Thirteen male zebra finches (~100 days old) were kept in outdoor aviaries at the University of Nevada, Reno with no previous exposure to ALAN. When they were ~ 140 days old, we moved them to individually housed indoor 47 cm × 31 cm × 36 cm cages and entrained them to 12 h light and 12 h dark (12 L,12D) for 4 weeks. For daylight, we used 1.4-Watt 5,000 K light emitting diode (LED) rated at 95 Lumens lights at 0:00 (zeitgeber time (ZT) 0) and lights off at 12:00 (ZT 12). Birds were given food and water ad libitum. Each cage contained a mechanized perch that relayed hop activity to MATLAB every minute. Cages had individual light-occlusion shades and constant white noise in the background to limit visual and acoustic cues.

We video recorded 30 min of behavior 90 min before perfusion and activity via automated perches (Alaasam et al., 2018). An observer blind to the treatments determined time spent eating/drinking, grooming, hopping, or no movement for the video recordings. We conducted a power analysis based on previously collected behavior data from control and ALAN exposed birds and determined that at least 3 birds were needed per treatment group (Power = 0.8, α = 0.05, effect size = 2, number of groups = 3).

Birds were randomly assigned to one of 3 conditions: control night (12 h light: 12 h dark; 12 L:12D sacrificed at dark night: ZT 14, n = 4), control day (12 L:12D sacrificed at day: ZT 10, n = 4), and experimental ALAN (12 h light: 12 h dim light; 12 L:12Ldim sacrificed at night with artificial light: ZT 14, n = 5). We chose the control day timepoint as close as possible to the night timepoints to be certain in capturing awake birds but also avoiding larger differences in circadian activity. As determined by One-Way ANOVA, groups did not differ in initial mass (p = 0.62). After the 4-week entertainment period, we sacrificed the control night group during the dark period (ZT 14) and the control day group during the light period (ZT 10). We sacrificed individuals in the ALAN group 2 h after the bird’s 1^st exposure to ALAN (ZT 14), to obtain peak protein expression and avoid overlap from the light period. ALAN was standardized to around 1.5 lux ±0.01 from a 20 cm × 1.5 cm 5000 K broad spectrum LED strip. This was done with an Extech Easyview Digital Light Meter (model EA13) and lux was calculated using a mean measurement at perch height and two opposing base corners. For a full-spectrum description of the lights, please see (Alaasam et al., 2021).

We anesthetized birds with 0.1 ml of anesthesia made from 30 mg Ketamine HCl, 105 mg Xylazine, and 8.25 ml saline. After no response to a hard toe pinch, weight was taken, and we perfused birds with 1X PBS for 5 min and 4% paraformaldehyde in 1X PBS (PFA) for 13 min. Brains were removed, left in 4% PFA for 24 h, switched to a 15% sucrose solution for 4–12 h, followed by a 30% sucrose solution overnight, and then flash frozen with powdered dry ice and stored in −80°C until slicing.

We cut the left hemisphere of the brain sagittally at 45 μm thickness in six series. Series 1 was stained for imaging (total slices analyzed: ALAN = 90, control day = 67, control night = 67) and 2–6 were stored in cryoprotectant (3.3% sucrose, 0.01% Polyvinyl-pyrrolidone (PVP-40), 30% ethylene glycol in 0.1 M Phosphate Buffer) in −80°C.

We incubated brain slices in a blocking solution (4% BSA, 0.4% triton, 0.05% Na-Azide in 1x PBS) for 3 h and then with primary antibody diluted in blocking solution (c-Fos anti-rabbit polyclonal from ABCAM (ab190289) diluted 1:1000, ZENK anti-mouse monoclonal received from Dr. Keays’ lab (Nordmann et al., 2020) diluted 1:300) in 4°C for ~46 h. We washed slices 3 times in 1X PBS for 25 min each at room temp and incubated overnight at 4° C with secondary antibodies (anti-rabbit 488 (from ABCAM ab150081) diluted 1:1000 and anti-mouse 594 (from ABCAM ab150116) diluted 1:1000), protected from light. We then incubated slices with DAPI for 15–25 min at room temperature and washed them in 1X PBS 3 times. Slices were mounted with antifade mounting medium (VECTASHIELD®) on slides. We imaged tile scans of full slices within 1 week of mounting on a Leica TCS SP8 confocal microscope.

We analyzed images on ImageJ. We determined brain regions using anatomical locations with DAPI staining and a reference atlas from zebrafinchatlas.org. Cells were determined positive for cFos or ZENK if they were three times the mean brightness and overlapped with DAPI. We divided the number of positive IEG cells by the total DAPI cell count to determine expression percentage in representative areas measured over several slices.

We performed statistical analyses in R, version 4.1.2 (R Development Core Team, 2011). We ran generalized linear mixed-effect models to assess if IEG expression levels were affected by the treatment group as a fixed effect (lme4 package). Slice number and bird ID were included as random effects. We used a Kruskal-Wallis test to analyze the interaction of behaviors and treatment groups. We ran a correlation matrix for all brain regions in each treatment for both cFos and ZENK (Supplementary Figure S1).

All procedures were conducted in accordance with the National Institute of Health Ethical Use of Animals and approved by the University of Nevada, Reno Institutional Animal Care and Use Committee.

There was a significant difference in the interaction between behavior and treatment group (Kruskal-Wallis test: Grooming: chi-squared = 6.22, p = 0.05, Feeding: chi-squared = 11.51, p < 0.01, Hopping: chi-squared = 6.45, p = 0.04, Inactive: chi-squared = 7.20, p = 0.03). Hop activity measured via perch recordings also showed that birds had significantly lower nocturnal activity (control night) than daytime activity (control day) and nocturnal activity under ALAN (Kruskal-Wallis test: chi-squared = 6.47, p = 0.04, Figure 1).

The ALAN group was significantly different from the control night (cFos p = 0.027, ZENK: p = 0.037) but not the control day (cFos: p = 0.17, ZENK: p = 0.66) when all 24 brain regions were analyzed together (Figure 2A). We broke down the analysis by looking into two major pathways—motor and visual—as well as additional areas. There was no significant difference between cFos and ZENK expression between the ALAN group and either control in all combined areas analyzed in the motor pathway (cFos-Day: p = 0.40, cFos-Night: p = 0.14, ZENK-Day: p = 0.72, ZENK-Night: p = 0.14). Similarly, we saw no significant difference for all areas analyzed in the visual pathway (cFos-Day: p = 0.08, cFos-Night: p = 0.07, ZENK-Day: p = 0.61, ZENK-Night: p = 0.11). However, individual areas in both pathways were significantly different (Table 1; Supplementary Figures S2, S3).

To determine if the expression was based on activity, we reanalyzed expression with birds separated into only two groups of active (n = 7) or inactive (n = 6; total minutes of activity <1 min) 90 min before perfusion. Active birds included the control day group and non-active included the control night, with the ALAN group split between the two, based on activity. There was no significant difference in cFos or ZENK expression overall between active and non-active birds (cFos: z = 1.18, p = 0.24, ZENK: z = 1.70, p = 0.09). Additionally, there was no significant difference between active and non-active birds in the whole motor (cFos: z = 1.28, p = 0.20, ZENK: z = 1.81, p = 0.07) or visual (cFos: z = 0.65, p = 0.51, ZENK: z = 1.37, p = 0.17) pathways.

In the visual pathway, the ALAN group showed significantly higher cFos expression in the striatum adjacent to the core of the entopallium, posterior hyperpallium (Figure 2B), and ventral mesopallium adjacent to the core of the entopallium than the control night group, and significantly higher ZENK expression in the nidopallium adjacent to the core of the entopallium but lower in the core of the entopallium. The ALAN group also showed significantly higher cFos expression than the control day group in the striatum adjacent to the core of the entopallium and posterior hyperpallium areas (Table 1).

In the motor pathway, the ALAN group showed significantly higher cFos expression than the control night group in the anterior mesopallium dorsal (Figure 2B) and anterior mesopallium ventral regions and significantly higher ZENK expression in the anterior mesopallium dorsal and nidopallium caudolateral regions (Figure 3). The ALAN group also had significantly higher levels of cFos expression compared to the control day group in the anterior mesopallium dorsal and lateral int arcopallium and higher ZENK expression in the nidopallium caudolateral. However, the ALAN group had significantly lower ZENK expression in the anterior striatum, nidopallium adjacent to the basorostral nucleus, and ventral mesopallium adjacent to the basorostral nucleus (Table 1). There was no significant difference between active and non-active birds in the anterior mesopallium dorsal, nidopallium caudolateral, or nidopallium adjacent to the basorostral nucleus.

The ALAN group also showed higher cFos expression in the area parahippocampalis, medial dorsal mesopallium, entopallium (Figure 2B), and lateral ventral mesopallium and higher expression of ZENK in the caudal striatum, medial dorsal mesopallium (Figure 2C), entopallium (Figure 2C), and lateral ventral mesopallium as compared to the control night group, but lower levels of ZENK expression in the hippocampus (Figure 2C) and septum. The ALAN group also showed higher levels of cFos expression in the entopallium and higher ZENK expression in the area para-hippocampalis, medial dorsal mesopallium, and entopallium as compared to the control day group (Table 1).

In the visual pathway, control night birds (LD sacrificed during the night) were significantly different from control day (LD sacrificed during the day) and ALAN birds (LLdim sacrificed during the night). These large differences are to be expected as LD control night birds were inactive. However, we still found two areas had significantly stronger cFos expression for ALAN birds than both control groups: posterior hyperpallium and striatum adjacent to the core of the entopallium. ALAN was a novel visual stimulus for the birds, likely employing a visual neuronal response.

The entopallium, the most prominent area to emerge, was significantly different from both controls and both IEGs. The entopallium is involved in visual pattern recognition (Watanabe et al., 2008, 2011). Surprisingly, we found that birds exposed to ALAN had different IEG expression in visual pathways compared to day controls. We see that even very dim levels of ALAN (around 1.5 lux) elicit a clear response in recognizing this visual input.

Out of the seven regions of the motor pathway analyzed, ALAN birds were significantly different from the day controls in either cFos or ZENK in six of them. However, when accounting for activity, the ALAN group remained significantly different with increased expression in the anterior mesopallium dorsal and nidopallium caudolateral and decreased in the nidopallium adjacent to the basorostral nucleus. Although the nidopallium caudolateral has additional functions, the anterior mesopallium dorsal and nidopallium adjacent to the basorostral nucleus are differentially activated under ALAN and not associated with hopping. These areas may be picking up movement we did not track, such as head turns and flapping wings, or associated with other functions we are unaware of.

We found birds exposed to ALAN were significantly different from both controls in areas associated with learning and memory. The ALAN group had significantly higher IEG expression than the day and night controls in the area para-hippocampalis and medial dorsal mesopallium, which are involved in spatial and object recognition and associative learning, respectively, (He et al., 2010; Damphousse et al., 2022). The ALAN birds also had significantly lower IEG expression than the night controls in the hippocampus, which is involved in spatial memory and learning (Bingman et al., 1990; Mayer et al., 2013). Dim ALAN dampens behavioral measures of learning and memory which have also been correlated with structural alterations in the hippocampus (Taufique et al., 2018, 2019; Liu et al., 2022). Lower nocturnal IEG expression in the hippocampus may partially explain why dim ALAN suppresses gene expression in the hippocampus (Taufique et al., 2018, 2019). It is believed that sleeping activates the hippocampus for memory consolidation (Klinzing et al., 2019). Indeed, we see higher IEG expression in our control night birds than day. A nocturnal suppression of hippocampal activity may impair memory consolidation and learning under ALAN.

ALAN treatment birds had significantly higher IEG expression in the nidopallium caudolateral than either of the controls. This aligns with previous research that has found dim ALAN alters the neuroarchitecture of the nidopallium caudolateral, the avian equivalent of the prefrontal cortex (Gunturkun, 2005; Gunturkun and Bugnyar, 2016; Taufique et al., 2019). The nidopallium caudolateral has been implemented in mimicking prefrontal area structures by having the same receptor architecture as the Brodmann Area 10 in humans, which is involved in many processes including reward and conflict, working memory, and pain (Herold et al., 2011; Peng et al., 2018). IEG activation in areas associated with memory support previous findings that ALAN impairs learning and memory (Liu et al., 2022). Additionally, the avian nidopallium caudolateral along with the entopallium have been shown to display attentional mechanisms (Johnston et al., 2017), implying an alert state in our ALAN exposed birds.

Another association to emerge was pain processing. Dim ALAN has been shown to alter pain reception in mice (Bumgarner et al., 2020). ALAN treatment birds had significantly higher activity in the caudal striatum from night controls and significantly higher activity in the nidopallium caudolateral from both controls. Although not much is known about the avian caudal striatum, this area is related to anxiety and pain in mice (Jin et al., 2020). Additionally, the nidopallium caudolateral has been associated with the Brodmann Area 10 in humans, also involved in pain reception (Herold et al., 2011; Peng et al., 2018).

Birds under ALAN had significantly decreased IEG expression compared to the control night birds in the septum and hippocampus, which directly regulate hormones leading to downstream physiological changes (Carreras et al., 1984). The hippocampal-septal pathway regulates hormones involved in stress and immune function including; corticotropin-releasing hormone (Vale et al., 1981; Nagarajan et al., 2017), thyrotropin-releasing hormone (Fabris et al., 1995; Ballard et al., 1996), and corticosterone (Garces et al., 1968; Nyakas et al., 1979).

Our results show that ALAN typically increases IEG expression in differentially activated areas compared to both controls. However, reduction of ZENK expression in the septum and hippocampus implies reduced neuronal activation in co-regulated functions—such as hormonal control. This is supported by previous research that ALAN alters hormone production (Ouyang et al., 2018; Moaraf et al., 2020).

In summary, through fine analyses of IEG expression, we found that initial ALAN exposure activates brain areas involved in vision, movement, learning and memory, pain processing, and hormone regulation, which may be differentially regulated under prolonged sleep loss or long-term exposure to ALAN. Additionally, first time exposure to ALAN at a different time in the night may produce differential responses from those we observed. Although ALAN may not be eliciting changes through circadian regulation, we still see substantial responses across brain areas that warrant further study. ALAN creates a unique brain state that is significantly different from day or nighttime brain activity. Dim light creates a novel environment, different from birds active in the day or sleeping at night, which produced widespread differential brain activity.