Post-weaning social isolation increases reward-seeking behavior in mice

Abstract

Social isolation is a growing public health concern. Although isolation at any age is harmful, previous studies have shown that isolation during adolescence, correlating with critical periods of brain development, can impair cognitive function and increase the risk for psychiatric illness later in life. In this study, we utilized a mouse model of social isolation (SI) during adolescence (postnatal day 21–35) and compared performance of isolated and group-housed mice on a touchscreen-based continuous performance test (rCPT) and fixed ratio/progressive ratio (FR/PR) tasks in adulthood. SI improved performance in the rCPT and the improvement in performance was consistent across time bins within the 45-minute testing session. There were no effects of SI on reaction times or reward retrieval latencies. A possible confound for performance in the rCPT would be SI-induced changes in reward-seeking or motivation for the strawberry milk reward. We next compared the SI mice to their group-housed littermate controls on both PR and FR schedules of reinforcement and found that SI mice had higher breakpoints on a PR4 schedule and earned significantly more reinforcers on an FR1 schedule of reinforcement compared to their group-housed littermates, suggesting that high performance in the CPT may be due to increased motivation for food rewards. These data indicate that SI during adolescence has significant effects on reward-seeking behavior in adult mice and may provide a useful behavioral model for studying the link between SI and risk for neuropsychiatric disorders.

1. Introduction

Social isolation (SI) is a stressor with significant acute and chronic effects, including increased risk for developing psychiatric disorders. Specifically, there is a clear link between SI and anxiety, depression, substance use disorders, and schizophrenia [1]. Despite the confirmation of SI as a risk factor, the biological link between SI and psychiatric disorders is not completely understood.

The relevance of SI as a source of environmental stress has increased dramatically over the past few years due to the ongoing effects of the COVID-19 pandemic. It is too early to definitively assess the long-term societal and economic impacts of the increase in SI due to social distancing measures, but its role in public health is being closely monitored [2–4]. A deeper understanding of the biological effects of SI will be critical to address the long-term effects that follow the sharp increase in SI within the population.

There is extensive literature on the effects of SI on older people (Reviewed in [5]), but recent studies have identified increases in the prevalence of SI and loneliness among children and adolescents [6]. SI in younger people is particularly important for disorders with significant neurodevelopmental components, including schizophrenia and substance use disorders [7]. Given the increased plasticity of the brain during childhood and adolescence [8], SI may have different or greater effects during these developmental periods compared to SI in adults.

SI has also been studied in animal models, and both chronic and acute isolation have been associated with anxiety- and depression-like behavior in rodents [9]. SI during postnatal days 21–35 produces deficits in myelination in the prelimbic and infralimbic cortices through an IL-6-dependent mechanism [10,11]. Additionally, SI during this period also enhances the activity of interneurons in the prelimbic and infralimbic cortices [12].

It is not clear why SI during postnatal days 21–35 produces such drastic effects on the structure and function of the brain. Interestingly, postnatal days 21–35 is a period of significant amounts of social play behavior in mice and isolation during this period can lead to deficits in social interactions in adulthood [13,14]. Moreover, the social behavior deficits are directly related to abnormal interneuron function in the prefrontal cortex [14]

We previously reported that SI during the 21–35 day period is characterized by increased levels of ΔFosB, a transcription factor associated with chronic stress[15]. ΔFosB has also been implicated in studies of reward-seeking behavior[16–18]. Given the developmental trajectory of the cortex and connected reward circuits [19], these studies suggest there may be a link between SI, ΔFosB, and cognitive function, particularly reward-seeking behavior.

Touchscreen-based behavioral tasks provide a flexible, sensitive, and translational platform for investigating cognitive function [20]. Sustained attention is a cognitive domain disrupted in neuropsychiatric disorders [21,22]. The rodent continuous performance test (rCPT) is a widely-used and translational behavioral assay sensitive to changes in sustained attention [23–26]. Fixed ratio (FR) and progressive ratio (PR) tasks can similarly detect changes in motivation and reward-related behavior [18]. Here, in a mouse model of adolescent social isolation, we used a cognitive testing battery consisting of rCPT, PR, and FR assays to interrogate the effects of adolescent SI on attention and motivation in adulthood. We found that SI during adolescence significantly affects performance across all tasks. SI mice demonstrated better performance in the rCPT, which is usually indicative of improved attention function. However, SI mice also have higher breakpoints and total responses in the PR and FR tasks, indicating increased reward-seeking behavior. Together, these results suggest that SI consistently increases reward-seeking and motivation for highly palatable food and this effect may serve as a confound for behavioral tasks designed to test other cognitive constructs that utilize food as a primary reinforcer.

2. Material and methods

2.1. Mice

Breeding pairs of C57BL/6J mice were purchased at 6–7 weeks old from The Jackson Laboratory (Bar Harbor, ME, USA). Only one litter from each breeding pair was used in these experiments. Pups were raised according to an isolation protocol adapted from the protocol described by Makinodan and colleagues [10]. Pups were weaned at postnatal day 21 (P21) and housed either in isolation or in a group of three same sex littermates. Isolated mice were rehoused with another same sex isolated littermate at P35 for the duration of the experiment (Figure 1a). Separate cohorts were used for the rCPT and the FR/PR experiments.

Figure 1. Isolation rearing schematic and behavioral testing battery.

a) Isolation rearing. Mice are weaned at postnatal day 21 and assigned toeither group housing or isolation. At day 35, isolated mice are rehoused with another isolated littermate. Testing began at postnatal day 63. b) Continuous performance test (rCPT) progression and scoring. c) Motivation test progression. “Minimum sessions” is the number of sessions above criteria required to advance to the next stage.

2.2. Handling, food restriction, and habituation

At P63, mice were placed on food restriction and briefly handled for three days. Food amounts were set to keep the animals between 85–90% of their free feeding weight based on growth charts provided by The Jackson Laboratory. During the initial handling days, mice were also introduced to a Bussey-Saksida testing chamber (Lafayette Instrument, Lafayette, IN, USA) for habituation. Mice were placed in the chamber and 1 mL of reward (strawberry Nesquik, Nestle, Vevey, Switzerland) was placed in the feeding tray. The habituation schedule consisted of a 30-minute session where the screen was responsive to touch, but touches were not rewarded. Mice were considered habituated when they had undergone at least three habituation sessions and had fully consumed the reward during at least one session.

2.3. Continuous performance test (CPT)

CPT training consists of four stages (Figure 1b, adapted from [23]).

In Stage 1, mice were trained to touch a visual stimulus (white square). The square was displayed (stimulus duration) for 10 seconds. The stimulus duration was followed by a 0.5 second limited-hold (LH) period during which the screen was blank, but a touch would still yield a reward. Upon interacting with the stimulus, a one-second 3 kHz tone would sound, a small amount of reward (20 μL) would be dispensed, and the reward tray would be illuminated. Head entry into the reward tray was detected by an infrared (IR) beam, and upon head entry the intertrial interval (ITI) of 2 seconds would begin. If the mouse did not interact with the stimulus, the ITI would begin immediately following the LH period. The next trial would begin immediately following the ITI. The criterion for advancement to Stage 2 was obtaining 60 rewards within a single 45-minute session.

In Stage 2, the white square pattern was replaced with either horizontal or vertical bars – the mouse’s assigned positive stimulus, or S+. The S+ was counterbalanced within groups. The SD was reduced from 10 seconds to 2 seconds, and the LH was increased from 0.5 seconds to 2.5 seconds. The criterion for advancement was the same as Stage 1.

In Stage 3, a negative stimulus (S−; snowflake pattern) was added. Each trial had a 50% chance of being either an S+ or S− trial. SD and LH were identical to Stage 2, but ITI was increased to 5 seconds. Interacting with S− during the SD or LH would not yield a reward and would start the ITI. The criteria for advancement to Stage 4 were a minimum of seven sessions, during which at least two consecutive sessions had a discrimination index (d’) score of 0.6 or higher. A discrimination index is a measurement derived from signal detection theory [27] used as a composite measure of performance that takes into account both errors of omission (misses) and errors of commission (mistakes). D’ was calculated as follows:

$$d^{\prime }=z\left(\frac{\mathit{correct}}{\mathit{correct}+\mathit{miss}}\right)-z\left(\frac{\mathit{mistake}}{\mathit{mistake}+\mathit{rejection}}\right)$$

In Stage 4, the snowflake S− pattern from Stage 3 was replaced with four new S− patterns. The S+ had a 50% chance of appearing, and the remaining 50% was split among the four S− images. Interacting with S− in this stage triggered correction trials, which consisted only of S− trials until the mouse withheld response. Parameters and scoring criteria were identical to Stage 3, but animals were tested for twenty sessions.

2.4. Fixed ratio (FR)/progressive ratio (PR) testing

FR/PR testing consisted of several stages of varying difficulty (Figure 1c). Fixed ratio 1 (FR1) required a mouse to touch a white square, similar to CPT Stage 1 without the stimulus duration timer. The session ended after 45 minutes, or when a mouse obtained 30 rewards within one session. After a mouse obtained 30 rewards within a single session, they progressed onto the next stage. FR2 and FR3 followed the same rules as FR1, but with the respective number of touches per reward. FR5 consisted of multiple sessions in which the mouse had to touch the square five times to obtain a reward. The same timing and cutoff rules from the previous stages applied, but mice were required to obtain 30 rewards during each of at least three sessions before advancing to progressive ratio testing. Progressive ratio 4 (PR4) followed the same parameters as the previous FR stages, but the number of touches required for a reward increased by four each time a reward was obtained. The initial PR4 phase consisted of seven sessions, and during each session the break point of each mouse was recorded. The break point was defined as the highest number of touches committed within one session to obtain a reward.

Following the initial PR4 phase, mice were again tested on FR1, but the 30-reward cap was removed. The number of rewards obtained per session was recorded. After the free-feeding sessions, mice underwent another seven-session PR4 phase.

3. Data analysis

Data from the CPT was analyzed using either two-way ANOVAs with sex and housing as the independent variables or three-way ANOVAs with sex, housing and session number as the independent variables and Greenhouse-Geisser corrections, when needed. Discrimination indices from the last ten sessions of stage 4 were used in analysis. We chose to use the last ten sessions because individual performance was stable by this point in testing. The average break points from the fourteen total PR4 sessions for each mouse and the number of rewards earned during the unlimited FR1 session between PR4 blocks were used for our analyses. A two-way ANOVA with sex and housing as the independent variables was used for both the break point and FR1 analyses. Sphericity was not assumed and the Geisser-Greenhouse correction was used, when needed. Where appropriate, Tukey’s post hoc test for multiple comparisons was used. ANOVA main effects, ANOVA interactions, and post hoc pairwise comparisons were deemed statistically significant with a p-value < 0.05. All data are presented as mean ± SEM. All statistical analyses were conducted using GraphPad Prism (Version 10.4.1; Graphpad Software, LLC; San Diego, CA, USA).

4. Results

4.1. SI did not affect progression through the early stages of the CPT, but impaired female performance in Stage 3

Neither housing nor sex affected acquisition of the task rules in Stage 1 (Housing: F_1,23 = 0.5490, p = 0.4662; Sex: F_1,23 = 0.5957, p = 0.4481; Housing X Sex: F_1,23 = 0.9759, p = 0.3335, Table 1). In stage 2, all mice reached criterion in a single session. Similar to Stage 1, neither housing nor sex affected the number of sessions required to reach criterion in Stage 3 (Housing: F_1,23 = 0.0820, p = 0.7772; Sex: F_1,23 = 0.0820, p = 0.7772; Housing X Sex: F_1,23 = 4.018, p = 0.0569). Additionally, neither housing nor sex alone affected the d’ score upon reaching criterion (Housing: F_1,23 = 0.9956, p = 0.3288; Sex: F_1,23 = 0.1988, p = 0.6599). However, there was a significant Housing X Sex interaction (F_1,23 = 11.07, p = 0.0029) where the female SI group had a significantly lower d' score than the female group-housed cohort (p = 0.0296).

Table 1.

Sessions to completion for the first three stages of CPT. Mean (± SEM) sessions required to reach criterion in each of the training stages of the continuous performance test for each rearing condition.

	Male		Female
	Group	SI	Group	SI
Stage 1	4.67 ± 0.80	4.86 ± 0.99	4.83 ± 0.94	3.50 ± 0.27
Stage 2	1 ± 0	1 ± 0	1 ± 0	1 ± 0
Stage 3	9.67 ± 2.90	7 ± 0	7 ± 0	9.00 ± 0.94

4.2. SI improves performance in CPT stage 4 compared to group-housed controls

Following completion of the three training stages, we tested mice for twenty sessions using Stage 4 parameters (Figure 1). As previously shown [23], performance (d’) improves across Stage 4 sessions (F_{6.751, 155.3} = 29.28, p < 0.0001; Figure 2a). Additionally, there were main effects of session (F_{6.702, 154.1} = 11.98, p < 0.0001 and housing (F_{1, 23} = 6.524, p = 0.0177) and a significant Session X Housing interaction (F_{19, 437} = 1.630, p < 0.0457) on the hit rate with SI mice having higher hit rates than group-housed mice (Figure 2b). With respect to the false alarm rate, there were main effects of Session (F_{7.714, 177.4} = 15.69, p < 0.0001), Sex (F_{1, 23} = 16.74, p = 0.0004, and Housing (F_{1, 23} = 8.177, p < 0.0089) with the false alarm rate decreasing across sessions, female mice having higher false alarm rates, and SI mice having higher false alarm rates compared to group-housed mice (Figure 2c). Given the large increase in d’ across the 20 sessions, we decided to focus our additional Stage 4 analyses on the last ten sessions, where performance was stable (F_{2.994, 68.85} = 2.040, p = 0.1164; Figure 2d). During these last ten sessions, there was a significant main effect of housing (F_{1, 23} = 8.572, p = 0.0076), but no effect of sex (F_{1, 23} = 3.118, p = 0.0907) on d’. There were no significant interactions. Interestingly, isolated animals exhibited better performance compared to their group-housed littermates.

Figure 2. Performance in Stage 4 of the rCPT is affected by SI.

a) Discrimination index (d’) over the twenty sessions of Stage 4. b) Hit rate over the twenty sessions of Stage 4. c) False alarm rate over the twenty sessions of Stage 4. d) Discrimination index over the last ten sessions. e) Hit rate (HR) over the last ten sessions. f) False alarm rate (FAR) over the last ten sessions. n = 6 group-housed males, 7 SI males, 6 group-housed females, and 8 SI females. Data are mean ± SEM.

To determine what contributed to the higher sensitivity in isolated mice, we also analyzed the hit and false alarm rates. We found that SI increased the average hit rate (F_{1, 23} = 10.27, p = 0.0039; Figure 2e). There was no effect of sex (F_{1, 23} = 0.1547, p = 0.6977) on hit rate or interaction effects. Interestingly, there was no effect of housing on the false alarm rate (F_{1, 23} = 3.120, p = 0.0906; Figure 2f), but there was a main effect of sex (F_{1, 23} = 8.455, p = 0.0079) with female mice having significantly higher false alarm rates. There were no interaction effects on the false alarm rate.

4.3. SI did not affect reaction time in males or females

Reaction time (RT) was measured between S+ presentation and screen touch (correct RT), S− presentation and screen touch (mistake RT), and screen touch and reward collection (reward latency). Neither sex nor housing significantly affected hit RT (Figure 3a, Housing: F_1,23 = 0.1155, p = 0.7371; Sex: F_1,23 = 2.440, p = 0.1319). False alarm RT did not show a housing effect (F_1,23 = 0.2791, p = 0.6023), but did show a sex effect, with females being slower to produce a false alarm response than males (Figure 3b, F_1,23 = 5.915, p = 0.0232). Reward collection time like correct RT was not significantly affected by housing (F_1,23 = 1.801, p = 0.1927) or sex, even though a few males had much higher than average latencies (Figure 3c, F_1,23 = 4.277, p = 0.0501).

Figure 3. Reaction time (RT) and reward latency in the rCPT are unaffected by sex or rearing.

a) RT to S+ (hit) in State 4 of the rCPT. b) RT to S− (mistake) in Stage 4 of the rCPT. c) Reward latency after hits in Stage 4 of the rCPT. n = 6 group-housed males, 7 SI males, 6 group-housed females, and 8 SI females. Data are mean ± SEM.

4.4. Time bin analysis of CPT

To determine if there were any time-course effects of housing or sex on performance during the last 10 sessions of Stage 4, we analyzed CPT data in three 15-minute bins (0–15, 15–30, and 30–45 minutes; Figure 4). In the first 15-minute period, we found that there was a significant effect of sex (F_1,23 = 7.853, p = 0.0101) with males producing higher d’ values. There was no significant effect of housing (F_1,23 = 2.572, p = 0.1224). In the second time bin, both sex (F_1,23 = 4.876, p = 0.0375) and housing (F_1,23 = 4.789, p = 0.0391) affected performance with males performing better than females and SI mice performing better than group-housed mice. In the third time bin, performance increased across sessions (F_4.143,92.99 = 2.931, p = 0.0234) and SI mice performed better than their group-housed littermates (F_1,23 = 5.283, p = 0.0310), suggesting that SI mice are able to maintain higher levels of performance over time. There were no significant sex X housing interactions in any of the time bins.

Figure 4. Time bin analysis of performance during Stage 4 of the rCPT.

a) 0–15 minutes. b) 15–30 minutes. C) 30–45 minutes. n = 6 group-housed males, 7 SI males, 6 group-housed females, and 8 SI females. Data are mean ± SEM.

4.5. SI increases reward-seeking behavior in progressive and fixed-ratio tasks

After observing that SI mice exhibited better performance, as measured by d’, in the rCPT, a test sensitive to changes in attention, we tested males and females in both progressive (PR4) and fixed-ratio (FR1) tests sensitive to changes in motivation and reward-seeking behavior. These tests allowed us to determine whether the observed changes in rCPT performance were being driven by an increase in attentional reserves or changes in reward-seeking behavior. We found that housing status significantly affected the breakpoint (F_1,39 = 9.735, p = 0.0034) with SI mice having consistently higher breakpoints than group-housed mice. There was no effect of sex (F_1,39 = 1.239, p = 0.2725) or sex X housing interactions (F_1,39 = 0.4941, p = 0.4863). Similar effects were seen in the FR1 task. SI mice earned and consumed significantly more food rewards compared to group-housed mice (F_1,39 = 14.38, p = 0.0005). There were no effects of sex (F_1,39 = 2.086, p = 0.1566) or sex X housing interactions (F_1,39 = 0.7659, p = 0.3868). These data suggest that significant changes in reward processing and motivation may contribute to the altered rCPT performance we observed in SI mice.

5. Discussion

Adolescent social isolation causes measurable behavioral changes that persist through adulthood. Using the rCPT, we found that SI increased sensitivity measures (d’), usually interpreted as improved sustained attention, in mice. SI also caused increases in reward-seeking behavior as measured by FR and PR tasks, indicating the effects in the rCPT may not have been specific to the domain of sustained attention. These results together form a consistent, robust profile of the effects of adolescent SI on reward-seeking behavior.

5.1. Adolescent SI improves performance in the rCPT

This study investigated the effects of adolescent SI on sustained attention via a CPT, which has established usage in rodent studies and human neuropsychological testing [28]. SI mice performed significantly better on Stage 4 of the rCPT than group-housed mice.

To further characterize how mice engaged with the task, we analyzed the CPT session data using 15-minute time bins. Male mice scored significantly higher than females during the first bin. In the second bin both males and SI mice scored higher than females and group-housed mice respectively and in the third bin SI mice performed better than group-housed mice. These findings support the hypothesis that SI mice are able to maintain higher performance over time. It is not clear what drives the sex difference during the first two time bins, but it may be related to the higher false alarm rate in females at baseline. For both sexes, mistakes decrease across the session.

Higher scores in the rCPT are usually associated with improvements in sustained attention [23], but many studies of SI indicate impaired attention as a key feature of the phenotype [29]. In response to this finding, we considered performance in the CPT as an interaction between the mouse’s willingness to engage in the task and their interest in the food reward. If SI impaired attention but also increased reward-seeking behavior, this may present as an increase in d’.

5.2. Adolescent SI increases reward-seeking behavior

To further investigate this interaction between attention and motivation, we utilized a fixed-ratio/progressive-ratio regimen. These tasks measure effort committed to obtain a reward without the cognitively demanding rules of the rCPT. SI mice demonstrated increased effort in the PR task, represented as higher breakpoints, and higher numbers of rewards earned in the FR task. These data support our hypothesis that changes in CPT performance caused by SI are due at least in part to significant differences in reward-seeking and motivation in SI mice compared to their group-housed littermates. This finding is in agreement with the known role of isolation as a risk factor for substance use disorders in humans. [19]

Although clinical literature links adolescent and childhood stress to increased drug use, the preclinical literature displays mixed results [30]. With respect to high sugar/high fat food reinforcers and natural rewards, one study found that early social isolation (ESI) stress did not alter the acquisition of sucrose self-administration on a fixed-ratio schedule, nor change the motivation for sucrose on a progressive-ratio schedule [31]. Another study similarly found that adolescent-onset social isolation did not impact acquisition of operant responding for sucrose pellets on a fixed-ratio schedule, nor motivation for sucrose pellets on a progressive-ratio schedule [32]. While contradictory to our findings, key methodological differences exist between all three studies. Our study isolates mice solely during early adolescence from PN21–35, while Singh et al. and Fosnocht et al. isolate mice for virtually all of adolescence from PN-21-PN56 and PN21-PN60–80, respectively. Furthermore, animals in Fosnocht et al. were individually housed at the start of operant training (PN60–80), meaning animals that were already isolated from PN21-PN60 could continue to be isolated into adulthood, depending on when testing commenced. The effect of social isolation on a given physiological and behavioral outcome is highly dependent on the age of onset and length of exposure [33,34]. As such, with varying isolation periods, discrepancies in results can be expected.

In rodents, reward processing is inferred by the measurement of certain behaviors, most prominently responsiveness to reward, approach motivation, and learning [35]. Responsiveness to reward is often measured with the sucrose preference test, using reward consumption as a proxy for liking. Various studies of sucrose preference in rodents subjected to post-weaning social isolation demonstrate increased hedonic consumption [36–40]. The effects of post-weaning social isolation on approach motivation, measured by conditioned place preference tests or progressive ratio assays as conducted in this study is often more nuanced, however, evidence indicates social isolation during an almost identical timeframe to our study (PN21–36) enhances responding for increasing concentrations of sucrose [41]. With respect to learning, post-weaning social isolation results in cognitive inflexibility, as measured with reversal learning and contingency degradation tasks [35]. Multiple studies utilizing contingency degradation paradigms indicate that post-weaning social isolation shifts decision-making strategy from goal-directed action to habit-like behaviors at the expense of flexible goal seeking, implicating changes in the prefrontal cortex and striatum [42–44]. As such, the results we see in socially-isolated mice can be attributed to increased reward sensitivity (from the increased rewards earned in the FR task) and increased approach motivation (from the increased breakpoint in the PR task). The PR task, with reward delivery being directly determined by responding, has a high correlation between response and reward rates and thus maintain goal-directed action [45]. Random interval, contingency degradation, or reversal learning tasks promote habit formation due to a low correlation between response and reward rates [45] and can be used to examine more complex decision making. Our PR results cannot inform how social isolation during adolescence mediates learning to parse between goal-directed and habit-forming behavior, but another study found that results in a contingency degradation and PR task were positively correlated, such that greater contingency insensitivity was associated with higher breakpoints in the PR task, though results depended on reinforcement schedule history, ethanol exposure history, and sex [45]. Further experiments are needed to determine correlation between high breakpoints and contingency insensitivity in socially-isolated animals. All together, these studies on social isolation and reward processing give valuable insight into hedonic experience and illustrate how motivation is a multidimensional behavior.

5.3. Behavior effects in the greater context of adolescent SI effects

One of the largest remaining questions about the SI phenotype is how an organism translates the experience of SI stress to the observed behavioral effects on reward processing. Social isolation during adolescence evokes widespread structural and functional changes in the brain, notably in the reward circuitry within the cortex and striatum [46]. With respect to serotonin systems, 5-HT neurons concentrated in the dorsal raphe project to regions essential for reward processing, and interplay with dopamine signaling [47]. Studies investigating the effects of social isolation during adolescence on 5-HT signaling suggest that both basal- and stimulus-induced 5-HT dynamics are altered in a region-specific manner [48–54]. With respect to dopamine systems, the literature supports increased dopamine activity in the nucleus accumbens following social isolation during adolescence, and decreased dopamine activity in the prefrontal cortex [35,46,47]. This combination of high dopamine activity in the nucleus accumbens and low dopamine activity in the cortex potentially contributes to the reward processing findings in socially isolated animals described above, wherein socially isolated animals show increased approach motivation behavior but deficiency in aspects of reward learning that require greater executive control [35].

Beyond changes in reward circuitry, numerous studies have targeted genes involved in the exhibited altered reward processing by rodents isolated during adolescence. In the prefrontal cortex, differentially expressed genes resulting from adolescent social isolation have been identified, including those involved in transcriptional regulation, stress signaling, sleep disorders [55], metabolism, mitochondrial function, oxidative stress, neurotransmitter release, transcription factors, cortical neurogenesis, neuronal migration, and synapse formation [56]. Further studies in the prefrontal cortex have identified differentially expressed genes involved in addiction and reward processing, including dopaminergic, serotonergic, and peptidergic function [57].

Another molecular target that is well established within addiction and reward processing is ΔFosB [58]. A previous study from our group [15] focused on the molecular effects of SI during adolescence and found that the transcription factor ΔFosB was upregulated. Those findings about ΔFosB combined with the behavioral effects in this study are new pieces of the potential biological pathway between SI stress and the observed changes. Future experiments will directly test the relationship between SI, ΔFosB, and reward-seeking behavior. Interestingly, increased reward-seeking behavior has also been observed following overexpression of ΔFosB, providing evidence for a causal link. [17,59]

Our findings that transient isolation during adolescence may provide critical information as we continue to face fallout from COVID-19-related SI. Although almost all people experienced increased isolation due to the pandemic, we may see significant variability in downstream effects. Our data suggest that the long-term effects may not be immediately observable or evenly distributed among the population, but long-lasting and significant nevertheless.

6. Conclusions

This study tested the effects of postweaning SI on CPT and FR/PR performance in mice. SI exerts significant effects on performance in the tasks utilized in this study SI mice demonstrating increased reward-seeking and motivation for the palatable food reinforce across all tasks. These results are in line with previous studies on the effects of SI across species. Future studies will address the potential molecular and circuit intermediates that connect SI to behavioral changes.

Figure 5. SI increases reward-seeking behavior as measured by progressive and fixed ratio behavioral tasks.

a) Average breakpoint over seven progressive ratio (PR) 4 sessions is higher in SI mice compared to group-housed littermates. b) Reward intake during an unlimited fixed ratio (FR) 1 session is higher in SI mice compared to group-housed littermates. n = 12 group-housed males, 14 SI males, 9 group-housed females and 8 SI females. Data are mean ± SEM.

Highlights.

• Social isolation improves performance in the rodent continuous performance test

• Social isolation increases reward-seeking behavior

• The effects of isolation are similar for male and female mice