Drugs of abuse and their associated CSs modulate memory consolidation, but it is possible that these effects are influenced by the mode of administration (i.e., active vs passive). Furthermore, if memory modulation by a drug CS is indeed a conditioned response, then it should display extinction like other conditioned responses. To explore these questions, this study employed a self-administration/yoked design in male Sprague-Dawley rats: one group was trained to intravenously self-administer infusions of 0.05 mg/kg heroin paired with the activation of a light CS, and the yoked group received an equal number of infusions and light CS presentations independent of their lever responses. Self-administration/yoked training was followed by extinction during which responses on the active lever had no consequences, and by a cue-induced reinstatement test during which active lever responses resulted in the reactivation of the light CS, but no heroin delivery. The post-training effects of self-administered/yoked heroin, extinction and reinstatement sessions on memory consolidation were tested using the object location memory task (see Fig. 1). It was found that post-sample heroin enhanced object location memory in yoked, but not self-administering rats. However, post-sample extinction and reinstatement sessions equally modulated memory in both groups. Taken together, these data in rats support the perspective that mode of administration impacts the cognitive consequences of exposure to drugs and environmental stimuli linked to their reinforcing effects.
The current study generated two primary findings that appear relevant to addictive behaviours. First, it was found that when yoked rats received heroin on session 12, they displayed enhanced object memory location 72 hours later (Fig. 2, panel C). This observation is consistent with other demonstrations of passive heroin- or morphine-induced enhancement of various types of memory when administered post-training [8,10,28−31]. Interestingly, however, when rats were able to self-administer heroin, there was no evidence of enhanced object location memory, an observation inconsistent with the memory consolidation hypothesis of drugs of abuse [11]. It could be argued that heroin self-administered in the current study may have produced cognitive impairments [32]. However, if a decline in cognitive functioning was indeed a pharmacological consequence of repeated heroin exposure, then it should have been observed in yoked animals as well.
Although the reasons underlying differential memory modulation by self-administered and yoked heroin are unclear, it is possible to postulate a role of prediction error and associated DA activation. In fact, it is known that a prediction error is generated when there is a discrepancy between what occurs and what is expected[33–35] and that the neurobiological correlate of this cognitive response is activation of mesolimbic DA [36–39]. It is also known that DA plays a key role in memory consolidation[40–42], and our laboratory has demonstrated that DA antagonists can indeed block the facilitation of memory consolidation by cocaine and nicotine when passively administered [43]. Accordingly, it is possible that once drug-taking behaviour becomes well established and the outcome of lever pressing is anticipated by the subject, the experience of engaging in those behaviours and the pharmacological effects of the drug become expected, the dopamine signal is reduced/lost, and the experience can no longer modulate memory consolidation. Within this context, the experience of yoked rats was likely different. That is, because no behavioural response reliably produced drug effects, it is possible that, other than the information provided by the context, yoked rats could neither predict, nor anticipate, when the drug would be administered or the amount they would receive. Such interpretation is consistent with measures of DA concentrations in the NAc revealing that both passive[44–46] and self-administered[47] morphine/heroin elevate DA, but that the DA response is lost after repeated heroin self-administration [46]. Clearly, it will be important to directly explore whether the memory enhancing function of post-sample passive heroin is dependent on DA activity.
Second, it was found heroin paired CSs enhanced object location memory equally in both experimental groups. In fact, in those that self-administered heroin, exposure to contextual cues during the first extinction session (Fig. 3, panel A) and exposure to the contextual cues plus response-contingent light CS during reinstatement (Fig. 4, panel A) elevated responding on the active lever and enhanced object location memory (Figs. 3 and 4, panel C). Similarly, although no comparable effects were observed on active lever responding, the effects on object location memory were identical in the yoked group. Overall, these findings confirm our previous observation that a heroin-paired CS can significantly modulate memory formation in rats [10], and further indicate that the mode of administration plays a minor role in this effect, presumably because Pavlovian conditioning occurs in both self-administering and yoked groups. This assertion is further supported by the observation that context-induced modulation of object memory was not observed after 6 extinction sessions (Fig. 3, panel D). That is, repeated exposure to the contextual heroin CS in the absence of heroin not only significantly decreased operant responding in self-administering rats, but also eliminated the CS effect on memory in both self-administering and yoked groups. As far as we know, this is the first demonstration that conditioned memory modulation follows Pavlovian principles such as extinction [48].
This is the first study in rats that attempted to establish a role for the mode of administration in the effects of heroin and heroin-paired cues on memory consolidation. Admittedly, object location memory could have been tested at other stages of drug exposure, other memory tasks could have been employed revealing different opiate effects on memory consolidation [49–53], different heroin doses or schedules could have been used, and the study could have also included female animals. Notwithstanding these limitations, our data reveal an interesting interaction between the pharmacological effect of heroin and environmental cues in determining their shared effects on cognitive process. That is, heroin’s action on memory formation may be dependent on how the drug is received and may change over repeated administration. However, our data also clearly indicate that a significant level of conditioning occurs to a variety of cues present during drug exposure regardless of how the drug is administered, and these cues have reliable and predictable effects on memory consolidation processes.