Binge drinking (BD) definition has yet to find an international consensus but it is characterized by an intense and episodic alcohol consumption, with recurring alternations between intense intoxication episodes and abstinence periods (Lannoy et al., 2019). BD is particularly worrying because of its health consequences (Kuntsche et al., 2017; Rolland and Naassila, 2017; Tavolacci et al., 2019) and its contribution to alcohol dependence vulnerability (Tavolacci et al., 2019).

Many factors have been associated with BD such as decision-making (DM) and impulsivity, two hallmarks of alcohol dependence. First, DM refers to a fundamental cognitive process allowing us to select a particular option among several alternatives, in order to select advantageous choices over disadvantageous one in everyday life [for review, (Ernst and Paulus, 2005)]. A recent meta-analysis showed a strong association between poor DM capacities using the Iowa Gambling Task (IGT), a tool widely accepted as a direct assessment method for measuring affective DM, and Alcohol Use Disorders (AUD) (Kovács et al., 2017). Furthermore, it has been shown that affective DM moderates the effects of associations on alcohol (Cappelli et al., 2017). Much less is known regarding BD, and the current results from the scientific literature are scarce. Previous studies have associated high and stable BD in college students with less advantageous choices in the IGT, hypersensitivity to reward and impaired reversal learning, without correlation to impulsivity or working memory capacities and academic school performances (Goudriaan et al., 2007; Johnson et al., 2008; Yoo and Kim, 2016). Impaired DM performances and response inhibition have also been reported as a predictor of future problematic use of alcohol (Goudriaan et al., 2011). Other studies have however found no relationship between BD and DM, suggesting that impairments of this cognitive function are associated with more severe forms of alcohol consumption (Bø et al., 2016; Carbia et al., 2017). Secondly, higher impulsivity scores have been associated with the maintenance and intensity of BD (Leeman et al., 2015; Adan et al., 2017), and correlated with drinking episode frequency and the number of drinks per episode (Doumas et al., 2017) (Ashenhurst et al., 2016). Overall, most previously cited studies have shown that BD is associated with impaired impulse control and DM abilities, but the link between them is poorly understood. Weak DM capacities could play a role in the vulnerability to BD, which in turn would lead to DM impairments and thus perpetuate BD. Most studies are cross-sectional and there is a clear need for further longitudinal studies. In a cross-sectional study we previously showed using a Rat Gambling Task (RGT), a protocol inspired from the IGT and described as an animal model of affective decision-making (de Visser et al., 2011), that a history of voluntary BD is associated with impairments of DM abilities in male rats (Jeanblanc et al., 2019). Longitudinal preclinical studies are lacking in the BD research field with the possibility to better control for various factors than in clinical studies (Jeanblanc et al., 2018).

Both DM and BD behaviors are influenced by sex. For instance, findings supported the subdivision of binge drinkers according to gender and personality dimensions (Gierski et al., 2017). Sex differences in IGT performances are however poorly understood, some human studies highlighting males outperforming females in gain (Evans and Hampson, 2015), while others found no difference (Hooper et al., 2004). It has been non-etheless suggested that women may use a different choice strategy, and are more sensitive than men to punishment frequency and occasional losses in the IGT long-term advantageous decks (van den Bos et al., 2013). Sex differences in DM may be attributable in part to interactions between gonadal hormones and dopamine signaling (Becker and Hu, 2008). Previous work have also shown that males and females differ in their responses to dopamine manipulations that could involve basal differences in extracellular levels of dopamine, dopamine receptor levels, and/or dopamine D2/3 autoreceptors control, all of which are modulated by estradiol (Becker and Hu, 2008).

Alteration of dopamine signaling in the ventral striatum is involved in the effects of BD and DM processes. Acute alcohol intake increases tonic dopamine concentrations in the NAc (Imperato and Di Chiara, 1986; Di Chiara and Imperato, 1988) and is involved in alcohol rewarding effects (for review, (Spanagel, 2009)). Alcohol displays biphasic effects on evoked phasic dopamine release in the NAc in vivo (Budygin et al., 2001; Robinson et al., 2005; Jones et al., 2006; Pelkonen et al., 2010) and ex-vivo (Budygin et al., 2001; Mathews et al., 2006) as well as an increase in the frequency of phasic dopamine transients (Robinson et al., 2009). Effects of chronic alcohol exposure and especially BD are much less investigated. Only a handful of studies assessed the consequences of a binge-like exposure on the dopamine mesolimbic system, but reveals possible impairments of the mesolimbic dopaminergic system, both at baseline and in response to alcohol, depending on age, dose and the withdrawal time (Pascual et al., 2009; Philpot et al., 2009; Shnitko et al., 2016; Zandy et al., 2016). Dopamine D2/3 receptors play a crucial role in both DM and alcohol reinforcing effects. Increasing dopamine release by either ventral tegmental area stimulation, or blockade of the D2/3 autoreceptors is associated with increased risky choices (Stopper and Floresco, 2015). In the same vein, the administration of a D2 receptor antagonist in pathological gamblers increased the rewarding effect of gambling and the desire to gamble (Zack and Poulos, 2007). Regarding addiction, numerous studies have shown that a low D2/3 receptor binding is associated with AUD in both animals and humans (Trifilieff and Martinez, 2014). Overall, phasic dopamine signaling in the NAc is involved in many central processes of reward and DM, and its alteration by alcohol could possibly constitute one of the main factors underlying its effects on both BD and DM capacities.

We posited that BD would decrease DM capacities associated with alterations in ventral striatum dopamine signaling and we therefore used the RGT with a longitudinal approach to investigate changes in DM capacities after a history of BD. We also explored whether BD could impact memory and anxiety-like behavior, since both could potentially interfere with DM capacities. We hypothesized that DM deficits induced by BD may be associated with changes in dopamine release potentially due to alteration in D2/3 receptors. In brain slices from the same animals, we tested the sensitivity of dopamine transmission to quinpirole (D2/3 agonist), due to the involvement of D2/3 receptors in DM. We decided to focus on the core part of the NAc due to its involvement in both DM (Sugam et al., 2012) and the approach and treatment of motivational stimuli (Dreyer et al., 2016). The NAc core plays a more significant role than the shell in behaviors regulated by cues, such as lever pressing for a reward or a cue-induced reinstatement of reward seeking (McFarland et al., 2003; Di Ciano et al., 2008). Different procedures to study BD behavior in animals are used mainly consisting in repeated passive exposures to alcohol and during adolescence (Pascual et al., 2009). More relevant models exist such as the one we recently developed based on the “happy hour” session in which the animal can freely self-administer a 20% ethanol solution during daily sessions of only 15 min (Jeanblanc et al., 2018). We used males and females with an operant self-administration procedure to evaluate inter-individual vulnerability (Jacobs et al., 2003; Jeanblanc et al., 2019).

A time line of the study is presented on Figure 1.

All the statistical analyses are provided in Supplementary Table S2.

The present study outlines that a chronic voluntary BD exposure impairs DM capacities specifically in male rats. Those DM impairments were not associated with an increase in cognitive deficits or in anxiety-like behavior. They were also not associated with impairments in dopaminergic transmission in the core of the NAc. In females, despite no effects on choice behavior, BD history altered impulsive control and increased sensitivity to the inhibitory effect of quinpirole on dopamine release thus suggesting adaptation in D2/3 receptor functioning.

Sex differences in DM capacities were seen before the onset of BD. Although we found no significant statistical differences, it seemed like most females were faster in acquiring the task which may be explained by better attention skills, despite making more omissions. As expected, sex differences were obvious regarding preferences for options: i) no poor decision makers in females; ii) males significantly preferred the optimal P2 option and iii) females preferred both P1 and P2 options. The P1 option, preferred by females, can be seen as a “risk-averse” phenotype, as it is associated with the lowest probability of punishment, as previously suggested (Georgiou et al., 2018). This observation is in line with human studies in which women use a different choice strategy than men, displaying more sensitivity to punishment frequency and occasional losses in the IGT (van den Bos et al., 2013).

Voluntary BD impaired choice behavior specifically in males, with a decrease of the optimal P2 option choice (50% decrease), although without effects on the overall advantageous choices made. BD history in males made their choice behavior become “random” (around 25% choice for each option), as they could not effectively select the advantageous options in the task anymore. Our results are in line with those showing impaired DM in conflictual and risky situations after both forced (Nasrallah et al., 2011; Boutros et al., 2015) and voluntary (McMurray et al., 2014; 2016) binge-like procedures in male rodents. Clinical studies have also shown impairment of DM abilities in adolescent bingers (Goudriaan et al., 2007; 2011; Johnson et al., 2008). Male rats did not significantly increase the less advantageous choices (more immediate reward with more punishment), suggesting that they did not display a state or trait of hypersensitivity to reward or increased risk-taking as previously suggested in humans (Johnson et al., 2008). Instead, rats were unable to discern options for their value and adapt from feedback, as described in alcohol addiction (for review, (Verdejo-Garcia et al., 2018)), but not yet in the case of BD. A previous preclinical study using low dose i.p. ethanol during the acquisition of the task found similar results (Spoelder et al., 2015). Our results show that DM impairments were not linked to changes in anxiety-like behavior in the LDB or cognitive deficits in the NOR. BD also notably increased motor impulsivity in both sexes (increased premature responses) in line with heightened motor impulsivity previously described in humans (Sanchez-Roige et al., 2014). It is also important to note that the second RGT measurements were performed during acute ethanol withdrawal, but we did not notice any acute withdrawal symptoms that may have affected behavior in the task.

One of the most striking result of our study is that BD history had no effect on DM capacities in female rats. Many human studies have shown that women are more sensitive to the effects of alcohol than men, but sex differences regarding BD remain poorly investigated (for review, (Wilsnack et al., 2017)). Some studies have found worse performances in global executive functioning and DM in male bingers (Parada et al., 2012), while others have found similar impairments in both sexes (Goudriaan et al., 2007; Carbia et al., 2017). In the present study, it is noteworthy that before BD onset, the female group displayed more good decision makers than the male group, and no poor decision makers. In addition, the males also displayed a higher level of anxiety than females, a trait that has already been associated with poor DM (Miu et al., 2008). Those results could suggest that good baseline DM performances are less sensitive to the effects of alcohol as shown in well-trained animals (Spoelder et al., 2015). In agreement with this observation, we showed in a previous study (Jeanblanc et al., 2019) using a different RGT procedure [as described in (Rivalan et al., 2009)], that alcohol could preferentially impair individuals with neutral or poor performances, while not affecting those with good performances. On the contrary, individuals with poor DM abilities could be more vulnerable. For instance, it has been shown that rats with poor DM abilities in the RGT display extreme scores in risk taking, reward seeking, behavioral inflexibility and motor impulsivity (Rivalan et al., 2013). Working on female rats also raises the question of the role of the hormonal cycle. However, our females were housed individually in ventilated racks and therefore did not synchronize their hormonal cycle. As such, there is little risk of a statistically significant effect of the phase of the cycle on our results. Moreover, we have previously confirmed in our laboratory that the hormonal cycle does not affect ethanol consumption in our BD procedure (data not shown), and previous studies have shown that it has no effect on the IGT in humans (van den Bos et al., 2013), and the RGT (Georgiou et al., 2018) in rats.

No study has yet analyzed whether DM abilities can predict the vulnerability to consume ethanol in individuals who have never used ethanol before. The use of a longitudinal design allowed us to assess the vulnerability to consume ethanol in the BD procedure, but the analysis revealed no correlation between baseline DM abilities (advantageous choices: P1 + P2) and ethanol consumption in the IA2BC and operant self-administration procedures (Figure 3). Interestingly, we found a positive correlation in females (but not males) between premature responses and active lever presses/ethanol intake during the FR3 15 min self-administration sessions. As premature responses in the task reflect motor impulsivity, this could mean that female animals with a high base level of impulsivity could end up drinking more in our task. We did not find any correlation between the other parameters in the RGT, the BD procedure and the FSCV data (Supplementary Table S3). It is however important to note that our sample size was not optimal for such analysis, and would probably require more animals. Thus, our results overall do not support the widespread intuition that poor DM abilities may increase vulnerability to drink ethanol, but rather that chronic ethanol intake is responsible for DM impairments in a sex specific manner. Some results in the literature indicate that cognitive deviations and personality traits (impulsivity, sensation seeking, risk taking…) accompanying addiction, rather than drug consumption in itself, may explain DM impairments (Kovács et al., 2017). The relation between drug consumption and preexisting impulsivity, risk taking or DM impairments has however been shown in a very mild and inconstant manner for alcohol and other drugs, with a remaining uncertain causality (Ahmed, 2018).

Our results on dopamine transmission demonstrate that BD does not affect basal dopamine signaling in the NAc core, regardless of sex, but reveal sex-dependent changes in the sensitivity to the inhibitory effects of quinpirole. Baseline dopamine signaling (release and uptake) was similar in both sexes in control groups. Our results are in line with those of a recent meta-analysis on 39 microdialysis studies showing that there are no sex-dependent differences in basal dopamine levels within the NAc (Egenrieder et al., 2020). It is noteworthy that other studies have however suggested sex-dependent differences in dopamine signaling (Walker et al., 1999; Walker et al., 2006). Thus, our results on basal dopamine signaling cannot explain our behavioral results showing sex-differences in baseline training and testing in the RGT, as we initially hypothesized. Using a bigger sample size, it would however be interesting to compare baseline dopamine signaling between different DM groups. In a previous work, we indeed found a significant difference in DA release between male rats with good and poor DM levels (although using a different RGT protocol and having a possible confounding effect of ethanol) (Jeanblanc et al., 2019).

Baseline dopamine signaling was not altered in both sexes after a history of BD exposure. Our result is in line with those of a recent study that used a forced BD exposure, although in adolescent male rats, and showed no changes in baseline dopamine signaling using in vivo FSCV at adulthood (Shnitko et al., 2016). To the best of our knowledge, no studies have analyzed the effects of a BD protocol on striatal dopaminergic signaling using ex vivo FSCV. It may be that basal phasic dopamine signaling is only impaired using protocols to induce alcohol dependence, but not BD (Karkhanis et al., 2015). While quinpirole reduced dopaminergic signaling in males and the history of BD had no effect, in females, we found that the history of BD may have induced neuro-adaptations making the dopaminergic transmission (release but not the transporter since the V_max is unchanged) sensitive to the inhibitory effect of quinpirole. It seems that autoreceptors are less sensitive to quinpirole in females with no BD history. A previous study using FSCV in female rhesus macaques showed that 1 year of daily ethanol self-administration induced greater dopamine uptake rates and sensitivity to D2 autoreceptors in the core of the NAc, thus driving a hypodopaminergic state (Siciliano et al., 2016). Altogether, these results suggest sex differences in striatal dopamine signaling and that alcohol exposure history may also differently affect release mechanisms depending upon sex-related factors. Previous findings indicate that dopamine neurotransmission is differently regulated in male and female rats (Walker et al., 2006; Georgiou et al., 2018). Contrary to what we would expect from our results, preclinical studies supporting sex differences in D2 receptors expression show that females have a higher D2 density in the striatum (Williams et al., 2021). Age could also be a factor in our results, especially the lack of effect of our quinpirole dose in female controls, as our animals were far into adulthood and it has been shown that D2 receptor density is declining in the striatum, although seemingly with a greater exponential decline in males than females (Williams et al., 2021). Finally, it is important to note that the fact that our control group did not undergo any behavioral procedure could influence our dopamine recordings and the responses to pharmacology. Phasic DA release in the NAc is involved with associative learning during operant conditioning (Roitman et al., 2004; Owesson-White et al., 2008), and active administration of drugs is associated with neuroadaptations in specific cognitive processes (Jacobs et al., 2003).

It is interesting to note that our behavioural results cannot be clearly explained by our FSCV data. As only male rats showed DM impairments with BD, we expected to observe phasic signaling impairments specifically in them, but only female rats were affected by BD in their sensitivity to quinpirole. It seems like impairments in DM processes are not explained by modifications in basal dopamine signaling or pre-synaptic D2 activity. The effects of D2 modulation on DM using the same RGT protocol than we did remains inconclusive. It has been shown that the selective D2 dopamine receptor antagonist eticlopride increased advantageous choices specifically in males, while quinpirole increased advantageous choices specifically in females (Zeeb et al., 2009). Other studies have found no effect at all on DM (Di Ciano et al., 2015) or only with a concomitant serotoninergic modulation (Di Ciano et al., 2018). Although the NAc core is important in value-based reward, being limited to it in ex vivo FSCV may explain the lack of a causality link between behavioral and neurobiological data. It will be important to use whole brain setups such as in vivo FSCV in the future to better understand how DM processes are affected by chronic ethanol.

Overall, our study emphasize that BD exposure impairs both DM processes and dopamine signaling in the core of the NAc in a sex-related manner, further suggesting that these effects may play a role in the vicious cycle leading to BD perpetuation and the early onset of AUD and dependence. We further advocate for the use of our operant model of BD, which shows better face validity and leads to changes in behavior without negative effects (behavioral suppression, inflammation, stress…) compared to classic passive exposures (injection, gavage…). So far, clinical results on the sex differences in the BD field of research looking at brain and cognitive deficits are still inconclusive and need more investigation. Thus, the overall difference in BD exposure raises the important question about the BD history that is also often matter of debate in human studies because it is difficult to have homogeneous population of binge drinkers with the same BD history.