The Effect of Economy Type on Heroin and Saccharin Essential Value

Abstract

According to behavioral economics, reinforcer value should be lower in an open economy than in a closed economy. An animal model was used to determine how economy type affected the value of heroin and saccharin. In a first phase, separate groups of rats worked for heroin or saccharin. The price of these reinforcers increased over sessions. For rats in the open heroin or open saccharin economies, the work period of each session was followed by a post-work period where a cheaper source of heroin or saccharin was available for three hours. For rats in the closed economies, the work period was their only opportunity to obtain the reinforcer. Rats in the open saccharin economy worked less hard to defend consumption of saccharin as price increased than rats in the closed saccharin economy. That is, opening the saccharin economy reduced its essential value. In contrast, economy type had no effect on heroin’s essential value. In a second phase, rats were allowed to choose between heroin and saccharin. The majority of rats strongly preferred saccharin over heroin regardless of economy type. The finding that economy type changed the essential value of saccharin, but not heroin, adds to previous findings suggesting that the value of drug reinforcers is unaffected by future drug availability. The difference in effect of economy type on drug vs. non-drug reinforcers could be relevant to addiction.

Essential value (EV) is a measure that reflects inelasticity of demand for a reinforcer (Hursh & Silberberg, 2008). As price increases, subjects work harder or pay more to defend consumption of reinforcers of high EV than they do to defend consumption of low EV reinforcers. Several recent studies have shown that in both animals and humans the EV of drug reinforcers is associated with addiction-related behaviors. For example, inelastic drug demand (i.e., high drug EV) in rats is associated with continued cocaine taking despite negative consequences (Bentzley et al., 2014), choice of cocaine (James et al., 2018; Kearns et al., 2017) or heroin (Schwartz et al., 2017) over non-drug alternatives, and reinstatement of drug seeking after extinction (Bentzley et al., 2014). Furthermore, long or intermittent access to cocaine (Bentzley et al., 2014; Christensen et al., 2008; James et al., 2018) or heroin (Lenoir & Ahmed, 2008; Schwartz et al., 2017) produces an increase in the drug’s EV. In humans, high drug EV is associated with drug-related problems and measures of drug dependence (Aston et al., 2015; Lemley et al., 2016; Murphy et al., 2009; O’Connor et al., 2016). These findings linking drug EV to addiction-like behavior are consistent with the hypothesis that addiction involves a “reinforcer pathology,” characterized by an overvaluation of drug reinforcers, where overvaluation is defined in terms of demand inelasticity (Bickel et al., 2011; Jarmolowicz et al., 2015).

Identifying the factors that influence drug EV might provide insight into addiction. The behavioral economics literature indicates that economy type is one variable that contributes to reinforcer value (Hursh, 1980, 1984). In a closed economy, there is only one source of the reinforcer. For example, if the only food that a subject obtains is that earned by lever pressing during experimental sessions, the economy for food is closed. If, on the other hand, the subject earns food during experimental sessions and is also given additional food by the experimenter after the session, the food economy is described as being open. The behavioral economic prediction is that a reinforcer should have higher value, and subjects should work harder to defend consumption of it, in a closed economy than in an open economy. This is because the extra-session reinforcers available in an open economy are thought to substitute for the reinforcers that can be earned during the session. In a closed economy, there are no such substitutes.

Several studies have found evidence consistent with the behavioral economic prediction. For example, Hursh et al. (1989) found that monkeys worked harder on a series of ascending fixed-ratio (FR) schedules for food when their only food was that obtained during the session than when they also had access to cheaper (i.e., available on a FR-1 schedule) food after the session. Lower food value in an open economy than in a closed economy has been reported in additional studies in monkeys (Carroll et al., 2000; Nader & Woolverton, 1992) and mice (Soto et al., 2016). Similar effects of economy type on reinforcer value have been observed in rats working for saccharin (Kim et al., 2018) or water (Ladewig et al., 2002), in monkeys working for ethanol (Carroll et al., 2000), and in humans working for cigarettes (Mitchell et al., 1994, 1998), coffee (Mitchell et al., 1995), or access to a video or video game (Roane et al., 2005).

In contrast to the results described above, which are in agreement with the behavioral economic prediction, there have been instances where reinforcer value appears to be unaffected by economy type. Carroll et al. (2000) reported that post-session access to phencyclidine (PCP) had no effect on monkeys’ progressive-ratio responding for PCP. This was observed under the same conditions that resulted in the expected effect of economy type on the monkeys’ responding for food and ethanol. More recently, Kim et al. (2018) found that post-session access to cocaine had no effect on rats’ demand for cocaine, whereas, under the same conditions, post-session access to saccharin made demand for it more elastic. Similarly, Banks and Negus (2010) found that post-session access to cocaine had no effect on monkeys’ preference for it over food. This contrasts with Nader and Woolverton’s (1992) finding that post-session food reduced monkeys’ preference for it over cocaine. It is notable that in the instances where economy type had no effect, the reinforcer was a drug of abuse. Perhaps for some drug reinforcers, access to the drug in the future does not substitute for current access.

The primary goal of the present experiment was to test the effect of economy type on the EV of heroin in rats. Rats were trained to lever press for intravenous heroin infusions during a daily “work” period where the price (number of responses per infusion) of heroin increased over sessions from FR 1 to FR 64. For rats in the closed economy, this work period was their only opportunity to obtain heroin. Rats in the open economy were given a second three-hour period, immediately after each work period, where they could press another lever to obtain heroin infusions that were always available on a FR-1 schedule. If future, cheaper heroin substitutes for current, expensive heroin, then heroin demand should be more elastic in the open economy than in the closed economy. To compare with heroin, the effect of economy type on the non-drug reinforcer saccharin was also evaluated here using the same procedures. Based on recent results (Kim et al., 2018), it was expected that opening the economy for saccharin would make demand for it more elastic. Finding that rats work less hard for heroin when they have post-session access to it would indicate that economy type is a determinant of heroin’s EV. On the other hand, finding that heroin’s EV is unaffected by economy type, whereas saccharin value is affected by post-session access, would suggest a potentially important difference between heroin and saccharin.

A secondary goal of the present experiment was to determine the effect of heroin and saccharin economy type on choice between these reinforcers. In a second phase, the work period was replaced by a choice period wherein rats made a limited number of mutually exclusive choices between heroin and saccharin. For rats in the open economies, this choice period was followed by three hours of access to heroin or saccharin on a FR-1 schedule. The behavioral economic prediction was that post-choice availability of, for example, saccharin would decrease preference for saccharin. This is because essentially freely available saccharin after the session could substitute for any saccharin forgone in favor of heroin during the choice session. On the other hand, previous studies investigating rats’ choice behavior have shown that it can be quite difficult to change their usually strong preference for saccharin over heroin (Lenoir et al., 2013).

Method

Subjects

Forty-four adult male Long-Evans rats (Envigo, Frederick, MD), weighing approximately 380–390 g at the beginning of experimental sessions, served as subjects. Seven other rats began the first phase of the experiment, but did not finish it due to catheter failure, illness, or failure to acquire stable lever pressing. Rats were individually housed in plastic cages with wood-chip bedding and had unlimited access to rat chow and water in their home cages. The colony room where the rats were housed had a 12-h light:dark cycle with lights on at 08:00 h. Training sessions were conducted five days per week during the light phase of the light:dark cycle. Throughout the experiment, rats were treated in accordance with the Guide for the Care and Use of Laboratory Animals (National Academy of Sciences, 2011) and all procedures were approved by American University’s Institutional Animal Care and Use Committee (protocol #1707, “Addiction as Maladaptive Choice of Drugs over Non-Drug Rewards”).

Apparatus

Training took place in 20 Med Associates (St. Albans, VT) operant test chambers. Each chamber measured 30.5 × 24 × 29 cm and had aluminum front and rear walls with clear polycarbonate side walls. Three Med-Associates retractable levers were located on the front wall of the chamber. A Med-Associates small cup liquid receptacle was located directly above the middle lever. Saccharin reinforcers (0.3 ml of 0.2% saccharin solution) were delivered to this liquid receptacle. A 100-mA cuelight was located above each lever. A speaker located in the center of the front wall provided a tone stimulus (4500 Hz and 80 dB). A 100-mA houselight was located above this speaker. Heroin (provided by the Drug Supply Program, National Institute on Drug Abuse, Bethesda, MD) in a saline solution at a concentration of 0.1 mg/ml was infused at a rate of 3.19 ml/min by 10-ml syringes driven by Med-Associates (St. Albans, VT) syringe pumps. Tygon tubing extended from the 10-ml syringes to a 22-gauge rodent single-channel fluid swivel (Instech Laboratories, Plymouth Meeting, PA) and tether apparatus (Plastics One, Roanoke, VA) that descended through the ceiling of the chamber. Heroin was delivered to the subject through tubing that passed through the metal spring of the tether apparatus.

Surgery

Before beginning operant training, all rats were surgically prepared with chronic indwelling jugular vein catheters, using procedures described in detail elsewhere (Thomsen & Caine, 2005; Tunstall & Kearns, 2014). In brief, approximately 3.5 cm of Silastic tubing was inserted into the right jugular vein. From this insertion site, an additional 8 cm of Silastic tubing passed under the skin to the midscapular region where it connected to the 22-gauge stainless steel tubing of a backmount catheter port (Plastics One, Roanoke, VA) that was implanted subcutaneously. The spring tether in the chamber was attached to the threaded plastic cylindrical shaft of the port that protruded through an opening in the skin. All surgery was conducted under ketamine (60 mg/kg) and xylazine (10 mg/kg) anesthesia. Rats were given 7–10 days to recover from surgery. Catheters were flushed daily with 0.1 ml of a saline solution containing 1.25 U/ml heparin and 0.08 mg/ml gentamicin.

Procedure

Overview.

See Figure 1 for a procedural schematic. Rats were randomly assigned to one of four groups prior to beginning training: the heroin open economy group (Heroin Open; n = 10), the heroin closed economy group (Heroin Closed; n = 11), the saccharin open economy group (Saccharin Open; n = 12), or the saccharin closed economy group (Saccharin Closed; n = 11). During Phase 1, demand for heroin and saccharin was measured. Each session consisted of a 1-h “work” period followed by a 3-h post-work period during which rats in the open economies obtained additional access to one of the reinforcers. Rats in the closed economies remained in the chamber during this period, but did not have access to heroin or saccharin. The FR in effect during the work period increased over sessions from 1 to 64. A FR-1 schedule always operated during the post-work period for open economy rats. In Phase 2, rats’ preference between heroin and saccharin was measured. During the first hour of each Phase 2 session, rats in all groups were given mutually exclusive choices between heroin and saccharin. This was followed by a 3-h post-choice period during which rats in the open economies had additional access to heroin or saccharin on an FR-1 schedule while rats in the closed economies had access to neither reinforcer.

1.

Schematic diagram of procedures used in Phase 1 and Phase 2. In the second phase, all groups were trained on the same discrete-trials choice procedure.

Phase 1: Demand.

The work period of the session began with illumination of the houselight and insertion of the left lever. For rats in the two heroin groups, presses on the left lever resulted in a 0.03 mg/kg infusion of heroin followed by a 10-s timeout period signaled by the tone and illumination of the cuelight above the lever. Lever presses were recorded, but had no consequences during the 10-s period. For rats in the saccharin groups, presses on this lever were followed by delivery of 0.3 ml saccharin, illumination of the 100-mA bulb inside the liquid receptacle for 1.67 s (the time it took to fill the cup), and a 10-s timeout period signaled by illumination of the cuelight above the lever. Initially, a FR-1 schedule operated during the 1-h work period of the session. Rats were trained on FR-1 for a minimum of 14 sessions and until there were three consecutive sessions during which the number of reinforcers obtained did not differ from the rolling three-session mean by more than 20%. Once meeting this stability criterion, the FR increased over two-session blocks according the following sequence: 2, 4, 8, 16, 32, 64.

The 3-h post-work period of the session began with retraction of the left lever. For rats in the open economy groups, the middle lever was inserted. For rats in the closed economy groups, no levers were inserted. For the Heroin Open group, presses on the middle lever resulted in a heroin infusion (same 0.03 mg/kg dose as before) and a 10-s timeout accompanied by the tone and cuelight above the left lever as previously. For the Saccharin Open group, presses on the middle lever resulted in delivery of 0.3 ml of the saccharin solution (plus illumination of the bulb inside the receptacle) and a 10-s timeout signaled by the cuelight above the left lever. The FR was always 1 during the post-work period throughout the experiment. That is, even when the response requirement on the left lever during the work period was FR 64, the FR on the middle lever during the post-work period was still 1.

Phase 2: Choice.

After the final demand session, rats were trained to press the right lever for the reinforcer that they had not yet experienced. During the time when the work period normally happened, the right lever was inserted and rats in the heroin groups pressed this lever for saccharin while rats in the saccharin groups pressed this lever for heroin. Reinforcer deliveries were accompanied by the same cues described previously except the cuelight that signaled the 10-s timeout was located above the right lever (i.e., the currently available lever). The 1-h period of pressing the right lever was followed by a 3-h period during which the usual open or closed economy conditions prevailed. That is, for rats in the open economy groups the middle lever was inserted and rats could press it on an FR-1 schedule for heroin or saccharin. All rats except one learned to press the right lever in three sessions on this procedure. One rat in the Saccharin Open group required an additional four sessions.

After rats learned to press the right lever, choice sessions began. The choice portion of the session occurred during the hour when the work period previously occurred. A discrete-trials procedure was used where choices were separated by a 5-min inter-trial interval (ITI). A 5-min ITI was used here to minimize the potential for heroin-saccharin interactions. A previous choice study found that cocaine-saccharin interactions occurred when no ITI was used, resulting in exclusive preference for cocaine (Vandaele et al., 2015). The first four trials of each session were forced-choice trials where only one lever, the left or right one, inserted. There were two such trials with each reinforcer, with the order randomized. These forced-choice trials ensured that rats sampled the reinforcers available on each lever twice at the start of each choice session. A single press on the inserted lever resulted in delivery of the reinforcer, presentation of the associated cues, retraction of the lever, and the start of the ITI. After the four forced-choice trials, free-choice trials began. Now, both the left and right levers inserted simultaneously and a single press resulted in delivery of the selected reinforcer, presentation of the associated cues, retraction of both levers, and the start of the ITI. During free-choice trials, if the rat did not make a response within 2 minutes, both levers retracted and a 5-min ITI started. Such trials were scored as omissions. No limited hold was used during forced-choice trials to ensure that rats experienced each reinforcer twice at the start of the session. When 1 h had elapsed since the start of the session, the choice portion of the session ended and the 3-h post-choice portion of the session began. The 3-h post-choice period was just like that during demand sessions. That is, the middle lever inserted and rats in the open economy groups pressed on a FR-1 schedule for heroin or saccharin. Choice sessions continued until for three consecutive sessions the percentage of trials on which heroin was chosen did not differ from the rolling three-session mean by more than 20 percentage points.

Data Analysis

For the demand data, the number of heroin infusions self-administered and the number of saccharin reinforcers earned were averaged over the training sessions at each FR (for FR 1, the mean of the last two sessions was used). Individual and group mean consumption data were fit by Hursh and Silberberg’s (2008) exponential-demand equation:

$$\text{log}Q=\text{log}{Q}_0+k(e^{–a{Q}_0}{}^C–1),$$ (1)

where Q is quantity consumed, Q₀ is consumption as price approaches 0, k is a constant defining the consumption range in log units (k = 2 here), α determines the rate of decline in consumption, and C is cost (FR size).

The primary measure of interest from the demand phase was EV, which is inversely related to α and reflects inelasticity of demand. EVs of heroin and saccharin were calculated for each subject according the formula given by Hursh (2014):

$$\text{EV}=1/\left(100*\alpha *k^{1.5}\right)$$ (2)

Normalized group mean demand curves are also presented below. To normalize consumption, the number of cocaine infusions and saccharin reinforcers obtained was expressed as a percentage of Q₀. Price was normalized by converting it to the number of responses required at a particular FR to obtain 1% of Q₀.

For all statistical tests, α was set to 0.05. Non-parametric statistics were used because some key variables were not normally distributed. Kruskall-Wallis tests, followed by Mann-Whitney tests where appropriate, were used to evaluate the reliability of between-group differences. Friedman’s two-way analysis of variance by ranks tests, followed by Wilcoxon signed-ranks tests where appropriate, were used to evaluate the reliability of within-subjects effects. The Benjamini-Hochberg (1995) procedure was used to keep α ≤ 0.05 for collections of multiple related pairwise comparisons.

Results

Phase 1: Demand

The number of sessions that each group had on the FR-1 schedule during the work period ranged from 15.8 to 17.2 (no group difference, Kruskall-Wallis H[3] = 0.5, p > 0.9). The top two panels of Figure 2 show the group mean numbers of heroin and saccharin reinforcers obtained during the work portion of the session across the range of FRs tested as well as the exponential demand model fits to the group mean data. The bottom panels present the same data in normalized terms. For heroin, it is clear that demand was unaffected by economy type. The slopes of the heroin demand curves are virtually the same across economy types. In contrast, demand for saccharin was more elastic in the open economy than in the closed economy. That the slope of the saccharin demand curve is steeper in the open economy is especially apparent in the normalized curves (lower right panel).

2.

The top panels present group mean consumption of heroin (left) and saccharin (right) for rats in the open (circles) and closed (triangles) economies plus fits of the exponential demand model. Note that the scale of the Y-axes differ for heroin and saccharin. The inset tables provide mean (± SEM) essential value (EV), Q₀, and R² for the fits of the model to individual subjects’ data. The bottom panels show normalized consumption as a function of normalized price as well the exponential demand model fits. Consumption was normalized by expressing the numbers of reinforcers earned as a percentage of Q₀. Normalized price is the number of responses required at a particular FR to obtain 1% of Q₀. Note that the scale of the X-axes differ for heroin and saccharin.

Figure 3 presents the mean EV and the mean Q₀ derived from fits of the exponential model to individual subjects’ data. In general, these results parallel those in Figure 2. As the top panel of Figure 3 shows, the EV of heroin did not differ by economy type, but the EV of saccharin was lower in the open economy than in the closed economy. Overall, the EV of heroin was lower than that of saccharin, regardless of economy type. The bottom panel of Figure 3 shows that Q₀ did not differ by economy type for either heroin or saccharin. Regardless of economy type, the Q₀ of saccharin was higher than that of heroin.

3.

Mean (± SEM) essential value (top panel) and Q₀ (bottom panels) for heroin (left) and saccharin (right) based on fits of the exponential model to individual subjects’ data for the rats in the open (white bars) and closed (black bars) economies. * indicates p < 0.05. *** indicates p < 0.001.

Statistical analyses confirmed that for EV there was an effect of group (H[3] = 22.9, p < 0.001). Subsequent Mann-Whitney U tests indicated the EV of saccharin was significantly lower in the Saccharin Open group than in the Saccharin Closed group (U[11,12] = 27.5, p < 0.02), but the heroin groups did not differ from each other (U[10,11] = 45.0, p > 0.45). The EV of heroin in each of the heroin groups was significantly lower than the EV of saccharin in either of the saccharin groups (all Us ≤ 25.0, all ps < 0.025). There was a significant effect of group on Q₀ (H[3] = 32.9, p < 0.001). There was no difference in Q₀ across economy types for heroin (U[10,11] = 46.0, p > 0.5) or for saccharin (U[11,12] = 42.5, p > 0.1). Both of the saccharin groups had significantly higher Q₀ than either of the heroin groups (all Us = 0, all ps < 0.001).

The filled circles in the top panel of Figure 4 show that the mean number of post-work period heroin infusions obtained by the Heroin Open group increased over two-session blocks from about 12 to over 20. This increase was significant (Friedman’s test, X²[6] = 34.2, p < 0.001). The open squares in the top panel of Figure 4 show the mean combined number of heroin infusions obtained during the work period and the post-work period for the Heroin Open group. There was a significant increase in total infusions per day (Friedman’s test, X²[6] = 16.0, p = 0.014), indicating that the increase in post-work period heroin consumption over sessions more than compensated for the decrease in work period heroin consumption as price increased. For ease of comparison, the total heroin infusions obtained by the Heroin Closed group, which only had heroin access during the work period, are presented here too (triangles in Figure 4). These are the same data presented in the top panel of Figure 1. The bottom panel of Figure 4 shows the corresponding data for the saccharin groups. Rats in the Saccharin Open group significantly increased post-work consumption of saccharin as the work period FR increased (Friedman’s test, X²[6] = 35.8, p < 0.001). However, total (i.e., combined work and post-work) daily saccharin consumption was fairly constant for the Saccharin Open group, beginning and ending at about 80 saccharin reinforcers per day. There was no significant effect of session block for total saccharin reinforcers (Friedman’s test, X²[6] = 5.8, p > 0.4). This indicates that for this group the amount by which saccharin consumption decreased during the work period as FR increased was equivalent to the amount by which post-work period consumption increased.

4.

Mean (± SEM) number of reinforcers obtained during the post-work period (circles) and mean (± SEM) combined number of reinforcers (squares) obtained during the work and post-work periods for the Heroin Open group (top panel) and the Saccharin Open group (bottom panel). Triangles represent mean (± SEM) total number of reinforcers obtained for rats in the closed economy groups, which only had access to the reinforcer during work periods. Note the difference in Y-axis scales across panels.

Phase 2: Choice

Catheter failures or illness prevented a number of rats from completing the choice phase. This reduced the ns for the Heroin Open, Heroin Closed, Saccharin Open, and Saccharin Closed groups to 8, 10, 7, and 9, respectively, for this phase. The mean number of sessions to meet the stability criterion on the choice procedure ranged from 8.1 to 9.7 across groups (no group difference; Kruskall-Wallis H[3] = 2.0, p > 0.55).

Figure 5 shows group mean and individual subjects’ percent choice of heroin over saccharin averaged over the final 3 choice sessions. On average, rats chose heroin on approximately 15–25% of trials and there was no difference in this measure across groups (H[3] = 3.7, p = 0.3). Mean numbers of omissions ranged from 0 to 0.5 over groups. Five out of 34 rats (15%) chose heroin more often than saccharin. There was much variability in preference within groups. For example, in the Saccharin Open group, scores ranged from 0% to 100% heroin choice. Similarly, in the Heroin Closed group, scores ranged from 0% to 67% heroin choice. During the 3-h post-choice period of the session, the Heroin Open group self-administered a mean of 27.0 (± 4.0 SEM) heroin infusions per session. The Open Saccharin group earned 97.0 (± 16.2) saccharin reinforcers. Neither of these means was significantly different from mean post-work period consumption at the end of Phase 1 (Wilcoxon signed-ranks tests, both Zs ≤ 1.2, both ps > 0.2).

5.

Mean (± SEM) and individual subjects’ percent choice of heroin averaged over the final 3 choice sessions for the Heroin Open (circles), Heroin Closed (squares), Saccharin Open (upward-pointing triangles), and Saccharin Closed groups (downward-pointing triangles). The dashed line represents indifference (50% preference).

Discussion

The main finding of the present experiment was that opening the saccharin economy reduced saccharin’s EV, whereas opening the heroin economy under the same conditions had no effect on heroin’s value. These results systematically replicate and extend to a new drug class those of Kim et al. (2018), who found that saccharin’s EV was lower in an open economy than in a closed economy, but economy type had no effect on cocaine’s value. Regardless of economy type, saccharin’s EV in the present experiment was significantly higher than that of heroin. This outcome confirms the results of Schwartz et al. (2017) who found the same effect using somewhat different procedures. The finding in the present experiment that only a small minority of rats (15%) preferred 0.03 mg/kg heroin over 0.2% saccharin when given a choice is also consistent with other choice studies in rats using the same heroin dose and saccharin concentration (Lenoir et al., 2013; Madsen & Ahmed, 2015).

The behavioral economic account of the effect of opening the economy on reinforcer value is that future, cheaper reinforcers substitute for current, more expensive reinforcers (Hursh, 1984). This requires the subject to integrate current and future reinforcers and to adjust its current behavior in anticipation of future availability. There is evidence suggesting that rats in the Saccharin Open group did this. Throughout the demand phase, rats in this group consumed a consistent number of saccharin reinforcers per day when combined across the work and post-work periods. They did this by increasing their consumption during the post-work period by an amount that was approximately equal to the amount that saccharin consumption decreased during the work period as the FR increased. This pattern might be expected if post-work period saccharin substituted for work period saccharin. For rats in the Saccharin Closed group, there was no future saccharin that could substitute for current saccharin. According to the behavioral economic explanation, this is why the Saccharin Closed group worked harder to defend work period consumption than the Saccharin Open group.

That economy type did not affect heroin’s EV here might suggest that animals do not integrate current and future heroin reinforcers in the same way that they integrate current and future saccharin reinforcers. The lack of an effect of economy type on heroin’s EV is the third reported instance of a failure of economy type to affect how hard animals work for a drug reinforcer. Carroll et al. (2000) previously found that opening the PCP economy did not affect monkeys’ progressive ratio breakpoints for it. Kim et al. (2018) found that opening the cocaine economy did not alter cocaine’s EV in rats. In all three cases, the failure of economy type to affect the drug’s reinforcing value was found in the same study where, using similar procedures, opening the economy did reduce the value of a non-drug reinforcer (food in Carroll et al., saccharin here and in Kim et al.). Perhaps the failure of future availability to influence current behavior is a characteristic of abused drugs like heroin, cocaine, and PCP.

Alternatively, it is possible that another process counteracted the effect of opening the economy on heroin’s EV. Opening the heroin economy entailed increased exposure to heroin. Previous studies have found that increased exposure to heroin results in an escalation of intake and a decrease in the elasticity of demand for heroin (Lenoir & Ahmed, 2008). In the Heroin Open group here, there was a small, but significant increase in total heroin intake across the demand phase. Such an increase in total intake was not observed in the Saccharin Open group. Furthermore, Schwartz et al. (2017) found that exposure to 3-h heroin self-administration sessions on a FR-1 schedule for just 7 days resulted in an increase in heroin’s EV. Importantly, similar exposure to saccharin had no effect on saccharin’s EV. The additional 3 h of heroin access that the Heroin Open group had each day compared to the Heroin Closed group was the same amount of heroin exposure found to increase heroin’s EV in the Schwartz et al. study. For this reason, plus the modest escalation of total heroin intake over days in the Heroin Open group, it might have been expected that heroin demand would be more inelastic in the Heroin Open group than in the Heroin Closed group during the work period. Instead, no difference was observed. This lack of a difference may have been the result of two opposing processes: the additional heroin exposure in the Heroin Open group acted to make demand for heroin less elastic, but the substitution of future, cheaper heroin for current heroin acted to make demand for it more elastic. It is possible that these opposing processes also explain the lack of an effect of opening the economy on cocaine (Kim et al., 2018) or PCP (Carroll et al., 2000) value.

Some have questioned whether the behavioral economic substitution account of economy type effects is plausible for animals like rats. As Posadas-Sanchez and Killeen (2005) wrote, “explanation in terms of deferred consumption assumes foresight, planning, and some measure of self-control on the part of animals that fail to show even modest self-control over delays of dozens of seconds.” They suggest instead that differences in satiation across open and closed economies can account for the effects observed in many economy type studies. The design and reinforcers used in the present study were chosen so that satiation could be ruled out as a potential explanation. Additional reinforcers in the open economy were available only after the work or choice periods, which were of the same length in both economies, and the time courses of both saccharin and heroin would not allow consumption on one day to affect satiation level at the start of the next session. The half-life of saccharin in rats is only approximately 30 minutes (Renwick, 1980; Sweatman & Renwick, 1980). This is about the same or even shorter than the half-lives of the active metabolites of heroin, 6-acetylmorphine (Gottås et al., 2013) and morphine (Cicero et al., 1997; Gottås et al., 2013). Furthermore, satiation to saccharin, which has no calories, is controlled by immediate oral sensory stimulation, rather than longer term post-ingestive consequences such as overhydration or fullness of the stomach (Mook et al., 1981). For these reasons, satiety seems unlikely to account for the effect of opening the saccharin economy observed here.

Another potential explanation of the economy type effect is that it is a form of behavioral contrast (Williams, 2002). In a study concerned with contrast, Williams (1992) trained rats on a multiple schedule consisting of an initial 10-min component where lever pressing was reinforced by food according to a variable interval (VI) 1-min schedule. In one condition, this VI 1-min component was followed immediately by a second 10-min component wherein lever presses were reinforced by food on a FR-1 schedule. In another condition, extinction was in effect during the second component (i.e., lever presses had no consequences). The session ended after the second component. Williams found that rats responded more slowly during the VI 1-min component when it was followed by a FR-1 component than when it was followed by an extinction component. The contrast in reinforcement density between the successive schedules was thought responsible for the result. When the VI schedule was followed by a richer schedule, rats responded more slowly than when it was followed by a leaner schedule.

The design of the present experiment parallels that of Williams’s (1992) experiment in some respects. For the open economy groups, rats responded during a first component (i.e., the work period) on relatively lean FRs. This was followed by a second component (i.e., the 3-h post-work period) where a richer FR-1 schedule was in effect. Rats in the closed economies experienced the relatively lean FRs in the first component, but they had no access to reinforcement in the second component. The economy type effect observed for saccharin could be a contrast effect, like that observed by Williams (1992) with food reinforcement. Why no effect was observed when heroin was the reinforcer is unknown. Previous behavioral contrast experiments have not used opioids as the reinforcer. Grove and Schuster (1974) performed a behavioral contrast experiment with cocaine as the reinforcer in monkeys and found weak, if any, evidence of a contrast effect. Perhaps contrast is less likely to occur with intravenous drug reinforcers than it is with food or saccharin reinforcers. Very little research relevant to this question has been performed to date.

Another possible explanation of the economy type effect is that opening the economy degrades the contingency between responses and reinforcers. Rescorla (1968) recognized the importance of contingency in Pavlovian conditioning when he showed that presentation of the unconditioned stimulus (US) without the conditioned stimulus (CS) reduced responding to the CS. The same outcome has been shown in operant conditioning where presentation of response-independent reinforcers reduce performance of the operant response (Colwill and Rescorla, 1986; Dickinson and Charnock, 1985; Hammond, 1980). In a closed economy where a subject earns all its food by lever pressing, there is a strong correlation, or contingency, between responses and food. In an open economy where, for example, the experimenter provides post-session supplemental food, there is a weaker correlation between responses and food. The weakening of the contingency between responses and reinforcers may explain why subjects work less hard in open economies. Another way of framing this account is in terms of the feedback relation between the subject’s total daily consumption of the reinforcer and its behavior (Hursh, 1978). Working harder produces more reinforcement in a closed economy, but does not necessarily do so in an open economy. This contingency account seems mostly applicable to experiments where the open economy extra reinforcers are given response-independently. In the present experiment, rats in the open economies had to respond to obtain their extra reinforcers, which meant that there was still a strong correlation between behavior and delivery of reinforcers. This was true for both heroin and saccharin and so it seems unlikely that this account could explain why an effect was observed for saccharin but not heroin.

A view advanced by Collier and Johnson (1990, 1997), related to optimal foraging theory, may be relevant for understanding economy type effects in animals. They argued that rats act to maximize profitability in their patch-exploiting behavior. According to this view, if a dense and a less dense patch are alternately available, the rat could minimize costs and maximize benefits by foraging mostly in the denser patch while rejecting opportunities to forage in the less dense patch. In contrast, if only the less dense patch is available, the rat should be more likely to accept foraging opportunities there. Johnson and Collier (1989) performed an experiment relevant to this hypothesis in an operant conditioning situation. They trained rats to first press a “search” lever which resulted in the illumination of a light over one of two “patch” levers, randomly selected. Rats could press the active patch lever to obtain food reinforcers, or wait at least 10 minutes for a chance to forage in the other patch. In one condition, Patch 1 was associated with a FR-10 schedule and Patch 2 was associated with a FR-40 schedule. In another condition, the two patches were both associated with the FR-40 schedule. Rats often rejected Patch 2 and waited for Patch 1 in the first condition. In contrast, rats accepted Patch 2 in the second condition since it had the same reinforcement density as Patch 1. Johnson and Collier argued that it was the relative profitability of the two patches that determined rats’ behavior. A subsequent study from Collier’s lab (Morato et al., 1995) showed that rats performed similarly when the alternating periods differing in relative profitability lasted 24 h or longer. The results of this later experiment suggest that under certain conditions, rats’ time horizons may be much longer than the short ones suggested by results of delay discounting studies.

This foraging account can be applied to the present experiment. The open economy groups received training similar to that of rats in Johnson and Collier (1989) in that they experienced alternating “patches” that differed in relative profitability. Rats in the open economies could increase overall profitability by forgoing the less dense patch – i.e., the work period – and waiting for the richer patch – i.e., the post-work-period of FR-1 availability. In contrast, the closed economy groups had no reason to reject the less dense patch and therefore they would be expected to work harder than the open economy groups during the work periods. Why this hypothesis was supported when saccharin was the reinforcer, but not when heroin was the reinforcer, is unclear. Collier’s work suggests that when food is the reinforcer, rats integrate costs and benefits over relatively long time frames. Rats maximized profitability in the long run by minimizing costs now and making up for lost food later. The present results suggest that similar factors control behavior when saccharin is the reinforcer. Perhaps the costs and benefits of intravenous heroin, under the conditions used here, are not integrated over time in the same way by rats. Timberlake et al. (1987) suggested that the time horizon over which current and future costs and benefits are integrated depends on the particular behavior system (among other variables) being studied. The difference between saccharin and heroin reported here is consistent with this suggestion.

The EV of heroin (about 2.5 here) was quite low relative that of saccharin (about 6 to 9 here), suggesting the possibility that a floor effect may have obscured differences between the heroin groups. There are reasons, however, why a floor effect is unlikely. EV reflects the slope of the demand curve, with lower EV indicating steeper slope. In looking at Figure 2, it is clear that a steeper slope would be possible for either heroin group. Nothing prevented rats from self-administering fewer infusions at the intermediate FRs, for example, which would have resulted in a steeper slope and lower EV. Further evidence suggesting that a floor effect was not responsible for the lack of a difference between the heroin groups comes from other studies that have found significance differences between groups or conditions with EVs even lower than 1 (e.g., Reed et al., 2016; Tucker et al., 2017; Yates et al., 2017).

It is possible that different results would have been observed here if there were a delay between the work period and the post-work availability period in the open economy groups. For the open economy groups here, the additional heroin or saccharin reinforcers were made available immediately after the work period. This is the most commonly used method for creating an open economy in those studies where an effect of economy type has been observed (for review, see Kearns, 2019). Further, previous studies with food reinforcement indicated that a stronger effect of opening the economy was observed when the extra reinforcers were available sooner after the work period rather than later (Hursh et al., 1989; Timberlake, 1984; Timberlake et al., 1987). Our goal here was to test for economy type effects under conditions most conducive to finding them.

It remains possible, however, that the ideal temporal parameters for observing an economy type effect differ for saccharin and heroin. It may be the case that the difference between heroin and saccharin in economy type effects is a quantitative one, rather than a qualitative one. That a current cheaper source of heroin should substitute for a current more expensive source of heroin seems to be an uncontroversial assumption. Perhaps approximating this situation by presenting the extra heroin infusions in the open economy closer in time to when the rats were working for heroin might have produced an economy type effect. Developing a procedure that accomplishes this without introducing potentially confounding variables could be challenging, but perhaps not impossible. For example, during work periods when rats lever press for heroin on a high FR, another lever could be presented occasionally and permit rats to obtain heroin on an FR-1 schedule. (Care would be needed in distinguishing satiety effects from substitution effects in such a concurrent access design.) Future research will be required to define the boundary temporal conditions under which an economy type effect for heroin might be observed.

Despite finding that opening the saccharin economy made demand for it more elastic, there was no effect of economy type on choice between heroin and saccharin. This may have happened because even though saccharin’s EV was reduced in the Saccharin Open group, it was still much higher than heroin’s EV. Saccharin’s EV was 5.6 in the Saccharin Open group, but only 2.4 and 2.8 in the Heroin Open and Heroin Closed groups, respectively. The magnitude of the difference in EV between saccharin and heroin may have made it difficult to produce a change in preference.

A limitation of the present study is that only a single heroin dose and single saccharin concentration were used. The 0.03 mg/kg/infusion heroin dose and the 0.2% saccharin concentration were chosen because those were used in previous heroin vs. saccharin choice studies (Lenoir et al., 2013; Madsen & Ahmed, 2015). This limitation may especially apply to the choice phase where rats in all groups showed strong preference for saccharin. If a weaker saccharin concentration or a higher heroin dose were used, heroin preference in the closed economy groups during the choice phase might have been closer to 50%, potentially making it easier for differences between these groups and the open economy groups to be observed. Interpretation of results from the demand phase may be less affected by the single dose and concentration limitation. Hursh and Silberberg’s (2008) exponential demand model was designed to isolate elasticity of demand while accounting for variables like reinforcer magnitude. When Hursh and Silberberg retrospectively applied their model to animals’ consumption of an opioid (alfentanil) or food, they found that the EV of each was independent of reinforcer magnitude. If heroin and saccharin EV are also independent of reinforcer magnitude, then using different doses or concentrations would not be expected to have led to different outcomes here. As this independence has not yet been established, the use of a single dose and concentration must be acknowledged as a limitation.

A related limitation is the use of single unit prices (responses/amount of reinforcer; DeGrandpre et al., 1993) for heroin and saccharin during the choice phase. Unit price can be varied by changing the number of responses per reinforcer or the magnitude of the reinforcer. To fully characterize heroin vs. saccharin preference, a complete function involving choice at multiple combinations of unit prices of the two reinforcers would be required. At higher unit prices of saccharin or lower unit prices or heroin, increased preference for heroin would be expected. Perhaps at different points along the choice function, an effect of economy type on preference would be observed. As the present study only used a single unit price of each reinforcer, conclusions regarding preference must be limited to the particular unit prices tested.

In conclusion, the present results add to others suggesting that economy type affects drug and non-drug reinforcers differently when tested under the same conditions. It is not yet clear why this may be the case. Differences in the way that current and future reinforcers are integrated could explain why economy type affects the value of some reinforcers but not others. Alternatively, differences in the way that exposure to the substance affects motivation for that substance could account for differences in the effect of economy type. For non-drug reinforcers like sucrose- or saccharin-sweetened water, additional exposure does not result in escalation of intake (Anker et al., 2010) or increased EV (Schwartz et al., 2017). For drug reinforcers like heroin and cocaine, increased access results in escalation of intake (e.g., Ahmed & Koob, 1998; Lenoir & Ahmed, 2008; Lenoir et al., 2013) and demand becomes less elastic (i.e., EV increases; Bentzley et al., 2014; Christensen et al., 2008; Lenoir & Ahmed, 2008; James et al., 2018; Schwartz et al., 2017). Because opening the economy results in increased exposure to the substance, the effect of economy type on reinforcer value may critically depend on whether increased exposure to the substance affects its value. Future research will be needed to decide between potential explanations of the apparent insensitivity of drug reinforcer value to economy type. The results of these studies could provide valuable information on a potentially important difference between drug and non-drug reinforcers that could help explain into how the reinforcer pathology that characterizes addiction develops and is maintained.

Public Health Significance:

This study investigated, within an animal model, factors that might contribute to opioid addiction. Results indicated that, unlike responding for non-drug rewards, current heroin-seeking behavior was not diminished by future availability of the reward. This difference between heroin and non-drug rewards could be relevant for understanding why addiction to opioids develops.