Effect of cocaine exposure on estimates of reward value and neurobiological correlates

Abstract

Drug addiction is characterized by an inability to correctly adapt behavior in a changing environment. The orbitofrontal cortex (OFC) is implicated in such adaptive responding by signaling information about expected outcomes. Several studies have shown that this function may be disrupted by drug-induced neuroadaptations in OFC. A recent study has demonstrated that the OFC is also crucial for integrating information to imagine novel outcomes. This ability to signal an outcome never before received is fundamental for interpreting reality and make adaptive decisions in the present, but also for learning when those imagined outcomes are not received to improve decision-making in the future. Here, I explored whether cocaine exposure may alter this function. To address this question, I trained animals to self-administer cocaine, and then tested their performance in a Pavlovian over-expectation task. This task consists of three phases: conditioning, compound training, and extinction testing. In conditioning, rats are trained that several cues predict reward. Subsequently, in compound training, two of the cues are presented together, still followed by the same reward. Typically, this results in increased responding to the compound cue. Subsequent test trials with the cues by themselves reveal a reduction in conditioned responding to the compounded cues. I found that prior cocaine self-administration impaired both the ability of imaging new likely outcomes and learning when those outcomes turned out to be incorrect. These data are consistent with a drug-induced disruption of OFC's role in outcome signaling but might also be explained in other ways. To test the hypothesis more directly, I recorded single unit activity in the OFC of cocaine-experienced rats. Consistent with the hypothesis, OFC signaling in cocaine-exposed animals did not reflect a real-time integration of the cue-evoked expectations for reward, necessary to generate novel estimates about future outcomes. These results are consistent with proposals that cocaine disrupts the normal ability of networks in OFC to signal estimates about future outcomes, normally used for both adjusting current behavior and driving learning. Such deficit results in inflexible behavior and impaired learning and it may account for the complex pattern of maladaptive behaviors associated with drug abuse.