Adolescence, the period of ∼12–20 years old in humans and from weaning [postnatal day (PND) 21] to ∼8 weeks in rodents, is characterized by high levels of playful social interactions, cognitive development, and increased risk-taking behaviors (Spear, 2000). During this period, brain volume and white matter increase linearly with age (Giedd et al., 1999; Sowell et al., 2002; Giedd, 2004) until age 20 (Pfefferbaum et al., 1994). In the prefrontal cortex (PFC), neurons also undergo dramatic structural reorganization and synaptic remodeling (Giedd, 2004). In the prefrontal and orbitofrontal cortexes, up to 50% of spines are pruned from early to mid-adolescence (Bourgeois et al., 1994; Huttenlocher and Dabholkar, 1997; Shapiro et al., 2017). Simultaneously, synaptic connections are being strengthened and refined in an activity-dependent manner (Shapiro et al., 2017; Hinton et al., 2019). In addition to changes in white matter and synaptic refinement, a shift in connectivity among mesocortical and mesolimbic regions occurs (Spear, 2015) with increased connectivity between limbic regions and cortical regions taking place (Yetnikoff et al., 2014). This increasing top–down control from cortical regions during the transition from adolescence to adulthood is associated with greater planning, behavioral flexibility, and executive control (Munakata et al., 2012). Cognitive changes in the developing adolescent include heightened reward sensitivity, sensation seeking, impulsivity, and diminished inhibitory self-control, all of which contribute to increased participation in risky behaviors, including the initiation and escalation of alcohol use (Spear, 2000; Lees et al., 2020). Binge drinking can be particularly harmful in adolescence as ethanol may delay or disrupt critical ongoing neurodevelopment, with profound consequences in adult behaviors, brain structure, connection, and function (De Bellis et al., 2005; Medina et al., 2008; Bava et al., 2013; Jacobus et al., 2013).

Beginning in adolescence, males and females switch their primary focus from family centered relationships to peer relationships (Nelson et al., 2005). In humans and rodents, adolescents share a marked increase in social interactions that serve as a learning experience and shape adult behaviors (van Kerkhof et al., 2013). Peer to peer interactions peak during adolescence and are thought to be crucial for social and cognitive development (Vanderschuren et al., 1997; Spinka et al., 2001). Adolescent rats spend more time engaged in social interactions and social play, and find social experiences more rewarding than adult rats (Trezza et al., 2010; Spear, 2011). Compared to adults, social interactions have a greater impact on decision making and behavior of adolescents (Gardner and Steinberg, 2005). Beneficial and adverse social experiences during adolescence influence anxiety-like (Weiss et al., 2004; Voikar et al., 2005; Koike et al., 2009; McCool and Chappell, 2009; Makinodan et al., 2012; Chappell et al., 2013; Yorgason et al., 2013; Skelly et al., 2015; Lander et al., 2017) and depressive-like behavior (Koike et al., 2009; Amiri et al., 2015), as well as cognition in adulthood (Koike et al., 2009; Zhang et al., 2016; Cao et al., 2017; Lander et al., 2017; Pais et al., 2019). Social isolation can be particularly disruptive at this time as it alters proper neural development and function (Arain et al., 2013; Butler et al., 2016) and increases risk for alcohol use (Lopez et al., 2011; Sanna et al., 2011). Thus, the age at which an individual is deprived of social interactions is critical for ongoing brain and behavioral development.

During adolescence, mesocortical connectivity increases and begins to exert top-down control of limbic regions (Spear, 2015). Adolescents have fewer mesocortical dopamine (DA) fibers, but similar levels in mesolimbic areas than adults (Naneix et al., 2012), indicating a developmental difference in the maturation process of these regions. As mesocortical signaling strengthens, higher cortical control of behavior occurs as the brain transitions from adolescence into adulthood.

Broadly, organizational hormonal actions and hormone receptor expression seems to drive region-specific sex differences in the hippocampus, amygdala, and prefrontal cortex, with females maturing slightly faster than males (Premachandran et al., 2020). Male rodents have larger corticolimbic hippocampal, amygdala, and PFC volumes than females, driven by the size of specific subregions within these structures. Several studies have also shown that the GABA switch from excitatory to inhibitory neurotransmission occurs a few days earlier in females, giving males a longer window of excitatory GABAergic activity in the hippocampus (Premachandran et al., 2020). Though sex differences are less investigated in the PFC, limited evidence suggests that gonadal hormones induce changes in GABAergic maturation, a major factor in the development of the excitatory/inhibitory balance of the PFC (Drzewiecki et al., 2020; Premachandran et al., 2020). Little is known about when sex differences occur in amygdala GABA signaling, only that its transmission is mediated by estrogen (Premachandran et al., 2020). Sex differences in dendritic spine development vary between subregions in the corticolimbic system, particularly the PFC, likely due to external excitatory inputs and GABA maturation (Premachandran et al., 2020). There are limited studies regarding transient sex differences in dopamine circuitry development, but they suggest DA fibers innervating the prelimbic and infralimbic layers of the PFC differ between males and females at PND 7 (Lolier and Wagner, 2021), but not after PND 25 (Willing et al., 2017; Premachandran et al., 2020). Gaps in the developmental time course and dearth of female subjects makes pinpointing the duration of sex-specific critical periods difficult during the maturation of the corticolimbic system (see Premachandran et al., 2020 for more detail) and highlight the need for further studies investigating such differences.

Disruptions to these systems by social isolation and/or binge ethanol can stunt cortical development and maintain use of the more juvenile limbic system (Crews et al., 2016). Juvenile social isolation or social stress leads to structural and molecular changes throughout the brain including reduced neuronal excitability in the dentate gyrus (Sanna et al., 2011; Talani et al., 2016), enhanced DA and serotonin (5-HT) function in the basal ganglia (Yorgason et al., 2013, 2016), alterations in glutamatergic and GABAergic transmission (Pisu et al., 2011b; Talani et al., 2016; Lander et al., 2017; Wang et al., 2019), reduced PFC myelin-related gene (Makinodan et al., 2012), and protein expression (Hinton et al., 2019) and impaired plasticity (Quan et al., 2010; Medendorp et al., 2018; Wang et al., 2019). Isolation stress also leads to altered mesolimbic DA signaling including enhanced presynaptic DA and 5-HT function in the nucleus accumbens (Yorgason et al., 2013, 2016; Karkhanis et al., 2014), deficits in DA and 5-HT function in the PFC (Bibancos et al., 2007), and widespread changes across the brain with an altered structural connectome (Liu et al., 2016). Similarly, adolescent binge ethanol also induces structural changes, including reduction of myelin content in the frontal cortex (Vargas et al., 2014), myelin gene expression (Montesinos et al., 2014; Wolstenholme et al., 2017), and disrupts DA signaling (Coleman, He et al., 2011; Karkhanis et al., 2014; Trantham-Davidson et al., 2017). These effects appear specific to adolescents – when binge ethanol was administered to adult DBA/2J mice, myelin-related gene expression remained unaltered, indicating that adolescents are particularly susceptible to binge ethanol (Bent et al., 2022). A lack of social experience or exposure to ethanol during this critical brain developmental period likely alters the necessary stimulation from cortical brain regions, and, by impacting brain structure, molecular function, and neural connections, gives rise to impaired development and behavior that can persist into adulthood (Figure 1).

The glucocorticoid (GC) system undergoes development throughout adolescence. As there is a high density of GC receptors in the hippocampus, medial PFC, and amygdala (McCormick and Mathews, 2010), this system is particularly vulnerable to lasting effects of stress, including social stress. In adolescence, circulating corticosterone (CORT), the rodent analog to human cortisol, increases and then levels off in adulthood (Goldsmith et al., 1978; Meaney et al., 1985; Laviola et al., 2002; Silveri and Spear, 2004; Gunnar et al., 2009a; Stroud et al., 2009). Temporarily high CORT concentrations occupy both low-affinity GC receptors and high-affinity mineralocorticoid receptors to coordinate high rates of cortical dendritic spine turnover and pruning during adolescence (Liston and Gan, 2011). While the number of GC receptors in the brain does not differ between adolescents and adults, adolescents do not tend to habituate (measured as a blunted CORT release) to repeated stressors as compared to adults (Vazquez, 1998; Romeo et al., 2006). Single housing in adolescent rodents disrupts hypothalamus-pituitary axis (HPA) function (Butler et al., 2014a; Hinton et al., 2019), changes corticosterone response (Weiss et al., 2004), increases locomotor (Silva-Gomez et al., 2003; Voikar et al., 2005; Bianchi et al., 2006; Bibancos et al., 2007; Varlinskaya and Spear, 2008; McLean et al., 2010; Chappell et al., 2013; Butler et al., 2014a; Skelly et al., 2015; Lander et al., 2017; Medendorp et al., 2018; Wang et al., 2019) and anxiety-like behaviors (Weiss et al., 2004; Voikar et al., 2005; Koike et al., 2009; McCool and Chappell, 2009; Veenema, 2009; Makinodan et al., 2012; Chappell et al., 2013; Yorgason et al., 2013; Skelly et al., 2015; Lander et al., 2017). Social isolation in adolescent mice also leads to glucocorticoid insufficiency, increased density of dendritic spines in the PFC, and increased PSD-95 immunostaining suggesting that social isolation leads to improper synaptic pruning and retainment of overabundant dendritic spines (Hinton et al., 2019). Notably, adolescent ethanol-exposed mice show similar alterations in HPA function and anxiety-like behavior (Varlinskaya et al., 2015). The CORT response in adults with a history of adolescent binge ethanol (PND 25–45, 3.5 g/kg, i.g.), did not habituate to repeated restraint stress as would occur in ethanol naïve adults, suggesting a dysregulation of the HPA axis (Varlinskaya et al., 2016). Thus, it is likely that social isolation or binge ethanol in adolescence delays or prevents the age-appropriate maturation of the HPA axis. The ongoing refinement of circuitry between brain regions (i.e., PFC, nucleus accumbens, hippocampus, and amygdala) and the HPA axis, as well as the ongoing development of related signaling systems such as 5-HT, DA and GC, may help explain the biological reasons that social isolation or exposure to binge ethanol is particularly damaging to the development of an adolescent. Social isolation stress and/or ethanol exposure during this time disrupts the trajectory of cortical development and can lead to behavioral problems later in life, including attention, social, and cognitive deficits. Below, we review the areas of convergence and divergence in social isolation and binge drinking on social, ethanol-related, anxiety-like, and cognitive behaviors in rodent models (Tables 1, 2). We also discuss structural changes in brain regions important for these behaviors. We focus specifically on rodent studies using single housing as the method of social stress as these studies are widely used to model social isolation stress. Additionally, our review is largely confined to studies using primarily adolescent (∼PND 28–50) ethanol exposure.

There is an age-dependent proclivity toward social interaction in both humans and rodents, with adolescents interacting more with a novel conspecific than adults. Adolescent social isolation tends to increase social interactions in adolescence and in adulthood (Table 1). Five days of social isolation in adolescent rats increased pro-social behavior in an age-dependent manner as compared to group housed rats (Varlinskaya and Spear, 2008). Single-housed rats displayed increases in social investigation frequency, contact behavior, and play fighting behavior compared to group housed rats and this effect was greatest in young adolescent (PND 28) rats (Varlinskaya and Spear, 2008). Socially isolated male and female C57BL/6J mice also showed increased social interaction following isolation from PND 28–70 (Rivera-Irizarry et al., 2020). Conversely, group housed C57BL/6 (Liu et al., 2015) mice or PLP-eGFP mice (Makinodan et al., 2012) were more interactive with a novel conspecific than socially isolated mice.

Social recognition, the ability to remember and discriminate a previously encountered conspecific, is an important component of social behavior. The multi-trial social recognition task and the 3-chamber social preference task are frequently used to assess social memory and social approach in rodents. The 3-chamber social preference task consists of two phases to measure social approach (preference for an empty chamber versus a chamber with a novel stimulus mouse) and social memory (preference for a novel versus familiar stimulus). In the 3-chamber social preference task, Thy1-GFP males and females (Medendorp et al., 2018) and socially isolated C56BL/6 males (Zhang et al., 2021) did not differ in their social affiliation as compared to group housed males and displayed a typical preference for a stimulus animal over an empty cylinder. Thy1-GFP mice are on a C57BL/6 genetic background and were used to allow for imaging of pyramidal neurons. Socially isolated CD1 males, however, spent more time with the stimulus mouse during both phases of the test (Liu et al., 2019). During the second phase to assess social memory, adolescent social isolation decreased the interaction time with novel mouse over the familiar, suggesting poor social memory in both male and female Thy1-GFP mice (Medendorp et al., 2018), C57BL/6 males (Zhang et al., 2021), and CD1 males (Liu et al., 2019). Similarly, in a 5 or 9-trial social recognition task, postweaning social isolation in CD1 males and C57BL/6 males and females increased social interactions to the stimulus mouse during the habituation phases and decreased their ability to recognize a novel mouse during the dishabituation phase, suggesting a deficit in social memory (Kercmar et al., 2011; Liu et al., 2019; Zhang et al., 2021). In a modified social interaction task, socially isolated males retained higher frequencies of social interaction and did not habituate to repeated exposures to the same stimulus mouse; mice were not tested for the dishabituation phase (Koike et al., 2009). Together, when tested for social interaction, social play, or in social memory-related tasks, adolescent social isolation in rodents typically increases overall social interactions and this increased interaction may preclude the ability to form a memory of the novel animal.

Most studies investigating ethanol effects on sociability have measured social behavior with acute ethanol administration, while ethanol was still present in their system (Table 2). In response to acute low doses of ethanol, juveniles show social facilitation and are more sensitive to ethanol’s rewarding effects than adults, increasing the incentive value for adolescents (Varlinskaya et al., 2014). In rats, low ethanol doses (0.25–0.75 g/kg) increased social activity and social preference for a novel rat, while high doses (3–4 g/kg) decreased social activity and increased peer avoidance (Varlinskaya and Spear, 2002). Adolescent rats are more sensitive to social facilitation and are less sensitive to the suppression of social interactions compared to adults; early adolescents (PND 28) show even more social facilitation than mid to late adolescents (PND 35 and 42) (Varlinskaya and Spear, 2004). Similarly in mice, acute high doses of ethanol (1.6 g/kg) impaired social approach in adult mice, while low doses (0.25 g/kg) alleviated social anxiety-like avoidance and high doses did not suppress social interactions in adolescents (Raymond et al., 2019). Adult male rats with a history of ethanol (PND 28–42, 3.5 g/kg, i.g., every other day) showed enhanced sensitivity to the social-facilitating effects of ethanol (Varlinskaya et al., 2014). Adult males with a history of ethanol were also more social with a conspecific than ethanol-naïve males, suggesting that the males retained an adolescent-typical responsiveness to alcohol; females were insensitive to the social modulating effects of ethanol (Varlinskaya et al., 2014). Ethanol’s effects on social memory in the social recognition task or 3-chamber social preference task have not been investigated.

Adolescents are less sensitive than adults to the social impairing, motor disrupting, aversive, and sedative effects of higher doses of ethanol (Varlinskaya and Spear, 2002; Varlinskaya et al., 2016), allowing adolescents to consume more ethanol than their adult counterparts (Ho et al., 1989; Tambour et al., 2008; Maldonado-Devincci et al., 2010; Strong et al., 2010; Metten et al., 2011). Prior exposure to ethanol tends to increase drinking later in life and this effect is stronger when binge ethanol exposure occurs during adolescence. Early life binge drinking, but not adult binge drinking, increases adult ethanol intake and preference (Moore et al., 2010; Strong et al., 2010; Metten et al., 2011; Montesinos et al., 2016; Carrara-Nascimento et al., 2017; Lee et al., 2017; Younis et al., 2019; Wolstenholme et al., 2020; Lesscher et al., 2021). As compared to animals that first received ethanol in adulthood, adolescent ethanol exposure tends to increase later voluntary consumption in adulthood (Younis et al., 2019; Lesscher et al., 2021). In a set of studies to determine the optimal age for adolescent ethanol exposure to increase adult intake, Metten et al. (2011) tested six ages of HS/Npt male mice, a genetically heterogeneous mouse strain, and found that ethanol exposure beginning in mid-adolescence (PND 28) is most effective at increasing ethanol intake in adulthood (Metten et al., 2011). Similarly, mice selectively bred for high ethanol consumption, HDID-1 mice, first exposed to ethanol DID at 4 weeks, consumed significantly more ethanol than HDID-1 mice first exposed to ethanol at 6 or 9 weeks of age (Metten et al., 2011). In C57BL/6J males, DID from PND 28–36 increased adult DID as well as adult ethanol intake and preference in the 2-bottle choice model (2BC), whereas DID exposure in adults (from PND 73-81) did not alter ethanol intake or preference 36 days later (Younis et al., 2019). The number of ethanol exposures also plays a role in whether an animal increases ethanol consumption later in life. In C57BL/6J male mice, daily DID sessions from PND 28–36 increased adult intake, while intermittent, every other day DID sessions only moderately increased adult intake (Younis et al., 2019). In adult mice, repeated DID cycles during adulthood increased subsequent voluntary ethanol intake and preference, and this effect became more robust with increasing DID cycles (Cox et al., 2013), further suggesting that a minimum number of ethanol exposures may be needed to predispose animals to consume more ethanol upon re-exposure. However, adolescent exposure to ethanol does not always lead to increased intake in adulthood. Male Thy1-GFP mice given 2 weeks of 2BC ethanol on weeks 4–6 and then 4 cycles of chronic intermittent ethanol vapor (CIE) did not increase their ethanol intake or preference when tested again for 2BC drinking in adulthood (∼10 + weeks) (Jury et al., 2017). Adult mice exposed to 2BC drinking (8–10 weeks) and CIE increased their intake relative to air-exposed controls (Jury et al., 2017). This lack of increased intake in adolescents with a history of CIE may be due to a ceiling effect in mice on a C57BL/6 background, as suggested by the authors (Jury et al., 2017). Together, these studies indicate that age of exposure and perhaps a minimum number of ethanol exposures in adolescence, as well as genetic background, are all critical factors that influence the ability to detect a reliable increase adult drinking (Tables 2, 3).

Using a variety of voluntary drinking models, adolescent social isolation without a history of adolescent drinking leads to increased drinking in adulthood in 2BC (Chappell et al., 2013; Butler et al., 2014a; Lesscher et al., 2015; Skelly et al., 2015), DID (Sanna et al., 2011), and operant responding models (Deehan et al., 2007; McCool and Chappell, 2009). In the few studies that have directly investigated drinking behavior in single versus group housed rodents, social isolation usually increases ethanol consumption behaviors in rats (Schenk et al., 1990; Wolffgramm, 1990; Chappell et al., 2013; Skelly et al., 2015) and mice (Advani et al., 2007; Lopez et al., 2011; Sanna et al., 2011). These effects appear to be limited to adolescence, as increased ethanol preference does not occur in isolated adult rodents (Schenk et al., 1990; Lopez et al., 2011; Chappell et al., 2013), suggesting that early timing of social isolation is crucial to produce significant changes in the sensitivity to ethanol in adulthood (Tables 1–3).

Not all studies have found similar effects of social housing on ethanol drinking. When tested in adolescence, single housed adolescent rats and group housed rats drank the same amounts, and single housed rats showed a decreased preference for ethanol in a 2BC continuous access model (Pisu et al., 2011b). In a 3-day 24 h access model, single and pair-housed rats consumed similar amounts of ethanol (Thorsell et al., 2005). Using a DID model, social isolation did not produce any differences in ethanol intake, preference, quinine-adulterated ethanol intake and preference, or sucrose preference in male or female mice (Rivera-Irizarry et al., 2020). Interestingly, high levels of social activity in adolescent male rats and high levels of social anxiety-like behavior in adolescent female rats were associated with elevated social drinking, suggesting that males ingest ethanol for its socially enhancing properties, while females may ingest ethanol for its social anxiolytic effects (Varlinskaya et al., 2015).

In general, female rodents, consume greater amounts of ethanol than their male counterparts (Vetter-O’Hagen et al., 2009; Lopez et al., 2011; Alfonso-Loeches et al., 2013; Lopez and Laber, 2015). A few studies have directly investigated sex differences in adult drinking following adolescent social isolation (Advani et al., 2007; Lopez et al., 2011; Butler et al., 2014a,b; Lopez and Laber, 2015; Rivera-Irizarry et al., 2020). Male and female mice with a history of social isolation in adolescence increased ethanol intake as compared to group housed mice, or mice isolated in adulthood, with effects stronger in males (Lopez et al., 2011; Lopez and Laber, 2015). When social isolation occurred in adult females, they consumed moderately less ethanol than group housed females (Lopez et al., 2011). Social isolation in adolescent male (Butler et al., 2014a) and female (Butler et al., 2014b) rats increased ethanol intake relative to group housed rats and effects were again stronger in males. In rats, single housed adolescent males and adult male and female rats consumed more sweetened alcohol under social circumstances than when individually housed, while single housed adolescent females consumed more ethanol than group housed females (Varlinskaya et al., 2015). Male rats exposed to ethanol during early adolescence retained an adolescent-like ethanol-induced social facilitation not normally seen in adults (Varlinskaya et al., 2014). Single housed male and female mice had increased ethanol preference during their first week of drinking at PND 60 which was maintained in adult males (Advani et al., 2007). But in adulthood, single housed females decreased their intake relative to pair housed mice (Advani et al., 2007). Socially isolated male rats showed robust increases in anxiety and ethanol intake compared to group housed males (Chappell et al., 2013; Skelly et al., 2015).

The effects of adolescent social isolation on anxiety-like behaviors are mixed and may be confounded by hyperlocomotion (Table 1). Social isolation in adolescent rats increases anxiety-like behavior in the elevated plus maze (McCool and Chappell, 2009; Pisu et al., 2011b; Chappell et al., 2013; Yorgason et al., 2013; Skelly et al., 2015). In the open field task, overall locomotor activity was increased following social isolation in rats and mice (Bibancos et al., 2007; Varlinskaya and Spear, 2008; McLean et al., 2010; Skelly et al., 2015; Liu et al., 2016; Medendorp et al., 2018; Wang et al., 2019). Anxiety-dependent measures in the open field, such as time in the center, did not differ between housing groups (Medendorp et al., 2018; Wang et al., 2019). In male mice single housed in late adolescence (PND 38), time in center of the open field was lower, without a change in total locomotor activity (Lander et al., 2017). However, when tested directly for locomotor activity in a longer 30-min task, single housed mice were hyperactive (Lander et al., 2017). Time in the open arms was also lower in single housed ICR males, suggesting social isolation increased anxiety-like behavior (Koike et al., 2009). Elevated locomotor activity is not always found, as no differences in locomotor activity in a novel arena have also been reported between single and group housed animals (Koike et al., 2009; Liu et al., 2015; Hinton et al., 2019). Similarly, in the elevated plus maze task, no differences were found between adolescent single and group housed animals for time spent in the open arms (Hinton et al., 2019). Adolescent social isolation increased the latency to enter the light side of the light-dark box compared to group housed mice, but the time in the light did not differ (Pais et al., 2019).

Following bouts of binge ethanol, animals may display withdrawal symptoms and this negative affective state has been assessed using anxiety-like behavior tasks. Adult animals typically display increased anxiety-like behavior in the elevated plus maze or the light-dark box while in ethanol withdrawal (Lee et al., 2018; Szumlinski et al., 2019), potentially in an intake-dependent manner (Lee et al., 2015). Twenty-four hours after the last adolescent ethanol binge, rats only sometimes display withdrawal-induced anxiety phenotypes (Pandey et al., 2015; Kyzar et al., 2016), some of which persist long-term (Vetreno et al., 2016). However, similar to adolescent social isolation, binge ethanol in adolescence does not always affect anxiety-like behavior (Table 2). In adolescent mice and rats, withdrawal-induced increases in anxiety-like behavior were not observed in the light-dark box or the elevated plus maze (Scheidt et al., 2015; Lee et al., 2016, 2018; Marco et al., 2017; Wolstenholme et al., 2017; Salling et al., 2018). It is possible that these inconsistencies may be due to procedural differences, differences in route of administration, or that binge ethanol in adolescence only produces lasting withdrawal-induced anxiety in rats and not mice. One possibility, as some have suggested, is that adolescent mice are resilient to early ethanol withdrawal (Lee et al., 2016, 2018). A recent review of the literature in adult models of negative affective behavior during ethanol abstinence has highlighted similar highly variable findings and has promoted more data driven approaches are needed toward understanding this complex interaction (Bloch et al., 2022).

Overall, the results of social isolation or binge ethanol in adolescence have given inconsistent results on anxiety-like behavior. As discussed above, locomotor differences can confound the interpretation of anxiety-phenotypes as these tasks typically rely on locomotor activity. While testing for anxiety 24 h after the last binge, ethanol exposed DBA2/J mice traveled farther in the light–dark apparatus than controls (Wolstenholme et al., 2017). In adulthood, a priming injection of 2 g/kg ethanol increased locomotor activity in females with a history of binge ethanol while testing for ethanol-induced anxiolysis (Wolstenholme et al., 2017). Adolescent rodents are also more sensitive to low dose ethanol’s locomotor stimulating effects than adults (Lopez et al., 2003; Hefner and Holmes, 2007; Stevenson et al., 2008; Melon and Boehm, 2011). Swiss mice exposed to ethanol as adolescents, especially at high doses, have enhanced locomotor stimulant effects as compared to controls and these effects persisted for 3 weeks after the last ethanol exposure (Quoilin et al., 2012, 2014). It is possible that the general increased locomotor activity of socially isolated or ethanol exposed adolescents may overshadow the display of anxiety-like phenotypes in tasks that require locomotor activity. This heightened activity has the potential to confound the experimenter’s ability to interpret anxiety-like behaviors following social stress or exposure to binge ethanol. Alternatively, as adolescence may be a period of heightened vulnerability to the deleterious behavioral and neurobiological effects of ethanol or social isolation, these insults may increase stress reactivity (i.e., locomotor activity in a novel environment). Clearly, new approaches or methods for analyzing these complex data are needed to fully understand the interaction between social stress, ethanol and negative affect.

There is a considerable amount of convergence in the effects of adolescent social isolation and binge ethanol on cognitive behaviors (Tables 1–3). Both adolescent social isolation and binge ethanol cause long-term deficits in recognition memory and cognitive flexibility. Relying on a rodents’ innate preference for novelty, the novel object recognition (NOR) test can be reliably used to assess both working and long-term memory. In this single trial task, an animal is allowed to explore a novel and a familiar object in a test arena. An intact short term or working memory is interpreted as higher preference for the novel object (see Antunes and Biala, 2012 for review). Adolescent social isolation decreased preference for the novel object in both rats and mice (Voikar et al., 2005; Bianchi et al., 2006; Koike et al., 2009; McLean et al., 2010; Pais et al., 2019). Following six weeks of social isolation, adult rats were able to discriminate between the original and novel objects in a 1 minute and 1 hour inter-trial interval (ITI), but not a 3.5 or 4 h ITI (McLean et al., 2010). Binge ethanol in adolescence also caused deficits in working or long-term memory when tested in adulthood (Pascual et al., 2007; Montesinos et al., 2014; Vetreno and Crews, 2015; Vetreno et al., 2016; Marco et al., 2017; Wolstenholme et al., 2017). Interestingly, binge ethanol in adolescence led to deficits in the NOR task primarily when mice were tested in adulthood and NOR deficits are not always observed in adolescence shortly after ethanol administration (Bent et al., 2022), suggesting a maturation of the brain circuitry in adulthood is necessary to observe these deficits. However, others have reported reduced performance in the NOR test in both late adolescent and adult rats after binge ethanol in adolescence (Pascual et al., 2007) and in late adolescent mice (Pascual et al., 2021). Of note, adult males and females that received the same binge paradigm as adolescent mice (4 g/kg, i.g.) did not display memory deficits in the NOR task (Bent et al., 2022). The combined effects of social isolation and binge ethanol have not been directly explored in the NOR task.

The Morris Water Maze (MWM) and Barnes Maze (BM) are two commonly used, multi-day tasks used to assess spatial memory and cognitive flexibility and acquisition of this task is largely hippocampal-dependent (Morris et al., 1982). The effects of social isolation beginning in adolescence are inconsistent for the acquisition of spatial learning. Long-term social isolation (6 months+) in Balb/c mice led to deficits in the acquisition and retention of spatial memory in the MWM, and this was accompanied by higher swim speeds (Wang et al., 2019), reminiscent of isolation-induced hyperactivity in the open field task. Five weeks of social isolation in male mice (Bagheri et al., 2021) also led to higher escape latencies and less time spent in the target quadrant in the MWM, when compared to socially housed mice. In the BM, 5 months of social isolation beginning at weaning impaired acquisition of both standard and reversal learning in male rats (Heimer-McGinn et al., 2020). In some studies, social isolation rearing did not impair (Quan et al., 2010; Han et al., 2011; Li et al., 2019) or improved spatial learning (Pisu et al., 2011a). In one study, adolescent-isolated adult rats had lower latencies to reach the MWM platform, and this was accompanied by a higher swim speed in the isolated animals and increased CORT levels on training day 5 (Pisu et al., 2011a), suggesting heightened anxiety-like or hyperactive behavior in socially isolated rats. During the reversal phase of the MWM to assess PFC-mediated cognitive flexibility, 2 weeks of social isolation in early adolescence led to reversal learning deficits in male rats, without housing effects on the spatial learning of the task (Li et al., 2019). Eight weeks of chronic isolation rearing in male rats did not affect the original acquisition learning, but impaired reversal learning in the MWM (Quan et al., 2010). Brief isolation in early adolescent male rats (isolated from PND 21–34 and re-socialized from PND 35–55) affected reversal learning without interfering with spatial learning in the MWM (Han et al., 2011). Attentional set-shifting was also impaired in isolation reared rats (McLean et al., 2010), and mice (Lander et al., 2017) further supporting PFC-related cognitive deficits from social isolation.

In most studies, animals with a history of binge ethanol are able to acquire spatial learning, suggesting that hippocampal-mediated spatial memory remains intact. Daily ethanol injections (2 g/kg, i.p.) immediately prior to testing in adolescent (PND 30) or adult (PND 60) male (Sircar and Sircar, 2005) and female (Sircar et al., 2009) rats impaired the acquisition of spatial reference memory in the MWM and the effects were stronger in adolescents. When tested up to 30 days after the last ethanol administration, ethanol-treated adolescents, but not adults, traveled longer distances to find the hidden platform than saline-treated controls (Sircar and Sircar, 2005), indicating that adolescent, but not adult, animals retained persistent deficits in spatial memory. However, in other studies, a history of ethanol exposure during adolescence did not lead to deficits in the spatial learning of the MWM (Coleman, He et al., 2011; Li et al., 2019) or the BM (Vetreno and Crews, 2012; Coleman, Liu et al., 2014). During the reversal phase, to test cognitive flexibility, adult male C57BL/6J mice with a history of adolescent ethanol (PND 28–37, 5 g/kg, i.g.) showed deficits in acquisition of a reversal Barnes maze (Coleman, Liu et al., 2014) and deficits in the reversal probe trial accompanied by a perseveration-like search strategy and more time spent in the standard MWM quadrant (Coleman, He et al., 2011). Rats exposed to CIE during adolescence (PND 28–42) were unable to alter their responses on a set-shifting task, indicative of impaired cognitive flexibility (Gass et al., 2014).

Similar to MWM and BM, acquisition of the fear conditioning task is not usually affected by adolescent social isolation, binge ethanol, or age. Contextual fear conditioning is primarily hippocampal dependent, whereas the mPFC (particularly the infralimbic cortex) is more associated with fear memory (Myers et al., 2006). Socially isolated rats continued to exhibit a fear-potentiated startle response and deficits in fear extinction learning, while the acquisition of contextual fear conditioning was unaffected by housing condition (Skelly et al., 2015). Isolation during early-mid adolescence produces deficits in extinction of contextual fear conditioning as far as 4 weeks after acquisition (Liu et al., 2015). Single housed mice eventually acquired extinction learning, but at a markedly slower rate than group housed conspecifics (Naert et al., 2011). Some studies report more hippocampal-mediated deficits in contextual fear conditioning 24 h after the initial training session following social isolation (Weiss et al., 2004; Voikar et al., 2005; Bagheri et al., 2021). Social isolation impaired both contextual and conditional fear memory in a fear conditioning test, indicative of impaired acquisition of the task (Ouchi et al., 2013; Liu et al., 2016). These changes could be mediated by altered structural connectivity among corticolimbic regions (Liu et al., 2016) or through the cholinergic system (Ouchi et al., 2013). Early adolescent exposure (PND 28–48) to ethanol impaired retention of the context in which animals had received a foot shock; late adolescent (PND 35–55) or adult (PND 70–90) exposure to ethanol resulted in impaired extinction of fear conditioning (Broadwater and Spear, 2013). Together, these data suggest that adolescent social isolation or binge ethanol leads to deficits in behavioral flexibility and increased perseveration of previously learned behaviors.