Augmented reinforcer value and accelerated habit formation after repeated amphetamine treatment

Abstract

Various processes might explain the progression from casual to compulsive drug use underlying the development of drug addiction. Two of these, accelerated stimulus–response (S–R) habit learning and augmented assignment of motivational value to reinforcers, could be mediated via neuroadaptations associated with long-lasting sensitization to psychostimulant drugs, i.e. augmented dopaminergic neurotransmission in the striatum. Here, we tested the hypothesis that both processes, which are often regarded as mutually exclusive alternatives, are present in amphetamine-sensitized rats. Amphetamine-sensitized rats showed increased responding for food under a random ratio schedule of reinforcement, indicating increased incentive motivational value of food. In addition, satiety-specific devaluation experiments under a random interval schedule of reinforcement showed that amphetamine-sensitized animals exhibit accelerated development of S–R habits. These data show that both habit formation and motivational value of reinforcers are augmented in amphetamine-sensitized rats, and suggest that the task demands determine which behavioral alteration is most prominently expressed.

Introduction

One of the defining characteristics of drug addiction is the occurrence of a compulsive pattern of drug use, which persists in the face of adverse social and health consequences (American Psychiatric Association, 2000, O'Brien and McLellan, 1996, Volkow and Li, 2004). Several influential hypotheses attempt to explain the progression from casual to compulsive drug use, which occurs during the development of addiction. According to the incentive sensitization theory, drug-induced adaptations in the neural circuits underlying incentive motivation cause drugs to gain an enhanced incentive value (i.e. exaggerated drug “wanting”, perhaps equivalent to drug craving), leading to excessive pursuit of drugs (Robinson and Berridge, 1993). A different theory argues that aberrant instrumental learning underlies the development of addictive behavior. Thus, instrumental learning for both drugs and natural rewards can progress through action–outcome (A–O) responding, during which performance of an action is determined by explicit knowledge of its consequences, to an automatic, stimulus–response (S–R) structure (commonly referred to as habits). Responses driven by an S–R structure are stimulus-driven and occur without the necessity for volitional control (Dickinson, 1994). In the case of drug abuse, accelerated and/or inappropriate S–R learning has been hypothesized to cause drug seeking behavior to become inflexible (Berke and Hyman, 2000, Everitt and Robbins, 2005, Tiffany, 1990).

Repeated exposure to drugs of abuse causes behavioral sensitization, a progressive and persistent increase in both the psychomotor response to drugs (Pierce and Kalivas, 1997, Robinson and Becker, 1986, Robinson and Berridge, 1993, Stewart and Badiani, 1993, Vanderschuren and Kalivas, 2000) and the reinforcing properties of drugs and natural reinforcers (Fiorino and Phillips, 1999, Lett, 1989, Piazza et al., 1989, Robinson and Berridge, 2003, Taylor and Horger, 1999, Valadez and Schenk, 1994, Vezina et al., 2002, Wyvell and Berridge, 2001). Behavioral sensitization is associated with an exaggerated dopaminergic response to drugs in both dorsal and ventral striatum (Kolta et al., 1985, Paulson and Robinson, 1995, Pierce and Kalivas, 1997, Robinson and Berridge, 1993, Robinson et al., 1988, Vanderschuren and Kalivas, 2000). Because of the involvement of ventral striatal dopamine in incentive salience attribution and the motivational influence of environmental stimuli on behavior (Berridge and Robinson, 1998, Cardinal et al., 2002, Schultz, 2002), this augmented release of dopamine has been hypothesized to critically contribute to exaggerated drug “wanting” (Robinson and Berridge, 1993, Robinson and Berridge, 2003). Interestingly, the dorsal striatum (Packard and Knowlton, 2002, White and McDonald, 2002, Yin and Knowlton, 2006), including its dopaminergic innervation (Faure et al., 2005, Packard and White, 1991), has been shown to be involved in S–R learning processes. Thus, though often regarded as mutually exclusive explanations for the development of drug addiction, both the increase in motivational value of drugs predicted by the incentive sensitization theory and the abnormal S–R learning proposed in the aberrant learning theory could be mediated by neural changes associated with drug sensitization.

We therefore hypothesized that amphetamine sensitization is associated with augmented incentive motivational as well as S–R learning processes. To test this hypothesis, we employed distinct schedules of reinforcement in operant conditioning paradigms and devaluation setups to elucidate responding based on S–R and A–O responding. Studies by Dickinson and colleagues have demonstrated that interval and ratio schedules of reinforcement in an operant conditioning paradigm rely on different associative structures (Dickinson, 1985, Dickinson et al., 1983). Ratio schedules, in which higher response rates are rewarded with more reinforcement in a linear fashion, promote responding based mainly on motivational value of the reinforcer. Animals trained on this schedule continue to demonstrate explicit knowledge of reinforcer value even after long periods of training, indicating that responding is determined by an A–O rather than an S–R structure. In contrast, interval schedules impose a maximum on the number of reinforcers that can be earned in a set time period, irrespective of response rate. Responding following extended training under an interval schedule can become independent of reinforcer value, demonstrating that it is based on S–R associations. Using the distinct properties of the two reward schedules, we tested the hypothesis that elevated response levels in sensitized animals are related to motivational value of the reinforcer using a ratio schedule. Increased response rates in amphetamine-treated animals would indicate enhanced motivational value of the reinforcer. Next, we investigated whether amphetamine-sensitized rats would show accelerated S–R learning using a satiety-specific devaluation paradigm in animals trained under an interval schedule of reinforcement. In this paradigm, the effect of decreased incentive value of a reinforcer can be tested by measuring rates of responding for a food reinforcer after animals have been prefed to satiation with that particular foodstuff before testing. After limited training, when responding is directed by A–O processes, responding is decreased by devaluation, but after extended training when an S–R relationship has developed, responding is insensitive to devaluation (Coutureau and Killcross, 2003, Dickinson, 1985). Earlier occurrence of insensitivity of responding to devaluation in amphetamine-treated animals would indicate accelerated S–R learning.