Wistar rats choose alcohol over social interaction in a discrete-choice model

The study addresses whether social interaction can serve as an effective alternative reward to reduce alcohol seeking and taking in rats, as reported for stimulants and opioids. Given the translational limitations of animal models of substance use disorders, integrating alternative non-drug rewards, particularly social rewards, may improve validity. Prior work showed most Wistar rats prefer saccharin over alcohol, and social interaction can suppress self-administration and craving for other drugs. The central question is whether such social choice effects generalize to alcohol. The authors evaluate choice between alcohol and social interaction across partner familiarity, interaction duration, housing conditions (short or chronic isolation vs. group housing), and sex, to test the robustness and boundary conditions of social choice against alcohol.

The authors situate their work within critiques of addiction models’ translational value, including failures of CRH1 antagonists in clinical trials. Choice-based paradigms show that access to non-drug rewards (food/sweet solutions) reduces drug taking in rodents and primates. Venniro et al. introduced a social choice model where social interaction reduced methamphetamine/heroin/cocaine self-administration and relapse-like behaviors. Prior alcohol work from the authors showed only ~15% of Wistar rats choose alcohol over saccharin, aligning with human AUD prevalence. However, whether social rewards similarly compete with alcohol was unknown. The literature also indicates complex social effects on alcohol use (social context can increase alcohol intake), and that social reward may be weaker than palatable food in choice contexts. Strain differences (Sprague-Dawley vs Wistar) and route of drug administration (oral alcohol vs intravenous stimulants/opioids) are noted as potential contributors to divergent findings.

Subjects: Adult male and female outbred Wistar rats (Charles River, Germany), total n=176 (48 assigned as social partners: 32 males, 16 females). Initial body weight 180–250 g (8 weeks). Group-housed (3–4 same-sex per cage) unless otherwise specified. Temperature 21 °C, humidity 45%, reversed 12 h light/dark cycle. Food and water ad libitum. All testing in dark phase. Ethics approved in Sweden. Apparatus and procedures: Operant conditioning chambers configured for alcohol self-administration and for social self-administration per published protocols. Alcohol reinforcer: 20% (v/v) ethanol solution delivered orally without saccharin fading. Social reinforcer: time-limited access to a partner rat via operant-controlled interaction (durations manipulated at 45 s, 30 s, 15 s). Training used fixed ratio schedules (FR1 then FR2), with discrete-trials choice procedures providing mutually exclusive access to alcohol or social interaction. Progressive ratio (PR) schedules were used to assess motivation (breakpoints). Discrete-choice sessions included sampling trials followed by choice trials. Inter-trial intervals (ITIs) manipulated in some phases. Effort cost manipulations: progressive increase of FR requirement for alcohol across sessions (e.g., FR3 up to FR32) while maintaining social at FR2 to determine indifference points. Experimental designs:

• Experiment 1 (n=32 males): Group-housed with cagemates used as partners. Alternate-day training to lever press for 20% alcohol or social interaction (FR1). Social interaction duration was systematically reduced (45 s for 10 sessions, 30 s for 4 sessions, 15 s for 4 sessions). Followed by discrete-trials choice between alcohol and social (FR1, 15 s social interaction for 6 sessions). Partner/test roles were switched at end; data pooled as no group differences.
• Experiment 2 (n=32 test males; 16 partner males): Novel partners housed separately. Phase 1: Short isolation manipulation—half isolated 2 h prior to social sessions; half remained group-housed; FR1 then FR2 alternating alcohol/social self-administration; choice sessions. Phase 2: Chronic isolation—previous short-isolation group moved to chronic isolation (≥2 weeks) while other remained group-housed. PR testing for alcohol and social. Discrete-choice with sampling analysis. Effort-cost manipulation: increased FR for alcohol across 33 sessions (≥3 per ratio; FR3 to FR32) vs social fixed at FR2, to assess indifference.
• Experiment 3 (n=32 test; 16 partners): Alcohol- and operant-naïve rats trained only on social self-administration to test whether prior alternation devalued social reward. 2×2 factorial: test rats group-housed vs chronically isolated; partner rats group-housed vs chronically isolated (n=8 per cell). Training: FR1 (11 sessions) then FR2 (19 sessions) for 30-min sessions. Choice test between social interaction and water (mutually exclusive). Then choice where alcohol concentration was increased across sessions (0% to 20%; ≥3 sessions per concentration) to assess at which concentration alcohol outweighs social. ITI manipulations (5 and 7.5 min) assessed.
• Experiment 4 (sex differences; n=32; 16 males, 16 females): Group-housed by sex; partners were unfamiliar same-sex rats. Alternate training on alcohol (20%) and social (30 s interaction), FR1 then FR2; PR testing; discrete-trials choice; effort-cost manipulation increasing FR for alcohol (FR3 to FR32) with social fixed at FR2. Outcomes and analyses: Active lever presses, reinforcers earned, percentage choice for alcohol or social across sessions, PR breakpoints, sampling counts, completed trials percentage. ANOVAs (two-way/three-way repeated measures) with post hoc Tukey HSD; eta-squared reported; significance typically p<0.05.

• Experiment 1: During alternate training (FR1), rats responded more for alcohol than social across sessions and earned more alcohol reinforcers irrespective of social-interaction duration (e.g., last three sessions: 34 ± 4.4 active presses for alcohol vs 8.2 ± 1.9 for social). Two-way RM ANOVAs showed main effects of session and reinforcer and significant interactions (e.g., presses: F(17,510)=12.66, p<0.001; reinforcer main effect p<0.001; interaction p<0.001). In discrete-choice sessions, alcohol choice increased from 54.7 ± 5.2% (session 1) to 79.7 ± 4.0% (session 6); 87% (28/32) were alcohol-preferring, 6% indifferent, 6% social-preferring.
• Experiment 2: Short (2 h) isolation before sessions did not affect responding or choice vs group housing. Chronic isolation increased social self-administration relative to group housing (e.g., last three sessions presses for social: 30.9 ± 4.3 isolated vs 13.5 ± 1.8 group-housed; main effect of housing p<0.001). PR: higher motivation for alcohol than social (reinforcer main effect F(1,30)=40.31, p<0.001), and chronic isolation increased overall motivation (housing main effect F(1,30)=4.66, p<0.05). In discrete-choice, both groups chose alcohol almost exclusively (last 3 sessions: 96.7 ± 1.7% and 96.2 ± 1.9% alcohol choice for group-housed and isolated, respectively). Increasing FR for alcohol reduced alcohol choice only at higher ratios; reduction began around FR20. Chronically isolated rats reached indifference near FR28; group-housed rats still preferred alcohol at FR32 though with fewer completed trials. Across increased FR, significant effects on choice and completed trials were observed (e.g., session main effect F(35,525)=6.99, p<0.001; completed trials housing main effect F(1,29)=13.08, p<0.01; session main effect F(35,1015)=20.29, p<0.001).
• Experiment 3: Social self-administration functions as a reinforcer, especially in chronically isolated test rats. Isolation increased social responding during training (housing main effect F(1,29)=10.86, p<0.01). Partner housing did not affect acquisition. At FR2, CI/GHP rats maintained higher social reinforcers (last three sessions: 17.3 ± 2.7 CI/GHP vs 10.5 ± 1.3 CI/CIP vs 7.0 ± 1.0 GH/GHP vs 8.0 ± 1.5 GH/CIP). In choice vs water, only CI/GHP showed robust social preference initially (79.2 ± 9.7% social) and chronic isolation of test rats promoted social choice (test housing main effect F(1,21)=4.56, p<0.05), while group housing of partner also promoted social choice (partner housing main effect F(1,21)=10.29, p<0.01). When increasing alcohol concentration (0–20%), social choice declined across groups (alcohol concentration main effect F(8,232)=12.60, p<0.0001). CI/GHP was the only group maintaining social preference at low alcohol (2%), but by 15% alcohol, social choice was significantly reduced across groups; CI/GHP first showed alcohol preference at 15% (social choice ~28.7%). Increasing ITI did not affect alcohol preference.
• Experiment 4 (sex): No sex difference in social self-administration; males earned nearly double alcohol reinforcers than females in training (last 3 sessions: 35.1 ± 1.9 vs 20.5 ± 1.3). PR showed higher motivation for alcohol than social (reinforcer main effect F(1,30)=18.87, p<0.001) with a significant sex × reinforcer interaction (F(1,30)=9.14, p<0.01); males had higher alcohol breakpoints than females. In discrete-choice, both sexes primarily chose alcohol (no sex main effect), with 93% of females and 87% of males alcohol-preferring. When increasing FR for alcohol, alcohol choice declined in both sexes but females shifted toward social at lower FR (differences at FR24, FR28, FR32; males remained >50% alcohol at FR32: 58.1 ± 13.3%; females 33.9 ± 9.0%). Overall: Across manipulations of partner familiarity, interaction duration, housing (short/chronic isolation), alcohol concentration, ITI, and sex, rats predominantly chose alcohol over social interaction. Rats persisted in choosing alcohol even when the effort cost for alcohol was 10–16× higher than for the social reward (range depending on sex and housing).

The results contradict prior findings with stimulants and opioids where social interaction reduces drug taking and relapse-like behavior, indicating that the social choice model does not readily generalize to alcohol. Several factors could contribute: (1) Route of administration differences (oral alcohol vs intravenous drugs) are considered but deemed unlikely to fully explain the divergence, given prior data showing strong preference for saccharin over alcohol under a similar choice procedure. (2) Strain differences (Wistar used here vs Sprague-Dawley in much prior social-choice work) may influence outcomes, although recent independent work (including Long Evans rats) also found alcohol preferred over social. (3) Social reward may be a weaker reinforcer than potent sweet solutions, making it less competitive with alcohol in choice contexts. (4) Alcohol may interact with social context differently than psychostimulants/opioids; social presence can increase alcohol intake and operant responding, suggesting potential bidirectional reinforcement between alcohol and social context. Chronic isolation increased social motivation (self-administration and PR) but did not translate to increased social choice over alcohol, potentially because isolation also enhances drug seeking and stress responsivity. Overall, findings highlight specific alcohol–social reward interactions that limit the protective effect of social alternatives observed with other drugs.

This study demonstrates that, in outbred Wistar rats, alcohol is robustly preferred over social interaction in a discrete, mutually exclusive choice paradigm. This preference persists across partner familiarity, interaction duration, housing conditions, and sex, and withstands substantial increases in response requirements for alcohol. Social interaction functions as a reinforcer—particularly in chronically isolated rats—but is generally insufficient to outcompete alcohol except against water or very low alcohol concentrations. These findings, replicated across laboratories and strains in related work, suggest the social choice model does not generalize to alcohol and that alcohol–social reward interactions are distinct from those of other drugs. Future research should: test across additional strains and settings; dissect neurobiological mechanisms underlying alcohol–social reward interactions; manipulate social reward magnitude/duration and session structures; and examine environmental or pharmacological interventions that might enhance social reward competitiveness against alcohol.

• Species/strain and generalizability: Only Wistar rats were used; prior social-choice studies often used Sprague-Dawley. Strain differences could influence outcomes.
• Route and modality: Alcohol was orally self-administered, whereas many social-choice studies used intravenous stimulants/opioids; modality differences may affect choice dynamics.
• Session structure: Social self-administration sessions were 30 minutes (shorter than some prior work using 1–2 h), which may limit the number of social reinforcers earned and underrepresent social reward value.
• Training history: Alternate training on alcohol and social in some experiments might have devalued social reward; although Experiment 3 mitigated this by training social-only, residual effects cannot be fully excluded across the whole study.
• Stress and housing: Chronic isolation enhances social motivation but also can increase drug seeking and stress, potentially confounding its effect on choice.
• Task parameters: While ITI and social-interaction duration were manipulated, other aspects (e.g., social partner characteristics beyond housing and familiarity, or extended access) may alter outcomes. The exact indifference point varied with housing/sex, indicating sensitivity to conditions.
• Potential textual/measurement constraints: Some reported statistics reflect trends or partial effects and the study relies on specific operant configurations; generalization to other operant setups requires caution.