The Hungarian Ethical Committee for Animal Research (RN: XIV/1248/2018 and XIV/1421/2023), in accordance with EU Directive 2010/63/EU, approved all experimental procedures. The study adhered to the guidelines outlined in the Animal Research: Reporting of In Vivo Experiments (ARRIVE 2.0).

Regarding the selective breeding process, we followed the same protocol as in our previous studies with Wistar and Wisket rats (8, 10, 38). Control and Lisket rats (both male and female) were weaned at three weeks of age. Lisket rats were then housed individually for four weeks and received intraperitoneal ketamine injections (30 mg/kg, 4 mL/kg) for five consecutive days, starting in the second week of isolation (12). Afterward, the animals were rehoused in groups of two to three per cage. Control (LE) rats were group-housed and received no treatment. At three months of age, Lisket rats exhibiting impaired behavioral performance in the tail-flick and Ambitus tests (see below) were selected for breeding (38). To minimize inbreeding, sibling mating was avoided. Typically, one male was paired with two females (resulting in six males and twelve females per generation) for a two-week mating period, after which females were housed individually to ensure undisturbed parturition. Animals from 8–11 generations were involved in the study. Control (LE, n=12) and Lisket (n=13) male rats were maintained under a 12-hour light/dark cycle at a controlled temperature of 22 ± 1°C.

On week 10, the acute nociceptive threshold was assessed using the tail-flick (TF) test, as previously described ( Figure 1B ) (38). The reaction time was measured by immersing the distal 5 cm of the tail in 48 °C hot water until a tail-withdrawal response occurred, with a cut-off time of 40 seconds. Since body weight influences nociceptive responses, TF latencies were normalized to body weight and expressed as relative TF (RTF) latencies for statistical analysis (39).

The Ambitus apparatus combines features of the reward-based Hole Board and maze corridor tests (8). It is a rectangular corridor constructed of clear plexiglas with a black floor ( Figure 1A ; www.deakdelta.hu). Each corridor contains four side boxes (16 in total) for food rewards (puffed rice, 20 mg). Infrared beams detect exploratory activity at each side box and locomotor activity in the middle of the corridor with a 1 ms time resolution. After placing the food rewards, trials began by positioning the rats at the starting point ( Figure 1A ), after which the experimenter immediately left the room. The animals were given 300 seconds to explore the corridor and collect the rewards. The apparatus was cleaned with 70% ethanol between animals. The Ambitus system allows for the detection and calculation of multiple behavioral parameters, including locomotor activity, exploratory behavior, and reward collection (eating), as well as cognition-related measures such as effective exploration, adequate exploration ratios, and learning capacity. Definitions, abbreviations, and calculations for these parameters are provided in Table 1 .

The HomeManner (HM) apparatus, developed in the authors’ laboratory, consists of six sets of operant cages housed in a separate room under standard conditions, with a house light positioned outside the cages ( Figure 1B ). Each cage measures 57 × 60 × 55.5 cm and has three levels, with steel wire grid forming the sidewalls and top ( Figure 1B ). Since a structured cage environment may be more beneficial for rodents than a large open floor area (40), the cage was divided into two sections: one side with a single floor and the other with three floors. The animals could access all floors by climbing the grid walls. The single-story section included a playing area designed to enhance well-being and sensory-motor stimulation. This area was equipped with a running wheel (for voluntary exercise), an abacus, and a plastic tube for environmental enrichment. At the front of the cage, a water bottle was provided for free access to drinking water, which could be reached from the playing area ( Figure 1B ). The first level of the three-floor section functioned as a shelter (28.5 × 60 × 11 cm), containing bedding material (nest box: sawdust, 3–4 cm thick), where animals could also hide marbles. The second and third levels, positioned above the shelter, were made of opaque plastic and steel grid, respectively. On the second floor, 20 cm from the water bottle, two food dispensers were attached to each cage. The food dispensers consisted of reward containers positioned outside the cage, each equipped with a small electric motor that delivered food pellets (45 mg, F0021, Bio-Serv™ Dustless Precision Pellets™ Purified Rodent Diet, BioServ, Frenchtown, NJ, US) into two small trays (12 × 20 mm) placed 7 cm apart inside the cage. LED lights were positioned near the containers to indicate reward availability.

The trays were fitted with infrared sensors to detect food delivery, reward consumption, and animal activity at the trays. Any contact with the trays was classified as exploration. The system was connected via an interface to a PC in a separate room, where custom software (developed by P.L.) controlled food delivery and recorded all events at both trays. The collected data were then extracted and analyzed using another software developed by L.K. At the start of the experiment, the pellet dispensers were filled with 12 g and 4 g of pellets for the large-dose (LD) and small-dose (SD) trays, respectively, maintaining these positions throughout the study. When the animals were placed in their cages, the LED lights were switched on, and the dispensers released three pellets on the LD side and one pellet on the SD side, following a paradigm based on a previous study (29, 41). Once the animals consumed the pellets from a tray, the LED light on that side switched off, and no further reward was available for 20 seconds (inter-trial interval, ITI).

After the ITI, the LED light switched back on, and if an exploration at the tray was registered (trigger stimulus), food pellets were dispensed. The rats were not time-restricted for initiating trials, allowing voluntary interaction with the trays throughout the experiment. Even in the absence of a programmed delay, the electric motors required 2–4 seconds to dispense the food pellets.

The trays remained accessible 24 hours a day, but once the animals consumed all the food available for that day (12 g + 4 g from the dispensers), the LED lights switched off until the next morning (between 9 and 10 AM), when the dispensers were refilled. At that point, the trays were rebaited with three pellets (LD) or one pellet (SD), and the LED lights were switched on again. In the event of a hardware malfunction, the LED lights were turned off.

The HM system allows for the detection and analysis of exploratory behavior, reward collection (eating), cognitive performance, and impulsivity-related parameters, including various types of explorations, learning capacity, and delay time. The definitions, abbreviations, and calculations of these parameters are detailed in Table 1 . Activity at the trays was characterized by the mean hourly duration of epochs, the mean number of cycles within one epoch, and the number of different types of explorations within one cycle (E1–E4). An epoch was defined as an active phase that began with a trigger stimulus and ended if the animal showed no activity at the tray for more than five minutes (300 seconds). One epoch could contain multiple cycles. Within an epoch, the first cycle started with a trigger stimulus and ended with the next LED_ON event. Subsequent cycles began with LED_ON and ended with the following LED_ON event.

Four different types of exploration were differentiated within one cycle. Adequate explorations (E1) were defined as the sum of the trigger stimulus (1) and the collection of rewards (1 or 3 pellets at SD or LD, respectively), making them standard at both trays. E2 captured anticipatory restlessness during the food delivery phase, encompassing the period marked by the food dispenser’s motor noise up to the onset of consumption; higher E2 values signified increased restlessness in anticipation of the reward. E3 denoted explorations occurring during the inter-trial interval (ITI), referred to as premature explorations. E4 represented explorations occurring during the delay period, categorized as incorrect explorations.

On week 10, TF latencies were recorded four times at 30-minute intervals ( Figure 1C ). In the following week, the animals participated in the Ambitus test, following the procedure used in previous studies (8, 9). Food restriction was applied 48 hours before the test, while water remained freely available to ensure adequate motivation. Two task types were used: Task 1 (Trials 1–2), in which all boxes contained rewards, was conducted in the morning, while Task 2 (Trials 3–4), in which only the inside boxes were baited, was performed three hours later ( Figure 1C ).

At least one week after the Ambitus test, the animals were transferred to the testing room, where they underwent two days of food restriction to maintain motivation for consuming the rewards in the HM. The animals were assigned to one of six cages in a pseudo-randomized manner and housed individually without visual contact for 13 days. Both groups were equally represented in the five experimental rounds required to obtain 15 animals per group. However, three control and two Lisket animals were excluded due to sensor system malfunctions in their cages over several days, resulting in a final dataset comprising 12 LE and 13 Lisket animals.

On Day 1 of the experiment, the animals were placed in the large cages between 9:00 and 10:00 AM without access to standard food, but both dispensers were filled with pellets (12 + 4 = 16 g). From the following day onward, standard food (10 g/day, which is about 40–50% of the daily required food amount) was provided on the first floor near the pellet-delivering trays. Thus, a moderate food restriction was maintained throughout the experiment in HM, but the required daily intake could still be achieved by consuming pellets from the dispensers.

During the training phase (Days 1–3), no delay was applied. In the testing phase (Days 4–13), a 10-second delay was introduced on the LD side for animals that had learned to prefer LD over SD rewards (with a preference at least 10% above random). The delay was gradually increased by 10 seconds each day, depending on the animal’s preference, following the method applied in previous studies (29, 35). Apart from the introduction of the delay, no parameters were modified throughout the experiment. Body weight was measured at the beginning and end of the experiment in the HM. Fluid consumption was recorded twice a week, with fresh water provided each time, and relative fluid consumption was calculated based on body weight. Pellet consumption was monitored daily, and the dispensers were refilled accordingly. Standard food consumption was assessed at the end of the experiment.

To minimize disturbance, the experimenter generally entered the room only once per day, except in cases of equipment malfunction. During these visits, the experimenter checked the functionality of the equipment, inspected the health of the animals, refilled the dispensers, provided standard food, and replaced the drinking water. While the HomeManner cages were not cleaned at all over the 13 days, only one animal was housed in each large cage (about 0.2 m³), and no extreme dirtiness was observed at the end of the study.

The Ambitus and HM systems recorded various behavioral parameters related to exploratory activity and reward collection ( Table 1 ). In order to compare the results of these different parameters on a common scale—despite differences in units and distributions—we normalized the raw data using z-score transformation. For each parameter, we subtracted the group mean (all animals in all trials or days) from each individual value and divided by the standard deviation. This standardization allows for the direct comparison of effect sizes across variables and test conditions (acute vs. chronic), as visualized in the summary figures. The method is widely used in behavioral neuroscience to unify diverse datasets for integrative analysis (42). Additionally, the coefficient of variation (CV = standard deviation/mean) was calculated to evaluate variability across trials or days.

Due to a software malfunction in the HM system, tray activity data from five days were lost (10 data points). Consequently, a total of 640 data points were analyzed across the two trays for the 25 rats over the 13-day period. Since tray activity was largely dependent on the time required to consume all food rewards (defined as eating time), several parameters were converted to the relative number of exploration events ( Table 1 ). One-way analysis of variance (ANOVA) was applied to analyze general observations (e.g., body weight and fluid intake), total z-scores, and covariance data, while factorial ANOVA was used for raw data in all tests. In the RTF and Ambitus tests, the factors were group (control vs. Lisket) and trial (Trials 1–4; Figure 2 ), whereas for the HM system, the factors included group, day, and the presence or absence of delay during the testing phase. When the global test was significant, Tukey post hoc tests were performed to evaluate the effects of different factors, as shown in the figures. Spearman correlation analysis was conducted to examine associations between the results of the Ambitus and HM tests. To visualize individual variability across all analyzed variables, a complex heatmap was generated using the z-scores from all tests for each animal. All data are expressed as means ± S.E.M., with significance set at p < 0.05. Statistical analyses were performed using STATISTICA 13.5.0.14 (TIBCO Software Inc., USA).

Fluid intake did not differ significantly between the two groups, as determined by one-way ANOVA (LE: 79 ± 3.7 ml/kg/day; Lisket: 78 ± 2.8 ml/kg/day). All standard food (10 g/day) was consumed by each animal in the HM. Regarding body weight, factorial ANOVA revealed significant effects of group (F_(1,138) = 100.12, p < 0.0001), age (F_(5,138) = 340.64, p < 0.0001), and their interaction (F_(5,138) = 2.35, p < 0.05). Lisket animals had significantly lower body weight than controls throughout the entire study period (LE: 315 ± 14.1 g; Lisket: 263 ± 11.7 g). However, body weight changes were similar in both groups, with slight weight gain observed by the end of the testing period (LE: 13.0 ± 2.29%; Lisket: 17.3 ± 2.06%).

Regarding RTF latency, factorial ANOVA revealed a significant effect of group ( Table 2 ), with Lisket rats exhibiting significantly longer RTF latency compared to control animals (LE: 1.9 ± 0.09; Lisket: 2.6 ± 0.13 s/100 g; z-score results shown in Figure 2A , left side), indicating lower acute heat pain sensitivity. The coefficient of variation across the four trials was similar in both groups ( Figure 2B , left side).

For most behavioral parameters in the Ambitus test, factorial ANOVA revealed significant effects of group, trial, and their interaction ( Table 2 ). Lisket rats exhibited reduced locomotor activity (LOCO), exploratory behavior (EXPL_TOT, EXPL_E, and EXPL_I), and cognition-related measures (A_E, E_E, and L_C) compared to their parent strain, as shown by their z-scores in Figure 2A , left side. Additionally, Lisket animals displayed a greater tendency for variance in most parameters, with significant differences observed in locomotion (LOCO), exploratory activity in the external boxes (EXPL_E), and effective exploration (E_E; Figure 2B , left side).

Regarding overall daily activity at the trays, control and Lisket animals performed 12 ± 0.2 epochs, 70 ± 1.2 cycles, and 869 ± 25.5 explorations. Analysis of z-scores over the 13-day investigation period revealed that Lisket rats showed impairments in six out of nine parameters, including eating behavior (EAT_T), various tray activity measures (Epoch_TOT, Epoch_BEF, Cycle number, Expl_BEF), and cognitive performance (L_C) compared to their parent strain, as shown in Figure 2A , right side. Additionally, variance was lower across all parameters in Lisket animals, with a significant difference observed for EAT_T ( Figure 2B , right side). Next, data were analyzed separately for the training phase and the test phase, both with and without delays.

The heatmap visualization of the z-scores revealed that, in addition to group differences, there were high interindividual differences in both groups ( Figure 5 ). Furthermore, considerable intraindividual variability was observed across different parameters, with Lisket animals showing a slightly higher coefficient of variance compared to control animals (CV: 0.6 ± 0.12 and 0.9 ± 0.20 for LE and Lisket animals, respectively, p = 0.3). To further examine interindividual and intraindividual differences, animals were categorized based on their z-scores for cognitive performance (L_C: Low performance – L vs. High performance – H), behavioral activity (locomotion and epoch duration: Inactive – I vs. Active – A) in the Ambitus and HM tests, and pain sensitivity (hyposensitive vs. normosensitive). If an animal had a z-score < 0 for a given parameter, it was classified as low performance, inactive, or hyposensitive, whereas a z-score ≥ 0 placed it in the high performance, active, or normosensitive category. This classification resulted in animals being assigned to one of 32 possible behavioral categories, and 13 of 32 remained empty ( Figure 5B ). The diagram clearly showed the high degree of separation between the two groups. Most control LE animals (10 out of 12) were categorized as normosensitive based on pain sensitivity and exhibited active behavior in at least one or both acute and chronic tests (7 out of 10). In terms of cognitive performance, 11 out of 12 LE animals showed high performance in at least one test. However, even animals with a high level of learning capacity exhibited variability in pain sensitivity or behavioral activity ( Figure 5B ). In contrast, most Lisket animals (10 out of 13) were categorized as low performance, with 8 out of 10 showing low performance in both tests. Additionally, most Lisket animals were classified as inactive (7 out of 13) and hyposensitive (8 out of 13). However, some Lisket animals with low learning capacity still exhibited normal pain sensitivity or behavioral activity ( Figure 5B ).