2.1 Animals

All experiments were performed in accordance with guidelines provided by the National Institutes of Health and approved by the Institutional Animal Care and Use Committee at Minnesota State University, Mankato. Wild-caught male green anoles (A. carolinensis; NCBITaxon_28377) with snout–vent lengths (SVLs) of 64 mm (± 0.5 mm) were purchased during the NBS (September) from Candy's Quality Reptiles (Louisiana, USA). These lizards were housed individually in 39.5 × 22 × 26-cm glass terraria with peat moss substrate, sticks, and a screen cover upon arrival. Lizards were fed calcium-coated crickets twice a week, and cages were misted with water daily. Water in plastic containers was also available ad libitum. Animals were allowed to acclimatize to the laboratory NBS conditions for at least 1 week before experiments began (as in Tao et al. [2022]).

NBS conditions were maintained by exposing the lizards to a 10:14 h light:dark cycle with ambient temperatures ranging from 15°C at night to 24°C during the day as in Cohen and Wade (2011). Overhead UV lights emitted UVB radiation, and heat lamps (10°C above ambient) required for basking were switched on for 4 h during the NBS light cycle. All behavioral experiments were performed at least 1 h after the heat lamps turned on.

All the experiments were performed during the refractory period of the NBS (between October and November) (Tao et al. 2022). Presurgery behavior tests were conducted 1–2 days before performing the surgery, while the postsurgery behavior tests were done 3 days after surgery (Figure 1A). Fifteen animals were injected with staurosporine into the AMY, out of which six missed the target. One animal was removed from the study as it received an injection in the lateral ventricle, causing extensive damage throughout the brain. Ten additional lizards were injected with saline.

2.2 Stereotaxic Surgery

To accomplish amygdala lesions, male green anole lizards were first sedated using an inhalation anesthetic, isoflurane (Covetrus), and then moved to a bed of ice to ensure complete sedation (similar to Greenberg [1982]). The lizards were placed into a stereotaxic apparatus (Kopf Instruments) where they remained on ice. The lizards’ heads were kept in place via three points of contact. First, two ear bars were gently but firmly placed against the tympanic openings. Next, a clamp was affixed to the lizard's snout. Once the lizard was secured in the equipment, coordinates were established using the parietal eye as a zero point. Holes (0.8 mm in diameter) were drilled bilaterally at the following coordinates: lateral (X-axis) = ±0.9 mm, anteroposterior (Y-axis) = −0.55 mm, and vertical (Z-axis) = −3.05 mm (modified coordinates for the ventromedial amygdala based on Greenberg [1982]). Using a Hamilton syringe (5-µL syringe, 26s gauge needle), 100 nL of 2 µM staurosporine (in saline [0.9% NaCl]; ALX-380-014; Enzo Life Sciences; RRID:SCR_003900) or saline (control) was injected on both sides. Pilot studies revealed that this injection of staurosporine ensured that cell death was restricted to a region approximately the size of the anole ventromedial nucleus of the amygdala (AMY). The wounds were then closed by applying petroleum jelly and vetbond tissue adhesive (3 M). Lizards were returned to their enclosures and monitored for 3 days before further testing. Lizard mortality rate and body weights before and 3 days after surgery were examined to determine surgery burden.

Variations in lesion location as determined by Nissl staining (see below) resulted in the formation of two experimental groups: (1) hit AMY, where there was at least a partial lesion to the AMY, and (2) miss AMY, where no lesions in the AMY were seen (Figure 2). The control group was injected with saline. One animal that was injected with staurosporine was removed from the analysis as the injection occurred in the lateral ventricles, resulting in widespread and nonspecific damage to the brain. Two animals that were injected with saline displayed mechanical damage to the AMY resulting from the needle and were thus included in the hit AMY group for behavior analysis.

2.3 Resident–Intruder Aggression Tests

In order to examine aggressive behaviors of lizards pre- and postsurgery, a resident–intruder behavior test was performed (Greenberg et al. 1984; Lailvaux et al. 2004). In brief, the resident (focal) male green anole lizards were first allowed to acclimatize to the testing platform (resident cage placed in front of the camera, away from other animals) for 5 min. Focal males were exposed to intruders (size-matched male conspecific lizards) undisturbed for 15 min, and the encounter was videotaped (Canon VIXIA HF R72) for later analysis. The focal males were exposed to the same intruder during pre- and postsurgery behavior tests. Intruders were unmanipulated males that were exposed to two focal lizards (total of nine intruders) or one focal lizard (six intruders) during the entire study.

The number of occurrences of behaviors (Table 1) was quantified using Behavioral Observation Research Interactive Software (BORIS, v.8.21.8; RRID:SCR_025700) (Friard and Gamba 2016) by experimenters naive to treatment. The sum of aggressive and assertion behaviors (as indicated in Table 1) was calculated. Assertion behaviors were defined as display behaviors such as dewlap extension, pushups, and headbobs, which can indicate a heightened arousal state (Greenberg et al. 1984). We defined aggressive behavior as assertion behaviors and other aggressive behaviors, such as biting, chasing, or approaching the opponent, eyespot development, and nuchal crest formation (Korzan et al. 2021; Lailvaux et al. 2004). The latency to first aggressive behavior was also measured. The latency for males that did not produce any aggressive behaviors for duration of the test was set to 900 s for analysis.

2.4 Tissue Preparation and Histology

Animals were sacrificed via rapid decapitation within 30 min of postsurgery behavior recordings. Brains were removed and flash frozen in cold 2-methylbutane (Thermo Fisher Scientific) before being stored at −80°C until use. To determine the breeding status of the lizards, testes color and vas deferens were visually examined. NBS individuals present with small yellow testes and clear vasa deferentia, while BS individuals present with large white testes and milky vasa deferentia. Additionally, testes were weighed, and the gonadosomatic index (GSI) was estimated as a proportion of gonadal weight to the total body weight, which is reduced in NBS compared to BS animals (Tao et al. 2022).

Using a cryostat machine (Leica Biosystems), frozen brain tissues were cut coronally into 20-µm sections and thaw-mounted to positively charged slides (Superfrost Plus, VWR) (as in Tao et al. [2022]). To visualize the lesion site, standard thionin staining (Nissl's staining) was performed (as in Beck et al. [2008]) and locations were determined based on the lizard forebrain atlas (Greenberg 1982).

2.5 Hormone Analysis

Trunk blood samples were collected after decapitation, centrifuged to obtain plasma, and stored at −20°C until use. Plasma T levels were quantified using an enzyme-linked immunosorbent assay (ELISA) kit (ADI-900-065; Enzo Life Sciences; RRID:SCR_003900). This kit has been previously validated in this species (Kang et al. 2020). Samples were run in duplicate on the same plate, and manufacturer's instructions were followed. The intraassay CV (coefficient of variance) was 2.6%. T levels are reduced in the NBS compared to breeding animals and thus can also be used to confirm the breeding state of the lizards (Lovern et al. 2004).

2.6 Image Acquisition and Morphometrics

Images of the stained tissues were taken using Gryphax (Jenoptik) software on a Zeiss Axiostar Plus microscope and analyzed in ImageJ (NIH; RRID:SCR_003070). The total volume of the amygdala was estimated using a truncated cone formula from area values obtained after manually outlining the AMY boundary (Tramontin et al. 1998). The volume of damage within the AMY was also similarly calculated after tracing the site of the lesion. This was then normalized to the estimated total volume of the amygdala, and the percentage of the AMY remaining from each animal was calculated.

2.7 Statistical Analysis

The body weights of pre- and postsurgery lizards were compared using paired t-tests or Wilcoxon signed-rank tests as appropriate. These tests were also used to compare the number of aggressive behaviors, number of assertion displays, and latency to aggression between pre- and postsurgery behavior tests. Paired t-tests were performed if the related sample differences were normally distributed (as reported by Shapiro–Wilk test, p > 0.05); otherwise, a Wilcoxon signed-rank test was used. Cohen's d was used to estimate effect sizes (0.2 for small, 0.5 for medium, and 0.8 for large effect sizes) (Cohen 1988). Analysis of variance (ANOVA) with a Tukey's HSD post hoc test was used to test differences among groups for GSI, T levels, bodyweights, presurgery aggression, and change in aggression. GSI and T levels were also compared with previously reported data (Tao et al. 2022; Shankey et al. 2024; Kang et al. 2020). Levene's test was used to check for equality of variance (p > 0.05). If the data did not pass homogeneity assumptions, then a Kruskal–Wallis test was performed. Partial eta squared (η_p²) was used to calculate effect sizes (0.01 for small, 0.06 for medium, and 0.14 for large effect sizes) (Richardson 2011). Lastly, correlations between the percentage of the AMY remaining and change in aggression, as well as the relationship between T levels and aggression, were tested using Pearson's r correlation test.

All statistical tests were performed using SPSS 28.0.1.0 (RRID:SCR_016479). All statistical comparisons with p ≤ 0.05 were considered statistically significant, and the data are presented as mean ± SEM. Grubb's outlier test was used for determining statistical outliers that were then removed during analysis. Final sample sizes are included in the figures.

3.1 Lizards Demonstrated Minimal Surgery Burden and Displayed NBS Gonadal and Hormonal Characteristics

All lizards that underwent stereotaxic surgery survived and displayed no significant differences in body weight between pre- and postsurgery in the hit AMY (t(7) = 1.93, p = 0.096, d = 0.681), miss AMY (t(5) = 0.76, p = 0.481, d = 0.311), and saline (t(9) = 0.86, p = 0.414, d = 0.271) groups (Figure 3A). Additionally, there were no significant differences in body weights across groups presurgery (F_2, 21 = 0.41, p = 0.670, η_p² = 0.037) and postsurgery (F_2, 21 = 0.22, p = 0.808, η_p² = 0.020).

We detected no significant differences in the GSI across groups, and the GSI resembled NBS GSI values in unmanipulated green anole lizards (Tao et al. 2022) (F_3, 28 = 0.81, p = 0.502, η_p² = 0.079). Our GSI data were significantly reduced compared to previously reported BS GSI values (Shankey et al. 2024) (H(3) = 20.33, p < 0.001) (Figure 3B). Similarly, plasma T levels of lizards in this study did not differ significantly among surgery groups (H(2) = 2.06, p = 0.357), resembled the NBS state (H(3) = 5.29, p = 0.152), and significantly differed from the breeding state (H(3) = 13.42, p = 0.004) as described in a previous study in unmanipulated anoles (Kang et al. 2020) (Figure 3C). These results confirm that the lizards in our study were in the NBS state.

3.2 Aggression Levels Were Significantly Reduced After a Partial AMY Lesion

Two animals injected with saline displayed mechanical damage to the AMY caused by the needle and, thus, were included in the hit AMY group for behavioral analysis. We did not find any significant differences in presurgery aggressive behaviors across groups with (H(2) = 2.99, p = 0.224) or without (F_2, 19 = 2.98, p = 0.075, η_p² = 0.239) these AMY-damaged saline-injected animals in the hit AMY group. Most of the animals in the hit AMY group had a partial lesion to the AMY (mean damage: 16.85 ± 6.24%; range: 1.5%–38.5%). However, there was one animal with a complete bilateral AMY lesion (less than 25% of the AMY remaining; animal VIII in Figure 2). This animal also behaved differently than the rest of the hit AMY group, with a 240% increase in aggression after surgery (Figures 4A and 5), and drives a negative relationship between the percentage of the AMY remaining and aggressive behavior (r²(8) = 0.55, p = 0.014). A Grubb's outlier test identified this animal as a statistical outlier. When that animal is removed from analysis, there is no relationship between the percentage of the AMY remaining and aggressive behavior (r²(7) = 0.02, p = 0.692; Figure 4B). It is possible that a threshold exists such that partial and complete lesions induce different behavioral responses (see discussion below). Therefore, we excluded the animal with a complete lesion from further analysis. We also examined whether the percentage of the AMY remaining influenced the latency to first aggressive behavior and found no relationship between these variables with (r²(8) = 0.01, p = 0.740; Figure 4C) or without (r²(7) = 0.001, p = 0.950; Figure 4D) the outlier.

The number of aggressive behaviors (Table 1) significantly decreased postsurgery in the hit AMY group (t(8) = 4.43, p = 0.002, d = 1.481), which was not observed in either the miss AMY (t(5) = 0.25, p = 0.809, d = 0.104) or saline (t(7) = 0.04, p = 0.967, d = 0.015) groups (Figure 5A). The number of aggressive behaviors remained statistically decreased postlesion in the hit AMY group when the two saline animals with mechanically damaged AMYs were removed from the analysis (t(6) = 4.21, p = 0.006, d = 1.590). The hit AMY group was also statistically different from the other groups when comparing the percent change in aggression with (H(2) = 9.29, p = 0.010) or without (H(2) = 6.6, p = 0.037) the two saline-injected animals with damaged AMYs.

The number of assertion displays (Table 1; Greenberg et al. 1984) was also found to be significantly reduced after the lesion in the hit AMY group (t(8) = 4.24, p = 0.003, d = 1.413), while there was no difference in assertion displays pre- and postsurgery in the other groups (miss AMY, t(5) = 0.36, p = 0.736; saline, t(7) = 0.0, p = 1.000) (Figure 5B).

The latency to the first aggressive behavior was not significantly different in each group before and after surgery: hit AMY (t(8) = 1.56, p = 0.159, d = 0.518), miss AMY (t(5) = 0.07, p = 0.944, d = 0.030), and saline (t(7) = 0.33, p = 0.748, d = 0.118) (Figure 5C).

Lastly, we did not find a relationship between T levels and aggression levels postsurgery in the hit AMY group with (r²(8) = 0.0, p = 0.734) or without (r²(7) = 0.0, p = 0.965) the outlier, the miss AMY group (r²(4) = 0.06, p = 0.636), and the saline group (r²(6) = 0.01, p = 0.783) (Figure 6).

4.1 A Partial Lesion of the AMY in Green Anoles Reduced NBS Aggression Levels

Our study found that a partial lesion (mean damage = 14.45 ± 5%) to the AMY resulted in a 74 ± 10.6% reduction in aggression levels on average in green anole lizards, suggesting that impairment of the AMY in NBS lizards negatively influences aggressive behavior. In particular, assertion displays were decreased by 84 ± 5.4% post AMY lesion, indicating that in addition to overall aggression, arousal-like behaviors were also severely impacted. Since most AMY lesion studies on seasonal breeders are conducted during the BS (Tarr 1977; Greenberg et al. 1984), it is not clear whether the AMY might have different functions depending on the breeding state of the species. For example, while previous work found that a damaged AMY during the BS resulted in no change in aggression (Greenberg et al. 1984), perhaps this result is due to compensation by other brain regions involved in BS aggression. In support of this idea, several studies have reported other SBN brain regions such as the ventromedial hypothalamus and the bed nucleus of the stria terminalis to also play a role in BS aggression (Halász et al. 2002; Kabelik et al. 2008, Kabelik et al. 2013, Kabelik et al. 2021; O'Connell and Hofmann 2011). Furthermore, Kabelik et al. (2018) noted functional connectivity between the AMY and other nodes of SBN in the brown anole lizard (Anolis sagrei) during the BS. Thus, it is possible that other structures that participate in the SBN might be able to overcome any functional deficits of a damaged amygdala during the BS. This compensatory mechanism might either be lacking and/or the AMY might be playing a bigger role during NBS aggression, which was impacted in our study. However, with a limited number of studies on NBS aggression available, additional research is required to test this hypothesis.

Alternatively, it is possible that our results differ from other work due to targeting different amygdala substructures and/or different cell types within the area, as these are known to have different influences on aggressive behavior (Levinson et al. 1980; McGregor and Herbert 1992; Wang et al. 2013). Targeting other AMY substructures might yield other outcomes, as previous work in the corticomedial AMY reported no aggressive behavioral changes post lesion (McGregor and Herbert 1992). Thus, further research into the role of specific AMY subregions is required to understand the neural control of aggression in this species. Also, as Emery et al. (2001) suggested, different lesion causative agents can result in varied outcomes. For instance, previous work utilized radiofrequencies within the AMY to cause a lesion and found no change in aggression (Busch and Barfield 1974; Greenberg et al. 1984). This technique has been reported to also damage the fibers extending from the AMY to other regions of the brain (Salinas et al. 1996; Emery et al. 2001), suggesting that the use of different lesion causative agents might contribute to the differences in behavioral outcomes observed in our study compared to others.

Additionally, we did not find any significant differences in latency to aggression post AMY lesion. This indicates that the cells within the AMY may be involved in regulating aggression but not motivation to initiate aggressive behaviors. Our result contrasts with previous work that demonstrated the AMY's role in regulating latency to aggression in lizards (Tarr 1977) and rodents (Vochteloo and Koolhaas 1987), suggesting that there might be either species, seasonal, or cell-specific regulation of latency to aggressive behaviors. Interestingly, Lin et al. (2011) reported that stimulating a different SBN node (the ventromedial hypothalamus) in rodents can promote aggression but also increases the latency to aggression, indicating that occurrences of aggressive behavior and latency to initiate said behavior might be differently regulated. Thus, the AMY might also differently regulate latency to aggression, but further research is needed.

4.2 A Complete Lesion of the AMY in Green Anoles May Increase NBS Aggression Levels

It is possible that the relationship between aggressive behavior and the AMY might change if the animal has a complete bilateral AMY lesion, such that AMY damage above a threshold induces a different response. Furthermore, green anole lizards have more neurons within the AMY during the NBS compared to the BS (Beck et al. 2008), and, as NBS lizards have decreased reproductive and aggressive behaviors (Lovern et al. 2004), it could be predicted that reducing neuron number in the NBS AMY might result in an increase in aggression levels. In support of these ideas, one animal in our study had over 75% AMY damage and displayed a 240% increase in aggression post lesion. This result suggests that the AMY might serve as a negative node in the SBN for aggressive behavior during the NBS and a complete lesion releases this behavior. More work is needed to explore this idea.

4.3 Plasma T Levels Did Not Correlate With Aggression Levels

The present study found that GSI and plasma T levels resembled NBS lizards (Tao et al. 2022; Kang et al. 2020) and the aggressive behaviors we observed did not correlate with plasma T levels. This contrasts with findings from many studies that found a positive association between aggression and plasma T levels and supported the challenge hypothesis (Wingfield et al. 1990) in lizards (Klukowski and Nelson 1998; Yang and Wilczynski 2002) and birds (Harding and Follett 1979; Van Duyse et al. 2002; Hunt et al. 2019). For example, a study on green anole lizards during the BS found that dominant lizards that win against conspecific opponents also have corresponding increases in plasma T levels (Greenberg and Crews 1990). However, many of these studies are based on aggression during the BS that might at least partly be controlled by circulating androgens. Our study supports other findings that postulate that aggression is differently regulated during the NBS. For example, studies that induced an increase in T levels in nonbreeding spiny lizards (Sceloporus jarrovii) (Moore and Marler 1987) and green anole lizards (Neal and Wade 2007) did not find a corresponding increase in aggression to BS aggression levels.

However, it should be noted that these results do not mean that T does not regulate NBS aggression at all. In the absence of mature gonads (site of T production), as is the case during the NBS, the brain can produce T in the AMY to modulate aggression (Pradhan et al. 2010; Spritzer and Roy 2020; Jalabert et al. 2024). Other hormones, such as estrogen, dehydroepiandrosterone, and neuropeptide Y, have been highlighted in previous work as possible modulators of NBS aggression (Quintana et al. 2021; Soma et al. 2002; Soma 2006).