Antipunishment effects of acute and repeated administration of phencyclidine and NPC 12626 in rats.

The effects of phencyclidine (PCP) and NPC 12626 on punished responding were examined using a modified Geller-Seifter procedure in rats. Both drugs are known to antagonize N-methyl-D-aspartate (NMDA) receptor mediated neurotransmission, albeit at different sites on the NMDA receptor complex. Rats were trained to lever press for food reinforcement under a multiple schedule, with responding in one component reinforced under a fixed-interval 60-sec schedule, while each response in the other component resulted in both food and brief electric shock. Both PCP and NPC 12626 produced selective increases in punished responding, although the effects were not as large as those produced by chlordiazepoxide. Repeated daily administration of each of these drugs for 6 days resulted in increases in punished responding during different portions of the treatment. A 5 mg/kg dose of chlordiazepoxide produced increases over the last 2 days of administration. PCP (2 mg/kg) produced an increase only during the second session, whereas NPC 12626 (30 mg/kg) produced increases for all but the first and fifth days of the 6-day regimen. Both competitive and noncompetitive NMDA antagonists can have antipunishment effects in this model.     

Discriminative stimulus properties of oxazepam in the pigeon

Five pigeons were trained to discriminate IM injections of oxazepam (4.0 mg/kg) from vehicle with responding maintained under a fixed-ratio 30 schedule of food delivery. Under test conditions, responding increased in a dose-dependent manner in all pigeons after the administration of other benzodiazepines including diazepam (0.01-1.0 mg/kg), temazepam (0.01-3.0 mg/kg), halazepam (0.1-56.0 mg/kg), and midazolam (0.1-1.0 mg/kg) as well as the barbiturate pentobarbital (2.0-8.0 mg/kg) and the non-benzodiazepine anxiolytic CL 218,872 (1.0-8.0 mg/kg). At the higher doses of each of these compounds, over 80% of responding occurred on the oxazepam-appropriate key. Cocaine (0.5-4.0 mg/kg), bupropion (3.0-56.0 mg/kg) and nortriptyline (3.0-56.0 mg/kg) failed to substitute for oxazepam even at doses that decreased rates of responding. The discriminative stimulus (DS) effects of the lowest doses of oxazepam and CL 218,872 that produced 100% drug-appropriate responding were blocked by the benzodiazepine antagonist Ro 15-1788. This antagonism was reversed by increasing the dose of the agonists. The DS effects of diazepam were antagonized partially by Ro 15-1788 (3 of 5 pigeons), and the antagonism was reversed by higher doses of diazepam in two of these pigeons. The DS effects of pentobarbital were antagonized by Ro 15-1788 in 2 of 5 pigeons, but the blockade was not reversed by higher pentobarbital doses.     

Decreases in nestlet shredding of mice by serotonin uptake inhibitors: comparison with marble burying

Methods for detection of anxiolytic-like behavioral effects of serotonin uptake inhibitors are limited. The present study introduces a new quantitative method that permits dose-effect analysis of these compounds. Male NIH Swiss mice were given 60-min access to a piece of cotton gauze and the amount of material not torn into nesting material was weighed. Other groups of mice were individually placed in containers with 20 marbles resting on top of sawdust bedding. The number of marbles buried (2/3) by sawdust after 30 min was counted. Mice were first placed on a 6-rpm rotorod and the number of mice falling off twice in 2 min was measured. Serotonin uptake inhibitors (clomipramine, citalopram, fluoxetine, and venlafaxine) dose-dependently suppressed nestlet shredding and marble burying at doses that were generally without effect on rotorod performance. The amine-based antidepressant agents imipramine and desipramine as well as the selective norepinephrine transport inhibitor nisoxetine produced similar qualitative effects on these behaviors. Anxiolytics (chlordiazepoxide, bretazenil, buspirone, and pentobarbital) produced effects in the nestlet assay that were similar to those reported using another anxiolytic assay in mice (punished responding), whereas these compounds were not active at non-motor-impairing doses in the marble burying assay. The antipsychotic agents chlorpromazine and risperidone generally demonstrated suppression of nestlet shredding and marble burying at doses that impaired rotorod performance. Although d-amphetamine suppressed nestlet shredding and marble burying at doses without effect on the rotorod, d-amphetamine but not fluoxetine stimulated locomotor activity. Both nestlet shredding and marble burying behaviors were generally consistent across five consecutive experimental sessions and fluoxetine did not produce any systematic trends over repeated testing. The methods should have utility in defining pharmacological effects of these compounds in vivo. Moreover, these data may be useful in the context of other behavioral effects when assessing the relevance of a compound for its potential therapeutic potential as an anxiolytic (nestlet shredding) or as an anti-obsessive-compulsive disorder agent (marble burying).     

Behavioral effects and benzodiazepine antagonist activity of Ro 15-1788 (flumazepil) in pigeons

The selective benzodiazepine receptor antagonist, Ro 15-1788, produced behavioral effects in pigeons at doses at least 100 times lower than those previously reported to possess intrinsic pharmacological activity in mammals. In contrast to its effects in mammalian species, in pigeons, Ro 15-1788 does not exhibit partial agonist activity. Key-peck responses of pigeons were studied under a multiple fixed-interval 3-min, fixed-interval 3-min schedule in which the first response after 3-min produced food in the presence of red or white keylights. In addition, every 30th response during the red keylight produced a brief electric shock (punishment). Under control conditions, punished responding was suppressed to 30% of unpunished response levels. Ro 15-1788 (0.01 mg/kg, i.m.) increased unpunished response rates by 33% without affecting rates of punished responding. Doses of 0.1 to 1.0 mg/kg Ro 15-1788 produced dose-related decreases in both punished and unpunished responding. As is characteristic of other benzodiazepines, midazolam (0.1 and 0.3 mg/kg, i.m.) markedly increased punished responding but had little effect on rates of unpunished responding. Ro 15-1788 antagonized the increases in punished responding and also reversed the rate-decreasing effects of higher doses of midazolam. However, the effectiveness of Ro 15-1788 as a benzodiazepine antagonist was limited by its intrinsic activity: rate-decreasing doses of Ro 15-1788 were unable to completely reverse behavioral effects of midazolam. Midazolam was an effective antagonist of the behavioral effects of Ro 15-1788 (up to 0.1 mg/kg) but midazolam did not influence the rate-decreasing effects of 1.0 mg/kg Ro 15-1788 across a 100-fold dose range. In the pigeon, the behavioral effects of relatively low doses of Ro 15-1788 (0.01-0.1 mg/kg) appear to be related to benzodiazepine receptor mechanisms, whereas other systems appear to be involved in the effects of higher doses.     

Opposite effects of two ligands for peripheral type benzodiazepine binding sites, PK 11195 and RO5-4864, in a conflict situation in the rat

The effects of two drugs acting at the peripheral type benzodiazepine binding sites, PK 11195 and RO5-4864, were examined in shock-induced suppression of drinking in rats. These two compounds have opposite effects : RO5-4864 (3.1-1205 mg/kg i.p.) enhanced whereas PK 11195 (25-50 mg/kg i.p.) decreased the punished responding, and PK 11195 (6.25 mg/kg, a dose which did not alter the punished responding) blocked the proconflict action of RO5-4864 (6.25 and 12.5 mg/kg). The effects of RO5-4864 and PK 11195 were not antagonized by RO15-1788, a selective antagonist of the central benzodiazepine site. In addition, PK 11195 (6.25 mg/kg) did not reverse the proconflict effect of two beta-carbolines : beta-CEE and FG 7142. AS picrotoxin did not change the punished responding, these data imply that the effects of RO5-4864 and PK 11195 on the one hand and those of chlordiazepoxide and beta-carbolines on the other hand are differentially mediated and suggest that the peripheral type benzodiazepine binding sites are involved in this conflict model.     

The benzodiazepine receptor inverse agonists beta-CCM and RO 15-3505 both reverse the anxiolytic effects of ethanol in mice

The antagonistic effects of the benzodiazepine receptor inverse agonist beta-CCM (1 mg/kg) and of the partial inverse agonist RO 15-3505 (3 mg/kg) on the anxiolytic properties of ethanol (1 g/kg) in mice confronted with a light/dark choice procedure and with the staircase test were investigated. Both drugs reversed the effects of ethanol on some of the behavioral parameters, but beta-CCM alone elicited anxiogenic intrinsic effects. RO 15-3505 induced seizures in mice treated with a subconvulsant dose of pentylenetetrazole, the most efficient doses being 3 and 6 mg/kg. These data indicate that beta-CCM and RO 15-3505 can reverse some of the anxiolytic effects of ethanol, acting probably to oppose GABA function via the benzodiazepine receptor.     

Effects of chlordiazepoxide and beta-carboline 3-carboxylic acid ethyl ester on non-suppressed and minimally-suppressed responding in the squirrel monkey.

Lever-pressing of squirrel monkeys was maintained under a multiple fixed-interval (FI) 5-min schedule of food presentation. In one component, responding was suppressed to various degrees by the presentation of electric shock following each 30th response. When responding was either substantially or minimally suppressed, intermediate doses of chlordiazepoxide (CDAP, 1-30 mg/kg) increased both suppressed and non-suppressed responding. Beta-carboline 3-carboxylic acid ethyl ester (beta-CCE, 0.1-3 mg/kg) had little effect at low to intermediate doses (0.1-0.3 mg/kg) and decreased both minimally-suppressed and non-suppressed responding to a comparable extent at higher doses. Repeated daily dosing with beta-CCE (up to 10 mg/kg) resulted in rapid tolerance to its rate-decreasing effects. As agonists do not typically exhibit rapid tolerance for anxiolytic efficacy, the current results suggest that some behavioral effects of inverse agonists may not be strictly opposite those of benzodiazepines.     

Do benzodiazepine receptors mediate the anticonflict action of pentobarbital?

The effects of the benzodiazepine antagonist CGS 8216 (2-phenylpyrazolo[4,3-c]quinoline-3(5H)-one) were examined in a thirsty rat conflict test in the presence and absence of pentobarbital. CGS 8216 (2.5-10 mg/kg i.p.) did not affect nonpunished responding, but doses of 5 and 10 mg/kg significantly reduced the rate of punished responding (i.e., the number of 3 second drinking episodes in a "shock" contingency). However, a dose of CGS 8216 which did not significantly alter punished responding (2.5 mg/kg) antagonized the anticonflict actions of pentobarbital. These observations suggest that while high doses of CGS 8216 may elicit an "anxiogenic" response in rodents, lower doses of CGS 8216 antagonize the anticonflict actions of a compound which has been shown to enhance benzodiazepine affinity in vitro. These data imply that the anticonflict actions of pentobarbital may be mediated through benzodiazepine receptors.     

Effects of thyrotropin-releasing hormone (TRH) and MK-771 on schedule-controlled behavior of squirrel monkeys, rabbits and pigeons

The effects of TRH (0.001-10.0 mg/kg) and a more potent TRH analog, MK-771 (0.001-5.6 mg/kg), were studied on comparable schedule-controlled performances of squirrel monkeys, rabbits and pigeons. Responding was maintained in the presence of different stimuli by a multiple fixed-ratio (FR), fixed-interval (FI) schedule of food presentation (monkeys and pigeons) or 0.25% saccharin solution (rabbits). Generally, TRH and MK-771 produced decreases in responding under both schedules and in all three species. TRH and MK-771 were roughly equipotent in the squirrel monkey, whereas in the pigeon and rabbit MK-771 was approximately 20 times more potent than TRH in decreasing responding to 50 percent of control levels. The duration of action of doses of TRH and MK-771 that reduced responding to 50 percent of control was approximately 3 hr in the squirrel monkey; recovery of performance occurred twice as fast under the FR schedules. With the pigeon, TRH effects that produced 50 percent decreases in responding lasted over 6 hours, whereas behaviorally comparable doses of MK-771 lasted about 4 hours. With few exceptions, TRH and MK-771 appear to produce similar effects of schedule-controlled behavioral performances of the squirrel monkey, rabbit and pigeon. Compared to the effects of other behaviorally-active substances under these procedures, TRH and MK-771 exert a distinctive array of effects.     

Effects of ethanol on punished and nonpunished responding under conditions of equated reinforcement rates and similar response rates

Lever pressing by pairs of rats was maintained under random-ratio (first subject) and yoked-interval (second subject) schedules of food presentation. The inter-reinforcement intervals generated under the ratio schedule comprised the interval values for the second (yoked) subject. This arrangement yielded nearly equivalent rates of food presentation for each subject pair. For the first rat of each pair a random-ratio schedule of shock presentation was added to the ratio schedule of food presentation. This manipulation resulted in similar rates of punished (first rat) and nonpunished (second rat) responding within subject pairs. Ethanol administration (0.25–1.5 g/kg) generally resulted in dose-related decreases in both punished and nonpunished responding. In general, punishment-specific effects were not obtained. These results suggest that ethanol may not be as effective as chlordiazepoxide or pentobarbital in increasing punished responding even when the effects of baseline response and reinforcement rates are controlled.     

Effects of thyrotropin-releasing hormone (TRH) and MK-771 on behavior of squirrel monkeys controlled by noxious stimuli

The effects of TRH (0.1-30 mg/kg) and an enzyme-resistant analogue, MK-771 (0.1-10 mg/kg), were characterized in squirrel monkeys on responding maintained in the presence of different visual stimuli by a multiple 3-min fixed-interval (FI), 30-response fixed-ratio (FR) schedule of stimulus-shock termination or by a multiple 5-min FI schedule of food or shock presentation. Under the termination schedule, the first response at the end of 3 min in the FI component or the completion of the 30-response requirement in the FR component terminated the visual stimulus in the presence of which shocks occurred (escape schedule). Under the schedule of food or shock presentation, the first response at the end of the 5-min FI produced food in the presence of red stimulus lights or shock in the presence of white lights. TRH and MK-771 produced large, dose-related increases in responding maintained under the FR stimulus-shock termination schedule whereas these peptides produced smaller increases or did not affect responding under the FI schedule. TRH and MK-771 also produced marked increases in responding maintained by shock presentation at doses that did not alter or decreased food-maintained responding in the same subject. Thus, performances maintained by noxious stimuli are uniquely sensitive to the rate-increasing effects of TRH and MK-771. These findings suggest that the behavioral effects of the neuropeptides, TRH and MK-771, can depend on the specific consequences of behavior and, as such, the effects of these substances are determined by many of the same variables that determine the effects of other behaviorally-active drugs.     

Interactions between the benzodiazepine receptor antagonist Ro 15-1788 (flumazepil) and the inverse agonist beta-CCE: behavioral studies with squirrel monkeys

The effects of Ro 15-1788 and ethyl-beta-carboline-3-carboxylate (beta-CCE) were studied alone and in combination on the behavioral performances of squirrel monkeys. Under one procedure, performances maintained by food were suppressed by electric shock presentation (punishment or "conflict" procedure). Under a second procedure, responding was maintained either by food or electric shock delivery under a 5-min fixed-interval schedule. Doses of beta-CCE between 0.1 and 3.0 mg/kg, i.m., produced graded decreases in punished responding which were reversed by pretreatment with Ro 15-1788 (1.0 - 10.0 mg/kg, i.m.). Low doses of beta-CCE (0.03 - 0.3 mg/kg, i.m.) increased responding of monkeys maintained by shock presentation, but did not affect food-maintained responding; higher doses of beta-CCE decreased responding under both schedules. These effects of beta-CCE are opposite those produced by the benzodiazepines under this procedure. Ro 15-1788 (1.0 mg/kg i.m.) antagonized the effects of beta-CCE, producing a shift to the right in the dose-response curves. These findings provide further support for the view that beta-CCE and Ro 15-1788 produce effects mediated by the same benzodiazepine receptor recognition site.     

Ro 15-1788 antagonizes the discriminative stimulus effects of diazepam in rats but not similar effects of pentobarbital

The benzodiazepine antagonist properties of Ro 15-1788 were evaluated in rats trained to discriminate between saline and either 1.0 mg/kg of diazepam or 10 mg/kg of pentobarbital in a two-choice discrete-trial shock avoidance procedure. When administered alone, 1.0 mg/kg of diazepam and 10 mg/kg of pentobarbital produced comparable amounts of drug-appropriate responding (> 84%), whether rats were trained to discriminate between diazepam or pentobarbital and saline. Ro 15-1788 (3–32 mg/kg, p.o.), administered 10 min before diazepam or pentobarbital, produced a dose-related blockade of the discriminative effects of diazepam in both groups of rats, but was completely ineffective in blocking the discriminative effects of pentobarbital. The dose-effect curve for the discriminative effects of diazepam was shifted to the right in a parallel fashion 3- and 13-fold by 10 and 32 mg/kg of Ro 15-1788, respectively, indicating that Ro 15-1788 acts as a surmountable, competitive antagonist of diazepam. When administered alone, Ro 15-1788 (32–100 mg/kg, p.o.) produced primarily saline-appropriate responding, although 100 mg/kg of Ro 15-1788 produced drug-appropriate responding in one out of eight rats. When administered orally 30 min after diazepam, Ro 15-1788 (32 mg/kg) completely reversed within 10 min the discriminative effects of diazepam. The blockade of diazepam's discriminative effects by 32 mg/kg of Ro 15-1788 appeared to last at least as long (approximately 2 hr) as the effects of diazepam alone.     

Behavioral effects of benzodiazepine antagonists in chlordiazepoxide tolerant and non-tolerant rats

Rats were trained to respond under 3-min fixed-interval schedules of food presentation, and effects of the benzodiazepine-receptor ligands, flumazenil, 2-(4-methoxy-phenyl)-pyrazolo[4,3-c]quinolin-3(5H)-one (CGS 9895), 3-carbo-t-butoxy-beta-carboline (beta-CCtB), and beta-carboline-3-carboxylic acid ethyl ester (beta-CCE) were assessed before and after the induction of tolerance to chlordiazepoxide. Before daily administration of chlordiazepoxide, none of the antagonists produced appreciable effects on rates of responding up to doses of 32.0 mg/kg i.p. beta-CCE was the only antagonist studied at a higher dose (100.0 mg/kg i.p.), which decreased response rates. After 23 days of daily chlordiazepoxide administration (oral doses started at 10 and increased to 100 mg/kg/day by the 17th day), dose-effect curves for chlordiazepoxide were shifted to the right by about one-half log unit. Subjects were also more sensitive to the flumazenil, CGS 9895, and beta-CCtB, however, since these drugs produced only small effects in non-tolerant subjects, precise estimates of the degree of the shift in dose-effect curves could not be estimated. However, there were differences in the changes in the dose-effect curves induced by chlordiazepoxide tolerance. These results suggest differences in mechanism of action of antagonists in tolerant and non-tolerant subjects, and further that the sensitivity that is induced to antagonists in tolerant subjects is not conferred equally to all drugs having benzodiazepine antagonist activity.     

Neuropeptide Y blocks anxiogenic-like behavioral action of corticotropin-releasing factor in an operant conflict test and elevated plus maze

Central administration of neuropeptide Y (NPY) produces anxiolytic-like behavioral effects in rat models of anxiety. Because previous evidence has suggested a relationship between NPY and corticotropin-releasing factor (CRF) in the brain, we have focused on the interaction of these neuropeptide systems in emotional responsiveness to stressful stimuli. Intracerebroventricular administration of CRF produced a marked response suppression in an operant incremental shock conflict paradigm. NPY [(1 microg, intracerebroventricularly (i.c.v.)] significantly antagonized the response-suppressing effects of CRF (0.75 microg, i.c.v.) on punished responding in the conflict test at doses that produced little or no behavioral effect when administered alone. Central administration of the CRF antagonist [D-Phe(12), Nle(21,38),C(alpha) MeLeu(37)]CRF (D-Phe CRF(12-41)) alone did not alter punished or unpunished responding in the conflict test. However, pretreatment with the CRF antagonist before a subthreshold dose of NPY (1 microg, i.c.v.) produced a significant potentiation of the release of punished responding relative to NPY alone and untreated controls. NPY also antagonized the "anxiogenic-like" behavioral effects of CRF in the elevated plus maze. These findings support the hypothesis that NPY and CRF may reciprocally modulate an animal's behavioral response to stressful stimuli.     

Naloxone blocks 'anxiolytic' effects of neuropeptide Y

Intracerebroventricular injection of neuropeptide Y (NPY) produces potent 'anxiolytic' effects in animal models of anxiety. Administration of opioid receptor antagonists suppresses NPY-induced food intake and thermogenesis. The present study examined whether the opiate antagonist naloxone would also suppress the 'anxiolytic' effects of neuropeptide Y. Following training and stabilization of responding in an operant conflict model of anxiety, rats were injected with either NPY or diazepam. Both NPY (veh., 2, 4, 6 microg, i.c.v.) and chlordiazepoxide (veh., 2, 4, 6 mg/kg, i.p.) produced a dose-dependent increase in punished responding in the conflict test. The 'anxiolytic' effects of NPY were not blocked by the administration of flumazenil (3, 6, 12 mg/kg, i.p.). The administration of naloxone (0.25-2.0 mg/kg, s.c.) antagonized the effects of NPY. Central administration of the selective mu opiate antagonist CTAP (1 microg, i.c.v.) partially blocked NPY-induced conflict responding. These results support the hypothesis that NPY may play an important role in experimental anxiety independent of the benzodiazepine receptor and further implicate the opioid system in the behavioral expression of anxiety.     

Caffeine-phenylethylamine combinations mimic the amphetamine discriminative cue

Although caffeine-phenylethylamine combinations are widely available as over-the-counter medications or as "legal" stimulants, little information is available concerning their behavioral pharmacology or abuse potential. In the present study, rats were trained in a food-reward, two-lever operant drug discrimination paradigm to differentially respond after saline or 0.5 mg/kg amphetamine injections. Tests for generalization to the amphetamine cue indicated only modest amphetamine-lever responding at various doses of caffeine alone or at various doses of ephedrine/phenylpropanolamine (PPA) combinations, but complete generalization to the training cue was found with higher doses of the triple combination (caffeine, ephedrine, and PPA) or with caffeine-ephedrine or caffeine-PPA combinations. All drugs produced response rate decreases at higher doses. These data clearly indicate that certain "legal" stimulants mimic the amphetamine cue and suggest that caffeine may interact additively with phenylethylamines to produce the cue.     

Evaluation of thyrotropin-releasing hormone as a potential antidepressant agent in the conscious dog

Results from preliminary clinical reports have indicated that thyrotropin-releasing hormone (TRH) produces improvement in depressed patients. In the present study, doses of TRH at least 25 times greater than those reported as clinically effective on a mg/kg basis were evaluated for antidepressant activity in the conscious dog. Clinically effective antidepressants, such as MAO inhibitors and tricyclics like imipramine, potentiate certain behavioral, autonomic, and cardiovascular responses produced by the indole alkaloid yohimbine, whereas general CNS stimulants such as amphetamine or cocaine do not potentiate these responses. Both 50 and 100 ug/kg doses of TRH failed to potentiate yohimbine effects. Certain gross similarities in effects produced by TRH and amphetamine observed in this study support the view that beneficial effects of TRH in depression may be related to general sympathetic activation produced by this hormone.     

Anxiolytic-like effect of neuropeptide Y (NPY), but not other peptides in an operant conflict test.

The peptide messengers neuropeptide Y (NPY), growth hormone-releasing hormone (GHRH), atrial natriuretic peptide (ANP) and beta-endorphin (BEND) were tested in an animal model of anxiety, the Geller-Seifter conflict test. Rats were subjected to a multiple schedule consisting of three components: in the first component, lever-pressing produced food-reward ('unpunished responding'). The second component was a time-out period, during which lever-pressing had no consequences. During the third component, lever-pressing produced food-reward, but was also punished by an incremental foot-shock ('punished responding'). After establishing a stable baseline of both unpunished and punished responding, animals were injected with various doses of NPY, GHRH, ANP, BEND, or with saline into the lateral cerebral ventricle, and testing was repeated. While changes in unpunished responding can reflect alterations in performance factors or motivational strength, increases in punished responding have previously been shown to be highly specific for anxiety-reducing drugs, such as the benzodiazepines. NPY markedly and dose-dependently increased punished responding. A smaller increase of unpunished responding was also seen. These results add further support to the hypothesis that NPY may be an endogenous anxiolytic. GHRH, ANP and END did not affect punished responding.     

Determinants of drug effects on punished responding.

The effects of drugs on punished responding depend on interactions among a large number of experimental variables. Among these variables are the drug history of the animal, the dose of the drug administered, the type of stimulus used to punish responding, the intensity and duration of the punishing stimulus, the schedule of presentation of the punishing stimulus, the control rate and pattern of punished responding, the schedule of positive reinforcement maintaining the punished responding, the species of animal, the deprivation state of the animal, the behavioral history of the animal, and the nature of the required response. Although it is not known how all of these variables interact to determine the effect of drugs on punished responding, there is evidence that many of these variables are important as determinants of drug effects. The task facing behavioral pharmacologists studying drug effects on punished responding is to determine under what conditions drugs produce their characteristic effects on punished responding.