Alcohol reshapes reward learning differently in male and female rats (2025)

New research shows how chronic alcohol exposure impairs flexible decision-making in male rats, while females maintain performance despite neural changes, pointing to critical sex differences in brain response to addiction.

Study: Chronic ethanol exposure produces sex-dependent impairments in value computations in the striatum. Image Credit:Gorodenkoff/ Shutterstock

In a recent study published in the journal Science Advances, researchers employed a novel dynamic probabilistic reversal learning (dynaPRL) task to investigate the impact of chronic ethanol (EtOH) dependence in rats, particularly the striatal encoding of reward learning signals.

The study revealed that EtOH induces significant alterations in exploration-exploitation trade-offs and several other behavioral metrics, with distinct neural adaptations in females and impairments predominantly in male rats. These behavioral deficits were long-term, persisting for more than 10 weeks after the last EtOH exposure.

These findings highlight that female and male rats differ in their neural encoding patterns and provide novel insights into the sex-specific effects of EtOH dependence in decision-making.

Background

Value-based decision-making is a principle- and knowledge-dependent cognitive process that selects the most beneficial action from several actions of varying personal benefit. Recent human and rodent studies have demonstrated that chronic ethanol (EtOH) exposure attenuates learning and cognition, altering brain function associated with reward processing and decision-making.

These changes are most profound in the dorsomedial striatum (DMS). Furthermore, previous studies and psychiatric disease records suggest substantial differences between males and females in their susceptibility and responses to alcohol-related disorders. Unfortunately, the mechanisms underlying alcohol’s disruption of striatal dynamics and the potential sexual dimorphism of male and female neural disruptions following chronic EtOH exposure remain understudied.

About the Study

The present study aims to address current knowledge deficits by leveraging novel dynamic probabilistic reversal learning (dynaPRL) tasks on male and female rats. The rats used in these experiments were 10-week-old wild-type Long-Evans (LE) rats (male n = 17, female n = 10).

Behavioral experiments were carried out in several sequential phases. All experiments were conducted in a custom-designed, soundproof modular operant chamber, which included a reward magazine fitted with two levers. Rodent entry into the reward magazine triggered lever insertion.

Pressing a lever resulted in rats being presented either with a reward (100-μl sucrose solution) or no reward. During the first (training) phase, pressing either lever resulted in reward delivery with a 100% probability. During the second (reversal task training) phase, only one lever provided the reward, while pressing the other gave no reward.

The standard probabilistic reversal learning (PRL) task experiment was carried out between one and three days following the above learning phase. The experiment consisted of pressing the correct lever, resulting in reward delivery 70% of the time. If the wrong lever was pressed, there was a 10% chance of the ‘reward’ lever becoming the ‘no reward’ lever and vice versa. Based on their performance in the PRL task, rats were separated into EtOH-exposed (male n = 8, female n = 5) and control (air; male n = 9, female n = 5) cohorts.

Exposure to either EtOH vapor or air was carried out for 14 hours, followed by 10 hours of withdrawal, with the procedure repeated over five consecutive days. Blood was routinely extracted from rodent tails to measure their blood EtOH concentrations (BECs). Several behavioral metrics (e.g., reversal performance) were used to record rodent behavior events during both the PRL and dynaPRL experiments.

Rats were further surgically implanted with custom-made electrode arrays to record their neural signals during behavioral experiments.

The novel dynaPRL task maintained the same average reward probability as the standard PRL task experiment but introduced unpredictable block transitions: the probability of pressing the correct lever delivering a reward and the likelihood of a block transition were assigned to a pseudorandomly chosen block. Each block had its own paired probabilities: 1. 45% reward, 45% lever swap, 2. 60% reward, 30% lever swap, 3. 80% reward, 10% lever swap.

These blocks varied in expected and unexpected uncertainty, as suggested by the authors, requiring rats to flexibly adapt their behavior to shifting reward contingencies. “The lever outcome contingencies were randomly reversed to one of the three blocks every 15 to 30 trials, except when a rat made four or more consecutive incorrect choices within certain trial blocks.”

Study Findings

The present study produces several notable findings: withdrawal from chronic EtOH exposure did not appreciably alter individual-specific performance in the standard PRL task. No differences were observed between male and female rats for either EtOH-exposed or controls.

The dynaPRL task, however, proved substantially more cognitively challenging for EtOH-exposed rats, with marked performance deficits noted compared to controls. While controls could adapt their standard PRL training to the dynaPRL task, EtOH-exposed rats (especially males) found difficulties doing so. Notably, female EtOH-exposed rats showed minimal behavioral deficits over time, while male performance did not improve even after 10 weeks of abstinence.

Support vector machine (SVM) analyses confirmed these findings, further revealing that male EtOH-exposed rats suffered from slowed adaptive learning, reduced exploration, and altered outcome-specific value updating, all of which affected their behavioral outcomes (obtaining the reward). SVM decoding was significant in males only; females were not classifiable by treatment group.

Neural recordings showed that male rats had reduced choice-related signaling in the DMS but heightened outcome encoding, whereas females exhibited distinct neural adaptations without corresponding behavioral deficits.

Reinforcement Learning Insights

Reinforcement learning models offered further insight: both sexes demonstrated enhanced value updating after rewards, but males exhibited enhanced forgetting of unchosen actions after no reward, which contributed to cognitive rigidity. These computational differences reflected changes in learning rates (α⁺) and forgetting rates (θ⁻), particularly in males, which may bias behavior toward previously rewarded actions.

Conclusions

The present study sheds light on the mechanisms underpinning value computation impairments in chronic EtOH-exposed rats. It highlights their sex-dependent effects (male performance impaired more strongly than female) and protracted effects in males, suggesting distinct sex-specific neural dynamics in value-based decision processing.

“These mechanisms identified in males may underlie some of the cognitive deficits observed in humans with alcohol use disorder. Low cognitive flexibility imposes a higher risk of alcohol use disorders. Behavioral and neural mechanisms for the changes we observed here after chronic EtOH might resemble those underlying loss of cognitive flexibility after chronic use of other drugs of abuse or other psychiatric disorders.”