Contact with the liquid significantly varied among treatments (ANOVA: F_33,2 = 11.76, p < 0.001). Chicks had more contact with umami (mean ± s.e.m.: 38.5 ± 4.63) than with bitter (14.5 ± 2.54; p < 0.001) or control (25.3 ± 2.98; p = 0.012), and more contact with control than with bitter (p = 0.03) (Figure 3A). The amount of liquid consumed was also significantly different among treatments (F_33,2 = 8.25, p = 0.001). Chicks consumed significantly more umami (0.05 ± 0.003) than control (0.03 ± 0.002; p = 0.041) or bitter (0.02 ± 0.003; p < 0.005) (Figure 3B). Head shaking was significantly different among treatments (F_33,2 = 3.47, p = 0.042). Chicks shook their head significantly more frequently after having contact with bitter (1.43 ± 0.22) than with umami (0.92 ± 0.14; p = 0.04) or control (0.84 ± 0.13, p = 0.02) (Figure 3C). Beak cleaning was also significantly different among treatments (Kruskal–Wallis: X² = 11.629, d.f. = 2, p = 0.002). Chicks cleaned their beak more frequently after having contact with bitter (1.83 ± 0.57) than with control (0.30 ± 0.12; p = 0.008) or umami (0.21 ± 0.12; p = 0.001) (Figure 3D).

All 36 brains (n = 12 in each treatment) were successfully stained for c-Fos (Figure 4). However, the accumbens region of one brain from the quinine group was damaged and was excluded from further analysis.

In nucleus taeniae of the amygdala there was a significant interaction between treatment and hemisphere (p = 0.008), meaning that in one of the hemispheres, treatment had an effect on c-Fos activation as revealed by ANOVA (Table 1). Only in the right nucleus taeniae was c-Fos activation higher in the umami group (Mean ± s.e.m: 1058.33 ± 90.92 cells/mm²) than in the control (813.75 ± 49.38 cells/mm²; p = 0.032), or bitter (742.81 ± 61.39 cells/mm²; p = 0.007) (Table 1). Such differences between treatments were not present in the left nucleus taeniae (control 851.45 ± 60.73 cells/mm²; umami: 952.08 ± 97.90 cells/mm²; bitter: 946.35 ± 86.28 cells/mm²) (Figure 5A). Moreover, an additional lateralised response was found in nucleus taeniae. c-Fos activation of the bitter group was higher in the left nucleus taeniae than in the right (p = 0.014). Control and umami treatments did not present such differences (Table 1).

In the other two brain regions (accumbens and lateral septum), there were no main effects of treatment or hemisphere and no interactions (Table1; Figure 5). However, patterns of activation qualitatively similar to the right nucleus taeniae of the amygdala were present also in right accumbens (control: 889.89 ± 122.50 cells/mm²; umami: 1104.99 ± 129.71 cells/mm²; bitter: 745.59 ± 82.49 cells/mm²), although the interaction was not significant at the alpha 0.05 level (p = 0.06, Table 1). The activation pattern in the left accumbens did not differ (control 848.30 ± 97.34 cells/mm²; umami: 986.27 ± 143.08 cells/mm²; bitter: 1030.36 ± 137.95 cells/mm²), nor did the left lateral septum (control: 536.97 ± 67.73 cells/mm²; umami: 535.52 ± 115.02 cells/mm²; bitter: 589.06 ± 103.74 cells/mm²), or right lateral septum (control: 563.67 ± 54.40 cells/mm²; umami: 562.57 ± 82.53 cells/mm²; bitter: 490.52 ± 62.03 cells/mm²).

There were no significant correlations between c-Fos activation and any of the measured behaviours in any of the brain regions analysed (Table S1).

We tested 36 four days old female chicks (Gallus gallus domesticus) of the Aviagen ROSS 308 strain. Fertilized eggs were obtained from a commercial hatchery (CRESCENTI Societ`a Agricola S.r.l. –Allevamento Trepola–cod. Allevamento127BS105/2). Eggs were incubated from embryonic day one (E0) to seventeen (E17) in complete darkness. On the morning of E18 to the evening of E19, eggs were light stimulated following (Lorenzi et al., 2019). This procedure has an important impact on the lateralisation of chicks visual system (Rogers, 1982; Rogers and Sink, 1988; Rogers, 1990; Rogers and Bolden, 1991; Rogers and Deng, 1999), on the maturation of visual responses (Costalunga et al., 2021) and on several aspects of chicks behaviour (Daisley et al., 2009; Salva et al., 2009). Chicks hatched on E21 in darkness at a constant temperature of 37.7°C and a humidity of 60%. On the day of hatching, chicks were placed in individual metal cages (28 cm× 32 cm× 40 cm; W × H × L), where they were housed for three days with food and water ad libitum at a room temperature of 30–32°C and light conditions of 14 h light and 10 h dark.

On the third day after hatching (24 h prior to the experiment), chicks were taken to the experimental room and housed individually in metal cages (Figure 1), identical to their home cages. The experimental cages differed to the home cage only by the addition of a black plastic wall (occluder) that allowed us to make two compartments in the cage. Chicks were located on the larger section of the cage (25 cm L); while the smaller section (15 cm L) was used for keeping the water dishes separated from the chicks until the start of the experiment, and prevented chicks from observing the experimenter (see below). The two compartments were connected by a closable opening (5 × 4.8 cm) in the plastic wall. Above every cage (76 cm), there was a video camera (Microsoft LifeCam Cinema for Business) for recording the experiment. Illumination of the cages was provided by top lights (25 W warm light) at 67.5 cm above the cage, while the rest of the experimental room was dark.

Chicks were left in their individual experimental cages for 24 h prior to the experiment with water and food ad libitum. Room temperature was 26.4–28.7°C, humidity 24–41%, and light conditions of 14 h light and10 h dark.

For the experiment, we assigned 12 chicks to each treatment. Treatments consisted of three experimental tastes (control, umami and bitter). We used water as a control, since chicks were habituated to drink it. As a palatable taste, we used umami in a combination of 2.5% monosodium glutamate (MSG, AJI-NO-MOTO >99%, CEE: E621, France) + 0.25% inosine 5′-monophosphate (IMP, ≥98%, Sigma-Aldrich I2879-1G diluted in water). Yoshida et al. (2018) showed that at this concentration, MSG is an enhancer of the umami taste (IMP) in chicks. As an unpalatable bitter taste, we used quinine (Alfa Aesar A10459 99%) 10 mM dissolved in water. The experimental solutions were prepared the day prior to the experiment and kept in the experimental room to adjust their temperature to the room temperature. Experimental tastes were colourless and were presented to chicks in identical containers with 80 ml solution in each. Before the experiment chicks were deprived of water for 45 min. After this period, the experimental taste was presented for 10 min, the behaviour of the chicks was recorded, and after that, the experimental container was removed. The containers with the experimental solutions were weighted before and after the experiment in order to estimate the amount of liquid consumed by each chick. We ran the experiment in 4 days, testing nine chicks per day. Every testing day, we assigned three individuals to each treatment in randomized order.

We analysed the videos for seven minutes after the first contact with the experimental taste. In this period, we recorded the total number of the following events: contact with the liquid (when chicks peck the experimental solution), head shaking, and beak cleaning. The category ‘beak cleaning’ includes two behaviours: beak wiping, when the animal wipes its beak on the floor and beak scratching, performed with a foot. Video analysis was conducted blinded to the experimental conditions using the VLC media player software (v. 3.0.12).

Seventy min after the first contact with the experimental taste, all chicks were weighted and received a lethal dose (0.8 ml) of a ketamine/xylazine solution (1:1 ketamine 10 mg/ml + xylazine 2 mg/ml). They were then perfused transcardially with phosphate-buffered saline (PBS; 0.1 mol, pH = 7.4, 0.9% sodium chloride, 4°C) and paraformaldehyde (4% PFA in PBS, 4°C). The head was post-fixed and stored in 4% PFA for at least 8 days until processing. All the following brain processing steps were performed blind to the experimental treatments. To ensure that the subsequent coronal brain sections would have the same orientation (°45), brains were removed from the skulls following the procedures described in the chick brain atlas (Kuenzel and Masson, 1988). The left and the right hemispheres were separated and processed independently. Each hemisphere was embedded in a 7% gelatine in PBS solution containing egg yolk at 40°C and post-fixated for 48h at 5°C in 20% sucrose in 4% PFA/PBS and further 48 h in 30% sucrose in 0.4% PFA/PBS solution. Four series of 45 μm coronal sections were cut on a Cryostat (Leica CM1850 UV) at −20°C. Two out of the four series were collected in cold PBS. The sections of the first series were used for labelling. The sections of the second series were kept as backup. For free-floating immunostaining, sections were incubated in 0.3% H2O2 in PBS for 20 min to deplete endogenous peroxidase activity. Between every of the following steps of the procedure the sections were washed in PBS. After the sections were treated with 3% normal goat serum (S-1000, Vector Laboratories, Burlingame, CA) in PBS for 30 min at room temperature, they were transferred to the c-Fos antibody solution (c-Fos antibody, 1:2000; rabbit, polyclonal K-25, Santa Cruz Biotechnology, Santa Cruz, CA) containing 0.1% Bovine Serum Albumin (BSA, SP-5050, Vector Laboratories) in PBS and incubated for 48 h at 5°C on a rotator. The secondary antibody reaction was carried out using a biotinylated anti-rabbit solution (1:200, BA- 1000, Vector Laboratories) in PBS for 60 min at room temperature on a rotator. The ABC method was used for signal amplification (Vectastain Elite ABC Kit, PK-6100, Vector Laboratories). Cells with concentrated c-Fos protein were visualized with the VIP substrate kit for peroxidase (SK-4600, Vector Laboratories), at a constant temperature of 25°C for 25 min. This produced a dark purple reaction product confined to the nuclei of c-Fos immunoreactive (-ir) cells. Brain sections of right and left hemispheres were mounted with the same orientation on gelatine-coated slides, enabling the coder to be blind to the hemisphere’s identity while counting. Slides were then dried at 50°C, counterstained with methyl green (H-3402, Vector Laboratories) and cover slipped with Eukitt (FLUKA).

Brain sections were analysed blind to the experimental condition and hemispheres. c-Fos-ir cells were stained purple-black and were easily discerned from the background and non-activated cells, which were stained light-green (Figure 4). Photos were taken with a Zeiss microscope (objective magnification: ×20 with a numerical aperture of 0.5) connected to a digital camera (Zeiss AxioCam MRc5) and the Zeiss imaging software ZEN. Exposure time of the camera and the light conditions of the microscope were kept constant for all photos, while the contrast was slightly adjusted for individual photos to match their visual appearance if it was needed.

To analyse brain activation within the regions of interest a spot with the highest number of c-Fos-ir cells was chosen, by visual observation under the microscope. Then a counting area (400 × 400 µm) was positioned over this spot by keeping a minimum distance of 10 µm from the borders. A snapshot was taken, cropped around the counting area and saved. Automatic counting of the c-Fos immunoreactive cells was performed using the “analyse particle” function of the ImageJ software (Schneider et al., 2012). All photos were analysed with a predefined macro, where the image was transformed into 8bit, threshold was set to 120, circularity of particles to 0.5–1.0 and particle size to 2–200.

For the analysis of nucleus taeniae of the amygdala five sections of both hemispheres were selected from an area corresponding to the A8.8 to A6.4 of the brain atlas (Kuenzel and Masson, 1988) (Figure 2C). To quantify c-Fos-ir cells in accumbens (Bálint et al., 2016), five sections of both hemispheres were selected from A10.0 to A9.2 (Kuenzel and Masson, 1988; Figure 2A). Lateral septum was measured on five selected sections from A8.8 to A7.6 (Kuenzel and Masson, 1988) in the lower half of septum starting from the ventral border of the lateral ventricle (Figure 2B). The lateral septum was further delineated from the medial septum based on anatomical landmarks that are visible after methyl green counterstain. The cell densities obtained from the different brain sections were averaged to estimate overall activity in each measured area. These individual bird means were employed for further statistical analysis.

All statistical analysis were conducted with R v. 4.0.4 (R Core Team, 2021). We analysed whether behavioural responses varied according to treatment. The response variables were contact with the liquid, liquid consumed per body weight, head shaking and beak cleaning. The raw values were divided by the number of contact with liquid episodes. This allowed to reveal how much the contact with liquid elicited the behaviours of interest. ANOVA was run on the first three variables with Fischer’s LSD tests for post-hoc analysis. Given that for beak cleaning the residuals were not normally distributed (Shapiro-Wilk test, W = 0.7382, p < 0.05), even after square root or logarithmic transformation, a non-parametric Kruskal–Wallis test was used. Significance between the comparisons was assessed with post-hoc Dunn tests.

In order to test whether c-Fos activation (response variable) varied according to treatment and hemisphere (factors), we conducted a repeated measures ANOVA for each brain region. We used a between-subject factors with three levels (Treatment: control, umami, bitter), and within-subject factor with two levels (Hemisphere: left and right). Given that for nucleus taeniae of the amygdala and accumbens residuals were not normally distributed (Shapiro-Wilk Test; nucleus taeniae: W₇₁ = 0.955, p = 0.013; accumbens: W₇₁ = 0.952, p = 0.009), we used a square root transformation. This procedure significantly improved the normality of residuals (nucleus taeniae: W₇₁ = 0.979, p = 0.272; accumbens: W₇₁ = 0.982, p = 0.393), and further analysis were conducted with transformed data. Between treatment comparisons were assessed with independent post-hoc t-tests, while between hemispheres comparisons for each treatment was assessed with paired t-tests. In order to test whether behaviour and motoric activity is correlated with c-Fos expression, we performed Pearson correlations between c-Fos expression in the brain regions analysed and the behaviours recorded for each of the treatments and hemispheres. All figures were created or postprocessed using Adobe Photoshop and Illustrator software.