International Journal of Comparative Psychology

This special issue of the International Journal of Comparative Psychology is based on presentations delivered at the Focus Session of the 2004 Winter Conference on Animal Learning and Behavior (WCALB) held in Winter Park, Colorado. The Associative Mechanisms and Drug-Related Behavior Focus Session began with an invited address by Shepard Siegel titled The Ghost in the Addict: Drug Anticipation and Drug Addiction. He described an impressive body of research showing that conditioning mechanisms underlie drug tolerance and withdrawal. Siegel's address underscored the important contribution of associative mechanisms to drug-related behavior and set the stage for the six papers presented in this issue. Siegel began by describing his landmark study that first demonstrated the "situational specificity of tolerance" (Siegel, 1975). In that study, tolerance to morphine was only observed when rats were injected with morphine in an environment where they had previously experienced morphine. In contrast, no tolerance to morphine was observed when rats were injected in a novel environment. This result demonstrated that environmental factors might be as important, or even more important, than pharmacological factors in the expression of tolerance to drugs. Siegel pointed out that these results were anticipated by Subkov and Zilov (1937) who demonstrated conditioned tolerance of epinephrine-induced tachycardia. Siegel hypothesized that this situational specificity of tolerance was mediated by conditioned compensatory responses (CCRs) that counteracted the analgesic effects of morphine. According to this conditioning account of tolerance, the environmental stimuli present before and during morphine (the unconditioned stimulus or US) administration should act as Pavlovian conditioned stimuli (CSs). Through these pairings, the CSs come to elicit a conditioned response (CR) that opposes the direct effects of morphine. Therefore, since morphine itself produces analgesia, the environmental CSs that are paired with morphine come to elicit hyperalgesia. These oppositional processes then summate to produce a zero net effect, which manifests itself as tolerance, when morphine is administered in the presence of cues previously paired with morphine.

A critical prediction of the CCR analysis of tolerance is that an effect opposite to the direct effects of morphine (e.g., hyperalgesia) should be observed if the morphine-paired CSs are presented without the morphine (e.g., saline injection substituted for morphine). This is because the full expression of the CCR should be elicited with nothing to counteract them. Siegel (1975) showed that hyperalgesia is indeed observed when previously morphine-paired stimuli are presented in the absence of morphine to morphine-tolerant rats. He has called these unopposed CCRs "withdrawal symptoms" (Siegel, 1999). Thus, for Siegel, tolerance and withdrawal are both manifestations of a CCR—tolerance is observed when the CCR is elicited in the presence of the drug and withdrawal symptoms are observed when the CCR is elicited in the absence of the drug (Siegel, 1999; 2002). Siegel proceeded to review numerous studies conducted over the past 30 years supporting the view that drug tolerance reflects the processes of classical conditioning. Principally, this evidence comes from studies showing that tolerance is affected by learning contingencies in the same way that other nondrug Pavlovian CRs are affected by these contingencies. This reveals generality of process through "functional contiguity" (Sidman, 1960). These learning phenomena include, but are not limited to, extinction, external inhibition, latent inhibition, partial reinforcement effects, blocking, sensory preconditioning as well as electrophysiological and pharmacological manipulations (Siegel, 1975, 1989, 1991; Siegel & Larson, 1996; Dafters et al., 1983; Siegel et al., 2000). So where is the ghost in Siegel's address? In describing his experience with opium addiction, Jean Cocteau wrote "the dead drug leaves a ghost behind. At certain hours it haunts the house" (Cocteau, 1958, p. 60). Siegel materializes the ghost by reframing it in terms of Pavlovian conditioning. For Siegel, the 'ghost' refers to the CRs elicited by drug-associated CSs resulting from extended drug experiences. The candidates for conditioned stimuli can be numerous and include the complex of stimuli present when drugs are taken, such as people, places, sounds and smells. He also posited that the CSs may be interoceptive in nature. Siegel presented his recent research on interoceptive drug-associated cues that logically extend his research on Pavlovian conditioning of exteroceptive cues. This work essentially shows that interoceptive cues can indeed acquire CS functions in ways similar to exteroceptive cues. He considered two types of interoceptive cues, those associated with self-administration and drug onset cues. Self administration cues are stimuli arising from the active administration of the drug (such as movement of the body and other proprioceptive stimuli). Evidence was presented that self-administration cues contribute to tolerance and symptoms of withdrawal (Weise-Kelly & Siegel, 2001; MacRae & Siegel, 1997). Siegel then described research demonstrating the CS function of drug onset cues. In a prototypical experiment, rats receive chronic injections of a large dose of morphine (50 mg/kg). On test days, a much smaller dose (e.g., 5 mg/kg) is given. The small dose of morphine precipitated opiate withdrawal as evidenced by the behavioral and thermic data. This finding is expected if the interoceptive stimulation produced by the small dose was similar to the early drug-onset cues associated with the administration of the large dose. In other words, the early drug-onset cues are analogous to exteroceptive morphine-paired stimuli and elicit CCRs (i.e., precipitate withdrawal) when presented without the US (see Sokolowska, Siegel, & Kim, 2002). Siegel's keynote address provided convincing evidence that drugassociated stimuli, environmental and internal, play a critical role in drug tolerance and withdrawal. The six papers presented in this issue are concerned with a variety of effects of drug-related stimuli, including place conditioning (Bevins), selective associations produced by cocaine-associated stimuli (Weiss, Kearns, Cohn, Panlilio & Schindler), conditioned tolerance to the ataxic effects of alcohol (Brooks), the drug as a CS (Tomie, Mohamed, & Poherecky), and the conditioned reinforcing properties of drug-paired stimuli (Shelton & Beardsley, Newman & Beardsley, and Bevins). Siegel's address and the spectrum of learning paradigms represented by these six articles confirm the central role learning and associative mechanisms play in drug-related behavior. They also illustrate that this is an active area of research that needs people with diverse backgrounds and interests including classical and operant conditioning, behavioral pharmacology and drug abuse. Clearly, people other than metaphysicians acknowledge that the "ghost" is alive and well, and worthy of study.

A major criticism of the place conditioning procedure for studying conditioned drug reward is that it is relatively insensitive to large quantitative shifts in drug dose (i.e., dose effects are all or none). Experiment 1 demonstrated this lack of sensitivity using a wide range of intravenous (IV) cocaine doses (0.1, 0.3, 0.45, 0.6, 0.9, or 1.2 mg/kg). Rats had cocaine repeatedly paired with one distinct end compartment of a 3 compartment apparatus; vehicle was administered in the other end compartment. In a subsequent drug-free choice test, the 0.45 to 1.2 mg/kg doses of cocaine conditioned a place preference. The magnitude of the effect did not differ. Experiment 2 used a modified version of this standard place conditioning method. In this alternative method termed reference-dose procedure, a fixed dose of cocaine (reference dose) was repeatedly paired with one end compartment (i.e., 0.45 mg/kg); the comparison dose of cocaine was administered in the other end compartment (vehicle, 0.6, or 1.2 mg/kg). Preference for the comparison-dose compartment increased with dose—a graded doseeffect curve. In contrast to the standard procedure, the reference-dose procedure revealed that the conditioned rewarding effect of 1.2 mg/kg of cocaine was greater than that of 0.45 mg/kg. This increase in sensitivity to conditioned reward with the reference-dose procedure will likely increase the utility of the place conditioning method as a preclinical model, as well as a procedure for studying processes mediating associatively-motivated choice behavior.

When a tone-light compound was a discriminative stimulus for cocaine-reinforced responding, the light gained most of the control over responding. In contrast, when the compound was an aversive SD for shock-avoidance, tone control increased. In previous studies, tone control also increased when the tone-light compound was made aversive by signaling food-absence. However, that was not the case in Experiment 2 where tone-light signaled cocaine-absence. Experiment 1 produced an interincentive (cocaine vs. shock) selective association with drug self-administration maintained behavior for the first time. This extends the generality of the selective association biological constraint on learning to self-administered drugs.

Four adult male rhesus monkeys were trained to lever press for cocaine under a daily two-component MIX PR (progressive ratio) schedule. During the first 10 min of experimental sessions, completion of progressive ratios resulted in 1-s presentations of brief visual stimuli (BS; colored lights) associated with cocaine infusions during the second component. Stimulus lights of different colors were associated with doses of 3, 30, and 300 μg/kg cocaine as the available self-administered infusate. A 5-min time out period followed the first component, which in turn was followed by a 60-min component during which completion of progressive ratios resulted in cocaine infusions and the associated visual stimuli. Once reinforcer rates had stabilized under each dosing condition in both components, break point tests were conducted separately for BS as the reinforcer and with cocaine + stimuli as the reinforcer. Break points for lever pressing maintained by BS alone increased as they were paired with increasing doses of cocaine. Break points maintained by actual cocaine delivery, however, demonstrated an inverted U-shaped function to cocaine dose. The results of this study suggest that the strength of cocaine-seeking behavior varies monotonically with the self-administered dose of cocaine and that the level of motivation to obtain cocaine may not be directly revealed by levels of actual cocaine self-administration.

This article reviews ethanol ataxic tolerance experiments with rats that investigate spontaneous recovery after extinction and how extinction-related cues reduce this recovery. Tolerance to the effects of many drugs including ethanol is partly the result of Pavlovian conditioning. Tolerance to the ataxic (and other) effects of ethanol depends critically upon the circumstances in which the drug is administered. Tolerance shows other characteristics common in Pavlovian conditioning, e.g.,. it can be extinguished and is subject to spontaneous recovery. The analogy of spontaneous recovery to instances of relapse in humans potentially makes such spontaneous recovery relevant to both researchers and clinicians. Recently, "extinction cues" have been found to reduce spontaneous recovery and other relapse- like effects in the animal conditioning laboratory. These cues may work in part by activating an association formed during the extinction process, and thus they may serve as memory retrieval cues. Research assessing spontaneous recovery using an ethanol ataxia method, as well as other Pavlovian conditioning methods, has contributed to an understanding of the properties and utility of extinction cues. These topics are addressed and the potential implications of this research for treating substance abusers is considered.

The reinstatement paradigm has been proposed as an animal model of human drug relapse. In most reinstatement studies, conditioned stimuli accompany drug infusions during self-administration, responding in extinction, as well as responding during reinstatement tests. The importance of these extinguished drug-paired stimuli during stress-induced reinstatement has not been examined. In this study, rats were trained to self-administer 0.5 mg/kg/infusion cocaine during daily, 2-h sessions until behavior stabilized. Each cocaine infusion was accompanied by a 6-s flashing light + tone conditioned stimulus (CS). In two groups of rats, responding during subsequent extinction and reinstatement had no scheduled consequences (CS Omitted). In two other groups of rats, responding produced the light + tone but no cocaine injections (CS Present). Footshock did not significantly reinstate cocaine seeking behavior in CS Omitted rats. Footshock significantly reinstated cocaine-seeking behavior over multiple test sessions in both CS Present groups, regardless of whether footshock reinstatement was examined on consecutive days or with trials spaced two days apart. These data show that environmental stimuli and stressors which are ineffective by themselves to occasion reinstatement of cocaine-seeking behavior can do so if concurrently present.

Previous studies of autoshaping of drinking report a positive relationship between experience with autoshaping procedures and drinking, but this effect was confounded with age, as the rats were older when they drank more. The present experiment evaluated the effects of the age of male Long-Evans hooded rats [90-days old (Younger group) vs. 135 days old (Older group)], at the beginning of the study, on drinking induced by Pavlovian autoshaping procedures. Autoshaping procedures consisted of pairings of sipper conditioned stimulus (CS) with food unconditioned stimulus (US). Rats were deprived of neither food or fluid, and sweeteners were not employed at any time during the study. For all rats (n = 32), during sessions 1-10, the sipper CS contained water. Thereafter, for rats in the Ethanol groups (n = 20), the sipper CS contained ethanol, with the concentration (1, 2, 3, 4, 5, 6%, v/v) increasing across autoshaping sessions. For rats in the Water groups (n = 12), throughout the experiment the sipper CS contained tap water (0% ethanol). Rats in the Younger group drank more ethanol and more water from the sipper CS than rats in the Older group, and across age groups there was more ethanol drinking than water drinking, an effect unlikely due to foraging for calories. Data support the hypothesis that ethanol’s pharmacological effect was to enhance autoshaping, resulting in a positive feedback loop inducing still more ethanol drinking. The younger rats were more vulnerable to autoshaping effects. Implications for models of addiction are discussed.