Introduction

Social interactions during early development play a fundamental role in the establishment of social skills1,2, emotional regulation3, and social cognition4. Early exposure to social stressors, such as isolation and neglect, has been associated with deleterious impacts on well-being that could endure into adulthood5,6,7,8,9. In humans, research has shown that social stress and a lack of social interactions during adolescence increases susceptibility for developing psychiatric disorders like depression and anxiety, which are often characterized by social withdrawal and impaired social behavior, later in life10,11,12,13,14,15,16. These consequences may arise from disruptions to the proper connectivity and function of brain neural circuits during a critical period of development17,18,19,20. Therefore, identifying a critical window when social stressors produce long-term changes in behavior will support future studies seeking to elucidate the neural mechanisms impacted by social stress during development.

Preclinical rodent models offer a valuable tool to pinpoint these critical periods by shedding light on when behavior and brain development are most vulnerable to long-term impacts of social stress, providing insight into the optimal timing for interventions in individuals exposed to social stress early in life. Previous behavioral studies in rodents have suggested that similar to humans, isolation can have enduring impacts on social and non-social behaviors. However, variable findings have emerged from these studies, making it difficult to form definitive conclusions regarding the impact of isolation. For example, social isolation in mice has been found to lead to increased aggression21 and reduced sociability22,23 in certain studies, while other studies reported an increase in social preference following isolation24,25. Such discrepancies are also apparent in rat studies26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48 (Table 1), with some reporting increased social investigation42 and social interaction49 following social isolation, whereas others have shown no effect27,29,32,45 or even a decrease in social contact and other deficits in social behavior26,32,34,38,50.

Table 1 Summary of studies on social isolation in rat models.
Full size table

Much of the inconsistency in social behavioral phenotypes reported following isolation can be attributed to differences in experimental design. Those include variations not only in species, strain, or sex used across studies but also in the isolation protocol employed. For instance, some studies implemented an acute isolation period ranging from 24 hours51,52 to 7 days53,54, while others imposed chronic isolation starting at weaning and lasting throughout the animals’ lifespan33,42,44,45,46,47,50,55. Importantly, studies that investigate juvenile social isolation (jSI) also varied in their design. Certain studies limited isolation to a specific developmental period and re-socialized animals before testing26,27,29,32,34, while others extended jSI into adulthood33,43,45,47,56,57,58, with this latter design making it challenging to determine whether behavioral consequences observed are due to stress during development, isolation throughout adulthood, or the stress of isolation housing experienced by animals tested while still in isolation. Therefore, behavioral phenotypes observed immediately following jSI without a period of resocialization27,33,36,38,39,42,43,44,46,49,59,60,61,62,63,64,65 leaves open the question of whether these behavioral changes are indeed long-term consequences of social deprivation during early development or a result of ongoing stress from the animals’ state of isolation housing.

To determine whether social interaction during the juvenile stage is a critical window of sensitivity to isolation resulting in social deficits in adulthood, we designed a comprehensive behavioral study that takes into consideration the variability issues observed across published studies and discussed above.

Methods

Animals

Male and female Sprague Dawley rats (Charles River Laboratories, Wilmington, MA, USA) were kept under veterinary supervision in a 12-h light/dark cycle (lights on 7:00–19:00 h) at 22 ± 2 °C and had access to food and water ad libitum, unless otherwise specified. Rats were divided into two cohorts, with each cohort undergoing a different isolation protocols. All experiments were conducted during the light phase of the light/dark cycle. Rats were handled by the experimenter and habituated to test chambers for three consecutive days before behavioral testing began. Rats were brought to the testing room at least one hour before the start of behavioral testing. All animal procedures were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee at the Icahn School of Medicine at Mount Sinai. The study is reported in compliance with ARRIVE guidelines.

Isolation protocols

jSI followed by rehousing (jSI-A)

In this protocol, rats were randomly assigned to either jSI (one rat per cage) or group housed (GH; two rats per cage) conditions at P21 and remained in these housing conditions until P42. At P42. At P43, all rats from both groups were rehoused with a novel age- and sex-matched rat until adulthood (approximately P60). Behavioral testing for these rats began at approximately P60. During the rehousing period and throughout the behavioral testing period, all rats were kept in pairs in group-housed conditions (Fig. 1).

Fig. 1
figure 1

Experimental design of the isolation protocols to assess the impact of juvenile social isolation on social behavior in adulthood. (A) jSI-A protocol: Rats were isolated from P21-P42 then re-housed with a novel and sex-matched cagemate at P42. Rats remained in group-housing until behavioral testing began at P60. (B) jSI-B protocol: Rats were isolated from P21-P42 and behavioral testing began at P42. Rats remained in isolation throughout the duration of behavioral testing. Parts of this figure was created with BioRender.com.

Full size image

jSI without rehousing (jSI-B)

In this protocol, rats were randomly assigned to either juvenile social isolation or GH conditions at P21 and remained in these housing conditions until P42. Behavioral testing for these rats began at P43, while remaining in isolation throughout the entire duration of the testing period (Fig. 1).

Behavioral tasks

Social/object vs. empty task

Rats were habituated to a test chamber (50 × 50 × 40 cm) which contained two empty compartments positioned at opposite corners of the arena, for 15 min, as previously described66. A social stimulus (a novel, sex-matched juvenile Sprague Dawley rat, 3–5 weeks old) or a moving object (toy rat) was then placed into one compartment, with the opposite compartment left empty. The location of the stimuli was alternated on each day and between animals for counterbalancing purposes. The stimuli were placed behind a wire mesh window in the compartment, allowing for exchange of visual, auditory, and olfactory cues while minimizing physical contact. The subject rat was then given 5 min to explore the chamber and interact with the stimuli (Fig. 2).

Fig. 2
figure 2

Timeline of behavioral assays. Parts of this figure was created with BioRender.com.

Full size image

Social/object vs. food task at satiety and after 48 h of food deprivation

Rats were habituated to the same test chamber, as described above. A social stimulus (a novel sex-matched juvenile Sprague Dawley rat, 3–5 weeks old) or a moving object (toy rat) was then placed in one compartment, while food (rodent chow pellets) was placed at the opposite compartment (with the location of each compartment counterbalanced across days). The subject rat was then given 5 min to explore the chamber. This task was repeated after 48 h of food deprivation. Novel juveniles were used on each day of the task, and the compartments containing stimuli were randomly rearranged to prevent habituation and learning (Fig. 2).

Open field test

Rats were placed at the center of a square arena (90 cm × 90 cm) and allowed to freely explore the arena for 60 minutes67. The orientation of the rat (the direction the rat was facing when placed into the chamber) was counterbalanced between rats to ensure there were no effects of preference for certain areas of the chamber (Fig. 2).

Behavioral analysis

The Social/Object vs. Empty Task and the Social/Object vs. Food Task were scored and quantified using TrackRodent, an open-source Matlab-based automated tracking system (https://github.com/shainetser/TrackRodent) that uses a body-based algorithm (WhiteRatBodyBased15_7_15 in the TrackRodent interface)68,69. Videos were de-identified in order to keep the experimenter blinded to the housing conditions while setting up the analysis. Behavioral data was then pulled from scored videos and analyzed using DeepPhenotyping, a custom Matlab-based graphical user interface (GUI) for visualizing and analyzing behavioral measurements acquired during social behavior tasks.

Investigation time: Investigation time was calculated as total investigation time of each stimulus during a testing session and in 1-min bins across the duration of the test session. Investigation time was automatically assessed based on active contact between the subject’s body and the stimulus’s chamber using the TrackRodent software69. Bout duration was defined as a continuous contact between the subject’s body and stimulus’s chamber with no gaps longer than 0.5 s.

Mean bout duration: Mean bout duration was defined as the average of all individual bouts of any duration during one testing session. Mean bout duration was calculated for each stimulus presented during testing.

Open field test: Raw data for the open field test were generated by scoring and analyzing videos using EthoVision XT (Noldus). The arena was virtually divided into a center zone and a peripheral zone. The total distance traveled during the test and time spent in the center zone were recorded.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 9.0 software (GraphPad Prism, San Diego, CA, USA). Data from both male and female rats were pooled and analyzed with sex as a factor to identify potential sex differences. A significant main effect of sex was observed in certain tasks; however, none of these tasks showed an interaction between housing condition and sex (Supplementary Table 1). We, therefore decided to present and visualize the behavioral data separately by sex to highlight meaningful differences in behaviors both between sexes and within housing conditions. Outliers were identified and assessed, and the data were tested for normality using appropriate statistical methods. The results from these statistical analyses, including analysis of the pooled data, are detailed in Supplementary Table 1. The image graphics were created using BioRender.com and Adobe Illustrator.

Results

To investigate whether social isolation during a critical developmental period has long-term effects on social behavior in adulthood, we assessed social behavior in two independent cohorts of Sprague Dawley rats, with each cohort subjected to a distinct social isolation protocol (Fig. 1). In the first protocol, named jSI-A, rats were isolated during the juvenile stage, from early to mid-adolescence (P21 to P42), then re-housed in pairs, and assessed for social behavior in adulthood (starting at P60). The rationale behind this design is to investigate whether isolation limited to this specific period (P21 to P42) has a long-lasting effect on social behavior. In the second isolation protocol, named jSI-B, rats were also isolated from P21 to P42, with testing beginning immediately after isolation, at P42. The rationale behind this design is to ensure that any absence of deficits following the previously outlined jSI-A protocol is not attributed to potential amelioration from undergoing re-housing and re-socialization within that protocol. For each cohort of isolated rats, we compared their behavior to that of littermates reared in group housing (GH). Finally, to examine any potential sex differences resulting from isolation, male and female rats were used throughout the study, with data from male and female rats presented separately.

Juvenile social isolation has no impact on social preference in adolescence or adulthood

To assess social preference, we tested both cohorts of rats on the Social vs. Empty task. In this task, a test rat is placed in an arena with two compartments located at opposite corners, then a sex-matched juvenile rat (the social stimulus) is placed in one compartment, while the other remains empty (Fig. 2). To determine whether any observed behavioral differences are specific to social interactions, we included an Object vs. Empty task, which follows the same design as the Social vs. Empty task but substitutes the social stimulus with an object (a moving toy rat).

On the Social vs. Empty task, we found that, rats in the jSI-A and jSI-B condition showed no significant difference in total investigation time (Fig. 3A), investigation behavior across time (Fig. 3B), mean bout duration (Fig. 3C), when compared to their GH littermates, regardless of sex. Similarly, rats in the jSI-B condition showed no differences in total investigation time (Fig. 3D) or investigation behavior across time (Fig. 3E). However, both males and females showed a significant increase in mean bout duration when interacting with the social stimulus compared to their GH littermates (Fig. 3F), suggesting increased interest in the social stimulus immediately after isolation. In the Object vs. Empty task, no significant differences were observed between the jSI-A rats and their GH littermates (Supplementary Figs. 1A-C), where all groups showed similar pattern of behavior toward the object or the empty compartment. Similarly, no significant differences were observed between the jSI-B rats and their GH littermates (Supplementary Figs. 1D-F); in this isolation protocol, both male and female rats in GH and jSI-B conditions showed greater investigation of the object stimulus (Supplementary Fig. 1D), with the only exception being the male GH littermates of the jSI-B group, which still exhibited a preference for the object stimulus though this preference did not reach statistical significance (Supplementary Fig. 1D, Top). This increased investigation of the object, observed in the jSI-B group and their GH littermates, which were both tested during juvenility, but not in the jSI-A rats and their littermates, which were tested in adulthood, suggests that an increased curiosity towards novel, non-social stimuli, may be characteristic of this developmental stage, regardless of the isolation condition. Collectively, our findings demonstrate that isolation limited to juvenility does not have a long-lasting impact on social preference in adulthood.

Fig. 3
figure 3

Social preference behavior in male and female rats is unaffected by duration and timing of isolation in the Social vs. Empty task. (A) Mean total time spent investigating a social stimulus or the empty following jSI at P21-P42 and subsequent re-housing until P60 (jSI-A). Male GH, n = 14; male jSI-A, n = 14; female GH, n = 17; female jSI-B, n = 16. (B) Mean total investigation time for the social stimulus and the empty chamber plotted in 1-min intervals over the 5-min testing period. (C) Mean bout duration representing the average length of time spent investigating the social stimulus or the empty chamber. (D-F) Data for social reward-seeking behavior observed immediately after jSI isolation at P21-P42, without rehousing (jSI-B), using the same analysis as in A-C. Male GH, n = 12; male jSI-B, n = 17; female GH, n = 13; female jSI-B, n = 10. Statistical significance indicated as ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 based on post-hoc tests following the main effect. Error bars represent standard error of the mean (SEM).

Full size image

Juvenile social isolation has no impact on social reward-seeking behavior in adulthood

To determine if social isolation during early development impacts social reward-seeking behavior, we used another social behavior task: the Social vs. Food task. In this task, test rats are placed in the same arena as the Social vs. Empty task (Fig. 2), but with food placed in the empty chamber to serve as a competing non-social rewarding stimulus. To examine if any behavioral differences we observed were specific to social reward-seeking, we conducted another independent task with the same set-up but replaced the social stimulus with an object (a moving toy rat). This task is referred to as Object vs. Food task.

We found that both male and female jSI-A rats showed no changes in total investigation time (Fig. 4A), investigation behavior across time (Fig. 4B), or mean bout duration for either the social or food stimulus (Fig. 4C), suggesting that isolation, when limited to P21–P42, does not have a long-term impact on social reward-seeking behavior in adulthood. To examine if any behavioral changes can be observed immediately after isolation during adolescence (from P21–P42), we tested the jSI-B cohort using the same behavioral task. In this cohort, we found that both male and female rats also showed no changes in total investigation time (Fig. 4D), investigation behavior across time (Fig. 4E), or mean bout duration towards the two stimuli (Fig. 4F). When the jSI-A and jSI-B cohorts were tested in the Object vs. Food task, where a moving toy rat was presented instead of a social stimulus, no significant differences were observed between the isolation groups and their respective GH littermates (Supplementary Fig. 2). The only exception was in the male jSI-A group, which displayed increased interest in the object (Supplementary Fig. 2A, Top); however, this difference was not notably distinct from their GH littermates, particularly when analyzing behavior over time (Supplementary Fig. 2B, Top). Collectively, our findings demonstrate that isolation limited to adolescence (P21-P42) has no impact on social reward-seeking behavior in adulthood.

Fig. 4
figure 4

At satiety, social reward-seeking behavior is unaffected by juvenile social isolation. (A) Mean total time spent investigating a social stimulus or a chamber containing food following jSI at P21-P42 and subsequent re-housing until P60 (jSI-A). Male GH, n = 15; male jSI-A, n = 15; female GH, n = 18; female jSI-A, n = 16. (B) Mean total investigation time for the social stimulus and a chamber containing food plotted in 1-min intervals over the 5-min testing period. C. Mean bout duration representing the average length of time spent investigating the social stimulus or the chamber containing food. (D-F) Data for social reward-seeking behavior observed immediately after jSI at P21-P42, without to re-housing (jSI-B), using the same analysis as in A-C. Male GH, n = 12; male jSI-B, n = 16; female GH, n = 12; female jSI-B, n = 8. Statistical significance indicated as ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 based on post-hoc tests following the main effect. Error bars represent standard error of the mean (SEM).

Full size image

Juvenile social isolation has no impact on the ability of rats to adapt social reward-seeking behavior in adulthood based on changing value of the presented reward.

To examine whether rats adapt their behavior based on the perceived value of social reward, we once again employed the Social vs. Food task but this time, testing took place after 48 h of food deprivation. Unlike testing under satiated conditions, food-deprived rats were expected to adapt their behavior by displaying increased interest in the food stimulus (demonstrated by increased time spent investigating the food stimulus), while still maintaining some interest in the social stimulus. However, time spent interacting with the social stimulus was anticipated to be reduced compared to the satiated state.

We found that in the jSI-A cohort, both GH and jSI-A male and female rats adapted their behavior to spend significantly more time investigating the food stimulus compared to the social stimulus (Fig. 5A), which was clearly reflected across the 5 min of testing (Fig. 5B), along with no differences between the housing conditions or the two sexes in mean bout duration (Fig. 5C). Notably, when we examined the jSI-B cohort and their GH littermates on the same task, we observed that while male and female rats raised in GH conditions exhibited a clear shift in investigation behavior, spending significantly more time with the food stimulus (Fig. 5D) (with male rats showing significantly higher mean bout duration when investigating the food stimulus (Fig. 5D, Top), both male and female rats in the jSI-B cohort did not display a similar shift and spent equal time investigating both the food and social stimulus despite food deprivation (Fig. 5D). This behavior could likely be due to the jSI-B rats’ sustained motivation in seeking the social stimulus following isolation. This behavior was also clearly reflected throughout the investigation behavior across the 5-min testing period in male and female rats (Fig. 5E). Additionally, the average bout duration for investigating the food stimulus was significantly higher in male GH rats (Fig. 5F), again indicating that male rats raised in GH conditions had shifted their behavior following food deprivation to spend longer average bouts investigating the food stimulus while male jSI-B rats did not display this same shift. Finally, to ensure that the observed behavioral changes in this task are specific to the social behavior, we tested both cohorts on the Object vs. Food task following 48 h of food deprivation and found no significant differences between any of the isolation groups and their respective GH littermates in either sex with the total average investigation time across all groups shifting toward increased investigation of the food stimulus (Supplementary Fig. 3). Together, these findings indicate that isolation during juvenility has no long-term impact on the rats’ ability to shift behavior between a social and non-social rewarding stimulus and that the state of isolation-housing itself may affect the animal’s motivation to seek social reward.

Fig. 5
figure 5

Adaptation of social reward-seeking behavior is affected by juvenile social isolation. (A) Mean total time spent investigating a social stimulus or a chamber containing food following jSI at P21-P42 and subsequent re-housing until P60 (jSI-A), following the same analysis detailed as in A-C. Male GH, n = 15; male jSI-A, n = 15; female GH, n = 18; female jSI-A, n = 15. (B) Mean total investigation time for the social stimulus and a chamber containing food plotted in 1-min intervals over the 5-min testing period. (C) Mean bout duration representing the average length of time spent investigating the social stimulus or the chamber containing food. (D-F) Data for social reward-seeking behavior observed immediately after jSI at P21-P42, without re-housing (jSI-B), using the same analysis as in A-C. Male GH, n = 8; male jSI-B, n = 10; female GH, n = 12; female jSI-B, n = 9. Statistical significance indicated as ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 based on post-hoc tests following the main effect. Error bars represent standard error of the mean (SEM).

Full size image

Juvenile social isolation does not result in changes in locomotor activity or anxiety-like behavior

Finally, to ensure that the observed changes in social behavior are not confounded by changes in locomotor activity or anxiety, all rats were examined in an open field test (Fig. 2) to analyze the distance traveled and time spent in the center zone of the arena. We found no differences in distance traveled between rats from either isolation protocol and their GH littermates, indicating that locomotor activity was not impacted by isolation or in time spent in the center zone (Supplementary Fig. 4). These findings provide further support that the differences in social behavior observed following isolation are specifically associated with social behavior and are not driven by changes in locomotor activity or anxiety potentially stemming from the experience of isolation.

Discussion

Rodent models have been widely employed to investigate the long-term effects of isolation on social behavior 19,70,71,72,73,74,75,76; however, variability in experimental design has impeded our understanding of how isolation during early development influences social behavior in adulthood. For example, in studies focusing on adolescent isolation in rats, some have employed isolation protocols limited to specific developmental periods26,27,28,30,32,33,34,35,38,48,77, while others have extended the isolation period into adulthood33,43,44,45,46,47,55,65,78,79,80. This inconsistency, along with differences in factors such as strains, sex, and age at behavioral testing, has resulted in contradictory findings in the literature, with some studies suggesting impaired social behavior following isolation26,31,32,34,46,77, others reporting no effect29,35, and some even reporting increased social interaction27,37,41,42,47,49. The objective of the current study is to investigate the long-term impact of adolescent social isolation by using a carefully designed approach that compares different isolation protocols within the same rat strain and across both sexes and minimizes variability while evaluating social behavioral outcomes across these protocols.

A commonly used protocol in studies on adolescent isolation involves housing rats in isolation starting during adolescence and continuing into adulthood at which point behavioral testing is conducted. Findings from these studies were inconsistent, likely due to variations in experimental conditions, including differences in rat strains. Some studies used Sprague Dawley, while others tested Wistars, which naturally exhibit lower levels of social interaction and play compared to Sprague Dawley rats81and have been found to differ in coping behavior and susceptibility to socially conditioned fear82, and therefore, may respond differently to social isolation. Even studies in Wistar rats using comparable isolation protocols have yielded mixed results. For example, a study by Oliveira et al. in male and female Wistar rats found no social preference following isolation when assessed using a social preference paradigm45. However, a separate study in male Wistar rats by Amiri et al. reported social deficits during free social interaction46, whereas Shirenova et al. found that a similar isolation protocol in females increased sociability on the free social interaction task but reduced preference for social novelty when assessed in the three chamber test47. Notably, this last study also demonstrated that extending the isolation period to 8 months resulted in social deficits and a significant decline in preference for social novelty47, indicating that changes in social behavior following isolation extending into adulthood are not uniform across the lifespan and can be dependent upon the duration of isolation. However, a study by Douglas et al., which tested Sprague Dawley rats of both sexes using an isolation protocol that extended from juvenility into adulthood and applying the social conditioned place preference task, found that socially-isolated adult rats spent more time engaging in social investigation compared to group-housed or adolescent rats in isolation42.

While the aforementioned observations are intriguing and warrant further investigation, it is crucial to recognize that using a protocol in which isolation extends into adulthood to study the lasting effects of adolescent isolation is not ideal as the extension of isolation into adulthood complicates the ability to determine whether the observed behavioral changes are truly long-term effects of adolescent social isolation, the result of adult isolation, a combination of both, or simply the direct effect of isolation housing. This uncertainty highlights the need for isolation protocols that restrict the isolation period to adolescence, which have also been used in the literature26,27,28,30,32,33,34,35,38,48,77 and within this study. Here, we specifically limited isolation to a defined adolescent period (P21-P42, juvenile period) by rehousing the rats at P42 and testing them in adulthood, using a protocol we termed jSI-A. We included both sexes to investigate potential sex differences, as some prior studies have highlighted distinct behavioral outcomes in each sex following isolation28,34,35,48,58 (Table 1). While a significant main effect of sex was observed in certain tasks, these effects were not driven by the isolation condition, as no interaction between housing condition and sex was found in any of these tasks (Supplementary Table 1).

We found that following jSI-A, rats’ social preference on the Social vs. Empty task and their social reward-seeking behavior on the Social. vs. Food tasks were similar to those of the GH rats. In agreement with our findings, previous studies that have incorporated resocialization into their isolation protocols have found that social experience can reverse altered behaviors detected following adolescent isolation. For example, Meng et al. reported that male Sprague Dawley rats in isolation displayed significantly higher social interaction during free social interaction compared to socially-reared rats. Similar to our findings, these differences did not manifest following resocialization27. The ameliorative effects of resocialization have also been detected in other behavioral readouts. For example, Seffer et al. found that one week of group-housing in male Wistar rats partially restored approach behavior in response to pro-social ultrasonic vocalizations after post-weaning isolation38. Similarly, Begni et al. demonstrated that returning Listar Hooded male rats to group-housing after adolescent isolation normalized increased locomotor activity33. These findings, along with the results of our current study, suggest that limited adolescent social isolation does not have a long-lasting impact on adult social behavior. However, it is important to acknowledge that several studies have reported persistent social deficits that resocialization failed to ameliorate26,31,34,77. As mentioned earlier, these discrepancies across studies may stem from variation in strain or sex used but could also result from differences in the isolation protocol itself, such as the duration of isolation and the length of resocialization used in these versions of jSI-A. For instance, Tulogdi et al. found an increase in aggressive behaviors in male Wistar rats that underwent 3 weeks of resocialization following 7 weeks of post-weaning social isolation starting at P2131. In contrast, P. Graf et al. reported that male and female Wistar rats isolated for just 3 weeks starting at P21 displayed no differences in social interaction compared to group-housed rats after approximately 7 weeks of resocialization35, suggesting that resocialization may only be effective in mitigating behavioral deficits if it takes place during an earlier developmental stage.

To determine whether the lack of social deficits in adulthood after limited juvenile social isolation was not merely due to the restorative effects of re-socialization, we assessed social behavior following the jSI-B protocol. Our findings revealed that while rats exhibited typical social preference in the Social vs. Empty Task and normal social reward-seeking in the Social vs. Food Task under satiety conditions, they displayed differences in behavioral adaptation in the Social vs. Food Task following food deprivation. This suggests an increased motivation for social interaction driven by the immediate effects of isolation housing, a phenomenon that became evident only after presenting rats with competing social and non-social rewards in the task. Previous studies have also observed an increase in social behavior immediately after adolescent isolation while animals are still in isolation housing, though using different social behavioral assays. For example, a study by Douglas et al. that employed a similar isolation protocol and assessed free social interaction with a novel juvenile rat found increased play fighting behavior in both male and female Sprague Dawley rats42. Similarly, several other studies using free social interaction reported increased social interaction following isolation during adolescence across different strains and both sexes27,37,39; notably, these studies also observed that isolated rats displayed increased aggressive behaviors towards the social stimulus. Considering the association between aggression and social stressors83, such as isolation70, the increased social interaction observed in these studies may be driven by aggression rather than affiliative, pro-social behavior. Importantly, some studies reported the opposite effect of adolescence social isolation on social behavior. For example, Seffer et al. found that male Wistar rats isolated for 4 weeks starting at P21 showed a lack of approach behavior specifically in response to pro-social vocalizations prior to rehousing38, while Dawud et al. observed that isolated male Sprague Dawley rats displayed an increase in avoidance behavior during free social interaction49.

In conclusion, our findings suggest that social isolation limited specifically to adolescence does not have a long-lasting effect on social investigation behavior and that the state of social isolation itself results in increased motivation for seeking social interaction. Notably, even acute social isolation lasting as little as 2 h has been shown to increase social investigation in Sprague Dawley rats84, highlighting how the state of isolation itself can trigger enhanced social behavior. This underscores the importance of rehousing when studying the effects of isolation during specific developmental periods and the need for caution when interpreting results from studies that purport to investigate adolescent-specific isolation but use isolation protocols that include both adolescence and adulthood. Additionally, our findings that no differences were detected in social preference on the Social vs. Empty task while changes in social reward-seeking were evident in the Social vs. Food task emphasizes the need for carefully choosing behavioral assays and using multiple tasks to address scientific questions and capture different features of social behavior. Finally, social behavior encompasses both affiliative (pro-social) and aggressive (anti-social or aberrant) behaviors and even with the use of multiple tasks in our study, there are still limitations, as we cannot definitively assert that the sustained social investigation behavior directly correlates with increased motivation for pro-social, affiliative social behavior rather than aggression, especially since the social stimuli in our tasks were confined to chambers where test rats could not engage in free social interaction. Moreover, studies in both mice85,86,87 and rats88 have suggested that aggression is also perceived as rewarding, further complicating the interpretation of increased social investigation as purely indicative of pro-social motivation. In light of these considerations, future studies should refine isolation protocols and incorporate a wider array of behavioral measures to gain a deeper understanding of how social isolation affects behavior, ultimately providing a more nuanced perspective on its impact on motivation for social interaction.