Factors that influence the conditioned reinforcing effects of a cocaine-associated stimulus

2.1. Subjects

Sprague Dawley rats (Envigo) served as subjects. For experiments 1 and 2, 8 adult male Sprague Dawley Rats were assigned to each condition (total, n = 80). For experiments 3 and 8–9 rats were included per condition (total, n = 67). For experiments 1 and 3, all rats were provided ad libitum access to standard rat chow. For experiment 2, which examined food restriction, half of the rats received 20 g of food per day, and the rest were free fed (see Procedure for details). All rats were housed in a temperature—(21–23 °C) and humidity-controlled colony on a 12-hour light/dark cycle with lights on at 0700 and were provided water ad libitum. The Institutional Animal Care and Use Committee (Protocol Number: 00011244) at the University of Michigan approved these procedures.

2.2. Surgery

This surgery protocol was reported previously.¹⁴ Each rat first had a catheter implanted into the femoral vein for i.v. infusions of cocaine. Rats were initially anesthetized using injections of ketamine (90 mg/kg i.p.) and xylazine (10 mg/kg i.p.). Once the rat was fully anesthetized, we made an incision 1 cm away from the leg and isolated the femoral vein. A catheter made from Micro-Renathane Tubing was inserted into the vein and tied to the vein via suture (MRE-040, Braintree Scientific Inc). The catheter was attached to a mesh backplate with a stainless-steel tube (22-gauge) for externalization that was sutured between the scapulae (313-000BM-15UP/1/SPC; P1 Technologies). Pre- and postoperative analgesia was managed via an injection of Carprofen (5 mg/kg s.c.) that was delivered before the surgery and 24 hours following surgery. Catheter patency was maintained daily via flushing it with 0.5 mL of heparinized saline (50 USP/mL).

2.3. Apparatus

Twelve standard Med Associates operant chambers (Med Associates Inc) situated in sound attenuating cubicles were used. During the Pavlovian conditioning phase, a 10-mL syringe with cocaine was attached to tubing and placed into a syringe pump (Med Associates, PHM 107). The tubing was attached to a 22-g metal tube on a swivel (Instech, 375/22PLS, Plymouth Meeting) attached to a drug delivery arm. A plastic tubing tether covered by a metal spring was attached to the swivel and secured to the externalized stainless-steel tube on a rat’s backplate. For Pavlovian conditioning, a house light was placed on the left side of the chamber and a speaker, which was used to generate an 80-dB white noise stimulus via a multipurpose sound generator (Med Associates, ENV-230) or by a programmable audio generator (Med Associates, ANL-926). The white noise was calibrated to 80 dB using the Med Associates sound pressure level measurement package (ANL-929 A-PC). No response manipulanda were available inside the chamber throughout the Pavlovian conditioning phase.

During the instrumental acquisition phase, the syringe pump was loaded with a syringe filled with saline, but was never operated. For the instrumental acquisition phase, 2 nosepoke manipulanda (ENV-114BM), each illuminated by a single LED light, were placed on the right wall of the chamber and located 9 cm apart. One nosepoke was active, and the other nosepoke was used as an inactive control. All responses made on both nosepokes were recorded.

2.4. Procedure

These procedures were modified versions of those reported previously.^14,15

Pavlovian conditioning. For all experiments, we assigned animals to either paired Pavlovian conditioning or the random control procedure. According to the paired Pavlovian conditioning procedure, rats received infusions of cocaine that were delivered at the same time as the presentation of a stimulus (white noise + house light). The random control procedure group received the cocaine infusions and stimulus presentations according to 2 different clocks that operated independently. Cocaine deliveries and stimulus presentations occurred following a variable time (VT) 15-minute schedule, under which the cocaine-stimulus pairings occurred on average every 15 minutes (range = 0 to 30 minutes) to approximate the half-life of cocaine. Infusion duration was based on the rat’s body weight and lasted around 2.0 ± 0.5 s.

2.5. Instrumental acquisition

Following the Pavlovian conditioning phase, for all animals across all experiments, we inserted 2 nosepokes into the chamber. During acquisition, we placed a syringe filled with saline on the pump; however, the syringe pump remained off during all sessions. Rats were attached to the tether and placed into the chamber. At the start of the session, both nosepokes were illuminated. A response was recorded when the animal inserted its nose into the nosepoke and broke an infrared beam. Interacting with one of the nosepokes (active response) resulted in the presentation of the white noise + houselight stimulus. For each session, the initial active response emitted produced a stimulus presentation. Following this, stimulus presentations followed active responding according to a random-ratio 2 schedule of reinforcement, under which each active response had a 50% chance of resulting in the stimulus presentation. The other nosepoke (inactive response) had no scheduled consequences and served as a within-subject control. Each instrumental acquisition session lasted for 1 hour. Sessions were conducted 7 days per week. The instrumental acquisition phase lasted for 42 days (experiment 1), 14 days (experiment 2), or 7 days (experiment 3).

2.6. Drugs

Cocaine hydrochloride was obtained from NIDA Drug Supply and mixed into sterile 0.9% physiological saline solution.

2.7. Data analysis

Statistical analyses were performed using IBM SPSS Statistics (Version 28). We analyzed data for each experiment using a mixed ANOVA. For each analysis, Mauchly’s test was used to assess violations of sphericity. We corrected for violations of sphericity using either Greenhouse–Geisser or Huynh–Feldt estimates following the conventions outlined by Field.¹⁶

3.1. Experiment 1. Effects of the dose of cocaine on the conditioned reinforcing effects of a cocaine-associated stimulus

Figure 1 shows active and inactive responding as a function of conditioning history for each dose of cocaine used during Pavlovian conditioning. Rats emitted more active responses than inactive responses, indicated by a main effect of response type, F(1, 42) = 145.35, P < .001, η_p² = 0.78. Responding was different across daily testing, which was indicated by a main effect of day, F(7.12, 298.91) = 37.68, P < .001, and η_p² = 0.47. The number of responses was higher for higher doses of cocaine, which was indicated by a main effect of dose, F(2, 42) = 4.48, P = .02, η_p² = 0.18. Rats in the paired Pavlovian conditioning group made more responses than animals in the random control group, which was indicated by a main effect of conditioning history that approached significance, F(2, 42) = 3.65, P = .06, η_p² = 0.08, a response type × dose interaction, F(2, 42) = 8.43, P = .001, η_p² = 0.29, and response type × conditioning history interaction, F(1, 42) = 11.00, P = .002, and η_p² = 0.21. Rats responded more when higher doses of cocaine were used during Pavlovian conditioning, which was indicated by a day × dose interaction, F(14.23, 298.91) = 2.78, P = .001, and η_p² = 0.12. Rats emitted more active responding across days following paired Pavlovian conditioning, which was indicated by a day × conditioning history, F(7.12, 298.91) = 3.07, P = .004, and η_p² = 0.07, and a response type × day interaction, F(7.86, 330.01) = 2.69, P = .008, and η_p² = 0.06. Animals assigned to paired Pavlovian conditioning with higher doses of cocaine maintained more active responding across days, which was indicated by a response type × day × dose interaction, F(15.72, 330.01) = 2.55, P = .001, and η_p² = 0.11, a response type × day × conditioning history interaction, F(7.86, 330.01) = 3.17, P = .002, and η_p² = 0.07, and a response type × day × dose × conditioning history interaction, F(15.71, 330.01) = 1.73, P = .04, and η_p² = 0.08. No additional main effects or interactions were detected (P ’s > .10).

Figure 2 shows the Cohen’s d effect sizes quantifying the magnitude of difference in preference score between animals assigned to paired Pavlovian conditioning vs random control procedure for each dose of cocaine used during the Pavlovian conditioning phase. Cohen’s d effect sizes are categorized as small (d = 0.2), medium (d = 0.5), or large (d = 0.8). The dotted line marks the cutoff for a medium effect size, and the dashed line marks the cutoff for a large effect size. The effect size analysis was used to better visualize the persistence of the conditioned reinforcing effects of the cocaine-associated stimulus by assessing the instances of medium and large effect sizes.

Generally, we found that Pavlovian conditioning using 100 μg/kg/infusion of cocaine produced mostly small preference score effect sizes. Pavlovian conditioning using 320 μg/kg/infusion of cocaine resulted in mostly medium and large effect sizes, with medium effect sizes persisting out to 42 days and large effect sizes persisting out to 38 days. Pavlovian conditioning using 560 μg/kg/infusion produced several medium and large effect sizes, with medium effect sizes persisting for 42 days and large effect sizes persisting out to 30 days.

3.2. Experiment 2. Effects of food restriction on the rewarding effects of a cocaine-associated stimulus

Figure 3 shows the number of active and inactive responses across days of testing. Generally, responding decreased across sessions, which was indicated by a main effect of day, F(6.16, 344.66) = 76.57, P < .001, and η_p² = 0.58. Food-restricted rats emitted more responses vs free-fed animals, which was indicated by a main effect of feeding condition, F(1, 56) = 10.21, P = .002, η_p² = 0.15, and day × feeding condition interaction, F(6.16, 344.66) = 2.70, P = .01, and η_p² = 0.05. Rats that received 320 μg/kg/infusion during conditioning emitted more responses than rats that received 100 μg/kg/infusion during conditioning, which was indicated by a main effect of dose, F(6.12, 342.56) = 8.95, P = .004, and η_p² = 0.14. The highest level of responding was observed across days of testing in food-restricted animals that were exposed to 320 μg/kg/infusion during Pavlovian conditioning, which was indicated by a day × feeding condition × dose interaction, F(6.16, 344.66) = 3.79, P = .001, and η_p² = 0.063.

Generally, rats made a higher number of active responses, which was indicated by a main effect of response type, F(1, 56) = 141.12, P < .001, and η_p² = 0.72, as well as a response type × day interaction, F(6.12, 342.56) = 4.94, P < .001, and η_p² = 0.074. Rats that received 320 μg/kg/infusion during conditioning emitted more active responses than rats that received 100 μg/kg/infusion during conditioning, which was indicated by a response type × dose interaction, F(1, 56) = 23.06, P < .001, and η_p² = 0.30, and a response type × day × dose interaction, F(6.12, 342.56) = 3.71, P = .001, and η_p² = 0.062. Rats that underwent food restriction emitted more active responses than inactive responses than free-fed animals, which was indicated by a response type × feeding interaction, F(1, 56) = 15.22, P < .001, and η_p² = 0.21. Rats emitted more active vs inactive responses after paired Pavlovian conditioning vs the random control procedure, indicated by a response type × conditioning interaction, F(1, 56) = 5.70, P = .02, and η_p² = 0.09, and response type × day × conditioning interaction, F(6.12, 342.56) = 2.10, P = .051, and η_p² = 0.036. Food-restricted animals emitted more active responses across days of testing than free-fed animals, indicated by a response type × day × feeding condition interaction, F(6.12, 342.56) = 3.97, P < .001, and η_p² = 0.066. No additional main effects or interactions were found (P’s > .12).

Figure 4 shows effect sizes for preference scores, which display the magnitude of the difference between responding by rats that underwent the paired Pavlovian vs random control procedure. The highest effect sizes observed were for free-fed rats that received 100 μg/kg/infusion of cocaine during conditioning. For other groups, we documented primarily medium and small effect sizes. This is likely due to the ability of food restriction, as well as 320 μg/kg/infusion of cocaine, to enhance responding across both paired Pavlovian and random control groups.

Figure 5 shows weight gain for rats that received 100 μg/kg/infusion (top panel) or 320 μg/kg/infusion (bottom panel) for each of the Pavlovian conditioning groups. We found that rats gained weight across sessions, which was indicated by a main effect of day, F(2.59, 145.07) = 233.29, P < .001, and η_p² = 0.81. Rats that were free-fed gained more weight than rats that were food restricted, which was indicated by a main effect of feeding condition, F(1, 56) = 324.001, P < .001, and η_p² = 0.85, as well as a day × feeding condition interaction, F(2.59, 145.07) = 55.29, P < .001, and η_p² = 0.50. No additional main effects or interactions were detected (P’s > .25).

3.3. Experiment 3. Effects of days of conditioning and number of pairings on the reinforcing effects of a cocaine-associated stimulus

Figure 6 shows active and inactive responding for animals that received either a total of 100 or 50 pairings across 5 or 10 days. Responses emitted changed across days of testing, which was indicated by a main effect of day, F(3.77, 222.51) = 38.80, P < .001, and η_p² = 0.40. Rats emitted more active responses than inactive responses, which was indicated by a main effect of response type, F(1, 59) = 105.17, P > .001, and η_p² = 0.64. Animals with a history of paired Pavlovian conditioning made more active responses than animals assigned to the random control procedure, which was indicated by a main effect of conditioning history, F(1, 59) = 10.65, P = .02, and η_p² = 0.15, a response type × conditioning history interaction, F(1, 59) = 12.58, P = .001, and η_p² = 0.18, a day × conditioning history interaction, F(3.77, 222.51) = 2.68, P = .04, η_p² = 0.04, a response type × day × conditioning history interaction, and F(4.18, 246.41) = 4.03, P = .003, and η_p² = 0.06. Receiving 20 pairings per day across 5 days resulted in a lower level of responding, which was indicated by a main effect of total pairings, F(1, 59) = 10.61, P = .02, and η_p² = 0.15. We found a response type × total pairings interaction, F(1, 59) = 5.70, P = .02, and η_p² = 0.09, with animals exposed to 20 pairings a day making the fewest active responses. Responses emitted changed across the day of testing, depending on the total number of pairings delivered during Pavlovian conditioning, which was indicated by a day × total pairings interaction, F(3.77, 222.51) = 5.61, P = .007, and η_p² = 0.09. Animals that underwent 10 days of conditioning showed an augmented responding, which was indicated by a day × days of conditioning interaction, F(3.77, 222.51) = 38.80, P < .001, and η_p² = 0.40. Rats emitted more active vs inactive responding and levels of responding changed across days of testing depending on the days of Pavlovian conditioning, which was indicated by a response type × day × days of conditioning interaction, F(4.18, 246.41) = 3.00, P = .02, and η_p² = 0.05. We also found that the number of active vs inactive responses depended on the days of conditioning, as well as if the rat underwent paired Pavlovian conditioning or the random control procedure, which was indicated by interactions between response type × days of conditioning, F(1, 59) = 3.57, P = .06, and η_p² = 0.06, and a day × conditioning history × days of conditioning, F(1, 59) = 2.22, P = .07, and η_p² = 0.04, that trended toward significance. No additional main effects or interactions were detected (P’s > .11).

Figure 7 shows effect sizes for preference scores to quantify the magnitude of difference between animals that experienced the paired Pavlovian vs a random control procedure for each experimental condition. Animals that were exposed to 10 pairings across 10 days or 5 pairings across 10 days showed a similar pattern of effect sizes, with most effect sizes falling in the medium or large cutoff, whereas animals that were exposed to 20 pairings across 5 days showed a slightly higher number of small effect sizes. Animals that were exposed to 10 pairings across 5 days showed the highest prevalence of small effect sizes. This is likely due to the observations that animals in this condition showed a similar level of active and inactive responding regardless of Pavlovian or random control conditioning history.

3.4. Average total active and inactive responding across experiments 1–3

Figure 8 shows the average total responses emitted during acquisition for all experiments.

For experiment 1, rats made more responses following Pavlovian conditioning with higher doses of cocaine, which was indicated by a main effect of dose, F(2, 42) = 4.48, P = .02, and η_p² = 0.18. Animals assigned to paired Pavlovian conditioning made more active vs inactive responding during acquisition than animals assigned to the random control procedure, which was indicated by a main effect of conditioning history that approached significance, F(1, 42) = 3.65, P = .06, and η_p² = 0.08, a response type × dose interaction, F(2, 42) = 8.43, P < .001, and η_p² = 0.29, and a response type × conditioning history interaction, F(1, 42) = 11.001, P = .002, and η_p² = 0.21. Rats made more active than inactive responses, which was indicated by a main effect of response type, F(1, 42) = 145.35, P < .001, and η_p² = 0.77. A Bonferroni post hoc test revealed significant differences between animals that received 100 vs 560 μg/kg/infusion (P = .02). No other main effects or interactions were detected (P’s > .32).

For experiment 2, rats made more active than inactive responses, which was indicated by a main effect of response type, F(1, 56) = 141.12, P < .001, and η_p² = 0.72. Rats made more responses following conditioning with a higher dose of cocaine, which was indicated by a main effect of dose, F(1, 56) = 8.95, P = .004, and η_p² = 0.14, and a response type × dose interaction, F(1, 56) = 23.60, P < .001, and η_p² = 0.30. Animals assigned to paired Pavlovian conditioning vs random control procedure made more active responses than inactive responses, which was indicated by response type × conditioning history interaction, F(1, 56) = 5.70, P = .02, and η_p² = 0.09. Responding was higher in food-deprived animals, which was indicated by a main effect of food restriction, F(1, 56) = 10.21, P = .002, and η_p² = 0.15, and a response type × food restriction interaction, F(1, 56) = 15.22, P < .001, and η_p² = 0.21. No other main effects or interactions were detected (P’s > .12).

For experiment 3, rats made more active vs inactive responses, which was indicated by a main effect of response type, F(1, 57) = 93.16, P < .001, and η_p² = 0.62. Rats assigned to paired Pavlovian conditioning emitted more active responses than rats assigned to the random control procedure, which was indicated by a main effect of conditioning history, F(1, 57) = 9.93, P = .003, and η_p² = 0.15, and a response type × conditioning history interaction, F(1, 59) = 11.64, P = 0.001, and η_p² = 0.17. Animals that received 5 pairings per day showed the highest number of active responses, which was indicated by a main effect of pairings, F(2, 57) = 5.23, P = .008, and η_p² = 0.16, and a response type × pairings, F(1, 57) = 3.27, P = .05, and η_p² = 0.103. A Bonferroni post hoc revealed that animals with 20 pairings per day showed the lowest responding compared to 5 pairings (P < .001) and 10 pairings (P = .007). Receiving 5 or 10 pairings resulted in a similar number of responses (P = .20). No other main effects or interactions were detected (P’s > .29).

4.1. Effects of dose

We found that the conditioned reinforcing effects of the cocaine-associated stimulus were dose-dependent, such that the lowest dose (100 μg/kg/infusion) produced the lowest conditioned reinforcing effects, whereas the higher doses produced higher conditioned reinforcing effects. This observation was supported by the effect size analysis. Rats in the paired Pavlovian conditioning group using 320 μg/kg/infusion produced more medium and large effect sizes that lasted across the entire 42-day testing period. As such, 320 μg/kg/infusion produced the most robust conditioning for animals assigned to paired Pavlovian conditioning without promoting incidental conditioning in the random control group.

As we have demonstrated for remifentanil,¹⁴ cocaine-associated stimuli acquire conditioned reinforcing effects dose-dependently. For both drugs, we found that an intermediate dose (eg, 3.2 μg/kg/infusion of remifentanil and 320 μg/kg/infusion of cocaine) produced the most robust responding for the drug-associated stimuli in the paired Pavlovian groups and produced minimal conditioning in the random control procedure. This information is critical, as it will allow future research to maximize conditioning in the paired Pavlovian condition and minimize incidental conditioning in the random control procedure.

Our findings are consistent with DiCiano and Everitt.¹⁰ The current findings add to DiCiano and Everitt in several ways. First, they tested 125 or 250 μg/kg/infusion of cocaine and found that these doses of cocaine produced similar responding for the cocaine-associated cue, whereas we tested a broader range of doses. Notably, 100 and 320 μg/kg/infusion of cocaine produced similar levels of responding to DiCiano and Everitt. Paired Pavlovian conditioning with 560 μg/kg/infusion of cocaine produced greater responding than the other doses; however, the effect dissipated after 2 days of acquisition. Second, they allowed animals to self-administer cocaine that was paired with a visual stimulus, whereas we used response-independent drug infusions and stimulus presentations during Pavlovian conditioning. Rats were allowed to self-administer 50 infusions per day for 8 days (400 possible infusions). We were able to demonstrate that a light + white noise stimulus took on conditioned reinforcing properties after only 100 pairings delivered noncontingently. Finally, DiCiano and Everitt showed that the conditioned reinforcing effects of a cocaine-associated stimulus were quite persistent, with robust responding for the cocaine-associated cue lasting 59 days; however, they only tested responding for the cocaine-associated stimulus intermittently. The current data demonstrate that the conditioned reinforcing effects of cocaine-associated stimuli are also persistent when tested daily.

4.2. Effects of food restriction

Generally, we found that food restriction augmented responding during acquisition regardless of conditioning history in a dose-dependent manner (see the Strengths/Limitations section for discussion about conditioning in the random control group). Taken together, these data support the conclusion that food restriction augments the conditioned reinforcing effects of a cocaine-associated cue in this procedure.

These findings are consistent with the literature across behavioral procedures that has documented that food restriction augments the rewarding effects of drug-associated cues across stimulants and opioids. Previous research has shown that food restriction enhances responding for cocaine- and heroin-associated cues in a reinstatement procedure.17, 18, 19, 20, 21, 22 In addition, previous work demonstrated that food deprivation enhanced conditioned place preference using amphetamine,²³ morphine,^24,25 and cocaine.²⁶ As such, previous research supports that food restriction enhances the rewarding effects of drug-associated cues across drug classes and we demonstrated this is also the case in cocaine-associated stimuli to the New Response Acquisition.

These findings are in contrast to those with Robertson and Jutkiewicz,¹³ who examined the extent to which food restriction altered the conditioned reinforcing effects of a remifentanil-associated stimulus and found that food restriction did not enhance responding for remifentanil-associated cues for animals assigned to paired Pavlovian conditioning. Due to the precedent in the literature for food restriction to enhance the conditioned reinforcing effects of drug-associated cues in reinstatement procedures, the disparate findings between the ability of food restriction to enhance the conditioned reinforcing effects of remifentanil- and cocaine-associated cues in the New Response Acquisition procedure are unexpected. It is possible that the 1 factor driving these disparate findings is the discrepancy between the duration of action of remifentanil vs cocaine, especially given that food restriction enhanced the rewarding effects of heroin- and morphine-associated cues in a reinstatement and conditioned place preference procedure, respectively. Future studies could use an opioid with a longer duration of action than remifentanil (eg, fentanyl) to examine the extent to which the time course of the drug interacts with food restriction to enhance active responding. More work is needed to better characterize the conditions under which food restriction may enhance the conditioned reinforcing effects of opioid-associated stimuli in the New Response Acquisition procedure.

4.3. Effects of days and number of pairings

We found that the number of pairings per day and the total number of pairings produced differences in active and inactive responding. Generally, animals that experienced 20 pairings each day for 5 days produced the lowest levels of active responding. According to the effect size analysis, receiving 10 days of Pavlovian conditioning produced the highest conditioned reinforcing effects, with rats receiving 5 pairings for 10 days producing more effect sizes in the medium to large range across acquisition compared with the other groups. This experiment was included to follow up on research that found that 100 pairings (20 pairings per day for 5 days) produced more robust conditioned reinforcing effects than 20 pairings (20 pairings for a single day).¹¹ This experiment extended this line of research by considering the total number of pairings, as well as the number of days of Pavlovian conditioning in a different drug class.

Ten pairings a day for 10 days produced a similar performance during acquisition as 5 pairings a day for 10 days. DiCiano and Everitt¹⁰ also investigated the extent to which 4 vs 8 days of Pavlovian conditioning resulted in differences in performance during acquisition across 3 days of testing. Notably, responding during acquisition days 1 and 2 was higher for rats that received 8 rather than 4 days of conditioning; however, by the third day of acquisition, responding was similar across the 2 groups. In the current experiment, we found that 10 pairings for 10 days produced comparable levels of responding during acquisition as rats that received only 50 total pairings. As such, the finding that a similar level of conditioned reinforcing effects can be achieved with half the number of pairings is consistent across these studies.

4.4. Strengths/limitations

These studies must be interpreted in light of their limitations. One limitation is that the stimulus in the random control procedure took on conditioned reinforcing effects, which is consistent with other studies.^11,14,15 That is, when animals were exposed to remifentanil during the random control procedure, it was found that the random control procedure produced some conditioning and that dose of remifentanil and food restriction could affect the extent to which the stimulus takes on conditioned reinforcing effects. In the current experiments, we also found evidence that the stimulus took on conditioned reinforcing effects following exposure to cocaine in the random control procedure.

Rescorla²⁷ suggested that the random control procedure (aka the “truly random control procedure”) is an ideal control procedure for Pavlovian conditioning experiments. The unique feature of this control procedure relative to alternative control procedures is that the contingency between the unconditioned stimulus and neutral stimulus is suspended—that is, experimental events (eg, stimulus presentations and drug infusions) are scheduled according to 2 clocks that are simultaneously active but independent. The contingency between the stimulus presentation and drug infusion is suspended in that one event does not depend on or predictably follow the other. It is possible that using psychoactive drugs as an unconditioned stimulus introduces complications that Rescorla did not consider. Rescorla based these suggestions on fear conditioning studies, in which the unconditioned stimulus was an electric shock. In this case, the unconditioned stimulus (eg, electric shock) is discrete event. Unconditioned drug effects, on the other hand, linger. Given that the half-life of remifentanil is estimated to be around 3 min,²⁸ whereas cocaine is estimated to be around 18.11–24.8 minutes,^29,30 it is possible that there is still temporal contiguity between the subjective effects of the drug and stimulus presentation. It should be possible to minimize these incidental pairings by lengthening the VT schedule to far exceed the duration of the subjective effects of the drug. Another possible solution is to replace the random control procedure with the explicitly unpaired procedure.²⁷ According to this procedure, stimulus presentations and drug infusions are presented in random order that do not occur in temporal proximity. For future studies, using a longer VT schedule in an explicitly unpaired procedure together should minimize possible conditioning. One possible downside is that these adjustments would increase the length of the conditioning sessions. Given that the sessions lasted ∼150 minutes in the current studies (eg, experiments 1 and 2), adding more time between stimulus deliveries and drug infusions would lengthen an already long conditioning session. Future researchers would need to consider strategies to equate the length of the procedure across the paired Pavlovian and control groups.

One important implication to take away from this series of experiments is that a variety of factors determined the conditioned reinforcing effects of a stimulus following paired Pavlovian conditioning and the random control procedure. As such, when considering strategies to maximize the conditioning following the paired Pavlovian conditioning and minimize conditioning in the random control procedure, one should consider the class of the drug, the dose of the drug, the feeding status (eg, free feed vs food restricted), number of daily pairings, and number of total pairings. Finding a control procedure that does not result in the stimulus taking on conditioned reinforcing effects is an important consideration for researchers interested in characterizing neurobiological changes underlying conditioned reinforcement. That is, one would want a control group that characterizes potential neurobiological changes of the drug and stimulus exposure—per se. Pavlovian conditioning likely results in neurobiological changes that are unique to this type of learning. The ideal control group would allow researchers to expose animals to the drug and stimuli in a manner that does not produce learning.

In addition to improvements made on the parameters of the random control procedure, at least 1 additional control group would strengthen the interpretation of these data. The experimental design of this study does not allow us to assess whether the stimulus might serve as a primary reinforcer. That is, one alternative explanation between active and inactive responding across paired Pavlovian and random control groups is that the light + white noise stimulus maintained responding due to the primary reinforcing effects of this stimulus. In particular, the observation that rats made more active vs inactive responses following the random control procedure suggests this is a possibility. Future researchers should account for this possibility and include a condition in which responding during acquisition produces a novel stimulus following both paired Pavlovian conditioning and the random control procedure. Comparing differences in active and inactive responding as a function of conditioning history for stimuli used during conditioning (eg, white noise) vs a novel stimulus (eg, pure tone) would allow future researchers to more confidently attribute differences in responding to Pavlovian conditioning rather than the unconditioned reinforcing effects of the stimuli under study.

Another limitation is the exclusion of female subjects. Previous researchers found sex differences in the New Response Acquisition, such that a stimulus paired with 32.0 μg/kg/infusion of remifentanil acquired conditioned reinforcing effects for females but not males.¹³ In light of these findings, there is precedent in the literature to include females in New Response Acquisition studies. In addition, previous research found that cue-induced reinstatement for a cocaine-associated cue differed for males and females, such that females showed weaker reinstatement at the highest and lowest doses of cocaine.³¹ Further, previous research has demonstrated that female rats showed greater reinstatement in a cocaine-conditioned place preference procedure than males.³² Taken together, future research should examine sex differences in responding to cocaine-associated cues in the New Response Acquisition.

Another limitation is that food intake in the free-fed animals was not monitored. Knowing how much food the animals assigned to the free feed consumed each day would allow for a stronger assessment of the suitability of using 20 g of food per day in the food-restricted condition. Given that the literature on the ability of food restriction to enhance the rewarding effects of drug-associated cues uses a variety of different food restriction procedures, such as acute or chronic restriction procedures, a systematic evaluation of different durations and magnitudes of deprivation in the New Response Acquisition is warranted.^17,20