; p = 0.002). Cats with greater central sensitization were more tolerant to the unpleasant sound. (B) Repeated exposure to ultrasound according to MI-CAT(V) total score. A moderate positive Spearman rank test correlation was revealed (ρS = 0.698; p = 0.022). Cats with greater functional impairments were more tolerant to unpleasant sound. Data of osteoarthritic cats (n = 10) are represented as mean for RMTS score, MI-CAT(V), and the repeated exposure to ultrasound. RMTS, response to mechanical temporal summation; MI-CAT(V), Montreal instrument for cat arthritis testing for use by veterinarians. Phase II: Neuromodulation by visual stimulation Effects of green light wavelength on neuro-sensitization assessed by QST Before light exposure (i.e., pre-exposure), healthy cats (n = 6) had a higher PWT and RMTS scores compared to OA cats (n = 24), respectively, by 45% (forelimbs PWT), 33% (hindlimbs PWT), and 100% (RMTS scores) (Tables 2 and S2). The RMTS score was correlated with forelimbs PWT (ρS = 0.435; p = 0.017; Figure 3A) and tended to be correlated with hindlimbs PWT (ρ = 0.333; p = 0.072; Figure 3B). Based on PWT measurements, 14 cats were classified as allodynic, and RMTS was used to assess neuro-sensitization for the remainder of phase II. In healthy cats, RMTS scores had an inter-day variability of 15%, which was used as the threshold to quantify a size effect (i.e., responder status). Following green light exposure, 54% of OA cats had a sufficient increase in RMTS score to be considered responders after 4 days ([15%–114%]; p = 0.059 Wilcoxon signed-rank test on raw RMTS score) and 50% after 8 days ([29%–150%]; p = 0.033 Wilcoxon signed-rank test) compared to pre-exposure. Non-allodynic cats (i.e., lower neuro-sensitization based on their PWT) had a 32% increase in their RMTS after 8 days of green light exposure compared to pre-exposure (80% of responders; p = 0.016 Wilcoxon signed-rank test), whereas only 29% of allodynic cats were responders (Table 2). Table 2Descriptive statistics of healthy cats and the cats affected by osteoarthritis chronic pain for their neuro-sensitization at pre-exposure and after green light exposureNRMTS(Number of stimulations)Pre-exposurePost-4 daysPost-8 daysHealthy626.25 (8.50–30.00)21.00 (12.50–30.00)22.83 (14.50–30.00)OA chronic pain2411.50 (4.50–26.00)11.00 (#) (6.00–30.00)12.25 (∗) (4.50–30.00)Allodynic149.25 (4.50–26.00)8.75 (6.00–30.00)8.75 (4.50–30.00)Non-allodynic1013.00 (6.50–23.33)14.25 (#) (6.50–27.00)17.75 (∗) (5.00–30.00)Median [min–max]––––RMTS, response to mechanical temporal summation; OA, osteoarthritis. Intragroup analyses using Wilcoxon signed-rank test: ∗ statistically different from baseline (p ≤ 0.033); # tended to be statistically different from baseline (p ≤ 0.059). Allodynia was defined based on a predetermined allodynic threshold, i.e., first quartile values for the forelimbs and hind limbs. Descriptive statistics of healthy cats and the cats affected by osteoarthritis chronic pain for their neuro-sensitization at pre-exposure and after green light exposure RMTS, response to mechanical temporal summation; OA, osteoarthritis. Intragroup analyses using Wilcoxon signed-rank test: ∗ statistically different from baseline (p ≤ 0.033); # tended to be statistically different from baseline (p ≤ 0.059). Allodynia was defined based on a predetermined allodynic threshold, i.e., first quartile values for the forelimbs and hind limbs. Correlations between central and peripheral sensitization (A) RMTS score according to forelimbs PWT (n = 30 cats). A weak positive Spearman rank test correlation was revealed (ρS = 0.435; p = 0.017). (B) RMTS score according to hindlimbs PWT (n = 30 cats). A tendency for a weak positive Spearman rank test correlation was observed (ρS = 0.333; p = 0.072). Data are represented as mean for RMTS score and PWT. RMTS, response to mechanical temporal summation; PWT, paw withdrawal threshold. Phase III: Mechanisms of neuro-sensitization and neuromodulation assessed by EEG Olfactory and visual stimulations: Feasibility of spectral analysis Cats (n = 11 OA, included in phase I) were exposed to GEO 0.1% or specific light wavelengths while undergoing simultaneous EEG recording. The GEO avoidance, as determined during phase I, was strongly and positively correlated with delta (r = 0.931, p < 0.001; Figure 4A) and theta (r = 0.881, p < 0.001; Figure 4B) band frequencies at the C4 electrode, as well as alpha (r = 0.873, p < 0.001) band frequency at F3. Cats exhibiting higher odor avoidance (corresponding to those with higher pain and neuro-sensitization, Figures 1B and 1C) displayed elevated delta- and theta-band frequency activity in the somatosensory cortex during GEO exposure. Regarding light wavelength exposure, spectral analysis of red light revealed moderate and positive correlations with MI-CAT(V) score: functional impairments were correlated with theta (r = 0.726; p = 0.011; Figure 5A), alpha (r = 0.747; p = 0.008; Figure 5B), and beta (r = 0.709; p = 0.015; Figure 5C) band proportions (Z score normalized). No significant correlations were found for the other light wavelengths (green and blue) or with neuro-sensitization assessed by RMTS. The avoidance of GEO smell according to the mean power of frequency bands at C4 electrode (A) Avoidance of the GEO smell on the delta frequency band. A strong and positive Pearson correlation was revealed (r = 0.931, p < 0.001). (B) Avoidance of the GEO smell on the theta frequency band. A strong and positive Pearson correlation was revealed (r = 0.881; p < 0.001). The EEG was performed under GEO 0.1% exposure while the avoidance score was obtained during phase I (n = 11 osteoarthritic cats). Data are represented as percentage of avoidance for GEO smell and mean of power spectral density for frequency band. Cats exhibiting higher odor avoidance displayed elevated delta and theta-band frequency activity in the somatosensory cortex during unpleasant smell. GEO, grapefruit essential oil. The functional impairment according to frequency band proportions during red light exposure (A) MI-CAT(V) score according to theta frequency band proportion (Z score normalized). A moderate and positive Pearson correlation was reported (r = 0.726; p = 0.011). (B) MI-CAT(V) score according to alpha frequency band proportion (Z score normalized). A moderate and positive Pearson correlation was revealed (r = 0.747; p = 0.008). (C) MI-CAT(V) score according to beta frequency band proportion (Z score normalized). A moderate and positive Pearson correlation was reported (r = 0.709; p = 0.015). Data are represented as mean for MI-CAT(V) total score and Z score normalized proportion for frequency band. MI-CAT(V), Montreal instrument for cat arthritis testing for use by veterinarians. Effect of light wavelengths on neuro-sensitization (QST and EEG) Since spectral analysis was sensitive and green light showed effectiveness after 4 days in phase II, a subgroup of OA cats (n = 6) was exposed to green light for a shorter duration (1 h for 4 consecutive days). As a comparison, cats were also exposed to white (corresponding to artificial light) and blue-enriched white light (corresponding to natural light). Neuro-sensitization of these cats was determined using RMTS. At pre-exposure, the RMTS median of stimulations accepted was 7.00 [6.00–9.00]. It increased by 85% [14%–250%] after 4 days of green light exposure compared to that of white (p = 0.046; Wilcoxon signed-rank test) and blue-enriched white light (p = 0.041; Wilcoxon signed-rank test), corresponding to 67% responders among 6 cats (Figure 6). The influence of light condition on central sensitization Data are represented as mean (SD) for RMTS score. Osteoarthritic cats (n = 6) exposed to green light for four days decreased their neuro-sensitization compared to white light (p = 0.046) and blue-enriched white light (p = 0.041) exposure using Wilcoxon signed rank test. EEG was recorded at pre-exposure and after the end of each light exposure. Analysis of the aperiodic spectral components (i.e., spontaneous cortical activity) revealed significant effects of both channel and light condition on the exponent and offset parameters (type III ANOVA, p < 0.001). Post hoc comparisons corrected by Tukey’s honestly significant difference test indicated that both the exponent and offset were significantly higher under blue-enriched white light and green light exposures compared with those under white light and the pre-exposure condition (p < 0.025). This increase in offset supported a uniform shift of power across frequencies, while the increase in exponent indicated higher inhibitory currents compared to the excitatory ones, supporting the robustness of the parameterization. For the periodic components (i.e., rhythmic and oscillatory potentials), normalized frequency band power (Z score relative to pre-exposure) showed significant main effects of light condition and a significant interaction between light condition and frequency band (type III ANOVA, p < 0.002), with no significant differences across channels. Post hoc analyses corrected by Tukey’s honestly significant difference test revealed an overall reduction in frequency band power for white, blue-enriched white, and green light exposures relative to pre-exposure (p < 0.043; Figure 7). Frequency band proportions for each light conditions recorded in osteoarthritic cats An overall reduction in frequency band power for white, blue-enriched white, and green light exposures relative to pre-exposure (p < 0.043) were revealed for osteoarthritic cats (n = 6). Post hoc analyses were corrected by Tukey’s honestly significant difference test. Specifically, beta and gamma band power were reduced following blue-enriched white and green light exposures compared to that of pre-exposure (p < 0.045). In contrast, delta and theta power increased following green light exposure relative to both pre-exposure and white light (p < 0.004). Delta power was also significantly increased for blue-enriched white versus white light exposure (p = 0.011) but not versus pre-exposure. No significant differences were observed for alpha band power. See Table 3 for all post hoc comparisons. Table 3Post hoc comparisons between the frequency band proportions recorded in osteoarthritis cats–WhiteGreenBlue-enriched whiteDeltaPre-exposureNSp < 0.001NSWhiteNSp < 0.001p = 0.011ThetaPre-exposureNSp = 0.004NSWhiteNSp < 0.001NSBetaPre-exposureNSp = 0.029p = 0.006WhiteNSNSNSGammaPre-exposureNSp = 0.045p = 0.004WhiteNSNSNSNS, non-statistically significant (p ≥ 0.05). Post hoc comparisons between the frequency band proportions recorded in osteoarthritis cats NS, non-statistically significant (p ≥ 0.05). Discussion This study has demonstrated that chronic OA pain can affect multiple sensory systems in cats. Importantly, sensory impairments were not related to deafness, and no concomitant health issues to explain any olfactory loss were identified. Cats in the mild MI-CAT(V) cluster perceived the unpleasant odor less repulsive than those in the moderate OA cluster, suggesting the olfactory testing as a sensitive tool for assessing OA pain severity. The first hypothesis was confirmed: similarly to humans,24,41 feline sensorial perception is measurable and correlates with central sensitization and OA functional impairment. Various reasons may explain the change observed in perception. Multisensory integration across tactile, auditory, visual, and olfactory stimuli was reported,42,43 and the involved cerebral areas overlap with those of pain perception.14,44 Higher sensory hypersensitivity scores were associated with greater odds of presenting two or more pain-related comorbidities. Of note, human OA patients exhibited higher scores than healthy controls.45,46 In a pilot study, feline response to pleasant odors was highly variable, leading to the use of an unpleasant odor. This is consistent with human findings: elderly individuals exhibit reduced hedonic perception of pleasant odors, but not of unpleasant odors, compared with young healthy individuals.47 In the phase I of the present study, the olfactory sensitivity increased with the OA severity, similarly to what was recently observed with PWT-assessed neuro-sensitization and the feline OA severity.48 Vasoactive-intestinal-peptide-expressing inhibitory interneurons present in both the olfactory bulb (inhibiting mitral cells that transmit olfactory information to higher cortical regions)49 and the olfactory cortex (inducing neuronal hyperexcitability) had increased activity in a rodent neuropathic pain model.50,51,52 Assessing the activity of these neurons in OA pain could clarify the influence of neuro-sensitization on olfactory sensitivity. The higher tolerance of neuro-sensitized cats to auditory stimulation observed during phase I can be explained by an inhibition phenomenon. Somatosensory stimulation (either nociceptive or not) suppresses auditory responses at several levels of integration, from the thalamic nucleus to the auditory cortex.53,54 Neuro-sensitized cats had a lower tactile perception threshold (PWT), meaning increased spontaneous transmission of tactile stimuli, which in turn can reduce auditory responses. This mechanism suggests a prolongation of the off response, referring to neural activity elicited at the end of a sound stimulus and originating from the brainstem (post-inhibitory rebound/facilitation) or the auditory cortex, during chronic pain.55 These findings can explain the variability in response among “nuisance” cats in their ultrasound sensitivity,36 potentially attributable to neuro-sensitization, as no health conditions were reported. Furthermore, presbycusis (i.e., age-related changes in hearing) does not seem to develop in older cats based on BAER results and available literature, reducing the likelihood of a confounding effect.56 The phase I provided evidence of olfactory and auditory changes in feline OA, suggesting innovative sensory-based diagnostic approaches. The first part of phase III demonstrated spectral analysis to be feasible and interpretable among conscious cats. Cats exhibiting higher odor avoidance were more neuro-sensitized, with increased delta, theta, and alpha band frequencies during GEO exposure. Interestingly, in human studies, increased arousal was reflected by reduced alpha oscillations in frontal and parietal regions,57 suggesting that feline EEG changes were not solely related to an unpleasant olfactory response but rather reflected a more complex pain-related phenomenon. Severe joint degeneration, for human knee OA, has been associated with higher alpha and beta frequency bands.11 However, published results concerning the theta activity were less clear, with either a negative or a positive correlation with pain severity in OA patients and negative correlation in horses with muscular tension along the spine.11,58,59 Also, both delta and theta were increased in the central region during low back pain in humans.60 Regarding the origin of these frequency bands, variations can be explained by the involvement of inhibitory-EC implying LC, PAG, and RVM.61 Notably, delta oscillations can originate from PAG,62 while LC norepinephrine neurons are involved in the generation of theta oscillations.63 A dysfunctional inhibitory-EC was associated with increased delta13 and theta12 activities in frontal and central brain regions in humans. Similarly, in humans,13 in a surgical OA rat model,64 and as suggested for cats and dogs,2,65 patients experiencing severe pain showed over time a reduced effectiveness of their inhibitory-EC mechanisms, reflecting a state of fatigue and resulting in greater pain facilitation.2,13,64,65 This supports the hypothesis that cats experiencing higher pain levels can exhibit dysfunctional inhibitory-ECs, which can be recorded during unpleasant olfactory stimulation, supported by increased delta and theta activity, suggesting that this test could be specific to the facilitatory/inhibitory-EC imbalance. The spectral analysis feasibility using visual stimulation revealed a positive correlation between feline OA pain and theta, alpha, and beta frequency bands following red light exposure. As cats are dichromats (no red photoreceptors),66 red light is considered neutral and likely reflects their condition without any analgesic effect. While delta and theta seemed associated with inhibitory-EC functionality, the beta band appears to be a marker of feline OA pain, with higher activity in cats with greater functional impairments (increased MI-CAT(V)). In phase II and III, 50%–67% of OA cats responded to narrow-band green light with an anti-hyperalgesia effect. This is consistent with results from a rat model of neuropathic pain, where green photoneuromodulation for 5 days reduced neuro-sensitization.27 In chronic migraine and most fibromyalgia human patients, green exposure for 10 weeks diminished pain perception, but neuro-sensitization status was not assessed.28,29 In the present study, considering neuro-sensitization status, 80% of non-allodynic cats responded positively, whereas allodynic cats’ response was much more limited. As mentioned, hypersensitized cats may have a dysfunctional inhibitory-EC, and a dysfunctional inhibitory-EC was predictive of poor analgesia and pain persistence after arthroplasty.67 Thus, the most sensitized cats probably experienced an inhibitory-EC fatigue that led to lower green light exposure effectiveness. This finding supports the development of personalized therapy according to inhibitory-EC functionality and degree of neuro-sensitization in feline chronic pain condition. Green light exposure was characterized by a decrease in beta and gamma frequency bands, suggesting reduction in pain, stress, and alertness,11,12,68 and an increase in delta and theta frequency band proportions reflecting an increase in relaxation compared to the white light and pre-exposure conditions.69 This increase in theta and decrease in gamma frequency bands also suggest a reduction in anxiety.70,71 Previous studies in humans affected by migraine72 and rats27 aligned with these findings and demonstrated a reduction in anxiety following narrow-band green light exposure. Cats are particularly prone to anxiety, which can complicate clinical assessment. Implementing green light in clinical settings could be of considerable interest in supporting fear-free practices. Blue-enriched white light was also associated with a decrease in beta and gamma frequency bands compared to pre-exposure, but probably not involving similar mechanisms (since it did not affect delta and theta waves) to those of the narrow-band green light.73 The reduction in beta band activity (proposed to be a pain marker) after green light exposure suggests its sensitivity to treatment, and it can be relevant to explore its relationship with MI-CAT(V) sensitivity. Interestingly, anti-hyperalgesia led to increased delta activity as observed following opioid administration.10 This is in line with green light exposure mechanisms, where the anti-hyperalgesia effect depended on descending inhibitory-EC (e.g., via opioid pathways).27,74,75 Synaptic plasticity, neuro-inflammation, and central sensitization were reduced,74,75,76 but no effect was noted on joint peripheral sensitization.30,31 Based on this mechanism of action, green light exposure may be sufficient to reverse neuro-sensitization by activating inhibitory-EC and counterbalancing facilitation. This phenomenon was demonstrated in the aperiodic spectral component analysis, with an increased exponent reflecting higher inhibitory activity.77 However, into the periodic spectral component analysis, delta and theta (proposed to be markers of dysfunctional inhibitory-EC) were increased and not decreased as expected. The green light anti-hyperalgesic effect could relieve the inhibitory-EC activity, thus leading to maintained delta and theta activities. Indeed, the analgesic effect of green light is mediated through the visual system rather than the integumentary system,27 and the effect depends mostly on the cones.78 Green light activates glutamatergic projections from the secondary visual cortex to GABAergic neurons in the anterior cingulate cortex and allows for the inhibition of glutamatergic neurons locally.79 This suggests that the green light exposure decreases the pain unpleasantness and may lead to an increase in relaxation reflected by an increase in delta and theta activity.69 Further studies are needed to elucidate the green light effect on inhibitory-EC and cortical integration using EEG tracing, as well as its potential as a unique or complementary therapeutic strategy for at-home, clinic, and shelter use.80 This study provided evidence of sensorial perception changes in cats with chronic pain, suggesting that OA should not be viewed merely as a joint disease. It supports updates in diagnostics, such as developing innovative sensorial tests, and management strategies. Olfactory assessment seems particularly promising in distinguishing inhibitory-EC functionality. Green light exposure emerges as an affordable translational non-pharmacological strategy for chronic pain management. Finally, the identification of delta and theta band frequencies as markers of inhibitory-EC functionality, and the beta frequency band as a potential marker of feline chronic pain, necessitates further investigation. Limitations of the study Limitations include the absence of healthy controls in the auditory and olfactory assessments. However, the range of neuro-sensitization among OA cats yielded significant results, supporting sensitivity. Age-matched controls could have elucidated the effects of aging on multisensory integration, and including a larger number of young OA cats would have been particularly relevant. Indeed, aging was associated with multisensory integration changes but often with incomplete health status descriptions.81 Olfactory impairments, considered prognostic factors for cognitive impairment, have been identified in human and canine aging studies.82,83,84 Additionally, an increase in delta and theta power has been associated with human aging and poor cognition.11 However, in dogs at risk of canine cognitive dysfunction, a positive correlation has been shown with alpha power.85 Canine musculoskeletal disease was correlated with impaired cognitive functioning, but neuro-sensitization was not tested, as human chronic pain neuro-sensitization showed no association with cognitive impairment.86,87 Assessing feline cognitive function during chronic pain and its association with sensorial perception and cortical changes would be an interesting avenue in future studies. Regarding light wavelength exposures, assessing the persistence of the green light’s anti-hyperalgesia effect would be beneficial, along with examining possible sex differences in the response, as observed in rats.75 Although all light illuminances were set to 100 lux, the blue-light condition delivered lower irradiance than the green and red lights. This supports the testing of the large-band blue-enriched white light with superior irradiance (Table S1). In future studies, irradiance may represent a more appropriate measure for comparing different wavelengths. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Eric Troncy (eric.troncy@umontreal.ca). Materials availability This study did not generate any new unique reagents. Data and code availability •All data reported in this paper will be shared by the lead contact upon request. All data reported in this paper will be shared by the lead contact upon request. •This paper does not report original code. This paper does not report original code. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Acknowledgments The authors thank all the employees at the ArthroLab Inc. and at the Farm Division and Animal Facilities (FANI) of the Faculty of Veterinary Medicine – Université de Montréal for their technical assistance and their tender loving care to the cats. They also thank the PIMUS (Plateforme Instrumentation Mécanique de l’Université de Sherbrooke) for providing their experimental infrastructure and technical resources. The research was supported by a grant from the Morris Animal Foundation (D22FE-301; A.C.). The GREPAQ is supported by a Discovery grant (RGPIN #05512-2020; E.T.) and a Collaborative Research and Development grant (RDCPJ #491953–2016; E.T.) supporting operations and salaries from the Natural Sciences and Engineering Research Council (NSERC) of Canada, as well as by an ongoing New Opportunities Fund grant (#9483; E.T.) and a Leader Opportunity Fund grant (#24601; E.T.), supporting pain/function equipment from the Canada Foundation for Innovation. A.D. was the recipient of a Fonds de recherche Québec – Nature et Technologies (FRQNT) doctoral grant (#350507). G.D. was supported by the Institute for Data Valorization, Montreal and the Canada First Research Excellence Fund (IVADO; #CF00137433), as well as the Fonds de recherche du Québec (FRQ; #285289). Author contributions Analysis and interpretation, A.D., A.C., G.D., and E.T.; conceptualization, A.D., A.C., F.S., J.M.-P., J.-P.P., and E.T.; data curation, A.D., A.C., G.D., C.O., M.H., and E.T.; funding acquisition, A.C., J.M.-P., J.-P.P., and E.T.; investigation and data acquisition, A.D., A.C., C.O., and E.T.; methodology, A.D., A.C., C.O., F.P., M.H., N.L., N.Q., P.G., M.M., and E.T.; project administration, M.M., J.M.-P., J.-P.P., and E.T.; resources, A.C., G.D., B.L., F.P., M.H., N.L., N.Q., P.G., J.M.-P., J.-P.P., and E.T.; supervision, A.C., J.M.-P., J.-P.P., and E.T.; validation, A.C., C.O., M.M., and E.T.; visualization, A.D.; writing – original draft, A.D., A.C., and E.T.; writing – review and editing, A.D., A.C., G.D., F.S., C.O., B.L., F.P., M.H., N.L., N.Q., P.G., M.M., J.M.-P., J.-P.P., and E.T. All authors have read and agreed to the published version of the manuscript. Declaration of interests The authors declare no conflict of interest. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERExperimental models: Organisms/strainsDomestic cats (Felis catus)ArthroLab Inc.N/ASoftware and algorithmsSigmaPlot© v.12.0Systat Software Inc, San Jose, CA, USAhttps://grafiti.com/sigmaplot-detail/IBM SPSS Statistics© v.31.0IBM Corp., Armonk, NY, USAhttps://www.ibm.com/products/spssMNE PythonGramfort et al.88https://www.python.org/ Experimental model and study participant details Animal selection and housing The study was approved by the Institutional Animal Care and Use Committee of ArthroLab Inc. (#A196-ART22F) and of Faculty of veterinary medicine – Université de Montréal (#CEUA-Rech-2114). Management of animals adhered to the regulations outlined in the Canadian Council on Animal Care, Guide to the Care and Use of Experimental Animals, Vol. 1, 2nd Edition (1993, revised in Feb. 2017), and complied with the US Dept. of Agriculture Animal Welfare Act (9 CFR Parts 1–3). All procedures also adhered to the ARRIVE guidelines for reporting animal research and the Committee for Research Ethical Issues of the IASP guidelines. Adult neutered healthy control cats (n = 6, 2 females; [1–4 years]; [3.57–5.15 kg]) or affected by OA chronic pain (n = 32, 22 females; [3–16 years]; [3.32–6.15 kg]) were enrolled in the study. All cats were healthy (except for OA diagnosis) based on complete blood count, biochemistry panel and normal physical and neurological examinations performed by the medical team, and a board-certified Veterinary Neurologist (A.C.), respectively. The OA diagnosis was based on radiographic screening of both forelimbs (carpus, elbow, shoulder) and hindlimbs (tarsus, stifle, hip) joints, as previously described.89 Radiographs were read and scored independently and blindly by a board-certified Veterinary Surgeon (B.L.). All cats were selected based on their behavioral compliance. Cats were group-housed in lighting-, temperature-, and humidity-controlled rooms with environmental enrichment (perches, beds, scratching posts, and toys). Cats were fed twice daily with veterinary quality commercial foods according to the manufacturer’s recommendations and fresh water was available ad libitum. Study timeline Cats were acclimated to the evaluators (A.D., A.C. or E.T.) and materials of all subsequent evaluations for a minimum of four acclimation sessions per outcome. A first cohort of OA cats (n = 11) was assessed using functional [MI-CAT(V)] and neuro-sensitization assessments (QST and EEG) exposed to innovant sensorial testing (Phase I). The aim was to characterize sensorial perception according to various degrees of neuro-sensitization, expected from a large amplitude of functional and radiographical alterations severity in this sample of cats. The second phase (n = 6 healthy and n = 24 OA) included neuromodulation by light narrow-band wavelengths tested by neuro-sensitization assessments (QST), while the third phase (n = 6–11 OA) aimed to investigate neuromodulation by spectral analysis (EEG) during sensorial stimulation (Table 1). Phase III included animals from Phase I, several days separated the phases to prevent habituation to stimuli. Functional [MI-CAT(V)], neuro-sensitization assessments (QST and EEG), innovant sensorial testing and neuromodulation evaluation are described in the following sections. Method details Functional and neuro-sensitization assessments Functional alteration - Montreal instrument for cat arthritis testing for use by veterinarians (MI-CAT[V]) The MI-CAT(V) is a reliable, specific, and sensitive feline OA pain scale.38,39,90,91,92 It is composed of four standardized sub-sections: body posture, gait, obstacles, and global distance assessments. The MI-CAT(V) total score is expressed in percentage of functional alterations. Discriminatory ability has been validated, with four clusters of OA severity: absent (<9%), mild (9–20%), moderate (21–35%), or severe (>35%).38,39 One evaluation was conducted by an expert in the field (E.T.) during Phase I on n = 11 OA cats. Peripheral sensitization - paw withdrawal threshold (PWT) Punctate tactile sensitivity was assessed through PWT as previously described using an electronic von Frey esthesiometer with a polypropylene probe (Rigid Tip 0.7 mm2 of surface 28G, IITC Life Science, Woodland Hills, CA, USA) applied to the plantar/palmar paw surface.2,40,89,93 A session included four PWT measurements conducted in duplicate (separated by a 4-min interval) in a randomized order to prevent habituation or anticipation. Measurements below 2 g were considered unreliable and discarded, and a cut-off of 200 g was imposed to prevent injury. A reliability number was attributed following a Likert scale (0: poorly reliable, 1: moderately reliable or 2: highly reliable). The lower the threshold, the more the cat is considered peripherally sensitized. One session was conducted during Phase II on n = 6 healthy and n = 24 OA cats. The average of the eight values, weighted by the reliability number, was used for data analysis. Allodynic (i.e., hypersensitized) and non-allodynic cats were determined based on a predetermined allodynic threshold, i.e., first quartile values for the forelimbs and hindlimbs (<89.15 g and 136.40 g respectively).40 Cats were classified as allodynic if at least two of their eight total measurements (forelimb and hindlimb) fell below this threshold. Central sensitization - response to mechanical temporal summation (RMTS) Mechanical temporal summation of pain was assessed using a custom-made mechanical stimulator (Top Cat®, Bespoke Measurement Systems, Cambs, UK)94,95,96 as previously described to assess spinal sensitization.2,4,93 The device applied repeated sub-threshold mechanical stimuli – each insufficient to elicit a pain response individually – at a fixed intensity (4 N), frequency (0.38 Hz) and duration of stimulation (1.3 s). The evaluator stopped the stimulus application upon observing avoidance behaviors such as vocalization, agitation, striking at the band, or withdrawal of the limb. If no avoidance behavior occurred, stimulation automatically ceased after a maximum of 30 repetitions. Each session was conducted in duplicate, with a 4-min interval between trials to ensure the cat remained in a positive, enriched environment. A reliability number was attributed following a Likert scale (0: poorly reliable, 1: moderately reliable or 2: highly reliable). The lower the number of accepted stimulations was, the more the cat was centrally sensitized. Two sessions, one per week, were conducted during Phase I on n = 11 OA cats, and one before and after each light wavelength exposure during Phase II on n = 6 healthy and n = 24 OA cats; the average of the four values, weighted by the reliability number, was used for data analysis. Cortical integration - electroencephalogram (EEG) The feasibility of EEG analysis during sensorial exposure was previously described using evoked potentials.37 Other EEG analysis corresponds to spectral analysis; one of the aims of the present manuscript was to determine spectral analysis during sensorial exposure. The acclimation and EEG evaluations of conscious cats were performed as previously described using Stellate Harmonie EEG unit (Stellate Systems Inc., Montreal, QC, Canada).37 The selected electrode sites were as follow: frontal [F3 (left) and F4 (right)], central [C3 (left), Cz (medial) and C4 (right)], and parietal [P3 (left), Pz (medial), and P4 (right)]. Two additional electrodes were placed on both sides of the neck, one for the ground (on the left) and the other as a reference for the recording. The cat was placed in a carrier, either sitting or lying down, with minimal restraint; a crocheted hat allowed to maintain the electrodes in place even during cat movements. The electrode impedance was maintained below 10 kΩ to ensure the recording reliability. EEG recording and analysis (see details below) were conducted during Phase III on n = 6–11 OA cats. Phase I – Innovant sensorial testing Olfactory stimulation A grapefruit (citrus paradisi, Divine Essence®, Montreal, QC, Canada) essential oil (GEO) was diluted (0.1%) in mineral oil (Sigma-Aldrich®, Oakville, ON, Canada) and presented in a 2 mL Eppendorf tube.37 It was placed in a cubic mesh cage (70 cm each side) for 30 s where the cat was free to move and a behavioral score was attributed a posteriori using video analysis.36 A maximum of 8 sessions was performed on different days within 4 weeks. The behavioral analysis was binary; either the cat was attracted (score of 0; smells the tube without subsequent avoidance) or repelled (score of 1; does not approach the tube or avoids it after smelling it) at each assessment, the total score was summated and divided by the number of evaluations performed, an avoidance percentage was assigned to each cat. This test was used during the Phase I on n = 11 OA cats. Auditory stimulation Brainstem auditory evoked responses (BAER): The BAER were conducted using Sierra II Wedge/EP® system (Caldwell Industries, Kennewick, WA, USA) to ensure that the cats could hear properly.97 A click sound (500 stimulations) was delivered at 11.33 Hz and 70 dB in conscious animals using ear inserts (pediatric foam ear tips, Caldwell Industries, Kennewick, WA, USA), except for n = 1 that received one intramuscular dose of 0.005 mg/kg of dexmedetomidine (Dexdomitor® 0.5 mg/mL, Zoetis Canada, Kirkland, QC, Canada). A masking sound at 30 dB was applied in the opposite ear tested. The BAER was recorded using subdermal needle electrodes (Rhythmlink® Paired 13 mm, Columbia, SC, USA) placed on the cranial vertex (recording), on the mastoid (reference, rostral to the base of the stimulated ear) and between scapula (ground). When no response was detected at 70 dB, the test was repeated at 90 dB. The impedance was maintained below 10 kΩ to ensure the recording reliability. This test was used during the Phase I on n = 11 OA cats to determine hearing disorder and prevent bias for innovant sensorial testing. Response to auditory temporal summation: A specific set up was used to deliver auditory stimuli in a controlled and reproducible manner, in accordance with previous recommendation.98 The stimuli were composed of sine waves bursts, generated using LabVIEW via a dedicated virtual instrument (VI) application to control frequency (between 100 Hz and 70 kHz), amplitude and duration of the sine burst. This VI controlled a function generator (National Instruments USB X-Series 6343, Austin, TX, USA) at a sample rate of 250 kHz, connected to a voltage amplifier (TS250 Accel Instruments, Irvine, CA, USA) to drive a large bandwidth loudspeaker (TangBand T1-1942SB, Taiwan, ROC) mounted on a 30 cm baffle and positioned outside of the cage. This baffled configuration ensured a semi-infinite acoustic field above 2 kHz, maximizing acoustic sound pressure level (SPL) in the cage. The system was calibrated using a broadband 1/8″ pressure microphone (G.R.A.S. 46DE, Holte, Denmark) placed at the center of the cage, ensuring that all stimuli were emitted at 80 dB SPL ±0.5 dB independently of the frequency. Additionally, SPL measurements were taken at eight other positions in the cage to characterize spatial variation, which ranged from 72.7 to 87.8 dB SPL depending on frequency. These measurements enabled the creation of frequency-specific acoustic maps to interpret the cat’s spatial behavior in response to auditory stimuli. The cat was placed in a cubic cage, free to move, and a behavioral score was attributed a posteriori using video analysis (unpleasant, neutral, or pleasant).36 An ultrasound considered unpleasant by most cats (25,600 Hz) was used, as determined in a pilot study assessing the responsiveness to 19 auditory frequencies from 100 Hz to 70 kHz. This was consistent with the literature where ultrasounds were used as repellent for foreign cats.36 The repeated exposure was performed as follows: the cat was free to move in the cage, the evaluator was not visible by the cat, but a video was recording in real time on the phone of the evaluator. The ultrasound was repeated at 0.25 Hz (1 s duration of sound) until an avoidance behavioral response was observed or the maximum number of stimulations (n = 30) was reached. Two sessions of evaluation were performed in duplicate with a 4-min interval, within a one-week delay. This test was conducted during the Phase I on n = 11 OA cats. Neuromodulation by visual stimulation Phase II – Effects of specific light wavelength on neuro-sensitization assessed by QST According to the literature on induced chronic pain rodent models,27 cats (n = 6 healthy and n = 24 OA) were exposed to 4 h of narrow-band green light during 8 consecutive days (from 6 a.m. to 10 a.m.) (Table S1). The light source was provided by a LED lamp (Konseen®, 240 LEDs, 45.7 × 15.2 × 3 cm) set at 4100 K with a 530 nm green filter (Lee Filters 139, Andover, United Kingdom). Cats were grouped by pairs in cages (in their hosting rooms). Neuro-sensitization assessments (RMTS) were performed in duplicate the day before the first exposure, at the fourth morning and at the eighth morning following light exposure. Phase III – Mechanisms of neuro-sensitization and neuromodulation assessed by EEG To investigate neuromodulation at the cortical level, spectral analysis of EEG traces was performed on OA cats (n = 11). EEG recordings were acquired during sensory exposure involving either olfactory or visual stimulation. Olfactory stimulation consisted of a single 20-s exposure to GEO (citrus paradisi, Divine Essence®, Montreal, QC, Canada) diluted at 0.1% (as previously reported37). Visual stimulations corresponded to a LED lamp set at 4100 K with one of the following three distinct color filters: blue (Roscolux filter R384, Rosco Laboratories Inc, Markham, ON, Canada), green (Lee Filters 139, Andover, United Kingdom), and red (Lee Filters 106, Andover, United Kingdom). These filters were selected to avoid wavelengths spectral overlap (Figure S1). Green and blue narrow-band light were chosen for their potential anti-hyperalgesic effects,27 whereas red light was used as a control, given that cats are dichromatic.66 The order of presentation was random; each stimulation last 20–30 s with 1–2 min intervals in-between stimulations.37 The LED lamp was positioned facing the cat to deliver an illuminance of 100 lux (as measured with the filters) in a dark room (see Table S1 for a comparison of lights characteristics). Three sessions were conducted; the three recordings of the same sensorial exposure were presented on the same tracing to optimize subsequent spectral analysis. Based on Phase II results, neuromodulation using green light was tested over exposure period of 4 and 8 days, demonstrating efficacy as early as day 4. A subgroup of OA cats (n = 6) was enrolled to determine if green light can induce central changes following a shorter exposure duration. Narrow-band green light effect was compared to large-band so-called natural (6000 K – blue enriched white) or artificial light (4100 K – white). Cats were exposed for 1 h (from 6 a.m. to 7 a.m.) during 4 consecutive days in the following order: white, green, and blue enriched white light (Table S1). Cats were placed by pairs in cages (in their hosting rooms to minimize stress), with illuminance again at 100 lux, and the exposure was performed in the dark. Neuro-sensitization QST and EEG assessments were performed in duplicate before (i.e., pre-exposure) and once on the fourth morning following the end of the exposure. Spectral analyses were performed on resting state EEG recorded for 1.62 [0.30–3.45] minutes before the start of the light exposure and then on the fourth afternoon after each exposure to a specific wavelength. Quantification and statistical analysis Descriptive statistics, using the median [min–max], individual percentage of change and coefficient of variation were used to evaluate the data dispersion. According to data distribution (linearity, homoscedasticity) and sample size, the non-parametric Mann-Whitney U test or Wilcoxon signed rank test, and either the Spearman rank test (RhoS, ρS) or Pearson (r) coefficient of correlation were used without correction. Intraclass correlation coefficient (ICC) estimates and their confidence interval at 95% (CI95) were calculated using a mean-rating, absolute agreement and 2-way mixed effects. The interpretation of the coefficient of correlation and ICC was as follows: value <0.50 = weak, 0.50–0.75 = moderate, 0.75–0.90 = good and >0.90 = excellent correlation or reliability.99,100 The chosen sample size (n = 6–24) was appropriate for an exploratory study, aiming to identify variables of interest and explore behavioral pattern when exposing cats to sensory stimuli.101 According to this limited sample size, only technical outliers were considered if necessary, and no subgroup analysis by sex or age was conclusive. The significance threshold was set as α = 0.05 indicated by an asterisk (∗) in the Table 2. Statistical details are indicated in Figure legends. Analysis was performed using SigmaPlot© v.12.0 (Systat Software Inc, San Jose, CA, USA) and IBM SPSS Statistics© v.31.0, (IBM Corp., Armonk, NY, USA) statistical softwares. Spectral analysis was performed with MNE Python.88 EEG data were notch-filtered at 60 Hz to suppress power line artifacts, down-sampled to 128 Hz, and visually inspected for noise and non-physiological artifacts. Power spectral density (PSD) was computed using Welch’s method on 2-s epochs with 50% overlap and averaged within canonical frequency bands: Delta (1–4 Hz), Theta (4–8 Hz), Alpha (8–12 Hz), Beta (12–20 Hz), and Gamma (20–60 Hz). PSD values were z-scored relative to either the red light or pre-exposure condition. Inferential analyses employed linear mixed-effects models with condition as a fixed factor and participant as a random intercept. Pairwise post-hoc comparisons were corrected using Tukey’s honestly significant difference test. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.117340. Supplemental information Document S1. Figure S1 and Tables S1 and S2 References Associated Data Supplementary Materials Document S1. Figure S1 and Tables S1 and S2 Data Availability Statement •All data reported in this paper will be shared by the lead contact upon request. All data reported in this paper will be shared by the lead contact upon request. •This paper does not report original code. This paper does not report original code. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.