Cocaine Cues Retain Silent Traces of an Excitatory History After Conversion into Conditioned Inhibitors: “The Ghost in the Addict”

Abstract

The present experiment investigated the extent to which the A+/AB− conditioned inhibition procedure could counteract an excitatory drug-related conditioning history. In two groups of rats, a light stimulus was established as a signal for the absence of cocaine. For the History Group, the light had previously been a discriminative stimulus (S^D) that occasioned cocaine self-administration and could thus be classified as a cocaine excitor. In comparison, the No-History Group first encountered the light during conditioned inhibition training. During conditioned inhibition training, both groups self-administered cocaine during tone as well as in click S^Ds, while drug seeking was eliminated in click-plus-light wherein cocaine was not available (A+/AB−). Drug seeking was essentially eliminated in both groups. Nevertheless, on a summation test the light reduced cocaine seeking occasioned by the tone S^D by 95% in the No-History Group, but by less than 50% in the History Group. This summation test result showed that the effects of a drug-related history persisted even after the light was converted into an effective conditioned inhibitor on the training baseline through the powerful A+/AB− procedure. Future research should seek procedures that produce even stronger conditioned inhibition that eliminates such residual “silent” drug excitation, the “ghost in the addict” (Siegel, 2002, 2005).

Introduction

Kearns and colleagues (2005) were the first to demonstrate the effects that a conditioned inhibitor (CI) for cocaine can have on cocaine seeking. In that study, rats self-administered cocaine during presentation of tone and click discriminative stimuli (S^Ds). A light was occasionally presented simultaneously with the click and signaled the absence of cocaine. This training established the light as a CI within a currently excitatory context by the A+/AB− procedure (Rescorla, 1969). When the CI for cocaine (the light) was presented with the tone S^D on a summation test, drug seeking occasioned by the tone was reduced by over 90%.

These findings suggest that the A+/AB− procedure could turn neutral stimuli into conditioned inhibitors that almost totally suppress drug seeking. However, it is unknown whether the A+/AB− procedure could convert a former drug cue into a CI powerful enough eliminate drug seeking. Previous research with non-drug reinforcers has shown that it can be difficult to eliminate the residual effects of a cue’s excitatory history. For example, Weiss & Schindler (1985) found that a previously shock-based excitor could not be converted into an inhibitor even after extensive training on an A+/B− differential conditioning procedure that appeared to have eliminated all traces of shock-related excitation on the training baseline. Despite the fact that the B stimulus no longer occasioned avoidance, in training or on the summation test itself, this stimulus completely failed to reduce avoidance responding occasioned by a still-active shock excitor on the summation test. In contrast, when the B stimulus did not have an excitatory history, the same A+/B− differential conditioning procedure created an inhibitor that reduced avoidance responding occasioned by the shock-related excitatory S^D by 50% on a summation test. It deserves to be emphasized that these summation test differences were revealed even though the rats with an excitatory shock-related light history were behaviorally indistinguishable from those without that history prior to the summation test. Both groups responded to postpone shocks at a steady moderate rate in tone and did not respond during the light.

The results of Weiss and Schindler’s (1985) shock avoidance experiment are consistent with other studies showing that extinguished shock cues (Hendry, 1982; Reberg, 1972), food cues (Kearns & Weiss, 2005), or cocaine cues (Kearns & Weiss, 2012) retain residual excitation that is only revealed on a summation test. These studies demonstrate that even when simple extinction eliminates behavior, there can still be “behaviorally silent” residual excitation. Clearly, a procedure that is more effective than simple extinction or A+/B− differential conditioning is necessary if the residual effects of an excitatory conditioning history are to be successfully overcome.

There is evidence that the A+/AB− conditioned inhibition procedure is more effective than either differential conditioning or simple extinction in neutralizing conditioned excitation. For example, Kearns et al. (2005) found that a CI created with the A+/AB− procedure produced significantly more suppression of cocaine seeking than an inhibitor created by the A+/B− procedure. That the inhibitor (Stimulus B) in the A+/AB− procedure is created against the strong excitation provided by Stimulus A has been used to explain such an outcome (Wagner & Rescorla, 1972). In contrast, with the A+/B− procedure, the inhibitor is created against the weaker excitation conditioned to the context. Further support for the power of the A+/AB− procedure comes from a study demonstrating that subjecting an excitor to A+/AB− training, rather than simple extinction, can eliminate context renewal (Rauhut, Thomas, & Ayres, 2001).

The present experiment sought to determine the degree to which the A+/AB− conditioned inhibition procedure might be able to overcome a cocaine-related excitatory conditioning history. This was accomplished by training a group of rats to self-administer cocaine whenever a light stimulus was present but not when it was absent. In a subsequent phase, that light was made the B− condition during A+/AB− conditioned inhibition training for this cocaine History Group. A comparison No-History Group of rats received similar conditioned inhibition training, but, the light in this group did not have an excitatory cocaine history prior to being trained as an inhibitor with the A+/B− procedure. Finally, the light was compounded with a cocaine excitor on a summation test session to determine the degree to which it would suppress drug seeking in the History and No-History Groups.

Method

Subjects

Eight adult male Long-Evans rats served as subjects. They were individually housed in plastic cages with cedar chip bedding in a colony room that had a 12-hour light/dark cycle (lights on at 0800). Cocaine self-administration training sessions lasting approximately 3 h were conducted between 1000 and 1800 h. Body weights were maintained at 85% of free feeding weights (approximately 400 g) by feeding rats approximately 15–20 g of rat chow following their training sessions. Water was available continuously in the rats’ home cages. Experimental sessions were conducted 5–6 days per week.

Apparatus

Training took place in 6 operant chambers that were enclosed in sound attenuation chests described by Weiss, (1970). Each operant chamber was 20 cm high, 23 cm long, and 18 cm wide and had aluminum front and rear walls, white translucent plastic side walls, and a grid floor. A response lever and food trough were located on the front wall of the chamber. The tone (2000 Hz and 85 dB) and click (10 clicks/min and 75 dB) stimuli were delivered through a speaker mounted 21.5 cm above the chamber and inside of the sound attenuation chest. The light stimulus was provided by two 15-cm, 25-W, 120-V tubular light bulbs located 10 cm outside of the chamber’s side walls. The level of illumination provided by these light bulbs was approximately 130 cd/m when measured at the center of the side walls. The apparatus and stimuli are described in greater detail elsewhere (Weiss, 1969).

Cocaine (National Institute on Drug Abuse, Bethesda, MD) in saline solution at a concentration of 2.56 mg/ml was infused at a rate of 3.19 ml/min by 10-ml syringes driven by Harvard Apparatus or MED-Associates syringe pumps located outside of the sound attenuation chests. Tygon tubing extended from the 10-ml syringes to a 22-gauge rodent single-channel fluid swivel and tether apparatus (Alice King Chatham Medical Arts, Hawthorne, CA) that descended through the ceiling of the training chamber. Cocaine was delivered to the subject through Tygon tubing that passed through the metal spring of the tether apparatus. This metal spring was attached to a plastic screw attached to the rat’s head in order to reduce tension on the catheter.

Experimental events were controlled by a MED Associates (St. Albans, VT) computer system located in a room adjacent to the one where the training chambers were located. Cumulative recorders used to monitor rats’ ongoing behavior were also located in this room.

Procedure

Lever-press acquisition

To magazine train rats, food was initially presented on a variable-time (VT) 120-s schedule. In addition, lever presses produced food on a fixed-ratio (FR) 1 schedule. The VT schedule was discontinued once a rat emitted at least 8 responses. The click stimulus was on continuously during these sessions. As responding developed, the FR requirement was gradually increased over sessions to FR 10. Once a rat displayed regular responding on the FR 10 scheduled, a catheter was implanted.

Surgery

Surgery was performed under ketamine (60 mg/kg) and xylazine (10 mg/kg) anesthesia using procedures described in detail elsewhere (Weiss et al., 2003). A catheter consisting of approximately 8 cm of vinyl tubing (Dural Plastics, 0.5 mm ID, 1.0 mm OD) connected to 4 cm of Silastic tubing was inserted into a rat’s right jugular vein. The vinyl portion of the catheter was passed beneath the skin around the shoulder and exited the rat’s back at a point between the shoulder blades. Dental acrylic was used to cement a 20-mm plastic screw to 6 stainless steel jeweler’s screws that were implanted in the rat’s skull to form the head mount that was attached to the spring tether.

Cocaine self-administration acquisition

After recovering from surgery for a minimum of 5 days, cocaine self-administration training began. With the click stimulus on continuously, rats were trained on an FR 1 schedule where each lever press was now followed by a 0.16 mg cocaine infusion (approximately 0.5 mg/kg/infusion) instead of a food pellet. As regular responding developed, the FR ratio requirement was gradually increased to FR 10 and the dose per infusion was reduced to 0.08 mg (approximately 0.25 mg/kg/infusion). Once a rat regularly responded on this FR 10 schedule, discrimination training began. These, as well as all subsequent training sessions, lasted approximately 3 h.

Phase 1: Click/no-Click and Tone/no-Tone Discrimination Training-All Rats

Table 1 presents a schematic of the successive discrimination training phases of the experiment. First, all rats were trained on a click/no-click discrimination. Click components lasting 60 s on average (range: 30–120 s) alternated with click-off components also lasting 60 s on average (range: 40–90 s). During click components, cocaine was available on a variable interval (VI) 30-s (range: 1–79 s) schedule, while responses went unreinforced [extinction (EXT)] during click-off components. In addition, a 10-s response correction contingency was added to the end of click-off components to help reduce responding during these components. A lever press during the response correction period delayed the presentation of the next click component by 10 s. Over sessions, the VI value was increased to 45 s (range 2–119 s) and the value of the response correction contingency was increased for individual subjects up to a maximum of 60 s. The dose per infusion remained 0.08 mg cocaine (~0.25 mg/kg/infusion). Rats were trained on this mult VI 45-s EXT schedule until response rates in click components were at least 4 times faster than in click-off components. The VI-45-s schedule was used, rather than a ratio schedule, in an effort to maintain steady response rates and to ensure a regular delivery of cocaine infusions during S^D components. On a ratio schedule, a rat could respond very fast early in a component, obtain multiple infusions, and then not respond again for several minutes. An interval schedule prevents this from happening because infusions are only delivered when a response is made after a required period of time has elapsed that varies unsystematically over intervals.

Table 1.

Schematic of the training conditions for rats in the No-History and History Groups. For each group, those stimulus conditions where cocaine was available for self-administration are denoted “Cocaine”. Those components where no reinforcement was available (i.e., extinction) are denoted (EXT). Dashed lines (--) indicate that the No-History Group did not have Phase 2 Light Discrimination Training.

Group	Phase 1: Initial Discrimination Training		Phase 2: Light Discrimination Training	Phase 3: A+/AB− Conditioned Inhibition 4-Component Terminal Baseline
	Tone	Click	Light	Tone	Click	C+L	All-Off
No-History	Cocaine (VI 45-s)	Cocaine (VI 45-s)	--	Cocaine (VI 45-s)	Cocaine (VI 45-s)	EXT	EXT
History	Cocaine (VI 45-s)	Cocaine (VI 45-s)	Cocaine (VI 45-s)	Cocaine (VI 45-s)	Cocaine (VI 45-s)	EXT	EXT

Rats were next trained on a tone/no-tone discrimination procedure like that described above except that tone signaled VI components instead of the click. The parameters of this VI 45-s EXT schedule were the same as those described above. Rats were trained on this schedule for a minimum of 2 sessions and until response rates were at least 4 times faster in tone components than in tone-off components. Rats were alternately assigned to the History Group (n = 4) or the No-History Group (n = 4) when they satisfied this discrimination criterion.

Phase 2: Light/No-Light Discrimination Training for History Group

Rats in the History Group were now trained on a light/no-light discrimination training procedure wherein cocaine was available according to a VI 45-s schedule for lever pressing during light components, while lever presses went unreinforced (EXT) in light-off components. The parameters of this mult VI 45-s EXT schedule were the same as those described above for the click/no-click and tone/no-tone discrimination training procedures. Rats in the History Group were trained on this procedure for a minimum of 5 sessions and until response rates in light components were at least 5 responses / min, and at least 4 times faster than in light-off components, for 3 consecutive sessions. Rats in the No-History Group did not have light excitor training, but instead went from their tone/no-tone discrimination training directly to the 4-component A+/AB− conditioned inhibition terminal baseline described below.

Phase 3: Four-component A+/AB− conditioned inhibition terminal baseline-All Rats

Rats in both groups were trained on a 4-component multiple schedule that incorporated an A+/AB− conditioned inhibition procedure designed to make the light an inhibitor of cocaine seeking. This four-component mult VI VI EXT EXT schedule consisted of tone components, click components, and click+light (C+L) components that were always separated by an all-stimuli-off component. During the tone components and the click components, cocaine was available for lever pressing on the VI 45-s schedule. During C+L components and all-stimuli-off components, cocaine was not available. Thus, click signaled cocaine availability except when the light was presented simultaneously with the click. Tone, click, and C+L components were equally likely to follow each all-stimuli-off component, with the restriction that the same condition not occur more than four times consecutively.

Rats were trained on this procedure for a minimum of 5 sessions and until the following stimulus control criteria were met for 3 consecutive sessions (or 3 out of 4 consecutive sessions, with the final two sessions satisfying the criteria): 1) response rates in tone and in click were at least 4 times faster than in all-stimuli-off components, 2) response rate in click was at least 4 times faster than in C+L, and 3) response rates in tone and in click did not differ from each other by a ratio of more than 2:1.

Stimulus Compounding (Summation) Test

Once the criteria listed above were met, a test was administered to assay the degree to which the light would suppress cocaine seeking occasioned by the tone. Prior to the test, rats in both groups were given an approximately 1–2-hr warm-up period on the 4-component terminal baseline schedule. The purpose of the warm-up period was to ensure that rats were responding stably on the terminal baseline schedule at the time that the test began. The test consisted of twelve presentations of click, tone, and tone+light (TL). Presentation of each test stimulus condition lasted 60 s and was followed by a 60-s period where all stimuli were off. The test stimuli were presented in 12 blocks, with each block consisting of one presentation each of tone, click, and TL. The order of presentation varied across blocks such that the three types of test stimuli occurred an equal number of times in each ordinal position within a block and each stimulus condition followed every other condition an equal number of times. Cocaine self-administration was discontinued during the stimulus compounding test (i.e., the test was conducted in extinction).

Data analysis

For all statistical tests, α = 0.05. The primary dependent variable was response rate in T+L as a percentage of responding in tone alone (i.e., [T+L rate/tone rate] × 100) during the summation test, to reflect the capacity of the light to reduce (inhibit) drug seeking. Because a Levene test for equality of variances performed on this measure was significant (F[1,6] = 7.6, p < 0.05), non-parametric Mann-Whitney U tests were used to compare groups.

Results

Rats in the History and No-History Groups were comparable in terms of the total number of Phase 1 discrimination sessions (click/no-click and tone/no-tone) required to meet the discrimination criteria before being assigned to groups, with the History Group having a mean (±SME) of 12.5 (± 2.5) sessions and the No-History Group having a mean of 13.5 (± 1.7) sessions. The History Group had 8.0 (± 2.0) light/no-light discrimination sessions in Phase 2. The History and No-History Groups were also comparable in the number of sessions on the 4-component conditioned inhibition procedure in Phase 3 prior to the summation test. These were 14.3 (± 3.2) and 13.5 (± 1.8) sessions, respectively.

Figure 1 presents mean (+ SEM) response rates (responses/min) in tone, click, C+L, and the all-stimuli-off components for both groups, averaged over the 3 criterion sessions prior to the summation test. In addition, light response rate from the final light/no-light discrimination session is presented for the History Group. (The No-History Group did not experience the light during this phase.) On the terminal baseline, both groups had mean response rates of 9 or more responses/min in those components were cocaine was available (tone and click). Mean response rates in those components where cocaine was not available (i.e., C+L and all-stimuli-off components) were 2.1 responses/min or less. Mann Whitney U tests indicated that there were no significant differences between the groups in response rate to any stimulus on the terminal baseline (all U[4,4]’s ≥ 4, p’s > 0.34).

Fig. 1.

Mean (+SEM) response rates (responses/min) in tone, click, C+L, and all-stimuli-off averaged over the last 3 training sessions prior to testing for the No-History Group and the History Group. The mean response rate in light during the final light excitor session is also presented for the History Group.

Though not significantly different across groups, the mean response rates in tone and in click were elevated in the No-History group, primarily because of one rat with response rates that were more than double those of any other rat. Response rates in click during the criterion Phase 3 sessions for rats in the History (H) and No-History (NH) groups ranged from 7.4 to 29.6 responses/min and were distributed as follows: 7.4 (NH), 8.8 (H), 9.9 (H), 10.1 (H), 11.5 (NH). 12.4, (H), 14.4 (NH), and 29.6 (NH). Similarly, response rates in tone ranged from 6.8 to 25.1 responses/min and were distributed as follows: 6.8 (NH), 7.8 (H), 8.7 (H), 9.4 (H), 10.0 (H), 10.4 (NH), 12.1 (NH), and 25.1 (NH). This analysis of individual subjects’ response rates provides further evidence that the groups did not systematically differ.

Figure 2 presents results from the summation test. Panel A shows the response rate during T+L expressed as a percentage of response rate in tone alone. The light suppressed responding occasioned by the tone almost completely (95%) in the No-History Group, but only by 47% in the History Group, a statistically significant difference (U = 0, p < 0.05). The right-hand portion of Figure 2 (panel B) presents mean response rates (responses/min) during tone, click, T+L, and all-stimuli-off block-randomized components on the summation test for inhibition. As during training, the No-History group had a mean response rate that appeared higher than that of the History group. However, this difference again was not significant (U = 4, p > 0.34).

Fig. 2.

Panel A presents mean (+SEM) responding in T+L expressed as a percentage of tone responding during the test ([T+L rate/Tone rate]*100) for the No-History and History Groups. Panel B presents mean (+SEM) response rates (responses/min) during tone, click, TL, and all-stimuli-off components on the block-randomized summation test for inhibition.

Discussion

The present experiment demonstrated that a CI that was previously a cocaine excitor was less effective in suppressing cocaine seeking on a summation test than an inhibitor that had no excitatory history. Nonetheless, the light in the History Group still reduced cocaine seeking by approximately 50%. This suggests that the A+/AB− procedure used here created an inhibitor that was able to at least partially counteract a still-active excitor, which is something that the inhibitor created with the A+/B− procedure used by Weiss & Schindler (1985) did not do. That would be anticipated from Rescorla (1969) since in the present experiment the conditioned inhibitor was created within the excitatory context of a reinforcement-associated S^D while in Weiss and Schindler’s A+/B− simple differential conditioning procedure it was not. The suppression produced by the light in the History Group of the present study is especially impressive because if it had not been “treated”, presenting it simultaneously with the excitatory tone would have increased cocaine seeking by 300% (Panlilio, Weiss, & Schindler, 1996).

Further comparison to Weiss and Schindler (1985) conditioning history study performed with the same apparatus and stimuli as the present systematic extension could be informative. The stimulus treated with just simple extinction in Weiss and Schindler’s (1985) No-History Group was only able to reduce avoidance by 50%. In comparison, the A+/AB− produced conditioned inhibitor in the No-History Group of the present experiment reduced drug seeking by almost 50% more, 95%. Likewise, the stimulus treated with just simple extinction in Weiss and Schindler’s History Group did not reduce avoidance at all on a summation, while the A+/AB− produced conditioned inhibitor in the History Group of the present experiment reduced drug seeking by almost 50%. That additional 50% reduction for both groups in the present experiment, compared to that of Weiss and Schindler (1985) is noteworthy and hopefully might stimulate additional research. Of course this comparison is only suggestive, but it should be viewed in the context where compounding stimuli essentially tripled rate when responding to the tone and to the light that were compounded was maintained by food (Weiss, 1971, Experiment 2), shock avoidance (Emurian & Weiss, 1972), water (Weiss, Schindler & Eason, 1988), cocaine (Panlilio, Weiss, & Schindler, 1996) or heroin (Panlilio, Weiss, & Schindler, 2000).

The 95% suppression observed in the No-History Group of the present experiment replicates the results of Kearns et al.’s (2005) A+/AB− Conditioned-Inhibition Group that used the same stimuli and training chambers as the present experiment. In that group, a light without an excitatory history was made into a signal for the unavailability of cocaine with the A+/AB− procedure used in the present experiment for the No-History Group. (Compare the white bars to the left in Figures 2 and 3.) It is worth noting that the Kearns et al. study included a group for which the light was presented non-differentially with respect to cocaine availability. In that Control Group the light had no effect on cocaine seeking during a summation test (see Figure 3, right bar). Therefore, there is no reason to doubt that the suppression produced by the A+/AB− conditioned inhibition procedure of the present study was due to conditioned inhibition and rather than to non-associative influences of the light (e.g., external inhibition).

Fig 3.

Results from the A+/AB− Conditioned Inhibition and Non-Differential Control Groups of Kearns, Weiss, Schindler, & Panlilio (2005). Data presented are mean (+SEM) response rates in T+L expressed as a percentage of tone responding during the summation test ([T+L rate/Tone rate]*100). The training and summation test results of this A+/AB− Conditioned Inhibition Group are comparable to the No-History Group of the present experiment presented in Figure 1.

It may appear from Figures 1 and 2 that response rates in the S^Ds were higher in the No-History group than in the History group. However, as noted in the Results section, ranking of individual subjects’ criterion baseline response rates over groups, plus statistical tests, indicated that these rates did not reliably or systematically differ across groups. Furthermore, even if tone and click response rates in the No-History group were higher than in the History group, this would only bolster confidence in the main conclusion -- that the light was a more effective conditioned inhibitor in the No-History group. In the No-History Group it reduced a higher response rate by 95% compared to only a 50% reduction for the lower response rate of the History group. A higher response rate in the tone in the No-History group should have made it more difficult, not easier, for the light to produce an almost total 95% degree of suppression on the summation test. So, with the light producing 45% less reduction in the History Group, whose drug seeking could have been weaker than that of the No-History Group, strengthens that a drug history impeded the light becoming a highly effective inhibitor even via the A+/AB− procedure.

The present results are consistent with previous studies demonstrating that residual excitation can be revealed by compounding stimuli that have been subjected to extensive extinction and no longer control behavior when presented individually (Hendry, 1985; Kearns & Weiss, 2005, 2012; Reberg, 1972; Weiss & Schindler, 1985). These studies can be thought of as returning the associative status of the excitor to zero by the end of simple extinction training. The present experiment, by converting the light into an inhibitor through the A+/AB− procedure, can be thought of as reducing the associative status of the light below zero because it acted to suppress the drug seeking occasioned by the excitatory click during training. The results of the summation test show that such a stimulus can still retrain traces of an excitatory history that are only revealed when presented in compound with another excitor.

An occasion-setting or context renewal account (Bouton, 1993, 2002) describes an alternative mechanism that could potentially explain the results of the present experiment. Rats in the History Group first learned that light, when presented alone, signaled cocaine. Then, during the conditioned inhibition phase, they learned that light did not signal cocaine when the light was presented simultaneously with the click. Thus, these rats learned two associations: light-cocaine, and light-no cocaine. The click served as a contextual cue, or occasion setter, that signaled when the light would not be paired with cocaine. On the test, the light was presented simultaneously with the tone and not the click. Because the second association learned (light-no cocaine here) is context-dependent (Nelson, 2002), it might be expected that the light-cocaine association would have been retrieved during the test. However, the light-no cocaine association would be retrieved if there were generalization between the tone and the click. The relative strength of each association could also determine which of the two would be more strongly retrieved (Miller & Laborda, 2011). The retrieval of both associations could potentially explain why the light suppressed responding in the History group, but to a lesser extent than in the No History group for which the light would have only retrieved a light-no cocaine association.

The approximately 0.25 mg/kg cocaine dose used here was chosen to generate a steady and moderate rate in S^D components without long post-infusion pauses. Previous studies have shown that rats regulate their cocaine intake by varying the length of post-infusion pauses (Lynch et al., 1998). If a higher dose had been used, generally lower rates of responding than those observed here would be expected. A high dose could have interfered with the development of stimulus control if, for example, a rat responded at a very low rate and therefore experienced relatively few infusions during S^D components. A lower dose would likely have produced higher rates. However, a risk with choosing a lower dose is that it could have been near or below the threshold dose needed for cocaine to be reinforcing under the conditions of the multiple schedules used here. Generally larger doses are needed to maintain responding when there are relatively long periods of drug unavailability on a schedule (Caine & Koob, 1994).

The phenomenon reported here highlights the challenges confronted when attempting to reduce the power of drug cues over drug seeking. Like reinstatement (Rescorla & Heth, 1975), renewal (Bouton & Bolles, 1979), spontaneous recovery (e.g., Rescorla, 1997), and rapid reacquisition (e.g., Weidemann & Kehoe, 2003), the effect described here demonstrates that excitatory associations are not erased by treatments that may appear to eliminate the control of such associations over behavior. Recognizing the persistence of such drug-related associations, even when evidence of their existence is not readily apparent in ongoing baseline behavior, is essential for the development of effective methods of reducing the control that drug-associated cues have over behavior. In the present experiment, this behaviorally silent excitation was only revealed on a summation test. Siegel (2002, 2005) has suggested these learned associations that energize the incentive-motivation for drugs can be likened to a “ghost in the addict” that returns periodically to “haunt” the individual even long after the last time the drug was used. If we are to get rid of the “ghost’, future studies investigating methods for neutralizing drug cues should also employ summation tests to detect potentially hidden residual excitation.