Altered performance in a rat gambling task after acute and repeated alcohol exposure

Abstract

Rationale: A bidirectional relationship between alcohol use disorder (AUD) and deficits in impulse control and decision making has been suggested. However, the mechanisms by which neurocognitive impairments predispose to, or result from AUD remain incompletely understood.

Objectives: The aim of this study is to gain more insight in the effects of alcohol exposure on decision making and impulse control. We used two modified versions of the rat gambling task (rGT) that differ in the net gain and the punishment magnitude associated with the different response options.

Methods: In experiment 1, we assessed the effects of acute alcohol treatment (0-0.8 g/kg) on rGT performance. In experiment 2, we determined the effects of alcohol on rGT acquisition (15 sessions, 0.6 g/kg). Next, these animals were challenged with alcohol (0-1.0 g/kg) prior to rGT sessions.

Results: Acute alcohol treatment suppressed baseline performance in both rGT versions but only modestly altered decision making. Treatment with alcohol during acquisition increased risky choices in the rGT version that involved larger punishment and blunted the reduction in win-shift behavior during acquisition in both rGT versions. Moreover, rats treated with alcohol during acquisition showed an increase in premature and perseverative responding upon subsequent alcohol challenges (0-1.0 g/kg) and were less sensitive to the behavioral suppressant effects of alcohol.

Conclusions: Our results show that repeated alcohol exposure alters decision making during rGT acquisition and reduces the ability to adjust choice behavior on the basis of feedback. In addition, repeated alcohol exposure unmasks its behavioral disinhibitory effects in the rGT. Impaired responsiveness to choice feedback and behavioral disinhibition may contribute to the development of AUD.

Fig. 1

Acquisition of choice behavior in GT1 (a) and GT2 (b). Choice behavior during the first five free-choice sessions differed between the two gambling tasks, in that rats showed a higher preference for the safe choice in GT2. Moreover, while animals in GT1 preferred the safe and optimal choice above the risky choice, animals in GT2 preferred the safe choice above the optimal and risky choice. Following five forced-choice sessions, rats in both rGT versions developed a preference for the optimal choice, which became more pronounced with increased training. Data are shown as the mean percentage choice + SEM

Fig. 2

Blood alcohol level (BAL) after an IP alcohol injection. The BAL was assessed in a separate group of animals at 30 min after IP injection of 0.6 and 1.2 g/kg alcohol (a). Investigation of the BAL over time after an injection with 0.6 g/kg alcohol showed maximal BAL with least variation at 15–30 min postinjection (b). Data are shown as the mean + SEM (a) or as mean and individual data points (b)

Fig. 3

The effect of acute alcohol treatment on stable choice behavior in the rGT (experiment 1). Alcohol significantly reduced the percentage of optimal choices. This effect of alcohol was independent of GT version. Hence, the data from both rGT versions were collapsed. Data are shown as the mean percentage choice + SEM. *Different from vehicle treatment (post hoc paired t test, p < 0.05)

Fig. 4

The effects of repeated alcohol (0.6 g/kg) or vehicle administration on the acquisition of choice behavior in the rGT, followed by ten sessions without treatment. Repeated alcohol administration during rGT acquisition increased risky choices in GT1 (c). Data are shown as the mean percentage choice + SEM. *Different from vehicle-treated animals (post hoc Student’s t test, p < 0.05), $ p < 0.062 compared to vehicle-treated rats

Fig. 5

The effects of repeated alcohol (0.6 g/kg) or vehicle administration on the percentage of shifts toward another choice after being rewarded or punished. Repeated alcohol administration during rGT acquisition tended to decrease lose-shift behavior after punishment on the risky choice in GT1 (a), but not in GT2 (b). Regardless ofchoice or GT version, vehicle-treated animals showed reduced win-shift behavior over sessions, whereas alcohol-treated animals did not (c). The percentage of lose-shifts was not different over sessions or between treatment groups (d). Data are presented in bins of five sessions (a, b) or sessions (c, d) and are shown as the mean + SEM percentage of lose-shift and win-shift behavior. *Different from vehicle-treated animals (post hoc Student’s t test, p < 0.05)

Fig. 6

The effects of acute alcohol treatment on stable choice behavior in the rGT in experiment 2. Alcohol had no effects on choice behavior, irrespective of pretreatment (alcohol or vehicle) or rGT version. Hence, the data from both pretreatment groups and rGT versions were collapsed. Data are shown as the mean percentage choice + SEM

Fig. 7

The effects of alcohol on behavior in the rGT during alcohol challenge sessions in alcohol-pretreated and vehicle-pretreated animals. In vehicle-pretreated animals, alcohol dose-dependently reduced total choices (a), premature responses (b), and perseverative responses (c), and increased omissions (d) and choice latency (e). In contrast, in alcohol-pretreated animals, alcohol had a biphasic effect on total choices, premature responses, perseverative responses (increase followed by decrease as the alcohol dose increased), omissions and choice latencies (decrease followed by increase as the alcohol dose increased). Alcohol pretreatment and alcohol challenges did not affect collect latency (f). The alcohol challenges had similar effects in GT1 and GT2, and the data were therefore pooled. Data are shown as mean + SEM. *Difference between pretreatment groups (post hoc Student’s t test, p < 0.05); #different from vehicle challenge (post hoc paired t test, p < 0.05)