Changes in social reward across adolescence in male mice

Abstract

In humans, adolescence is a time of dynamic behavioral and emotional changes, including a transient decrease in affect associated with being among family members. It is not clear if a similar change occurs in rodent species used to model human psychiatric disorders. Here, we investigated the developmental profile of the rewarding value of interactions with siblings across adolescence in male mice, using the social conditioned place preference task. We found that the reward value of social interactions followed a similar course to that in humans: high in early adolescence, it decreased in mid-adolescence and returned to the initial level in late adolescence. The observed change was specific to social interaction, as no age-dependent changes in preference for cocaine-conditioned context were detected. Taken together, these data show similarities between mice and humans in developmental changes in sensitivity to the rewarding effects of interactions with familiar kin.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-30755-2.

Introduction

Adolescence is a time of rapid behavioral and neural changes, as well as the peak onset age for many mental disorders^1,2. It is postulated that the emergence of psychiatric symptoms during adolescence results from alterations in typical developmental processes³. Adequate animal models are crucial in the dissection of causal links between adolescent changes in brain maturation, behavior and pathophysiology. Moreover, it is of great importance to understand to what extent the behavioral development of model animals parallels the features of human adolescence⁴.

In male laboratory mice, which descend from Mus musculus species, adolescence typically spans postnatal days (P) 30 to 60, corresponding to ages 11–21 years in humans⁵. Although no precise biological markers define the start or end of adolescence in either species, significant physiological and neural changes occur during this period. For example, rapid increase in serum testosterone levels and nonlinear changes in the expression of D1 and D2 receptors in the basal ganglia are observed^6–8. Despite these developmental dynamics, most behavioral studies treat adolescence as a uniform phase^9–12 or compare adults to a single adolescent timepoint, usually around P28-30^12–16, with some notable exceptions^17–21. Here, we hypothesize that social reward in adolescent mice may also develop nonlinearly.

One of the characteristic features of human adolescence is changes in social preferences. Whereas infants fully depend on parental care and display strong distress following separation from their mothers, mid-adolescents show a decrease in the time spent with their family members along with an increase in the time spent alone or with peers^22,23. This behavioral shift is accompanied by changes in emotions associated with relatives: early and late adolescents show positive affect in the company of their family members, while mid-adolescents report more negative emotional states²². However, it is not known whether affective changes resembling those observed in humans also occur in model animal species. Therefore, the goal of the present study was to assess the reward value of interactions with familiar kin in early, middle and late adolescence in male mice.

Results

Rewarding effects of interactions with siblings across adolescence

To investigate the possible changes in the rewarding effects of social interactions with siblings during adolescence in mice, we used the social conditioned place preference (sCPP) test (Fig. 1A)²⁴ with male mice representing early (around postnatal day 33 [P33]), middle (P38) and late (P43) adolescence stages (for information about experimental groups, see Tables S1 and S2). Animals were conditioned to associate one environmental context with group housing and another with social isolation, and then were tested to determine context preference. Mice aged 33, and 43 (but not 38) days at posttest showed a significant increase in the time spent in the social context from pretest to posttest, as revealed by significant interaction between age and conditioning (Fig. 1B, Table S3, Age F_2,63 = 0.2182, p = 0.8046; Pre-Post F_1,63 = 21.69, p < 0.0001; Age x Pre-Post F_2,63 = 3.253, p = 0.0453; post hoc P33 p < 0.0001, P38 p = 0.8958, P43 p = 0.0064). There was no effect of context (i.e., the type of bedding) on preference (Fig. S1, Table S4). It had no significant effect as main factor or in interactions with age or conditioning, while interaction between age and conditioning remains significant when context is added as variable (Context F_1,65 = 1.738, p = 0.1920; Age x Context F_2,65 = 0.6916, p = 0.5044, Pre-Post x Context F_1,65 = 0.1533, p = 0.6967; Age x Pre-Post x Context F_2,65 = 0.2435, p = 0.7846; Age x Pre-Post F_2,65 = 4.288, p = 0.0178). Thus, interactions with siblings had a lower reward value for mid-adolescent mice. The decrease in preference for the compartment associated with social contact in mid-adolescent mice was also clearly apparent in the social preference score (Fig. 1C). The score showed that the rewarding effects of social interactions were more than two times lower in mid-adolescent (P38) than in early-adolescent (P33) mice and returned to the early-adolescent level in late-adolescent (P43) mice (Age F_2,63 = 3.353, p = 0.0420). Importantly, motor activity was not significantly affected by age or interaction between age and conditioning; thus, it was not a confounding factor (Fig. 1D, Table S5, Age F_2,63 = 2.279, p = 0.1108; Pre-Post F_1,63 = 85.03, p < 0.0001; Age x Pre-Post F_2,63 = 0.5909, p = 0.5569). Taken together, the results reveal a transient decrease in the rewarding effects of interactions with related individuals in mid-adolescent mice.

Fig. 1.

Social conditioned place preference during adolescence. (a) Schematic representation of the experimental schedule. (b) Time spent in the social context. The graphs show the mean time spent on the social-conditioned context during the pretest and posttest (black dots and line, error values are s.e.m.). The gray points and lines show all individual mice in the corresponding age bracket as indicated above. (c) Preference of the social context. The points show mean values of the preference score, i.e., the difference between the time spent in social context and isolation context during posttest of mice at different adolescence stages as indicated below. (d) Distance traveled during the tests. Points represent mean distance traveled during pretest and posttest, error bars are s.e.m., gray points and lines represent individual animals. Significant post-hoc (ANOVA with Šidák correction) differences between mean values are shown as: ‘**’p < 0.01, and ‘***’p < 0.001.

Social contact

Next, we investigated if the change in social behavior in mid-adolescent mice was specific to the rewarding effects of social interactions or if contact seeking was also altered. To explore this possibility, we administered a test in which contact with another mouse is enabled through a transparent, perforated plexiglass wall (Fig. 2A)²⁵. The interaction partners were siblings reared in the same cage but isolated for one day before the test to match the conditions of the sCPP posttest. We observed no age-related changes in the time spent in the proximity to the partner’s compartment (Fig. 2B, Table S3, Age F_2,65 = 1.038, p = 0.3648) or the distance between the focal mouse and the partner’s compartment (Fig. 2C, Age F_1,65 = 0.7805, p = 0.4660). These results might indicate that the decrease in the rewarding effects of interactions with related individuals in mid-adolescent mice was not accompanied by a general decrease in social contact seeking.

Fig. 2.

Social contact seeking during adolescence. (a) A schematic representation of the task. Mice were tested for seeking social contact with a sibling after 24 h of isolation. (b) Time spent in the interaction zone. The points show mean time spent in the zone, error bars are the s.e.m. (c) Mean distance from the partner’s compartment. Group sizes: P33 n = 10, P38 n = 18, P43 n = 10.

Rewarding effects of cocaine

An alternative explanation for the observed decrease in the rewarding effects of social interactions could be a general anhedonia or impairment in associative learning. To test whether the decrease in the rewarding effects of social interaction reflected a stimulus-independent reduction in the expression of conditioned behaviors, the cocaine-induced CPP was assessed. As expected, a significant increase in time spent in the cocaine-paired compartment was observed but without the effect of age or interaction between age and conditioning (Fig. 3A, Table S3, Age F_2,41 = 0.6877, p = 0.5084; Pre-Post F_1,41 = 14.94, p = 0.004; Age x Pre-Post F_2,41 = 0.3454, p = 0.7100). In line with this interpretation, no significant effect of age on the place preference score was detected (Fig. 3B, Age F_1,41 = 0.9828, p = 0.3829). No significant effect of age or conditioning on locomotor activity was observed (Fig. 3C, Age F_2,41 = 0.8940, p = 0.4168; Pre-Post F_1,41 = 0.8799, p = 0.7682; Age x Pre-Post F_2,41 = 0.8897, p = 0.4186). These data show that no general impairment of reward-conditioned preference occurred during mid-adolescence.

Fig. 3.

Cocaine-induced conditioned place preference during adolescence. (a) Time spent in the cocaine-conditioned context. The points show mean time spent in the cocaine-paired context during the pretest and posttest, error values are s.e.m. The gray points and lines show individual mice in age brackets indicated above. (b) Preference of the cocaine-conditioned context. The points show mean values of the preference score, i.e. the difference between the time spent in cocaine-conditioned context and saline-paired context during posttest of animals at different adolescence stages as indicated below. (c) Distance traveled during the tests. Points represent mean distance traveled during pretest and posttest, error bars are s.e.m., gray points represent individual animals. A significant post-hoc (ANOVA with Šidák correction) difference between mean values is shown as: ‘**’p < 0.01.

Relationship between body weight and conditioned place preference results

Next, we hypothesized that the reduction in sCPP in mid-adolescence is a result of increasing aggression. If this hypothesis is correct, a reduction in sCPP would be more pronounced in smaller, potentially subordinate mice. Meanwhile, in cocaine CPP, body weight would explain less variance. To test this hypothesis, we analyzed the correlation between body weight in the pre-test and CPP score. No significant correlation was observed between these parameters neither in sCPP nor cocaine conditioning experiments (Table 1).

Table 1.

Pearson’s correlations between body weight in the pre-test and CPP score.

Group	sCPP			Cocaine CPP
	r2	p	df	r2	p	df
Early-adolescent	0.021	0.447	28	0.066	0.446	9
Mid-adolescent	0.14	0.1031	18	0.014	0.6203	18
Late-adolescent	0.0021	0.8665	14	0.214	0.214	11

Discussion

We found that the rewarding effects of interactions with familiar kin in male mice exhibit a transient decrease during mid-adolescence (around P36-40). This decrease was specific to social reward, as no change in social contact seeking, and the reward value of cocaine was observed during this period. Our results resemble human data reported by Larson and Richards (1991), who observed that the affect associated with time spent with family members was more positive in 10-year-old and 16-year-old boys than in 12- to 14-year-olds²². The age range of 11 to 16 years in male humans corresponds to P30-40 in male mice, and this period is considered “peripubertal”⁵. This may indicate that the phenomenon observed in our study, i.e., the temporary decrease in the reward value of interactions with familiar kin, is evolutionarily conserved^26–28. This finding may facilitate future research on the neuronal and physiological underpinnings of rebellious behaviors in adolescence.

Our results complement those of two previous studies that assessed the rewarding effects of social interactions using the sCPP paradigm at selected time points in adolescent mice^18,29. In contrast to the methodology presented here, both previous studies used animals that were familiar but were not specifically kept in sibling groups. Our previous research on the reward value of social interactions has shown that eight weeks of familiarization with nonrelated mice is not equivalent to being reared in the same cage before weaning, at least in females, i.e. female mice do not form an sCPP when conditioned as adults in groups coming from different litters and housed together since weaning²⁴. Moreover, both previous studies^18,29 used a paradigm with only two days of conditioning, which produces different results than the paradigm with six conditioning sessions used in this study^24,30. Bearing these differences in mind, we note that Cann and collaborators (2020) reported that social contact is rewarding in mice tested on postnatal day 29 but not on postnatal day 38, which aligns with our results. Conversely, in the study by Nardou and collaborators (2019), a decrease in the rewarding effects of social interactions during adolescence was not observed in male mice (although it was observed in females). However, the previous results were not interpreted in terms of the possible decrease in the reward value of social interactions during adolescence. Instead, Cann and collaborators conclude that sCPP, under the applied conditions, can be induced in early-adolescent, but not adult mice, implying that the observed lack of conditioning effects at the age of 38 days is considered permanent. Results presented in this work, together with a previous paper where adult mice were studied³⁰, suggest that sCPP can be induced in late adolescent and adult male mice, provided that the 6-day protocol is used.

Recent study by Murray and collaborators (2024) investigated developmental changes in the motivation to obtain social reward in laboratory rats³¹, a species in which the equivalent of P30-40 in mice is P42-55⁵. In this study, the operant conditioning protocol was employed and an unfamiliar conspecific was used as a stimulus animal. The authors found that social motivation is higher in pre-adolescent (P30) and mid-adolescent (P50) rats than in early-adolescent animals (P40). This might indicate that the rewarding effects of and motivation to obtain social reward differ depending on the familiarity of the stimulus animal, with mid-adolescence being a period of decreased familiar kin contact seeking, and higher motivation to interact with unfamiliar individuals. A study using the 6-day sCPP protocol on mice housed with familiar and unfamiliar conspecifics in different adolescence timepoints should be performed to test this hypothesis. However, it must be noted that Murray and colleagues use different labels for the developmental periods, and describe P40 as mid-adolescence and P50 as late adolescence³¹.

In contrast to social reward, there were no apparent changes in social contact seeking (assessed with the partition test) during adolescence. This finding indicates that the changes in social interactions in adolescent mice are qualitative rather than quantitative. We speculate that the amount of time spent with siblings may not change, but aggressive encounters may replace affiliative interactions. This interpretation is supported by earlier studies showing a profound decrease in passive social contact²¹ and play behavior³², along with an increase in fighting^20,21, in the second postnatal month of mice life.

Strikingly, the cocaine-induced CPP did not change during the adolescent period. This indicates that the decrease in social CPP in mid-adolescent mice cannot be explained by a general decline in learning abilities. Previous studies in rats regarding developmental changes in cocaine-induced CPP yielded inconsistent results. For example, Brenhouse and colleagues reported a greater cocaine (10 mg/kg)-induced CPP in adolescents (P44) than in adults (P105)³³. Notably, preadolescent rats (P27-37) showed similar levels of cocaine-induced CPP as adult animals^33,34, which suggests an inverted U-shaped relationship between cocaine reward and age. These results are consistent with previous reports of heightened sensitivity to other drugs of abuse in adolescent animals³⁵. Conversely, the results by Badanich and collaborators indicated the highest sensitivity to cocaine reward in pre-adolescent rats, because a low cocaine dose (5 mg/kg) induced CPP only in pre-adolescents (P35), but not in adolescents (P45) or young adults (P60 animals)³⁶. This suggests that some factors unidentified yet, for example the level of stress imposed by the experimental protocol, might differentially influence the rewarding effects of cocaine in different sub-periods of adolescence, and more studies are needed to resolve this issue. Irrespective of these inconsistencies, our results indicate that social and drug reward follow different developmental trajectories in the peri-adolescent period, which suggests that different neuronal processes may underlie the rewarding effects of cocaine and social contact³⁷. Indeed, our recent findings regarding modulation of sCPP expression in adolescent male mice by opioid receptor antagonists³⁸ suggest that sCPP is driven by negative reinforcement (avoidance of social isolation) in younger mice (around P34), whereas in older adolescent mice (around P41) it is driven by positive reinforcement (reward from social contact). Only the latter process would be comparable to cocaine action in terms of activation of the brain reward system. On the other hand, the transient sCPP decrease during mid-adolescence described herein could be the developmental stage when the mechanism of sCPP changes from negative to positive reinforcement, i.e. the first mechanism has already extinguished and the second has not been established yet.

Although the pubertal stage was not directly assessed in our study, previous research allows us to speculate on the physiological changes cooccurring with the observed behavioral shift. The time window analyzed in the present study (i.e. postnatal days 32–46) is a period of dramatic physiological changes associated with sexual development in mice. The first sign of sexual maturation in male mice—balano-preputial separation—appears between P28 and 32^39–41. Plasma testosterone levels rise rapidly from P30 to P40 and begin to decline slightly after that age^6,7. Studies show rapid brain development during adolescence, including transient changes in dopamine receptor expression in the dorsal and ventral striatum peaking around P35^8,42. Physiological changes during puberty are accompanied with behavioral changes, especially in the social domain. Adolescence is a period of social re-orientation, characterized by the emergence of sexual behavior, decreased interest in family members, and increased interest in peers²³. Social re-orientation towards unfamiliar individuals is considered one of the causes of emigration from natal environment (dispersal). Dispersal begins around postnatal day 30 and continues well into adulthood, with approximately half of the males having migrated from their natal environment by postnatal day 42⁴³. Our study suggests that a decrease in the rewarding effects of interactions with familiar kin occurring in mid-adolescence might serve as a proximate affective mechanism for the initiation of dispersal from the natal environment. At later stages, in contrast, dispersal might be driven by the search for sexual partners.

The main limitation of our study is that it was conducted exclusively on male mice.

Sex differences in the developmental trajectory of social and non-social reward valuation is a focus of our next study⁴⁴. Another limitation is that the social interaction test did not allow for direct social contact. It is possible that the results would have been different if the animals had interacted directly. Future studies should resolve this issue.

Taken together, our data show similarities between male mice and humans in the pattern of social reward development. Specifically, we show that the behavioral changes occurring during adolescence are nonlinear, and mid-adolescence seems to be a peculiar period. Similar conclusions emerge from some human studies^26,45,46. Hence, we believe that future animal studies should focus on dissecting the neural changes occurring in adolescent brain with higher temporal resolution. This could help understand the causes of mental disorders which typically begin in adolescence or early adulthood.

Methods

Animals

Experiments were performed with C57BL/6 male mice bred at the Maj Institute of Pharmacology animal facility. Mice were housed in a 12/12 h light-dark cycle (lights on at 7 AM) under controlled conditions: a temperature of 22 ± 2 °C and a humidity of 40–60%. After weaning, the mice used in sCPP and social interaction tests were housed with all littermates of the same sex. The average number of males in the litter was 3.94 (SEM = 0.11, range: 2 to 6). For cocaine CPP, larger litters (containing 5 to 6 male mice) were divided into two smaller groups (n = 2 to 3). Rodent chow and water were available ad libitum. Home and conditioning cages contained aspen nesting material and aspen gnawing blocks. Behavioral tests were conducted during the light phase under dim illumination (5–10 lx). Social conditioned place preference (sCPP) and social interaction tests were video recorded with additional infrared LED illumination. The age and weight of mice in each experimental group are summarized in Table S1.

Three age groups were studied: early (around postnatal day 33 [P33]), middle (P38) and late (P43) adolescence (for information about experimental groups, see Tables S1 and S2). The average age of animals during the pre-test and post-test of conditioned place preference tests is presented in Table 2. The characterization of puberal stages in our study was inspired by the review by Margaret Bell⁵. She defines a period between 18 and 29 days in male mice as juvenile, 30 to 40 days as peripubertal, and 41 to 60 as late-adolescent. We are aware that other researchers use slightly different definitions. For example, postnatal day 35 is considered the last day of pre-puberty/pre-adolescence by some authors⁴⁷ but the first day of puberty by others⁴⁸.

Table 2.

Age of the animals during the pre- and post-test of the sCPP and cocaine CPP tests.

Group	sCPP		Cocaine CPP
	Mean pre-test age	Mean post-test age	Mean pre-test age	Mean post-test age
Early-adolescent	26.97	33.97	28.18	32.17
Mid-adolescent	31.05	38.05	33.6	37.6
Late-adolescent	36.63	43.63	39.26	43.23

All behavioral procedures were approved by the II Local Bioethics Committee in Krakow (permit numbers 35/2019, 185/2020, 38/2021, 266/2020, 305/2020) and performed in accordance with the Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. The reporting in the manuscript follows the ARRIVE guidelines.

Social conditioned place preference

The test was performed to assess the rewarding effects of housing with siblings and followed the procedure described previously²⁴. The test consisted of three phases: pretest, conditioning, and posttest (Fig. 1A).

The pretest and posttest phases were performed in a two-compartment cage, as in previously published papers^{18,24,30,37,49}. Each cage compartment contained a novel context (context A or context B) defined by type of bedding and gnawing block size and shape. Bedding materials used were beech (context A, P.P.H. “WO-JAR”, Poland or PPHU Natur-Drew A. Czaja, Poland or Terrario Peak Wilderness, DMR Group, Poland) and cellulose (context B, Scott Pharma Solutions, cat no. L0107). In the home cages, aspen bedding was used (ABEDD, Latvia or Tapvei GLP, Estonia). Mice were allowed to freely explore the test cage for 30 min, and the time spent in each compartment was recorded. Animals that spent more than 70% of the pretest time in one of the contexts were excluded (Table S2).

After the pretest, animals were returned to their home cages for approximately 24 h. Then, mice were assigned to undergo social conditioning (housing with cage mates) for 24 h in one of the contexts used in the pretest followed by 24 h of isolate conditioning (single housing) in the other context. The length of the conditioning sessions was purposefully not matched between cocaine CPP (40 min) and social CPP (24 h). One day conditioning periods were previously used by us^24,30 and by several other research groups^18,29,49,50. Protocols with short, 15–30-min, conditioning sessions have also been employed using either an unfamiliar^51,52 or a familiar⁵³ mouse as a stimulus animal. These studies revealed that the rewarding effects of a brief interaction were dependent on familiarity, i.e. only contact with unfamiliar mice resulted in robust place preference^51,52. The effects of the brief contact with a familiar conspecific were reported to be strain-specific; CD1 female mice showed social context preference, while C57BL/6J did not⁵³. Since in this study we aimed to assess the rewarding effects of contact with familiar kin in C57BL/6 mice, we opted for long conditioning sessions.

Although the behavior of mice was not recorded after regrouping, it was usually observed for several minutes by the experimenter. Regrouping often induced transient fighting bouts among late-adolescent mice. Similar behavior was rarely observed among mid-adolescent mice and hardly ever observed among early-adolescent mice. Conditioning was performed in cages identical to the home cage, with ad libitum access to food and water. To prevent bias, the social context was randomly assigned such that approximately half of the animals received social conditioning in context A and half in context B. In cases where the final number of animals conditioned in each context was not equal (due to an unequal number of animals passing the 70% criterion or unequal number of animals in the litter), we pseudorandomly trimmed the larger group using a Python script (https://zenodo.org/record/8100281)³⁸. The exception from completely random selection was introduced to preserve a mean 50% initial context preference during the pretest, i.e., ascertain that the test was fully unbiased (Table S6). Analysis on the full data set (not trimmed) was also performed, to assess the potential interaction between the age and conditioning context on social reward (Figure S1, Table S4). The conditioning phase lasted 6 days (3 days in each context, alternating every 24 h), and then the posttest was performed.

Two measures of the rewarding effects of social interactions were used: (1) pretest vs. posttest comparison of the time spent in the social context, (2) score: time spent in the social context minus time spent in the isolation context during the posttest.

Social interaction in the partitioned cage

This test was carried out to assess social contact seeking with a sibling partner after 24 h of isolation. The procedure was performed in a rectangular cage (48 × 12 cm, 25 cm high) divided by a transparent, perforated plastic wall into two compartments: a smaller partner compartment and a larger focal animal compartment (Fig. 2A). One day before the test, the animals were weighed. Pairs of mice were chosen based on weight, i.e. if there were more than two male mice in the litter, mice with similar weight were paired, and the heavier animal from each pair was designated as the focal animal. Next, the animals were habituated to their respective cage compartments for 10 min. During habituation, only one mouse was present in the test cage. After habituation, mice were placed in separate home cages for approximately 24 h, after which time the focal animals were placed in the test cage for the second adaptation session (5 min). After adaptation, the partner was introduced for 10 min. Two measures of social contact seeking were used: time spent in close proximity to the partner’s compartment and distance to the partner’s compartment.

Cocaine-induced conditioned place preference

For the cocaine CPP paradigm, three-compartment cages were used (Med Associates, St. Albans, VT, USAMED-CPP-MSAT); the two peripheral compartments (that contained distinctive visual and tactile cues) were linked to the central compartment by guillotine doors. The test consisted of three phases: pretest, conditioning, and posttest. For the pretest and posttest phases, animals were introduced to the central compartment of the apparatus, and the doors between the compartments were lifted such that the animals could freely explore the apparatus for 20 min. Animals that spent more than 70% of the pretest time in one of the contexts were excluded. The less preferred of the two peripheral compartments was designated the cocaine compartment. The next day, the 3-day conditioning phase started. Each day, two 40-minute conditioning sessions were performed, separated by approximately 3 h, during which animals were kept in their home cages. Before the morning conditioning session, animals received an i.p. saline injection, while before the afternoon session, they were injected with cocaine hydrochloride dissolved in saline (10 mg/kg, 5 µl/g). Immediately after the injection, animals were placed in the respective cage compartment, while the guillotine doors separating the compartments were closed. The posttest was performed on the day after the last conditioning session. The measures used to assess cocaine reward were the same as for the social conditioned place preference.

Data analysis

The distance traveled and time spent in separate cage compartments in the sCPP and social interaction tests were analyzed automatically using EthoVision XT 15 software (Noldus, The Netherlands). In the social interaction test, the zone close to the partner’s compartment was outlined digitally. In the cocaine-induced CPP test, the position of the mouse was registered automatically by the Med Associates system. Comparisons of sample means were performed using analysis of variance (ANOVA) followed by Šidák’s post hoc correction or Student’s t test for cases with only two samples. The statistical significance threshold was set at p < 0.05. Before the analysis, the Grubbs’ test for outliers was performed. For sCPP results, the outlier test was performed on “score”, “index 1”, “index 2”, and “distance moved during post-test” parameters before the trimming of the data. While only the “score” measure is shown in the present manuscript, all other measured are available at: https://zenodo.org/records/15284680.

The test was performed separately for the context A and B data. Four outliers were detected in the sCPP data: one in the P33 group, two in the P38 group, and one in the P44 group. For the social interaction data, the outlier test was performed on the parameter “time in interaction zone”. No outliers were detected. For the cocaine experiment, the outlier test was performed on “score”, “index 1”, “index 2”, “exploration during posttest”, and “movement during posttest” parameters. No outliers were detected. Data concerning animals excluded from the analysis are presented in Table S2. The analysis was performed in GraphPad Prism 9.4.1. For full descriptive statistics and statistical test results, see Supplementary Material.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: ZH, BZ, RR, JRP; Methodology: ZH, KM, MK, MC and JRP; Investigation: ZH, MK, KM, MC, ŁS, AR and MKJ; Visualization: ZH, KM and JRP; Supervision: BZ, JRP; Writing—original draft: ZH and BZ; Writing—review & editing: ZH, BZ, RR, and JRP with help from all the authors.

Funding

This work was supported by the Polish National Science Centre [grant number 2016/21/B/NZ4/00198].

Data availability

All data are available at https://zenodo.org/records/15284680. This paper is available at bioRxiv⁵⁴.

Declarations

Competing interests

The authors declare no competing interests.