Behaviorally conditioned effects of psychoactive drugs in experimental animals: What we have learned from nearly a century of research and what remains to be learned

dopamine receptor antagonist haloperidol, morphine and antidepressant drugs. In each section, the drug under discussion is briefly introduced, followed by a detailed examination of conditioning features, including doses and dosing regimens, characteristics of the conditioning process such as test environments or specific conditioned stimuli, testing and conditioned response characteristics, possible extinction or reconditioning or reversal training, neural mechanisms, and finally, the potential clinical relevance of the research area related to the drug. We focus on key outcomes, delve into methodical issues, identify gaps in current knowledge, and suggest future research directions.


Introduction
Continuous treatment with drugs is crucial for managing conditions like chronic pain and neuropsychiatric disorders, including depression and schizophrenia.Drugs' effectiveness depends not just on their pharmacological action but also on the individual's psychobiological state, shaped by genetics, emotional context, and past experiences with similar treatments, which influence expectations of outcomes.These factors are recognized through placebo and nocebo effects.
Drug effects are often replicated via associative learning, like Pavlovian conditioning.Here, drug exposure (the unconditioned stimulus, US) triggers specific responses in the central nervous system (CNS) and physiology (the unconditioned response, UCR).When paired with a neutral stimulus during the acquisition phase, this stimulus can later elicit a drug-like effect (the conditioned response, CR) on its own during the evocation phase, that is, the stimulus has become a conditioned stimulus (CS).
Yet, clinical analyses of drug treatments rarely consider associative learning, which could optimize patient treatment and enhance trial sensitivity by highlighting drug-placebo differences (Enck et al., 2013).Understanding these learned drug responses is vital for addressing issues like the opioid crisis by developing strategies against drug-seeking behaviors and cravings (Humphreys et al., 2022).
Clinical studies show significant placebo effects, with rates up to 70% in disorders like depression and schizophrenia (Leuchter et al., 2014, Moncrieff et al., 2004).These effects, and the order of placebo or drug administration, significantly influence treatment efficacy, underscoring the role of conditioned pharmacological effects.
In practice, pharmacological conditioning might reduce drug doses while maintaining efficacy and minimizing side effects, offering a complementary strategy for treatment (Hadamitzky et al., 2020;Hadamitzky and Schedlowski, 2022).Moreover, understanding the neuropsychological and neurobiological bases of associative learning can help create interventions to disrupt drug-related behaviors and addiction.

Early observations
Nearly a century ago, experimental findings showed that drugs could induce conditioned pharmacological effects in animals (Pavlov, 1910;Collins and Tatum, 1925).Pavlov reported that dogs treated repeatedly with apomorphine (APO) and a tone exhibited CRs like restlessness or salivation upon hearing the tone alone.He also noted that morphine injections became associated with symptoms such as nausea and vomiting, previously caused by the drug, through the mere preparation for injection.Collins & Tatum detailed that for dogs (and a cat), routine morphine shots led to anticipatory responses like salivation when merely seeing the experimenter or syringe.
Subsequent research has expanded on these early studies, exploring various drugs, conditioning theories, neurobiological mechanisms, and their potential clinical implications.Despite the age of some studies (Lynch et al., 1976;Eikelboom and Stewart, 1982;Stewart and Eikelboom, 1987), recent interest has focused on these effects as a neuropsychological basis for the placebo effect (Hadamitzky and Schedlowski, 2022).
This review focuses on drugs influencing psychomotor activity in rodents, particularly psychostimulants like d-amphetamine (AMPH) and cocaine, and their effects through dopaminergic pathways.It also covers dopamine receptor agents such as APO and haloperidol, plus opioids like morphine, noting their analgesic and psychomotor impacts.While antidepressants affecting rodent psychomotor activity are discussed, conditioned place preference studies are excluded.
The review is structured by drug categories: psychostimulants (AMPH, cocaine), dopamine agents (APO, haloperidol), and opioids (morphine), concluding with antidepressants.Each section examines conditioning procedures, doses, test settings, CRs, neural underpinnings, and clinical relevance.The final chapter synthesizes these findings, discusses methodological aspects, and suggests future research directions.

Amphetamine
Experimental evidence strongly indicates that amphetamines are highly effective in inducing behaviorally conditioned drug effects across a variety of systemic doses.d-Amphetamine (AMPH), in particular, has been extensively studied due to its clinical use in treating narcolepsy and attention deficit hyperactivity disorder (ADHD).AMPH acts as a potent indirect agonist of monoamines, increasing extracellular levels of dopamine (DA), noradrenaline (NA), and serotonin (5-HT) primarily by promoting their release.It also reduces transmitter reuptake and enzymatic degradation through monoamine oxidase to a lesser extent.Neurochemically, AMPH is more potent than l-amphetamine, especially in its effects on DA compared to NA or 5-HT (Heal et al., 2013).Its behavioral effects are dose-dependent: lower doses, such as 0.5-1.5 mg/kg, result in increased locomotion and rearing in rats, while higher doses, like 3-5 mg/kg, lead to stereotypies (for review, see Antoniou et al., 1998).Another notable behavioral effect of AMPH is the production of ultrasonic vocalizations (USVs), specifically 50-kHz calls (Wöhr, 2022).Moreover, repeated AMPH injections can cause sensitization, a phenomenon extensively studied in the context of drug addiction and schizophrenia (Robinson and Becker, 1986).
The evidence primarily discussed below was obtained from studies in rats, which are the preferred species for this type of research.However, there is also evidence of effective behaviorally conditioned effects with AMPH in mice, as demonstrated in several studies (Hayashi et al., 1980;Mead et al., 1998Mead et al., , 1999;;Tirelli and Terry, 1998;Steckler and Holsboer, 2001).Moreover, the majority of these rat studies involved male subjects, with relatively little comparable experimental evidence available for females (Tilson and Rech, 1973).
The US -Effective drugs and dosage regimens: In the studies where AMPH was used as the US, a wide range of doses was often administered repeatedly under various environmental conditions, such as activity boxes, to study several conditioned and unconditioned drug effects, typically focusing on locomotor activity or stereotypies.Specifically, the doses of AMPH in many studies varied from as low as 0.1 mg/ kg (Herz and Beninger, 1987;Guillory et al., 2022) to as high as 5.0 mg/kg (Schiff, 1982), covering behaviors from increased locomotion/rearing to those predominantly characterized by stereotypies, without directly comparing the effects of different doses.Notably, studies using low, intermediate, or high doses of AMPH consistently reported behaviorally conditioned effects, suggesting that such outcomes are not dependent on the dose.The dosing regimens varied, including a single dose of 3.0 mg/kg (Ross and Schnitzer, 1963), two doses of 2.0 mg/kg (Mazurski and Beninger, 1987), or regimens involving repeated drug challenges, with some protocols using up to 21 injections (Beninger and Hahn, 1983;Poncelet et al., 1987).Pihl and Altman (1971) observed that three, as opposed to nine or fifteen dose/environment pairings (3.0 mg/kg), were insufficient to induce behaviorally conditioned effects in measures of gross behavioral activity.Only a few studies employed randomized blocked designs, where repeated blocks of drug injections were interspersed with vehicle tests (Mazurski and Beninger, 1991;Stewart and Vezina, 1991;Pickens and Crowder, 1967).
In the majority of studies, the route of administration was either intraperitoneal (ip) or subcutaneous (sc), with one experiment using intravenous (iv) administration (Crombag et al., 2001).The timing between injections and exposure to the test environment varied across studies.Some researchers exposed their subjects to the environment immediately following drug administration (Ross and Schnitzer, 1963;Mazurski andBeninger, 1987, 1991), while others allowed intervals of 30 minutes before exposure (e.g., Tilson and Rech, 1973;Drew & Glick, 1987).Pickens and Crowder (1967) observed more pronounced conditioned effects when AMPH was administered immediately prior to exposure to the conditioned environment, compared to administration 30 minutes before.Krank and Bennett (1987) placed the animals in the test environment (with red light and noise) for 15 minutes before administering AMPH, followed by another 60 minutes in the test box.
Behavioral effects and potential changes resulting from repeated drug administration were frequently, though not uniformly, reported.Some studies observed behavioral sensitization following repeated injections (Tilson and Rech, 1973;Hayashi et al., 1980;Mazurski and Beninger, 1987), while others did not find such effects (Hiroi and White, 1989;Mazurski and Beninger, 1991).This inconsistency in findings may be partially attributed to the types of behavioral readouts used.For instance, one study noted an increase in sensitized locomotion during the first four days of injections, followed by a decline, suggesting that the drug's effects qualitatively shifted over time from locomotion (typical of lower doses) to stereotypies (typical of higher doses), indicative of sensitization (Martin-Iverson and McManus, 1990).This suggests that behavioral assessments should not rely solely on a single measure but should ideally include a range of drug-relevant behavioral readouts.
Features of conditioning: Various devices have been utilized to examine the effects of behavioral conditioning, including activity boxes equipped with photocells (e.g.Ross and Schnitzer, 1963, Mazurski and Beninger, 1987, Stewart and Vezina, 1991, Jodogne et al., 1994), doughnut-shaped activity cages (Tilson and Rech, 1973), rotometers (Drew & Glick 1988) and translucent observation boxes (Poncelet et al., 1987).The widespread use of photocell apparatuses is likely attributed to the fact that many studies focus on measuring locomotion (and occationally rearing), which unlike stereotypies, can be easily quantified using such automated setups.
The use of stimuli other than contextual ones as an additional CS, such as sounds or odors, was not commonly reported.However, some studies did introduce a 1-minute tone into the test environment (Schiff, 1982;Krank and Bennett, 1987).In other research, animals were extensively habituated to the experimental apparatus (e.g., Tilson and Rech, 1973;Mazurski and Beninger, 1987) to minimize the effects of novelty.Nevertheless, habituation does not appear to be necessary, at least not when conditioning involves repeated drug treatments.
During a standard conditioning procedure, animals in the experimental group received AMPH paired with the conditioning environment, followed by an injection with a vehicle (typically saline) several hours later in their home cages (e.g., Guillory et al., 2022).Additionally, experimental designs often included a control group of 'pseudo-conditioned' animals (Fig. 1).These controls underwent a reversed treatment, receiving a vehicle administration before exposure to the conditioning environment and AMPH several hours later in the home cage.Therefore, 'pseudo-conditioned' controls received the same number of drug injections as the experimental animals, but these injections were not contingently paired with the conditioning environment.In these designs, it is posited that animals may have differing experiences, such as the environment/cues paired with the drug acting as a conditioned inhibitor (Lysle and Fowler, 1985;Guillory et al., 2022).Consequently, the outcomes of later saline tests in these control conditions may not necessarily mirror those of control animals that did not have contrasting drug experiences in another environment.
Features of testing and conditioned responses: Behaviorally conditioned effects are usually tested after the administration of a vehicle, applied after a retention interval of one or a few days (Schiff 1992, DiLullo & Martin-Iverson 1992).As a result, relatively little is known about whether the administration of a vehicle is necessary as a cue to elicit a CR.Given the high level of salience of the vehicle administration, typically done via injection previously also used for drug administration, it is tempting to assume that this cue is needed (e.g.Pavlov 1910, Collins andTatum, 1925), yet systematic studies are lacking.Krank and Bennett (1987), however, who consistently applied an initial 15-min no-injection phase in the conditioning context before administering saline or drug, successfully used this initial phase to show conditioned behavioral activation in the later test for conditioned drug effects.
Against the backdrop of the stimulus substitution assumption in associative learning, it was posited that the CR would mirror the UR when AMPH was used as the US: for instance, UR of locomotion leading to a CR of locomotion, or a UR of stereotypy resulting in a CR of stereotypy.The methods used to assess such responses varied significantly across studies.In some instances, behavioral data were scored and presented using an observation-based scaling system, which categorized behaviors ranging from inactivity, through locomotion and stereotypy, to dyskinetic movements (Beninger and Hahn, 1983).Other studies quantified different behaviors separately, such as locomotion versus rearing (Mazurski andBeninger, 1987, 1991), and identified conditioned effects in rearing but not in locomotion (Mazurski and Beninger, 1987;Stewart and Vezina, 1991), suggesting evidence of drug-induced sensitization.Segal and Mandel (1974) observed no effect in a saline Fig. 1.Conditioning design with pseudo-conditioned controls.In this experimental setup, the experimental group (illustrated in the upper half) is repeatedly exposed to the drug (indicated by a red syringe) paired with the test environment, and to the vehicle (indicated by a blue syringe) paired with the home cage.This is in contrast to the control group (illustrated in the lower half), which receives the drug paired with the home cage and the vehicle paired with the test environment.Subsequently, the subjects are tested in the test environment after receiving vehicle injections (shown on the right).In this between-subjects design, all subjects have received the same number of drug and vehicle injections.However, only the experimental rats are expected to exhibit a drug-like response (in this case, exemplified by rearing behavior) during the test.
test after 36 daily injections of 1 mg/kg of AMPH, noting that the animals were continuously exposed to the experimental environment without pairing drug treatment with a specific CS.In another study, the drug was administered intravenously to animals that had been in the testing environment for a considerable amount of time, without the administration being signaled by any discrete stimuli, handling, or other indicators.Under these conditions, no evidence of conditioned drug effects was observed (Crombag et al., 2001).
A few studies have explored additional behavioral responses such as defecation (DiLullo and Martin-Iverson, 1991) or turns in intact subjects (Drew andGlick, 1987, 1988).The latter was informed by observations that, when treated with a relatively low dose of AMPH (specifically, 1.25 mg/kg), animals exhibit not only locomotion or rearing but may also perform complete rotations, termed "net turns," showing moderate individual lateralization to one side as a CR.These rotations also became the primary conditioned readout in animals with unilateral lesions of the mesostriatal DA system, as detailed in a supplementary document titled "The Rotating Rodent."Martin-Iverson and McManus (1990) proposed that the UR of locomotion is not necessary for eliciting a later conditioned response, echoing Herz and Beninger (1987).They posited, "that locomotion is not the UR, and is probably not the CR either, even if that is the effect measured."Similarly, Swerdlow and Koob (1984) constrained rats from locomoting during the conditioning phase through physical restraint and discovered that such restriction did not inhibit conditioned locomotion during the subsequent test phase.This suggests that, "locomotor behavior during training is not required for an animal to learn to associate an environment with amphetamine".
More recently, 50-kHz USVs have been examined in the context of conditioned drug effects.Specifically, it is well-established that AMPH induces dose-dependent increases in 50-kHz calls, which are thought to reflect an appetitive emotional and motivational state (Wöhr, 2022), and are dependent on meso-limbic DA function.Similar to AMPH, the effect on calling can become sensitized with repeated injections (e.g., Simola and Morelli, 2015).Costa et al. (2020) provided preliminary evidence that such calling is increased in a vehicle test in animals with previous AMPH exposures in the test context compared to rats with prior vehicle exposures.However, it remains uncertain whether the outcome in the drug group truly reflects a CR, as this study did not include appropriate controls, such as a pseudo-conditioned control group.
Studies that explicitly compared the UR with the CR found that the CR depended on the dose of the US during acquisition (Tilson and Rech, 1973;Martin-Iverson and McManus, 1990) and was smaller than the UR (Mazurski and Beninger, 1987;Gold et al., 1988;Gold and Koob, 1989), for instance, one-third of it (Hiroi and White, 1989).These authors also noted that the CR was most prominent during the first 10 minutes of testing (Beninger and Hahn, 1983;Poncelet et al., 1987), while the UR peaked later, around 30 minutes after injection.To align the peak behavioral drug effect with the conditioning environment more closely, further rats were exposed to this environment only during the 35-45 minutes post-injection.This adjustment increased the CR to about two-thirds of the UR.
It has also been suggested that the purported conditioned locomotor response actually reflects a lack of habituation (Pickens and Dougherty, 1971;Gold et al., 1988;Damianopoulos and Carey, 1992), as the repeated drug/environment pairings may prevent behavioral habituation, leading animals in a subsequent no-drug test to behave as if the environment is novel.Several arguments support this hypothesis, including observations that the magnitude of the CR may not exceed that observed in naïve rats introduced to such an environment (Tirelli and Terry, 1998).
Extinction, re-conditioning, reversal training: The investigation of potential extinction of the CR with repeated unreinforced test trials (i.e., without the drug) was pursued only infrequently (Hayashi et al., 1980;Drew and Glick, 1988;Stewart and Vezina, 1991;Jodogne et al., 1994).When studies examined the relationship between extinction and sensitization, they found that the decrease in sensitization did not correspond with the extinction of the CR, such as in response to a low challenge dose of AMPH (Stewart and Vezina, 1991).This finding supports the hypothesis that conditioning and sensitization represent at least partly distinct mechanisms.Furthermore, it has been suggested that sensitization may not necessitate context-dependent conditioning (Browne and Segal, 1977).No studies involving re-conditioning or reversal training were identified.
Pharmacological manipulations, conditioned neural responses and possible neural mechanisms: The role of neurotransmitters, especially DA, in the context of behaviorally conditioned responses to AMPH has been extensively investigated.Various studies have utilized approaches such as employing DA receptor antagonists to inhibit synaptic transmission, drugs to reduce catecholaminergic transmitter availability, neurotoxins to destroy DA neurons in the brain, or analyzing neurochemical changes related to conditioned drug effects.In contrast, the investigation of other drugs or neurochemical mechanisms, such as the effects of clonidine, yohimbine, lithium, diazepam, or glutamatergic drugs, has been relatively rare (Poncelet et al., 1987;Mead et al., 1998Mead et al., , 1999)).
Conditioned stereotypic effects resulting from AMPH exposure could be inhibited by the DA receptor antagonist haloperidol (Schiff, 1982;Drew and Glick, 1988).The DAergic metabolite homovanillic acid (HVA) in neostriatal tissues or samples, including the olfactory tubercles and nucleus accumbens (NAcc), was found to increase after a vehicle test compared to pseudo-conditioned controls.This supports the view that "the conditioned response to cues previously associated with AMPH administration might be mediated by increased DA turnover".A microdialysis study (Dietze and Kuschinsky, 1994) corroborated this finding, showing enhanced striatal DA release concurrent with conditioned stereotypies in the test for conditioning compared to pseudo-conditioned controls.More recently, Guillory et al. (2022) did not observe evidence for enhanced DA release in the NAcc.However, in their study, animals had been in the dialysis chamber for several hours before the critical saline test, which may not have been sufficient to trigger a conditioned DAergic response.Conversely, increased levels of extracellular glutamate were detected in the same brain area, leading the authors to hypothesize that glutamatergic neurotransmission could be associated with conditioned behavioral effects, particularly in locomotion.Beninger and Hahn (1983) tested the DAergic antagonist pimozide, which blocks D2-family receptors.Administered during the acquisition phase, pimozide inhibited the acute effects of AMPH on behavioral activity and prevented subsequent conditioned effects.However, when given during evocation, weak CRs were still observable.Poncelet et al. (1987) reported similar outcomes with pimozide and also tested haloperidol and sulpiride, both of which inhibited the acute UCR and the CR of AMPH in locomotion.Additionally, the D1 antagonist SCH 23390 blocked both the UCR and CR of AMPH in locomotion.Martin-Iverson and McManus (1990) found that haloperidol (0.05 mg/kg) and SCH 23390 (0.02 mg/kg), whether given alone or in combination, did not block the UCR of AMPH in locomotion but did prevent the subsequent CR.They concluded that D1 or D2 receptors are not necessary for acquiring conditioned drug effects of AMPH.
Regarding neuronal mechanisms, Martin-Iverson and McManus (1990) posited that "The UCR must originate from an event in the central nervous system (CNS) that precedes the behavioral expression."In relation to the CNS event, they suggested that given amphetamine's direct action on the presynaptic release of DA, it initially seems plausible that the presynaptic release of DA becomes conditioned to environmental stimuli.During conditioning, in the presence of antagonists, this DA release is not inhibited.On the test day, the behavioral effects of this conditioned DA release are revealed in the absence of the antagonists.
Treatment with alpha-methyl-p-tyrosine (AMPT) inhibits the synthesis and thus the availability of DA and NA, while DSP4 selectively depletes NA.The administration of N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine (DSP4), either during acquisition or evocation, did not affect the UCR and CR of AMPH.In contrast, AMPT reduced the UCR but not the CR, as indicated by DiLullo and Martin-Iverson (1991), suggesting that NA is not essential for either the acquisition or expression of conditioning.Conversely, DA appears to be critical for the UCR to AMPH but not for the CR.In a subsequent study, DiLullo and Martin-Iverson (1992) examined the effects of AMPT and reserpine.Reserpine inhibits the neuronal vesicular monoamine transporter VMAT2, leading to decreased storage and, consequently, decreased release of DA and other monoamines.The authors replicated their earlier findings with AMPT, whereas reserpine did not impact either the UCR or the CR of AMPH in locomotion.This suggests that reserpine was ineffective, in contrast to AMPT, possibly because it depleted different cellular pools than AMPT.
In a different experimental setup, 6-hydroxydopamine (6-OHDA) was bilaterally injected into the NAcc to destroy DAergic cells either before or after conditioning with AMPH (Gold et al., 1988).These lesions significantly depleted DA levels, and to a lesser extent, NA levels in the NAcc.When applied prior to acquisition, the lesions diminished the UCR of AMPH in locomotion and inhibited the subsequent CR.Furthermore, a CR was not observed when the lesion was introduced before the evocation phase.Conversely, no evidence was found for the involvement of the NAcc (in contrast to the medial prefrontal cortex or amygdala) in the CR, using c-fos as an anatomical marker of neuronal activity in a mouse model.However, c-fos labeling did indicate increased activation in response to AMPH treatment itself (Mead et al., 1999).
Clinical relevance: Given that AMPH is a potent substance of abuse, many studies have framed their findings within the context of addiction (Gold and Koob, 1989;Stewart and Vezina, 1991;Jodogne et al., 1994).Specifically, the potential relevance of conditioned drug effects for clinical extinction programs has been underscored (Stewart and Vezina, 1991), with suggestions that conditioned behaviors might represent "a conditioned appetitive response to the incentive properties of the drug" (Gold and Koob, 1989).The role of learning in the drug treatment of schizophrenia has also been highlighted (Beninger and Hahn, 1983).However, numerous empirical and theoretical studies have concentrated their discussions on fundamental scientific issues, such as reward and reinforcement (Ellinwood and Kilbey, 1980;Mazurski and Beninger, 1991), learning and associative processes (Beninger and Hahn, 1983;Swerdlow and Koob, 1984;Martin-Iverson and McManus, 1990), and interactions with habituation (Gold et al., 1988).
Summary and discussion: AMPH is notably effective at inducing behaviorally conditioned drug effects across a spectrum of systemic doses.These effects are typically established by associating specific environments with the drug's action.Generally, multiple drugenvironment pairings are utilized, with the unconditioned behavioral effects -most notably locomotion and rearing-being dose-dependent and subject to change over time, often manifesting as sensitization.This phenomenon is observed to be distinct from the conditioning process itself.Many experimental setups have included 'pseudo-conditioned' animals as the most relevant controls.These are animals that received AMPH in identical frequencies and doses as the experimental group but without pairing with the subsequent test environment.The CR often resembles the UCR in quality but tends to be less intense and may diminish with repeated testing in the absence of the drug.Intriguingly, evidence suggests that typical unconditioned drug actions, such as increased locomotion, might not be necessary for subsequent conditioned effects.This finding challenges the traditional view that these overt behavioral reactions should be categorized strictly as UCRs.Concerning neurochemical mechanisms, the actions of AMPH have been associated with DAergic activity in the brain.Yet, it remains unclear whether the function of mesolimbic DA, including transmitter release and specific DA receptor involvement, is essential during the acquisition and/or evocation of conditioned effects.

Cocaine
Cocaine induces behavioral effects that are largely similar to those of AMPH.When administered in doses of 10-20 mg/kg to rats, cocaine acutely leads to increased locomotion, rearing, and stereotypies, such as sniffing and head movements (e.g., Gulley et al., 2003).These effects typically intensify, or sensitize, with repeated injections (Vanderschuren and Kalivas, 2000).Similarly to AMPH, cocaine also elevates extracellular levels of DA, NE, and 5-HT, primarily through its potent action as a monoamine reuptake inhibitor (e.g., Wolf and Kuhn, 1991).
Evidence for behaviorally conditioned effects of cocaine was documented over 90 years ago in dogs, rats, and monkeys (Tatum and Seevers, 1929;Downs and Eddy, 1932), but systematic research began in the late 1970s and peaked in the 1990s and early 2000s.Nearly all published studies reported successful conditioning using cocaine as the US (but see Post and Rose, 1976;Damianopoulos and Carey, 1992).The reasons for these negative findings are unclear, but Bridger et al. (1982) suggested that the repeated 20-minute habituation period before cocaine injection, used by Post and Rose (1983), might have resulted in extinction.The presence of a publication bias regarding cocaine -that is, unpublished studies with additional negative results-remains uncertain.
In terms of species, mice were more commonly used in studies involving cocaine compared to AMPH or APO (for example, Tirelli et al., 2003Tirelli et al., , 2005;;Hall et al., 2009;Smith et al., 2016).With respect to sex, the majority of experimental evidence on conditioned drug effects primarily involves male subjects, with the exception of the study by Hall et al. (2009), which tested both male and female mice.
The US -Effective drugs and dosage regimens: The doses of cocaine commonly used in studies with rats and mice ranged from 10 to 20 mg/kg, typically administered sc or ip.Only Reimer and Martin-Iverson (1994) and Michel and Tirelli (2002) included lower doses in their studies, and Damianopoulos and Carey (1992) tested a higher dose of 50 mg/kg.At a dose of 5 mg/kg, Reimer and Martin-Iverson (1994) found no evidence of conditioned locomotion in rats, as this dose did not produce significant psychostimulant effects.Conversely, Michel and Tirelli (2002) observed psychostimulatory effects at the same dose in mice but did not find conditioned locomotion.The number of drug injections/conditioning trials varied widely among studies, ranging from a minimum of 3 (Michel et al., 2003) to a maximum of 28 (Post and Rose, 1976).Most studies did not systematically investigate the impact of the number of trials, except for Michel et al. (2003), who reported no evidence of conditioned effects with compared to 6 or 12 daily injections in mice.Cervo and Samanin (1996) reported conditioned effects after 4 injections in rats, with several days between each drug administration.The bulk of research indicates that or 6 injections of 10-20 mg/kg are sufficient to establish conditioned drug effects in both rats and mice (e.g., Carey and Gui, 1998;Druhan and Wilent, 1999;Hotsenpiller et al., 2001;Michel et al., 2003;Hall et al., 2009;Smith et al., 2016).The drug was typically injected immediately before exposure to the conditioning environment (e.g., Cervo and Samanin, 1996;Panlilio and Schindler, 1997;Tirelli et al., 2005), or subjects were exposed to the test environment for 20 minutes prior to drug administration (Post et al., 1981).In the test phase, conditioned effects were observed not only after but also before saline injection, suggesting the significance of the environmental context in conditioning.Exposure times in the test environment after drug injection varied from 20 (Damianopoulos and Carey, 1992) to 60 minutes, with the most common durations being 30 (Brown and Fibiger, 1993;Panlilio and Schindler, 1997;Hotsenpiller et al., 2001;Hall et al., 2009) or 60 minutes (Beninger and Herz, 1986;Reimer and Martin-Iverson, 1994;Tirelli et al., 2005).
Interestingly, the examination of behavior during the conditioning phase, and any potential changes over repeated drug injections, was not systematically undertaken in the studies reviewed.When behavior during conditioning was analyzed, it was typically in terms of locomotion, which is a relatively consistent outcome following doses of 10 mg/kg or higher.Additionally, several studies observed sensitization, defined as an increase in locomotion across repeated trials (Post and Rose, 1976;Barr et al., 1983;Martin-Iverson and Reimer, 1994;Panlilio and Schindler, 1997;Hotsenpiller et al., 2001; but see Damianopoulos and Carey, 1992;Carey and Damianopoulos, 1994;Brabant et al., 2003).Post and Rose (1976), who quantified not only locomotion and rearing but also stereotypies, reported that sensitization also entailed a shift towards more stereotypical behaviors with repeated injections.Echoing research on AMPH, several researchers concluded that sensitization and conditioning are mediated by distinct neural mechanisms (Martin-Iverson and Reimer, 1994;Hotsenpiller and Wolf, 2002;Brabant et al., 2003;Michel and Tirelli, 2002;Tirelli et al., 2003).This conclusion is based on observations that either phenomenon can occur independently of the other or can be differentially affected by various pharmacological agents.
However, caution is required when drawing conclusions about sensitization, as this outcome is sometimes determined based on comparisons with saline-treated controls, who typically exhibit habituation over repeated trials (Martin-Iverson and Reimer, 1994).Sensitization might not be as readily observable when conducting within-group comparisons, that is, when comparing outcomes to previous cocaine trials (Damianopoulos and Carey, 1992).These methodological nuances are particularly important when examining the relationships between potential sensitization across drug trials and subsequent conditioning (see also Carey et al., 2008 for a discussion on the role of habituation).
Features of conditioning: The range of test apparatuses utilized for conditioning studies on cocaine effects mirrors those used in AMPH research.Glass aquariums (Bridger et al., 1982;Barr et al., 1983), various types of plexiglass chambers and cages (Post and Rose, 1976;Beninger and Herz, 1986), circular cages (Brown and Fibiger, 1993), and open fields (Damianopoulos and Carey, 1992) have been employed.Behavioral quantification typically depended on sophisticated infrared beam systems to measure locomotion (Beninger and Herz, 1986;Martin-Iverson and Reimer, 1994;Hotsenpiller et al., 2001), video image analyzing devices (e.g., Damianopoulos and Carey, 1992), or observational methods, particularly for stereotypies (e.g., Bridger et al., 1982;Barr et al., 1993;Damianopoulos, 1994).A notable and more recent deviation from these methods is the study by Smith et al. (2016), which, due to the necessity of local electrophysiological recordings in the brain, examined conditioned cocaine effects in head-restrained mice using a treadmill apparatus.
Some studies utilized the experimental environment itself as the CS without reporting the use of additional distinct stimuli.In contrast, others incorporated auditory stimuli (Bridger et al., 1982;Barr et al., 1993;Panlilio and Schindler, 1997;Druhan and Wilent, 1999;Hotsenpiller et al., 2001;Damianopoulos and Carey, 1992;Carey and Damianopoulos, 2006;Carey et al., 2008), visual stimuli like striped wall patterns (Brown and Fibiger, 1992) or lights (Panlilio and Schindler, 1997;Hotsenpiller et al., 2001Hotsenpiller et al., , 2002;;Damianopoulos and Carey, 1992;Carey and Damianopoulos, 2006;Carey et al., 2008), and olfactory stimuli using bedding (Brown and Fibiger, 1992;Smith et al., 2016).The timing of stimulus exposure typically matched the duration of exposure to the conditioning context.In the studies by Bridger et al. (1982) and Barr et al. (1983), however, drug injection and exposure to the environment were preceded by a 1-minute exposure to another environment along with a distinct tone.Many studies also included saline injections, usually before exposure to the home cage, in their experimental animals (e.g., Bridger et al., 1982;Beninger and Herz, 1986;Brown and Fibiger, 1992), and introduced pseudo-conditioned controls, which received saline injections paired with the conditioning environment and cocaine paired with the home cage (e.g., Bridger et al., 1982;Barr et al., 1993;Beninger and Herz, 1986;Brown and Fibiger, 1992;Reimer and Martin-Iverson, 1994; but see Post and Rose, 1976).Damianopoulos and Carey (1994) employed a unique approach, using two distinct open-field setups to which all animals were repeatedly exposed.After a no-drug test in one, experimental animals were given cocaine and then exposed to the other open field.As both open fields were used again for the subsequent conditioning test, this method facilitated within-group comparisons of behavior in a context previously paired with cocaine versus an unpaired context.This strategy effectively addresses potential concerns associated with 'pseudo-conditioned' controls treated with drugs in their home cages.(Fig. 1).
No study has systematically explored the influence of specific contexts and/or additional stimuli paired with drug effects on their conditioning efficacy.However, the broad spectrum and variability of procedures utilized suggest that behaviorally conditioned effects of cocaine are quite robust, manifesting across a wide array of experimental designs.
Features of testing and conditioned responses: Most studies commenced with several conditioning trials, followed by evocation tests.However, some employed alternating designs, wherein multiple blocks of 4 consecutive drug trial days were followed by a saline test, as applied in studies by Brabant et al. (2003), Hotsenpiller et al. (2002), and Carey et al. (2008) to explore alternative approaches.Damianopoulos and Carey (1992) examined the impact of novelty versus habituation in conditioning tests, noting that, depending on the experimental design, the saline experience in the test context could be novel for experimental animals if they had previously been exposed only to cocaine.Consequently, differences in locomotion between paired and unpaired animals during evocation might reflect novelty (in experimental animals) versus habituation (in controls) rather than conditioning per se.This complication can be mitigated through various strategies, such as conducting within-group comparisons of no-drug behavior during the test against a prior baseline, instead of between-group comparisons of experimental versus control animals during the test phase (Damianopoulos and Carey, 1994).
In many cases, the quantification of locomotion has been the primary measure of the CR (Cervo and Samanin 1996;Hotsenpiller et al., 2001; but see Beninger and Herz, 1986), with additional studies examining stereotypies and rearing (Bridger et al., 1982;Barr et al., 1993;Post et al., 1981;Michel and Tirelli, 2002;Damianopoulos and Carey, 1994), the spatial distribution of locomotor activity (Druhan and Wilent, 1999;Carey and Gui, 1997), pupil dilation (Smith et al., 2016), and qualitative patterns of turns during locomotion (Damianopoulos and Carey, 1992).Among these methods, analyzing locomotion not just quantitatively but also in terms of spatial distribution-particularly in larger test environments like open fields-has proven to be a valuable approach.Cocaine not only generally increases locomotion but also, and perhaps more so, in the center of the environment, a behavior that can also be conditioned (Druhan and Wilent, 1999; see also Carey andGui, 1997, 1998).Druhan and Wilent (1999) further suggested that this central tendency might reflect conditioned drug-seeking behavior rather than a decrease in anxiety or habituation effect influenced by experience and/or drug exposure.Hotsenpiller et al. (2001), who observed conditioned drug effects in terms of rearing, noted these effects were primarily seen in spatial proximity to cue presentation, specifically near a tone-producing metronome and a flashing light.However, since the locations of these cues were not systematically varied, it's possible these were arbitrary effects.
Conditioning outcomes were typically assessed after a saline injection, immediately followed by exposure to the prior conditioning context (e.g., Post and Rose, 1976;Michel and Tirelli, 2002).Exceptions include Hall et al. (2009), who did not use injections for testing, and Hotsenpiller et al. (2001), who performed "mock" injections, meaning only the plastic tip of the syringe touched the animal's abdomen.Carey and Damianopoulos (1994) also forewent injections before testing, suggesting that the conditioned locomotion they observed was solely induced by exteroceptive stimuli.In contrast, Carey and Gui (1998) posited that a "compound stimulus complex" of extero-and interoceptive cues was effective.This underscores the potential need for systematic studies on the role of injections as cues, similar to research on AMPH.
The duration of evocation trials varied but often mirrored the conditioning trials.Temporal response patterns during testing received limited attention, though Hotsenpiller et al. (2002) focused on the first 30 minutes of the evocation trial, where most CRs occurred, and Carey and Gui (1998) noted that conditioning effects were most pronounced during the first five minutes of testing.
The relationship between the strength of a CR and the number of conditioning trials was seldom compared.Michel et al. (2003) found a more pronounced CR after 12 compared to 6 trials, suggesting that the degree of the CR correlates with the number of US-CS pairings, supporting a Pavlovian interpretation.
The interval between acquisition and evocation varied; some studies conducted evocation tests the day after the last cocaine conditioning trial (e.g., Bridger et al., 1982;Druhan and Wilent, 1999;Tirelli et al., 2003), while others waited two (Brown and Fibiger, 1993;Cervo and Samanin, 1996) or more days (Reimer and Martin-Iverson, 1994;Damianopoulos and Carey, 1992;Carey and Gui, 1998).Tirelli et al. (2005) systematically examined these intervals (1, 8, or 28 days) and observed conditioning effects on days 1 and 8, with a reduced effect on day 28 for locomotion.This aligns with Hotsenpiller et al. (2001), who noted similar levels of conditioned locomotion 3 and 10 days post-trial.Bridger et al. (1982) reported conditioned locomotion and rearing after 15 days, whereas conditioned stereotypy dissipated sooner in a measure-dependent manner (e.g., grooming vs. sniffing).Barr et al. (1993) observed conditioned locomotion and rearing up to 15 days post-trial, but increases in sniffing and head bobbing were only noted up to 10 days.
Extinction, re-conditioning, reversal training: Bridger et al. (1982) investigated the possibility of extinction in response to repeated context-paired saline injections but found no evidence of extinction over four consecutive days in terms of locomotion, rearing, sniffing, and head bobbing.However, other stereotypic behaviors such as licking, chewing, and grooming did not show differences compared to pseudo-conditioned controls.Consequently, the authors suggested that certain stereotypic measures might not reliably condition.They further highlighted the importance of utilizing multiple and separate measures, rather than combined scores, since different measures may be indicative of different neural substrates, only some of which may be susceptible to conditioning.Barr et al. (1983) also conducted tests over four consecutive saline days and observed no extinction in locomotion, sniffing, and head bobbing.Conversely, Carey and Gui (1997) identified a conditioned cocaine effect in terms of increased center entries but noted that this effect dissipated across six extinction trials.Michel et al. (2003) explored extinction following 6 or 12 cocaine conditioning trials and reported quicker extinction after 6 trials, a finding that may affirm the Pavlovian nature of the CR.Hotsenpiller et al. (2001) demonstrated that sensitization to cocaine persisted even after the CR had been extinguished, underscoring that sensitization and conditioned drug effects, though both reliant on associative learning, are mediated by distinct mechanisms (see also Carey and Gui, 1998).
Pharmacological manipulations, conditioned neural responses and possible neural mechanisms: To better understand the neurobiological correlates of the CR, researchers have obtained a variety of brain measures.For instance, Barr et al. (1983) analyzed brain tissue samples shortly after conditioning and found no significant changes in DA metabolites in the neostriatum and nucleus accumbens/olfactory tubercle between conditioned and pseudo-conditioned subjects.Similarly, Carey and Damianopoulos (1994) detected no evidence of DAergic or NAergic effects in conditioned animals in post-mortem samples of the neostriatum or samples containing the NAcc, olfactory tubercle, and adjacent tissue.However, they observed higher levels of 5-HT and its metabolite 5-hydroxyindoleacetic acid in prefrontal tissue.These authors also reported no differences in plasma catecholamine or corticosterone levels between conditioned and pseudo-conditioned rats, suggesting a lack of stress level differences between groups (Damianopoulos and Carey, 1994).In-vivo microdialysis, used to measure extracellular DA in the NAcc, did not show differences in DA or its metabolites between conditioned and pseudo-conditioned rats, with increases observed in both groups to a similar extent during exposure to the test environment.These findings suggest "that conditioned stimuli associated with cocaine do not evoke similar neural states as the drug itself" (Damianopoulos and Carey, 1994).Nonetheless, these results do not exclude the involvement of the NAcc and DA within it, as demonstrated by Hemby et al. (1992), who showed that two local cocaine injections into the NAcc, which acutely increased locomotion, also induced conditioned locomotion but not conditioned place preference.
Glutamate transmission may play a significant role in the NAcc, as extracellular glutamate levels were found to increase in cocaineconditioned rats upon exposure to the conditioning environment.Furthermore, the systemic administration of an AMPA receptor antagonist reduced the expression of cocaine-conditioned rearing behaviors.Hotsenpiller et al. (2002) also investigated the immediate early gene c-fos as an indicator of neuronal activation in two experiments following the test for conditioned drug effects.They noted conditioned locomotion in both experiments but observed increased c-fos labeling in the NAcc (and prelimbic cortex) in the experimental group compared to controls only in the set of animals where the conditioning context was not paired with a flashing light cue.The authors suggested that variations in conditioning procedures might lead to "differences in terms of the expectation of receiving reward," alongside the activation of distinct neuronal circuits.While cocaine itself induced increased c-fos labeling in the NAcc and other structures, no such increase was observed (including in the piriform cortex and neostriatum) after evocation.However, heightened labeling in various other brain regions, such as the cingular cortex or amygdala, was detected post-evocation (Brown et al., 1992).This absence of a conditioning effect in the NAcc was interpreted as evidence against its involvement in the conditioned effects of cocaine.Investigating the amygdala's role through excitotoxic lesions with quinolinic acid did not alter acute or conditioned behavioral effects of cocaine but did impair cocaine-induced conditioned place preference (Brown and Fibiger, 1993).This suggests that conditioned drug effects and conditioned place preference may be mediated by different brain mechanisms.
Electrophysiological evidence supporting the involvement of medial prefrontal and ventral striatal neurons, including dorsal parts of the NAcc, in conditioned responses was presented in a study by Smith et al. (2016) with head-fixed mice.In this study, the strength of the conditioned pupil dilation effect (unlike treadmill locomotion) correlated with local field synchrony, suggesting that the neural effects may be related to the processing of cocaine-associated cues.
In addition to identifying neurobiological correlates of CRs, targeted experimental manipulations aimed to establish causality by blocking unconditioned versus conditioned behavioral responses, including neuropharmacological interventions.For instance, pimozide, an antagonist of D2, D3, and D4 DA receptors, did not inhibit cocaineconditioned rearing when administered prior to the test.However, when given during training days, it reduced acute behavioral effects without preventing subsequent conditioned drug effects (Beninger and Herz, 1986).Martin-Iverson and Reimer (1994) found that a low dose of haloperidol did not block cocaine-conditioned locomotion.Combining this treatment with the L-type calcium channel blocker nimodipine also failed to inhibit conditioned locomotion.Notably, when administered during the conditioning phase, haloperidol did not prevent conditioned locomotion, unlike nimodipine or the combination of both drugs.Cervo and Samanin (1996) examined various DAergic and glutamatergic antagonists.The D2 antagonist sulpiride, administered during the conditioning phase, did not completely prevent conditioned locomotion but partially reduced it at a higher dose when given prior to the conditioning test.A similar outcome was observed with the D1 antagonist SCH 23390.These results suggest "that conditioned locomotion by stimulants occurs even when locomotion is blocked during conditioning, " a finding consistent with amphetamine studies (Swerdlow and Koob, 1984).The NMDA antagonist MK-801, administered during conditioning, prevented a subsequent conditioned effect but was ineffective when given only prior to the test.Similarly, the AMPA/kainate receptor antagonist DNQX, injected into a ventricle, produced largely similar outcomes, with the CR being reduced at a higher antagonist dose.Systemic injections of another NMDA antagonist, CPP, yielded comparable results (Druhan and Wilent, 1999).
Genetically modified mice lacking DA, NA, or 5-HT reuptake proteins were also tested.No conditioned locomotion was observed in DA transporter knockouts, while reduced conditioned effects were seen in NA and 5-HT transporter knockouts (Hall et al., 2009).These knockout groups exhibited specific behavioral patterns during prior cocaine conditioning: DA transporter knockouts did not respond to the first cocaine challenge but showed increased locomotion in subsequent ones, whereas psychomotor activation in 5-HT transporter knockouts declined with repeated testing and remained relatively stable in NA transporter knockouts.
Clinical relevance: Given the significant role of cocaine as a substance of addiction, many researchers have interpreted their findings on behaviorally conditioned effects within the context of addiction.This includes discussions on drug-seeking behavior, craving, and the impact of drug-associated stimuli (Bridger et al., 1982;Barr et al., 1983;Brown and Fibiger, 1992;Panlilio and Schindler, 1997;Carey and Gui, 1998;Hotsenpiller and Wolf, 2002;Carey et al., 2008).Additionally, some have considered these findings in relation to schizophrenia, noting that chronic abuse of cocaine, similar to AMPH, can lead to psychotic symptoms (Beninger and Herz, 1986;Martin-Iverson and Reimer, 1994).This dual framing underscores the complex implications of cocaine's conditioned effects for understanding both addiction and potential psychiatric comorbidities.

Summary and discussion:
The findings from conditioning studies using cocaine, a classic tool in research on conditioned drug effects, show patterns that largely mirror those observed with AMPH.This similarity likely arises from the common neurochemical pathways these two psychostimulant drugs affect, leading to comparable behavioral outcomes such as increased locomotion and stereotypies in response to acute drug effects and sensitization.Additionally, similar conditioning procedures and behavioral setups have been employed in studies with both substances.However, research involving mice, as opposed to rats, has been more prevalent with cocaine studies.In both species, repeated conditioning trials are necessary to establish a reliable conditioned effect, which typically manifests as increased locomotion and/or specific stereotypies.The CR tends to emerge early and transiently during the test and may extinguish with repeated no-drug tests, though different behaviors may exhibit distinct extinction profiles.Given these nuanced behavioral outcomes, advanced methods for behavioral recording and analysis are recommended.Like with AMPH, sensitization and conditioning can be differentiated, suggesting they are mediated by partly distinct mechanisms.
In terms of the neural mechanisms underlying conditioned cocaine effects, current data suggest a role for glutamate but do not conclusively show that DA activity, for instance in the NAcc, is altered or required for the CR's execution.Furthermore, blocking DA receptors during conditioning does not necessarily inhibit a subsequent CR, even if acute behavioral responses, such as locomotion, are suppressed.This indicates that the execution of what might be considered an UR is not essential for establishing a CR.

Apomorphine
Apomorphine (APO) is a DA receptor agonist primarily acting on D2 receptors (Feldman et al., 1997) and, to a lesser extent, on D1, 5-HT, and adrenergic receptors.Clinically, APO is utilized in the treatment of Parkinson's disease to enhance motility during off-phases (Pessoa et al., 2018).In veterinary medicine, it is employed to induce vomiting in intoxicated dogs (Auffret et al., 2018), although such emetic effects are not observed in rodents.Unlike AMPH and cocaine, APO does not possess significant abuse potential.
The neural and behavioral impacts of APO on rodents are highly dose-dependent.At low doses, such as 0.05 mg/kg, APO predominantly stimulates D2 autoreceptors, which reduce DA release and synthesis (e. g., Zetterström and Ungerstedt, 1984).At higher doses, like 0.5 mg/kg or more, APO mainly activates postsynaptic DA receptors, effectively mimicking the postsynaptic actions of DA.These differential pharmacokinetic effects lead to varied behavioral outcomes: low doses typically result in hypomotility (sometimes manifesting as akinesia) and ptosis, while higher doses induce behaviors such as sniffing, licking, and gnawing, often in a stereotypic manner (Robertson et al., 1986).
While the focus here is on rats as test subjects, it is noteworthy that conditioned drug effects with APO were initially reported in dogs (Pavlov, 1910) and have also been studied in mice (Jodogne, 1990) and pigeons (Lindenblatt and Delius, 1987); these findings, however, will not be detailed further.
The US -Effective drugs and dosage regimens: Different doses of APO were explored, ranging from 0.05 mg/kg (Braga et al., 2009a) to 5.0 mg/kg (Schiff, 1982).Most studies involved repeated injections, from three to as many as ten (Schiff, 1982), typically administered daily and subcutaneously.Following repeated injections, some researchers (Möller et al., 1987b) noted changes in drug-induced responses; effects not only intensified (e.g., Möller et al., 1987a) but also occurred more swiftly and shifted from behaviors typical of moderate doses (such as sniffing) to those associated with higher doses (like licking), suggesting both sensitization and potentially some form of conditioning.
Welsch-Kunze et al. (1988) found that the CR was more pronounced following prior treatment with 2.0 mg/kg than with 0.5 mg/kg, and that fewer injections might be more effective than more.Unfortunately, the interval lengths between injection and environmental exposure, as well as subsequent exposure durations, were often not explicitly reported but are described as being longer than 20 minutes in most cases.These temporal aspects could be crucial: Hiroi and White (1989) observed that the acute effect of APO on stereotypies peaked around 15-25 minutes post-injection.Consequently, they exposed their subjects to the conditioning environment for 10 minutes, specifically during the 10-20 minutes after drug administration, and noted a substantially stronger CR compared to animals with exposure times of 45 minutes.This suggests that the US corresponds to the drug's peak effect in terms of DA receptor stimulation, resulting in the climax of stereotypies (UCR).Santos et al. (2015) demonstrated that postsynaptic doses of APO administered shortly after exposure to a distinct context were also effective in establishing a CR in locomotion, indicating that trace conditioning can be effective in the context of conditioned drug effects, at least with APO.Moreover, conditioned behavioral activation (i.e., horizontal and/or vertical activity) could be established using the D2 agonist quinpirole as well as the D1 agonist SKF 38393 (Mazurski and Beninger, 1991).
Features of conditioning: Various experimental setups were employed to pair with acute drug actions, including open fields of different sizes (e.g., Hiroi and White, 1989;Braga et al., 2009a), a "5-gallon aquarium" (Schiff, 1982), small metal wire cages (Möller et al., 1987a), and specific Faraday cages for studies requiring behavioral outcomes to be integrated with electrophysiological recordings (Kropf et al., 1991).Many studies did not add specific additional cues to these environments, while others combined a wire cage with a specific olfactory and a tone cue, demonstrating that this compound stimulus was more effective at inducing a strong CR than either tone or smell alone (Ferger and Kuschinsky, 1995).In experiments investigating the roles of individual cues, the same type of wire cages was used; however, animals were habituated to the cages beforehand to minimize the potential for associating the cage with drug effects -a procedure akin to latent inhibition.Typically, during acquisition, experimental subjects received APO paired with the conditioning environment and a vehicle (usually saline) several hours later in their home cages.Pseudo-conditioned animals, serving as major controls, received the opposite treatment: vehicle prior to exposure to the conditioning environment and APO several hours later in the home cage.Some studies involved single-housed versus group-housed subjects, but there is no conclusive evidence on how such housing conditions might influence APO-dependent conditioning.
Diverging from other studies that relied on systemic injections of APO, Pinheiro Carrera et al. (1998) explored whether intra-neostriatal injection of small APO doses paired with a specific testing environment could establish a CR.They found that repeated pairings effectively induced conditioned drug effects on locomotion.Notably, these repeated pairings of drug and context did not lead to sensitization of acute drug effects, further supporting the notion that conditioned effects of drugs like APO can be distinguished from sensitization mechanisms.
Features of testing and conditioned responses: Behavior was monitored through various methods, including visual observation techniques (e.g., Schiff, 1982;Möller et al., 1987a;Braga et al., 2009a), automated devices with infrared systems (Mattingly and Gotsick, 1989), or video tracking (e.g., Carrera et al., 1998).Automated systems offer advantages in quantifying locomotion, while stereotypies, such as sniffing or grooming, require meticulous observation by experienced observers to investigate specific behavioral details.In some studies, data on stereotypy (like sniffing, licking, or grooming) were presented separately (Schiff, 1982;Möller et al., 1987b), while others combined stereotypic behaviors into sum scores (Welsch-Kunze et al., 1988) or scored them based on durations (Möller et al., 1987b).These varying methodologies have led to findings based on either a single measure, typically locomotion, or multiple measures, presented either separately or in combination.
Most studies provided evidence for significant conditioning, not limited to a specific postsynaptic dose or behavioral measure (but see Mattingly and Gotsick, 1989).Schiff (1982) reported conditioned stereotypies only when combining several dose groups.All other publications found significant CRs within individual experimental groups.Often, qualitative similarities between the UCR and CR were noted (like Möller et al., 1987b), which, depending on the drug dose and methodology, were either stereotypies (e.g., Möller et al., 1987b) and/or hyperlocomotion (e.g., de Matos et al., 2010).Quantitatively, the CR was observed to be weaker compared to the UCR (see Hiroi and White, 1989).
Autoreceptor doses, such as 0.05 mg/kg, were tested less frequently and, when tested, did not produce a CR (Braga et al., 2009a, b).These doses of APO inhibit DA release and likely prevent conditioning of their behavioral outcomes, such as hypoactivity.Santos et al. (2015) interpreted this as evidence that endogenous DA activity is necessary for learning, in this case, acquiring a conditioned drug effect.Interestingly, a slightly higher dose, 0.18 mg/kg (Nowak and Kuschinsky, 1987), which might also affect postsynaptic receptors, led to conditioned effects with complex qualitative and temporal characteristics.Behavioral profiles varied not only among animals but also within subjects over time, showing alternation between low-dose features (like ptosis) and high-dose features (like sniffing).The authors speculated that "signs of DA receptor stimulation in one brain region might compete and alternate with those produced by stimulation in another area." More recently, USV emission has been explored as an additional measure of a CR, since this type of appetitive vocalization can be elicited by both AMPH and postsynaptic APO doses (2.0 and 4.0 mg/kg) in intact subjects (Simola et al., 2021).However, no evidence for conditioned vocalizations was found, contrasting with results from animals with unilateral lesions of the nigro-striatal DA system (Simola et al., 2021).
Extinction, re-conditioning, and reversal training: Extinction, reconditioning, and reversal training have not been regularly explored in studies on the conditioned drug effects of APO in intact subjects, primarily because many researchers conducted only a single no-drug test following APO treatment.Nonetheless, the limited evidence available suggests that extinction does occur with repeated retesting.Welsch--Kunze et al. (1988) reported gradual extinction over three repeated tests with vehicle injections, during which transient periods of drug-like behaviors were still observed.Moreover, the course of extinction did not follow a uniform pattern; conditioned behaviors often manifested earlier after vehicle injection, typically as sniffing, while gnawing episodes appeared later.The authors concluded that "extinction of conditioned dopaminomimetic effects is not a continuous process but involves phasic phenomena."De Matos et al. ( 2010) demonstrated rapid extinction of increased locomotion induced by repeated APO treatments, although sensitization to APO was still evident in a subsequent challenge with the drug.Furthermore, reversal training was shown to be effective.
Pharmacological manipulations, conditioned neural responses and possible neural mechanisms: Regarding potential conditioned neural responses, preliminary evidence indicated DAergic involvement in the CR, as observed by Schiff (1982).Post-mortem samples collected shortly after evocation revealed higher levels of homovanillic acid (HVA), a DA metabolite, in the neostriatum and olfactory tubercle/NAcc tissue of conditioned rats compared to pseudo-conditioned animals, suggesting the CR might involve increased DA turnover.However, since HVA is a DA metabolite, it has since been recognized that changes in HVA concentrations cannot definitively pinpoint a specific DAergic mechanism.A subsequent study analyzing neostriatum and NAcc tissue samples failed to find differences in DA levels or its primary metabolite DOPAC between conditioned and pseudo-conditioned rats.
The role of DA was further examined pharmacologically, showing that the DA receptor antagonist haloperidol (0.1 mg/kg) only partially reduced conditioned stereotypies, even though the same dose significantly mitigated APO's acute effects (Welsch-Kunze et al., 1988).Utilizing pimozide, an antagonist of D2, D3, D4, and 5-HT7 receptors, blocked both APO's acute effects and the CR, suggesting that the outcomes were not due to motor side effects and that the CR necessitates activation of postsynaptic DA receptors by increased DA release (Hiroi and White, 1989).The D2 antagonist sulpiride only blocked the expression of APO-conditioned locomotor stimulant effects at higher doses, whereas the D1 antagonist SCH-23390 could inhibit the CR at both lower and higher doses.Since both doses also reduced general locomotion, a nonspecific effect of D1 blockade could not be ruled out (Dias et al., 2010).These findings have yielded somewhat mixed results, likely due to variations in methodological factors such as APO dosage during conditioning or the choice of dependent measures, like stereotypies versus locomotion.
The role of glutamate was also explored in a repeated conditioning paradigm with 2.0 mg/kg APO, using two glutamate receptor antagonists in a dose-dependent manner: the selective NMDA antagonist MK-801 and MLV-6976, which also blocks kainate and quisqualate receptors, respectively (Welsch-Kunze and Kuschinsky, 1990).Administering both antagonists during acquisition did not prevent the later expression of the CR.However, MK-801, when injected before the evocation phase, inhibited the CR's expression in both locomotion and stereotypies, suggesting that glutamatergic mechanisms might not be crucial for CR acquisition with APO but could play some role in evocation.
Furthermore, in-vivo recordings of electrical brain activity during acquisition and evocation using APO as the unconditioned stimulus revealed changes during conditioning and the subsequent test: there were drug-induced increases in locomotion and stereotypies, accompanied by episodes of increased alpha power in the cortex (Kropf and Kuschinsky, 1991) as well as the hippocampus and neostriatum (Kropf and Kuschinsky, 1993), which subsided after repeated and unreinforced exposure to the conditioned stimulus.
Clinical relevance: Although APO is utilized clinically, particularly for the symptomatic treatment of Parkinson's disease, the potential clinical relevance of APO-induced behaviorally conditioned effects has been infrequently discussed in the aforementioned studies, which are largely categorized as basic psychopharmacological research.Connections to addiction have been sporadically explored (Welsch-Kunze et al., 1988).However, APO is not commonly abused by humans and exhibits relatively mild effects in animal models of addiction (e.g., Wise et al., 1976), where drugs with significant DAergic effects, such as AMPH or cocaine, are the primary subjects of research due to their clinical relevance.
Summary and discussion: APO, unlike AMPH or cocaine, acts directly on DA receptors in the brain.The stimulation of pre-or postsynaptic DA receptors by APO, which varies depending on the dose, results in rather distinct behavioral profiles.For behavioral conditioning, low doses appear inefficient, likely because they reduce DA release via stimulation of autoreceptors.In contrast, higher doses stimulate postsynaptic sites and are more effective in producing a CR.Often, such postsynaptic doses are administered multiple times, but achieving a conditioned outcome may not necessitate numerous drug/environment pairings.The timing of pairing the conditioning environment with specific drug actions seems crucial, and conditioning may occur even if the drug is administered shortly after exposure to a distinct environment.This suggests that APO studies have explored aspects of conditioning that have not been thoroughly examined in psychostimulant research.Moreover, many APO studies underscore the importance of detailed behavioral analyses, as different drug effects, such as locomotion or specific stereotypies, may be subject to conditioned effects or exhibit unique extinction profiles.
The involvement of brain DA in conditioned effects with APO remains uncertain: Neurochemical studies are relatively limited and have yielded inconsistent results.Furthermore, pharmacological investigations have not provided clear patterns, as DA antagonists produce variable effects during acquisition and evocation phases, indicating the importance of endogenous DA in the establishment and/or expression of conditioned drug effects.There is some evidence suggesting the involvement of glutamate, particularly the role of NMDA receptors in the expression of conditioned drug effects with APO.Clearly, further mechanistic research is necessary to elucidate the underlying mechanisms of behaviorally conditioned effects with APO.

Dopaminergic receptor antagonists
Conditioned drug effects have been demonstrated not only with DAergic agonists but also, though less frequently, with DA receptor antagonists, particularly the butyrophenone haloperidol, which predominantly blocks D2 receptors.Like DA receptor agonists such as APO, the behavioral effects of haloperidol are significantly dose-dependent, affecting pre-and postsynaptic DA receptor involvement and resulting in distinct behavioral profiles.For behavioral conditioning, lower doses, which reduce DA release via autoreceptor stimulation, seem ineffective, whereas higher doses that also stimulate postsynaptic receptors are more likely to elicit a CR.Most studies involving antagonist-induced CRs used male subjects, leaving open the question of whether similar effects would be observed in females.
Special attention is warranted for experimental catalepsy, defined as an animal's inability to correct an externally imposed posture.Unlike locomotion and stereotypy, which can be assessed through observation, catalepsy requires direct interaction with the animal, such as in the bar test, which measures how long it takes for the rat to step down from an imposed posture.The cataleptic effect, typically measured after a delay to allow the drug to act, requires specific testing conditions, not always detailed in scientific reports.
Studies have shown behaviorally conditioned drug effects with haloperidol, based on repeated treatments with doses ≥ 0.1 mg/kg.During acquisition, sensitization occurred, that is, increasing catalepsy scores with repeated drug injections, while exposure to the CS during evocation induced increased catalepsy compared to controls.However, detailed behavioral analyses have yielded exceptions, and specific features of the CR procedure were rarely investigated.Some studies varied exposure time to the conditioning context and the interval between injection and testing, finding that longer exposure times and shorter intervals led to more efficient conditioned catalepsy but did not affect sensitization similarly.
In contrast to the D2 antagonist haloperidol, conditioned hypoactivity was not induced by the D1/D2 antagonist chlorpromazine but was by the D1 antagonist SCH 23390, suggesting that intact function of D1 receptors is required for conditioned catalepsy.However, contradictory results regarding SCH 23390 between different studies and species suggest that methodological differences or species-specific responses may play a role.
Some studies have indicated that repeated treatment with haloperidol can lead to tolerance in catalepsy rather than sensitization, a context-dependent outcome not elicited by a subsequent challenge injection with haloperidol in a new context.This contrasts with studies showing sensitization and is likely due to substantial methodological differences, such as lack of behavioral testing during the acquisition phase and the frequent administration of relatively high doses.
Summary and discussion: DA receptor antagonists like haloperidol have been less studied for behaviorally conditioned effects compared to agonists such as AMPH, cocaine, or APO.Nonetheless, substantial evidence demonstrates behaviorally conditioned hypoactivity or catalepsy with haloperidol as a US, which can sensitize with repeated administration under certain conditions and exhibit CRs that may extinguish with repeated evocation trials.The fact that conditioned effects are achievable with both DA agonists and antagonists challenges the assumption that DA activity is crucial for conditioning, suggesting the need for further neurobiological investigations, including the role of glutamate.

Morphine (and other opiates)
The alkaloid morphine, the principal psychoactive component of opium, acts as a potent agonist at mu-opioid receptors in the brain and spinal cord (Devereaux et al., 2018).As a highly effective analgesic, morphine can also cause respiratory depression and constipation and possesses a significant potential for addiction.R.K.W. Schwarting et al.Effects on locomotion and stereotypies are commonly observed after systemic administration of morphine, with these effects being dosedependent and exhibiting complex qualitative and quantitative changes over time.A dose of 1 mg/kg in rats leads to moderately enhanced locomotion and increased rearing, along with significant increases in grooming, while higher doses result in sedation (5 mg/kg) or even catalepsy (20 mg/kg) (Fog, 1970).Higher doses of morphine (10 and 20 mg/kg) initially cause behavioral depression, followed by delayed excitatory effects around 2-3 hours after injection, illustrating the drug's complex behavioral profile (Babbini and Davis, 1972).Contrary to developing tolerance in behavioral activation, locomotion intensifies progressively over up to 48 days and begins to peak earlier post-injection.Ayhan and Randrup (1973) observed hyperactivity (locomotion, rearing, grooming, eating, drinking) with doses of 1 or 2 mg/kg, manifesting around 20 minutes after injection and peaking after about an hour, differing substantially from the effects of the psychostimulant AMPH, which at a dose of 1 mg/kg led to continuously increased psychomotor activation, in contrast to morphine's alternating bursts of activity and periods of sedation.
Regarding conditioned morphine effects, Lynch et al. (1976) summarized around 70 years of research, beginning with Edwin Stanton Faust (1900), who possibly published the first scientific report.Faust observed that dogs, after long-term treatment with subcutaneous morphine, displayed increased levels of arousal close to the typical injection time, behaving as if they eagerly anticipated the drug injection, including "lebhafte Freudensbezeugungen" ("lively expressions of joy").Pavlov (1910) reported conditioned nausea, salivation, vomiting, and sleep after repeated morphine treatment, i.e., presumably conditioned symptoms mirroring responses to the initial drug injections.
In addition to these historical accounts, Lynch et al. (1976) reviewed 28 studies, concluding that conditioned drug effects seem to require only a few morphine trials and can be established across diverse species and paradigms.These effects can be detected using various dependent measures, and the CR may not necessitate the same response as during acquisition or may even be opposite.Lynch et al. defined three research epochs: an initial phase using dogs, subcutaneous administration, and salivation as the main measure; a shift towards rats and the exploration of brain mechanisms in morphine conditioning; and a third phase focusing on conditioning's role in morphine addiction, particularly (conditioned) tolerance as a factor in addictive behavior, with rats and primates as prevalent experimental subjects.
Lynch et al. also highlighted unresolved questions, such as the impact of doses, trial numbers, and administration routes on the acquisition of different CRs, the relationship between these responses and addiction, extinction rates, enhancement or prevention of extinction, and the generalizability of findings across species.They noted the necessity for further research, including human studies, without providing specific methodological recommendations for these issues.Some relevant studies published before Lynch et al.'s review, as well as more recent publications on conditioned morphine effects, will be reviewed next, categorized by their primary behavioral variables: behavioral activation (locomotion, motility, stereotypy) and pain.

Behavioral activation
The US -Effective drugs and dosage regimens: The studies reviewed were conducted exclusively on male rats, with doses and methods of drug administration varying significantly across the research.The lowest effective dose reported was 1 mg/kg (Kamat et al., 1974), but most studies administered higher doses, typically ranging from 5 to 15 mg/kg (Mucha et al., 1981;Hayashi et al., 1980;Walter and Kuschinsky, 1989).These doses were delivered through various routes, including ip (Walter and Kuschinsky, 1989), sc (Kamat et al., 1974), iv (Perez-Cruet, 1976), or via sc catheters (Mucha et al., 1981).Additionally, the frequency and pattern of injections varied, such as 3, 6, or 10 daily injections (Mucha et al., 1981;Kamat et al., 1974;Hayashi et al., 1980) or 8 injections every other day (Walter and Kuschinsky, 1989).The duration of exposure to the environment post-injection ranged from as brief as 2 minutes (Mucha et al., 1981) to as long as 180 minutes (Walter and Kuschinsky, 1989;Hayashi et al., 1980).Typically, subjects were first introduced to the drug environment, then injected, and immediately re-exposed to the same environment (Kamat et al., 1974).The sc catheter technique employed by Mucha et al. (1981) facilitated drug administration without the need for manual injection.Kamat et al. (1974) observed hyperactivity 1-2 hours after drug injections (1 mg/kg; also reported by Hayashi et al., 1980), while Mucha et al. (1981) noted reduced locomotion on the first day, which did not persist.Walter and Kuschinsky (1989) reported hypo-activity during the initial two days of drug administration, followed by increased locomotion and gnawing thereafter.
Features of conditioning: For conditioning environments, smaller plastic or metal cages (Kamat et al., 1974;Walter and Kuschinsky, 1989) or small open fields (Hayashi et al., 1980;Mucha et al., 1981) were utilized.In addition, various specific cues were implemented as additional CS, such as flickering lights (Hayashi et al., 1980), tone/taste combinations (Kamat et al., 1974), or noise/smell combinations (Walter and Kuschinsky, 1989).Perez-Cruet (1976) employed a 30-second buzzer preceding repeated intravenous injections of morphine (10 mg/kg) or methadone.Mucha et al. (1981) did not incorporate such specific cues, and their catheterized injection procedure did not necessitate interaction with the test subjects, rendering their approach more contextual in nature.Behavioral outcomes were either assessed through observation (Walter and Kuschinsky, 1989;Mucha et al., 1981) or by using automated activity devices (Kamat et al., 1974;Hayashi et al., 1980).The details of behavioral outcomes across repeated drug injections were not consistently detailed.The management of control groups or procedures also varied, from studies without any control group (Kamat et al., 1974), to those with drug-treated groups either with or without specific cues (Hayashi et al., 1980), or including pseudo-conditioned controls (Mucha et al., 1981;Walter and Kuschinsky, 1989).

Features of testing and conditioned responses:
In most studies, the conditioned effect was tested after saline injections, with the examination occurring up to 7 days after the conclusion of drug treatment (Walter and Kuschinsky, 1989).In some cases (Kamat et al., 1974;Hayashi et al., 1980), the saline injection was preceded by a 30-minute no-treatment phase without any injection, followed by a longer post-injection test phase.These methods enabled the investigation of the effects of the context alone compared to those of the CS and context combined.
The CRs observed were consistent with the UCRs, specifically increased activity.However, unlike the UCR, this increase appeared earlier post-injection and lasted for a shorter duration (Kamat et al., 1974;Hayashi et al., 1980).Other researchers reported an increase in locomotion, rearing, and gnawing during the evocation phase (Walter and Kuschinsky, 1989).Generally, dose/response issues were infrequently addressed, with reports of dose-related effects in locomotion using the opioid etorphine as the US (Mucha et al., 1981).More recently, rat 50-kHz USVs were employed as outcomes following behavioral conditioning with morphine.Hamed et al. (2012) found that rats emitted 50-kHz USVs when exposed to a) a context and b) a subsequent saline injection two weeks after repeated morphine experiences in that context, interpreting this as an indication of drug reward anticipation.
Extinction, re-conditioning, reversal training: These factors did not receive specific attention except that Hayashi et al. (1980) reported extinction of increased motility over the course of five unreinforced CS exposure days.

Pharmacological manipulations, conditioned neural responses and possible neural mechanisms:
The available evidence on conditioned neural responses to morphine is limited and lacks systematic investigation.Perez-Cruet (1976) observed a post-mortem increase in striatal homovanillic acid (HVA) levels (and to a lesser extent, DA) following a saline/buzzer test, after a conditioning phase that involved iv morphine paired with a buzzer over several days.This increase, compared to rats with a history of morphine exposure not paired with the buzzer CS, was similar to the unconditioned effects of morphine, suggesting an increase in DA turnover.In contrast, Walter and Kuschinsky (1989) did not find evidence for conditioned changes in DA or its metabolite DOPAC in the neostriatum or nucleus accumbens.Vezina and Stewart (1984) provided evidence supporting the role of the DA system by administering morphine directly into the ventral tegmental area (VTA).They observed both unconditioned increases (and sensitization) in locomotion and conditioned locomotion following saline injection when paired with the previous conditioning context.

Clinical relevance:
The morphine studies mentioned above are typically viewed as foundational research, with little to no direct mention of explicit clinical implications.Only a few studies have contextualized their findings within the framework of addiction (Vezina and Stewart, 1984;Hamed et al., 2012).
Summary and discussion: This summary focuses on studies published after the review by Lynch et al. (1976).Research on the psychomotor effects of morphine presents a more complex scenario compared to drugs like AMPH, cocaine, or APO, as the acute effects of morphine are dose-dependent and exhibit intricate patterns over time post-injection.This complexity makes defining an unconditioned behavioral response challenging.Similar to AMPH, cocaine, or APO, sensitization can develop with repeated morphine administrations.Various doses of morphine, administered through different parenteral routes, have been shown to be effective in eliciting a behaviorally conditioned response.Some administration methods offered the benefit of requiring no manual interaction with the drug, thereby eliminating this methodological variable.Conditioning experiments were conducted in relatively small environments or cages, a choice possibly rooted in historical practices, as single housing has become less common.While specific CS were utilized, the evidence suggests that the context provided by the testing environment alone is sufficient.Data on the extinction of conditioned behavioral activation with morphine as the US are limited.Moreover, only a few neurobiological studies have concentrated on DA mechanisms, with preliminary evidence indicating the involvement of the meso-striatal system in learned behavioral activation.

Pain modulation
The research published since 1975 can be categorized into studies involving healthy subjects and those investigating a painful disease-like condition (Keller et al., 2018).Experiments with healthy animals were conducted on both rats and mice, while studies focusing on disease-like conditions exclusively used rats.The majority of these studies involved male subjects (although Guo et al., 2010Guo et al., , 2011;;McNabb et al., 2014 included female subjects); only a single study assessed both male and female subjects (Boorman and Keay, 2021).The role of sex clearly deserves more attention in future studies, because Martin et al. (2019) showed that male but not female CD1 mice displayed context-dependent conditioned pain sensitivity which was dependent on testosterone function (see also Trask et al., 2022).

Pain modulation in healthy animals
The fundamental concept of these studies was to determine whether experiences with morphine, when paired with a specific context or certain CS, would result in a CR of drug-like hypoalgesia upon reexposure to the CS during the evocation phase.
The US -Effective drugs and dosage regimens: Morphine was administered in doses ranging from 3 mg/kg (Valone et al., 1998) to 15 mg/kg (Nolan et al., 2012), either ip or sc.Dose-response relationships were explored by Valone et al. (1998), who reported an inverted U-shaped outcome regarding conditioned analgesic effects in a hot-plate test, indicating successful conditioning with 10 mg/kg, but not with 3 or 30 mg/kg.Typically, the drug was administered repeatedly, often around 3-4 times (Randall et al., 1993) up to 9 times during the acquisition phase (Krank et al., 1981).There is only one study that achieved successful one-trial conditioning using the highly potent opioid fentanyl instead of morphine (Bryant et al., 2009).Regarding experimental designs, within-group approaches were relatively common.In these designs, the drug was administered at one time in environment A, and a vehicle was given at another time in environment B, either on alternate days (Randall et al., 1993;Valone et al., 1998) or on an am/pm basis (Zhang et al., 2013).Pseudo-conditioned controls were not typically included in these designs (except in Valone et al., 1998;Krank et al., 1981).Additionally, the potential for stress-and fear-conditioned analgesia must be considered in rodent placebo studies, as the drug administration procedures themselves can have aversive/painful properties, potentially confounding the effects attributed to placebo analgesia.Thus, what is interpreted as placebo analgesia might actually be stress-/fear-conditioned analgesia, or a combination of placebo and stress/fear effects.
Features of conditioning: Various methods were utilized to implement conditioning regimens.While most studies used hot-plate tests for outcome measures, they varied in their approach during conditioning, with animals either a) not exposed to hot plates (Krank et al., 1981), b) exposed to unheated plates (Randall et al., 1993;Bardo and Valone, 1994;Valone et al., 1998), or c) exposed to heated plates (Siegel, 1975;Guo et al., 2010Guo et al., , 2011;;Zhang et al., 2013).Consequently, only a few studies assessed whether the administered dose of morphine had an unconditioned analgesic effect and if this effect changed with repeated injections.Interestingly, studies with repeated drug/pain pairings reported varying results.Decreasing (and context-dependent) analgesia, suggesting tolerance, was observed with repeated pairings (Siegel, 1975, in rats; also see Siegel, 1976), in contrast to constant levels of analgesia in studies without repeated pairings (Guo et al., 2010(Guo et al., , 2011, in mice;, in mice;Zhang et al., 2013, in rats).The reasons for these different outcomes are unclear, but methodological differences, such as the lower morphine dose used by Siegel (1975) compared to the other studies, might contribute.Additionally, none of the studies analyzed different behavioral responses during acquisition and evocation phases, thus lacking evidence on whether the drug approach led to hyperactivity, hypoactivity, or other behavioral changes that could influence pain-related outcomes during subsequent testing.An exception is the study by Boorman and Keay (2021), which also assessed locomotion and rearing under morphine in a model of lasting neuropathic pain.
The conditioning procedures varied significantly across studies.Some began with a 15-minute exposure to a specific test cage environment combined with a distinct odor, followed by injection, another odor/context phase (e.g., 30 minutes), and then brief exposure to an unheated hot plate (Randall et al., 1993;Valone et al., 1998).Others started directly with the drug injection, followed by context exposure (30 minutes), and concluded with exposure to a heated hot plate (Guo et al., 2010(Guo et al., , 2011;;Zhang et al., 2013).Miller et al. (1990) combined morphine administrations with a taste aversion approach, where rats developed not only an aversion to a saccharin taste paired with morphine but also conditioned analgesia (also see Bardo and Valone, 1994).
In a unique approach, rats were individually housed in cages within separate cabinets (Krank et al., 1981), with a conditioning trial starting by opening the cabinet doors (combined with overhead illumination and cabinet fans off for 15 minutes), followed by morphine injection and another 30-minute phase with open doors.The final hot-plate test was conducted outside the cabinet.In other studies, animals were exposed to environments different from their housing conditions, such as small steel cages (Randall et al., 1993) or commercial place compartments (Guo et al., 2010), often combined with distinct odors (Randall et al., 1993) or specific illuminations (Guo et al., 2010).Drugs were administered at various intervals, including every 48 hours (Siegel et al., 1975), between 24 and 72 hours (Krank et al., 1981), every other day interspersed by vehicle trials combined with different cues (Randall et al., 1993;Valone et al., 1998), or twice daily with alternations of drug and vehicle injections (Guo et al., 2010(Guo et al., , 2011;;Zhang et al., 2013).
Features of testing and conditioned responses: Tests for conditioned outcomes often occurred the day after the last conditioning trial (Randall et al., 1993;Valone et al., 1998;Guo et al., 2010Guo et al., , 2011;;Zhang et al., 2013) or 48 hours or 4 weeks later (Siegel, 1975).The exposure experience with the hot plate varied between studies; the hot plate was either completely novel (Krank et al., 1981), novel in that it was now heated (Randall et al., 1993;Valone et al., 1998), or entirely familiar (Guo et al., 2010(Guo et al., , 2011;;Zhang et al., 2013) during testing.The impact of such pre-exposure might be crucial for test outcomes, as Lee et al. (2015) demonstrated that pairing non-drugged, distinct environments with low (45 degrees Celsius) or high (50 degrees Celsius) heat experiences can lead to less pronounced pain responses to a high heat test stimulus if administered in an environment previously associated with low heat, interpreted as an effect akin to placebo analgesia.Xu et al. (2018) also explored the role of non-drugged context-paired low or high pain experiences, finding that, in contrast to Lee et al. (2015), the high pain experience led to conditioned hyperalgesia since their female rats exhibited shorter paw lick durations not only during repeated high heat exposures but also during subsequent conditioning tests.
The conditioned outcomes varied across studies, with two reporting hyperalgesic effects as CRs (Siegel, 1975;Krank et al., 1981), in contrast to more recent studies documenting analgesic CRs (Miller et al., 1990;Bardo and Valone, 1994;Randall et al., 1993;Valone et al., 1998;Bryant et al., 2009;Guo et al., 2010Guo et al., , 2011;;Nolan et al., 2012;Zhang et al., 2013;Yin et al., 2020).Krank et al. (1981) observed that the conditioned hyperalgesic response was similar whether rats underwent 3 or 9 conditioning trials, suggesting these responses might reflect conditioned compensatory reactions to morphine treatments.These outcomes could indicate conditioned tolerance (Siegel, 1975) or that environmental stimuli associated with morphine treatments lead to a CR that "induces a heightened sensitivity to nociceptive stimulation" (Krank et al., 1981).However, such a conclusion may not apply universally across morphine studies, as most found conditioned analgesia rather than hyperalgesia.Cho et al. (2021) highlighted the importance of specific social factors in rodent pain research, potentially relevant to placebo studies, such as the "within-order-of-testing effect": Subjects (in this case, mice) taken first from a group cage may exhibit a lower pain response than those tested later, likely due to some form of social transmission of their prior pain experience within the group cage, sensitizing the responsiveness of subjects tested subsequently.
Extinction, re-conditioning, reversal training: Extinction of conditioned hyperalgesia was observed over four consecutive daily saline tests (Siegel, 1975).Randall et al. (1993) provided evidence for the extinction of conditioned analgesia as early as the first retest with saline.
Pharmacological manipulations, conditioned neural responses and possible neural mechanisms: Neurobiological mechanisms underlying opioid-conditioned drug effects have not been extensively explored, primarily focusing on opioid pathways.Guo et al. (2010) demonstrated that the opioid antagonist naloxone, when administered systemically prior to evocation, blocked the conditioned analgesic effect of morphine (Nolan et al., 2012;Miller et al., 1990), unlike the conditioned analgesic effects observed with aspirin, a non-steroidal anti-inflammatory and moderately analgesic drug (Guo et al., 2010).Increasing opioid receptor sensitivity through chronic naloxone treatment (as opposed to acute treatment blockade) was found to enhance the conditioned analgesic effect (Miller et al., 1990).Zhang et al. (2013) reported that naloxone's effect was dose-dependent, exhibiting an inverted-U-shaped curve, and that this effect could also be triggered by direct injection into the anterior cingulate cortex, a critical component of the pain matrix.This intracranial effect was also observed with microinjection of D-Phe-Cys--Tyr-D-Trp-Orn-Thr-Pen-Thr-NH2, a selective mu-opioid receptor antagonist, but not with naltrindole or nor-binaltorphimine, which are selective sigma-and kappa-receptor antagonists, respectively.Interestingly, naloxone (and the DA receptor antagonist haloperidol) could block non-drugged experience-dependent analgesia (Lee et al., 2015), associated with reduced c-fos expression in the anterior cingulate cortex and an increased density of tyrosine hydroxylase-positive (presumably DAergic) cells in the VTA.Naloxone also negated the non-drugged experience-dependent hyperalgesia reported by Xu et al. (2018), suggesting that opiates may play a role in conditioning mechanisms rather than pain per se in these tests.Furthermore, Xu et al. (2018) demonstrated that conditioned outcomes could be inhibited by optogenetic stimulation and by suppressing glutamatergic neurons in the medial prefrontal cortex, using optogenetic manipulation techniques.
Clinical relevance: In the majority of studies, particularly those examining the psychomotor aspects of conditioned opioid effects, discussions were predominantly framed around fundamental scientific issues, such as the behavioral conditioning of opioid effects.Clinical implications were typically associated with factors like drug abuse, addiction, tolerance, and withdrawal (Krank et al., 1981;Mucha et al., 1981;Vezina and Stewart, 1984;Bryant et al., 2009;Hamed et al., 2012).

Pain modulation in disease-like conditions
In recent years, various techniques have been employed to analyze behaviorally conditioned analgesic effects in chronic neuropathic or inflammatory pain conditions.These techniques include unilateral ligation of a lumbar spinal nerve (McNabb et al., 2014;Zeng et al., 2018), unilateral constriction of the infraorbital nerve (Akintola et al., 2019), injection of complete Freund's adjuvant into one plantar hind paw (Yin et al., 2020), and unilateral constriction of the sciatic nerve (Boorman and Keay, 2021).
The US -Effective drugs and dosage regimens: Dose-response relationships were explored by Yin et al. (2020), who tested morphine at doses of 2.5, 5, and 10 mg/kg in a hot-plate test.The lowest dose failed to induce analgesia, the intermediate dose produced a "modest" effect, and the highest dose resulted in sedation, which likely interfered with the accurate measurement of analgesia.During the acquisition phase, drug administrations were typically paired with compounds of several CS (olfactory, visual, tactile, gustatory;McNabb et al., 2014;Akintola et al., 2019;Boorman and Keay, 2021) or different rooms (Lee et al., 2015).Zeng et al. (2018) identified the ip injection procedure itself as a conditioning stimulus, but this study did not include other controls commonly used in experiments on conditioned effects.Their "pseudo-conditioned control group" received vehicle treatment (unlike the typical drug treatment in standard pseudo-conditioning studies).

Features of conditioning and testing:
The majority of studies focused on morphine and other opioids, such as loperamide (McNabb et al., 2014), a peripherally acting mu-opioid receptor agonist, and fentanyl, a highly potent mu-agonist (Akintola et al., 2019).Additionally, Zeng et al. (2018) explored the effects of gabapentin, which targets the α2δ calcium channel subunit and is commonly used for neuropathic pain treatment (also mentioned by McNabb et al., 2014).These studies typically assessed drug-induced analgesia during conditioning trials and compared outcomes between affected and non-affected limbs/paws using mechanical (with von Frey hairs; McNabb et al., 2014;Zeng et al., 2018;Akintola et al., 2019) or thermal stimulation (Yin et al., 2020;Boorman and Keay, 2021), conducting acquisition 4-7 times.The sole study examining both male and female subjects (Boorman and Keay, 2021) identified several sex-dependent differences, including stronger analgesic drug effects in females and increased drug-induced locomotion in males, highlighting the need for more systematic investigation of both sexes in future rodent placebo research.Boorman and Keay (2023) recently provided evidence for the role contextual factors during conditioning in a sciatic nerve rat model, but this study did not investigate drugs as US.
Most studies found either no evidence for behaviorally conditioned drug effects (McNabb et al., 2014;Akintola et al., 2019;Yin et al., 2020) or observed conditioned analgesia in subsets of subjects (Zeng et al., 2018;Boorman and Keay, 2021), categorizing them as responders (less than 40% of all subjects tested) or non-responders based on quantitative criteria related to paw withdrawal responses to mechanical (Zeng et al., 2018) or thermal stimulation (Boorman and Keay, 2021).Boorman and Keay (2021) noted that their criteria for distinguishing responders from non-responders were established a priori, not post hoc, and their findings of responder proportions mirrored those in human placebo studies of analgesia.They speculated on reasons for the responder/non-responder dichotomy, particularly individual differences in morphine sensitivity or the encoding of drug and conditioning stimuli relationships, suggesting that some subjects might be more attuned to and responsive to neutral stimuli paired with unconditioned stimuli (Robinson and Flagel, 2009).Notably, they observed that responder subjects exhibited less acute locomotion and rearing in response to morphine, suggesting individual differences in morphine sensitivity and emphasizing the importance of acute drug effects in placebo analgesia studies.Yin et al. (2020) speculated that the lack of conditioned effects might be attributed to the severity of the inflammatory pain model used and that the conditioning paradigm might have elicited an expectation of treatment in the animals, which was then contradicted by the altered treatment on the test day.
Pharmacological manipulations, conditioned neural responses and possible neural mechanisms: One study investigating the neurobiological mechanisms of behaviorally conditioned analgesic effects found that ibotenic acid lesions of the prelimbic cortex, which did not alter the pain response itself, inhibited the conditioned effect (Zeng et al., 2018).Positron emission tomography using labeled deoxy-glucose, conducted shortly after vehicle injection and pain testing during the evocation phase, revealed differential and naloxone-dependent activity patterns in several brain areas, including the prefrontal cortex, NAcc, and periaqueductal gray, between responders and non-responders (Boorman and Keay, 2021).The researchers also examined the rostral ventromedial medulla, which is thought to be part of a final common pathway for placebo-induced analgesia, but found no placebo-dependent effects on c-fos labeling in this area.
Summary and discussion: Studies focusing on behaviorally conditioned analgesic effects in animal models with disease-like and chronic pain conditions have gained popularity in recent years.Unlike research involving healthy rodents, these studies have not consistently provided substantial evidence for behaviorally conditioned analgesia.This variability in findings could be attributed to several factors, such as the choice of pain model, conditioning regimens, or the type of analgesic drug used.However, some evidence supports the significance of subjectdependent factors (Cho et al., 2021) or individual differences, which warrant further investigation, as they could also play a crucial role in human subjects (Nir andYarnitsky, 2015, Graeff, 2021).Additionally, there may be a fundamental distinction between acute and chronic pain models.Acutely inflicted pain, being a novel experience, might be more susceptible to context and conditioning due to the formation of new memories that are more amenable to conditioning compared to chronic pain conditions, which may be associated with more entrenched memories.Nonetheless, such memories might still be susceptible to modification through drug conditioning if the procedure effectively influences the process of reconsolidation as discussed in case of addiction (Milton and Everitt, 2010).

Anti-depressant drugs
As previously discussed, research on DAergic drugs and opioids has important implications for understanding placebo effects, in particularly in the realm of conditioned analgesic effects.Such studies have led to critical insights into how conditioned drug effects can be harnessed to optimize treatment outcome by amplifying the placebo effect (Bingel et al. 2011;Kaptchuk et al., 2020).However, in comparison our understanding of placebo effects and their underlying mechanisms in neuropsychiatric disorders, such as depression remains limited.Despite evidence indicating that improvements observed in placebo groups of antidepressant studies constitute a significant portion of the anticipated drug effects (Rief et al., 2009;Kirsch, 2019) the mechanisms remained underexplored.Therefore, it is crucial to extend the knowledge acquired from studies on DAergic drugs and opioids to the broader context of neuropsychiatric disorders, benefiting from the insights gained.In contrast to DAergic drugs and opioids, other drug classes have received comparatively less experimental focus.This includes Herrnstein's often-cited study (1962) on scopolamine, a muscarinic cholinergic receptor antagonist.In this research, saline injections produced similar (though shorter-lasting) suppressive effects on lever-pressing behavior in an instrumental learning paradigm as scopolamine itself, with these saline-induced outcomes labeled as "a placebo effect" attributable to Pavlovian conditioning.Compared to subsequent rodent placebo studies, these somewhat pioneering findings were derived from a relatively unique experimental approach and were based on a limited sample of just two rat subjects.
Recently, initial attempts have been made to explore whether conditioned drug effects can be elicited in rodent models using antidepressants.In humans, placebo effects are quite common with antidepressant drugs (Moncrieff et al., 2004;Leuchter et al., 2014) and are sometimes even used as a benchmark when assessing the efficacy of potential therapeutic drugs.It's important to note, however, that the clinical (i.e., antidepressant) effectiveness of drugs such as MAO inhibitors, tricyclic antidepressants, or selective serotonin re-uptake inhibitors (SSRIs) requires considerable time to manifest, typically a week or more.This delay suggests that there might not be an immediate contiguity between drug administration and clinical outcome, rendering classically conditioned antidepressant effects rather improbable.Nonetheless, environmental factors have been shown to significantly influence antidepressant outcomes (Rief et al., 2016), potentially including conditioned effects, which may have deterred most researchers from investigating such drugs in rodent models of conditioned drug effects.
The introduction of ketamine as an antidepressant has somewhat altered this perspective.Ketamine, an antagonist of NMDA receptors, can exhibit clinical effectiveness within hours after administration.In rodents, rapid effects post-injection have been observed using tests such as the forced swim test among others (e.g., Maeng et al., 2008;Autry et al., 2011).Moreover, ketamine can induce dose-dependent increases in locomotor activity, which can be inhibited by DA receptor antagonists (e.g., Irifune et al., 1991).
To investigate whether the effects of ketamine could be conditioned, Krimmel et al. (2020) injected male and female mice with ketamine intraperitoneally every two weeks (for a total of three injections) and placed the subjects in an open field containing a small amount of chocolate.Various control groups were established, including repeated ketamine treatments without exposure to the open field.An open-field test with saline was conducted two weeks later, followed by two forced swim tests, one immediately after the open-field testing and another 24 hours later.As anticipated, ketamine increased locomotor activity in both sexes, but there was no evidence for conditioned locomotion (or other open-field parameters) in the saline test.Thus, these outcomes -namely, the absence of conditioned locomotion-differ from those observed with DAergic psychostimulants, potentially due to variations in the neural mechanisms between these drugs or differences in conditioning procedures, such as the relatively long intervals between injections used by Krimmel et al. (2020).Unlike open-field behavior, immobility time in the forced swim test, typically considered a crucial index in depression models (e.g., Kara et al., 2018), was reduced to a degree similar to that in the experimental animals compared to controls acutely treated with ketamine.This effect was observed in male but not female mice.It remains unclear why behavior, specifically locomotion, although consistently elicited by ketamine, was not conditioned, while behavior in the forced swim test, which had never been directly associated with ketamine's effects, was altered in a manner similar to ketamine's effects.Such an outcome, however, is plausible, as prior research with different types of drugs (as discussed above) has demonstrated that conditioned changes in behavior do not necessarily require similar experiences during conditioning.Possibly, the conditioning procedure established a state that could be evoked by relevant conditioned stimuli/context, manifesting behaviorally under appropriate situational demands.
In light of the importance of clinical placebo effects, ketamine, but also other and more traditional antidepressant drugs should receive more attention in the future with respect to possible conditioned drug effects.

General discussion
This review on conditioned drug effects has primarily concentrated on five psychoactive drugs.Four of these drugs (AMPH, cocaine, APO, and haloperidol) are distinguished by their dose-dependent psychomotor effects, specifically the stimulation and/or inhibition of locomotion or stereotypies.Meanwhile, the fifth drug, morphine, is noted not only for its psychomotor effects but also for its analgesic properties.For all these drugs, a considerable number of studies have provided substantial evidence for conditioning, which indicates that conditioning is a rather robust phenomenon.Much of the literature on conditioned drug effects dates back to before the end of the 20th century, which might suggest that most pertinent questions have already been resolved.However, this is not the case.In fact, recent years have witnessed a resurgence of interest in animal studies on conditioned drug effects, underscoring the need to revisit this body of work.Such a review is warranted to identify common themes as well as potential shortcomings and limitations.The following sections will detail key findings associated with the five drugs (AMPH, cocaine, APO, haloperidol, and morphine), leading into a comprehensive discussion that covers the most relevant aspects of animal research on conditioned drug effects, including subjects, pharmacokinetics, conditioning procedures, etc.This discussion will also address factors which had become apparent during the literature research for this review (like sex, strain, age, housing) but were only sometimes stated explicitly in the previous chapters for clarity's sake.
The subjects: The majority of experimental evidence on conditioned drug effects is derived from male subjects, both in rat and mouse models.Studies incorporating female subjects are relatively scarce, and comparisons between male and female subjects are even rarer.Consequently, it's challenging to confidently assert that findings from male subjects are universally applicable to females.Considering the higher incidence of affective disorders among females, this gap in research significantly warrants further investigation, ideally by including both sexes within the same experimental designs.Additionally, the impact of age has not been thoroughly explored, as the research predominantly

Box 1
The Rotating Rat -Investigating Conditioned Drug Effects with Unilateral Lesions.
Conditioning research that employs animals with unilateral lesions of the meso-striatal DA system helps to explore DA's role in Parkinson's disease.This method primarily utilizes neurotoxic models like 6-hydroxydopamine, injected into one brain hemisphere to study the differential effects between the lesioned and intact hemispheres (Schwarting and Huston 1996a,b).Such lesions often result in profound unilateral DA depletion, leading to distinctive turning behaviors that differ significantly between spontaneous and drug-induced conditions.Normally, untreated rats show moderate ipsiversive turning (away from the lesion), but do not exhibit contralateral turning unless induced by drugs.
Drugs like AMPH enhance ipsiversive turning by increasing DA activity in the intact hemisphere.In contrast, apomorphine (APO), a DA receptor agonist, strongly elicits contraversive turning due to the heightened sensitivity of DA receptors in the lesioned hemisphere-a similar effect is observed with L-DOPA, which converts to DA and preferentially activates these supersensitive receptors.
Urban Ungerstedt (1971a,b) was among the first to associate these behaviors with conditioned drug effects.He observed that lesioned rats displayed "paradoxical rotation" (explosive turning towards the non-lesioned side) when stressed or placed in new environments, initially thought to be caused by novelty but later linked to previous drug exposure, suggesting a conditioned response.Silverman and Ho (1981) empirically tested this by pairing APO injections with specific environments, leading to conditioned contraversive turns upon re-exposure, even months later and with lower APO doses.This response was absent in rats where APO and the environment were unpaired, indicating a conditioned rather than purely pharmacological effect.
Effective Drugs and Dosage Regimens (for further details see Table 4): APO and L-DOPA are consistently effective, impacting a range of dosages from autoreceptor levels (0.05 mg/kg) to higher postsynaptic doses (0.5 mg/kg or more).Studies often use multiple pairings of drug and environment to explore both conditioning and sensitization, observing more rapid and intense responses with repeated exposures.
Conditioning Features: Conditioning has been successful in various setups like rotometer bowls and open fields, typically without specific additional cues.Delays between drug administration and exposure to the environment vary, but even brief exposures (about 10 minutes) can establish strong conditioning, supported by designs using pseudo-conditioned controls and differential conditioning strategies.
Testing and Conditioned Responses: Conditioned responses (CRs) are clearer after a few days and can intensify over weeks ("temporal incubation").The intensity of CRs often does not correlate directly with the initial dosages; interestingly, lower doses may produce stronger CRs than higher ones.CRs initiate rapidly upon environment re-exposure but diminish quickly, typically within minutes, influencing the design of test durations and the choice of environments used for conditioning assessments.
Extinction and Re-conditioning: CRs extinguish quickly with repeated testing but can spontaneously recover, facilitating multiple phases of conditioning and testing within the same subjects.This model also allows for reversal training, where a conditioned effect established with one context is extinguished and then re-established with another.
Neural Mechanisms and Clinical Relevance: The precise neural mechanisms are not well understood but involve selective stimulation of D1 over D2 receptors.Clinically, this model is invaluable for exploring drug effects in Parkinson's disease, suggesting that unrecognized conditioning impacts could affect both the efficacy and side effects of treatments like L-DOPA.
In summary, the rotating rat model with unilateral DA lesions offers a unique method to study conditioned drug effects, providing both quantitative and qualitative insights into drug behaviors, significantly aiding the understanding of Parkinson's disease treatment dynamics.focuses on adult or young adult animals.
The influence of strain and stock on conditioned drug effects has not been systematically studied.Most available studies have utilized outbred rodent strains, such as Wistar or Sprague-Dawley rats, with investigations involving inbred strains being uncommon.Comparative studies are particularly lacking, which may be crucial for understanding conditioned drug effects, given that different strains of rats (and mice) can exhibit variations in learning, memory, motivation, and emotional responses.The potential variability in experimental outcomes due to these differences remains to be examined.Even within a single strain, individual phenotypes can exhibit systematic differences in behavioral traits linked to specific neural mechanisms.Typically, subjects are chosen more or less at random, without pre-selection for any specific individual dispositions, though such traits could significantly influence study outcomes.
For instance, Jodogne et al. (1994) pre-tested their animals for activity levels in response to a novel environment, leading to the categorization of rats as either high or low responders.In their study of AMPH conditioning, high responder rats displayed more locomotion than low responder rats in absolute terms, and evidence for conditioning (and sensitization) was primarily observed in high responders.These findings contribute to the understanding of the role of DA in conditioning, supported by a wealth of evidence indicating that high responder rats exhibit higher DA activity in the nucleus accumbens (e.g., Hooks et al., 1992).Lastly, genetically modified animals have yet to be extensively utilized in this area of research (with exceptions like Hall et al., 2009), likely due to the majority of studies being conducted before the advent of such genetic tools.
Housing conditions: Given the extensive history of research in this field, changes in housing conditions over the decades are not unexpected.Initially, many laboratories defaulted to single housing for their subjects.However, recognizing that rats and mice are inherently social creatures, group housing has become more common and is now even mandated by law in many countries, sometimes with the inclusion of enrichment materials.These housing conditions can significantly impact brain function and behavior, including interactions between drugs and the environment, a factor also considered in the context of human psychopharmacotherapy (Rief et al., 2016).Consequently, it is reasonable to speculate that housing conditions might influence experiments on conditioned drug effects.Although it has yet to be explicitly tested whether certain housing conditions facilitate or impede conditioned drug effects, there is considerable evidence that they can affect the unconditioned effects of drugs and their neural underpinnings.This is demonstrated in studies with the antidepressant fluoxetine, where housing conditions were shown to have a significant impact (Branchi et al., 2013;Alboni et al., 2017).Furthermore, the type of housing might play a crucial role in experimental designs employing pseudo-conditioned controls, where saline treatment in experimental animals is paired with the housing environment (and drug treatment in control animals).
Therefore, animals in group housing might experience control treatments within a social context in contrast to single housed animals from earlier studies who experienced control and conditioned treatments individually.Group housing could induce an unintended confounder.However, this issue might be mitigated if control treatments are not associated with the group cage but with a different environment, such as a smaller cage where the animal is placed after injection for a specific duration (e.g. 30 min) before returning to its group cage.Regardless of the precise methodology, it is advisable to maintain consistent social conditions for both experimental and control subjects during the treatment phase.

Factors of conditioning and testing:
Pharmacokinetic considerations play a crucial role in the study of conditioned drug effects.Typically, psychostimulants and other drugs are administered parenterally, particularly via sc or ip routes, and occasionally through iv methods.These administration routes result in   rapid neurochemical changes and behavioral effects, potentially facilitating effective pairing with specific environmental cues and, consequently, more efficient conditioning compared to the slower and more gradual effects of oral administration.The feasibility of utilizing oral applications for inducing conditioned drug effects in animal models remains an area for future research, especially since oral administration is a common clinical method for delivering psychoactive drugs.
A notable limitation of sc or ip injections is the necessity of inserting a needle into the body, often repeatedly, which can be a somewhat aversive procedure that may influence or even hinder the conditioning process.To mitigate this issue, it is recommended that experiments involving such procedures also include extensive handling of the subjects.Implementing mock injections as part of this handling process can help acclimate the subjects to subsequent drug or saline injections.Such procedural details, including the type of syringe and needle used, should be comprehensively documented in the methods section of research reports.Additionally, the specific drug dose (typically expressed in mg/ kg, calculated as either salt or free base), the solvent used (usually saline), and the volume injected per body weight (commonly in ml/kg) should be clearly stated.
Conditioned drug effects can sometimes be established with just a single drug/environment pairing, but more often, multiple pairings are necessary-typically three or more, usually conducted on alternate days.Across these repetitions, the drug-induced behavioral profile may evolve, frequently showing signs of sensitization, such as increased locomotion, or a qualitative shift, for example, from locomotion to stereotypy in the case of psychostimulants.Recognizing these dynamics is crucial when comparing the behavioral profile in tests for conditioned behavior with that observed during conditioning, i.e., when the drug's effects are directly experienced.
Environmental context, conditioned stimuli, behavioral variables: The literature review suggests that a wide range of procedures can effectively establish conditioned drug effects.Incorporating specific conditioned stimuli, such as particular tones or smells, can be beneficial and may even be recommended for certain research questions.However, for foundational experiments, such additional stimuli may not be necessary, as the overall testing context could suffice as a compound stimulus, provided it is sufficiently distinct from a control environment not associated with the drug.Alternatively, to distinguish between the effects of context and specific stimuli (e.g., a tone), one might expose animals to the conditioning context alone for several minutes before introducing the tone and then compare behaviors across these phases and against those of control animals.
In studies focused on psychomotor effects, the choice between using smaller environments, which might limit behavioral variability, and larger ones, like open fields, that allow for a broader range of behaviors, is crucial.The method of behavior recording also warrants consideration.Automated devices can facilitate rapid data collection but may overlook intricate behavioral nuances if the detection method is overly simplistic, such as using a single infrared beam to measure activity.Conversely, advanced video-image analyzing systems can offer detailed analyses, including a wide array of behavioral outcomes, temporal resolution, and dynamics over time.Generally, incorporating multiple measures, such as locomotion, rearing, and various stereotypies, is advisable.Particularly in larger testing environments, analyzing the spatial distribution of specific behaviors can also provide valuable insights.For experiments involving repeated drug/environment pairings, continuous behavioral monitoring is recommended to detect potential sensitization.
Studies involving opiates and pain represent a unique case.Unlike psychomotor-focused research, pain studies necessitate a pain assessment.This requirement introduces at least two significant considerations: the timing and method of pain testing throughout the experiment, and whether the testing situation should be novel or familiar based on previous pain/pain relief experiences.Both approaches have their merits, and the choice should be tailored to the       specific aims of the experiment.Additionally, opiates like morphine can produce complex psychomotor effects in rodents, which may interfere with pain assessments.Although there is no one-size-fits-all solution to this challenge, it is advisable to monitor not only pain responses but also changes in psychomotor activity, such as locomotion, to fully understand the drug's effects.Control conditions: Basically, one has the option to test for conditioned outcomes using within-or between-group comparisons, and an optimal design might include both options.The typical procedure for a between-group design (see Fig. 1) is one with a pseudo-conditioned control group, where experimental and control groups receive identical drug and saline dosing paired either with the conditioning or a different context (traditionally the home cage) and the outcomes are tested under saline in the conditioning context.Here, one has to consider that the experience of saline treatment in the conditioning context might be novel for the experimental group, therefore, conditioned effects might confound with responses to novelty.Alternatively, one can use two distinct environments, one for drug (A) and the other for saline treatment (B) in case of the experimental group (and vice versa in controls).Importantly, behavior is recorded in both environments, so that one can test unconditioned and conditioned effects in both, which allows within-group comparisons (saline effects in environment A versus B) but also between-group comparisons (saline effects in A in experimental versus control animals).This design also does not exclude the novelty problems, since saline in A is still a novel experience for experimental animals.To reduce the impact of novelty, one might start the experiment with a habituation phase to A (and B) in all animals, but this might lead to some latent learning which would have to be overcome by the training and testing procedures.
Extinction is a well-documented phenomenon in which conditioned effects diminish and may even vanish after repeated tests without the drug, and there is evidence suggesting that the rate and extent of extinction are influenced by the prior doses and frequencies of drug administration.Although these findings are indicative, the precise factors driving extinction-whether it is primarily due to repeated no-drug exposures or simply the passage of time-remain to be fully elucidated.Considering the growing interest and research into extinction (and reconsolidation) processes, particularly in the context of fear conditioning, it is recommended to incorporate systematic extinction (and reconsolidation) assessments in studies on conditioned drug effects.This approach would help determine whether the underlying mechanisms of extinction in drug conditioning share similarities with those observed in fear conditioning, among other questions.
Defining US, UCR, CS and CR In numerous conditioning studies, a presumed CS, such as a test cage, is repeatedly paired with a presumed US, like cocaine.Subsequent presentation of the CS alone often elicits a CR that mirrors the UCR, such as locomotion.This pattern aligns with Pavlov's stimulus-substitution theory, which posits that the CS effectively takes the place of the US (Domjan, 2010).However, several studies have demonstrated that performing an assumed UCR, such as locomotion, may not be necessary for developing a conditioned response (Swerdlow and Koob, 1984).Earlier research reviewed by Eikelboom and Stewart (1982) also indicated that the CR sometimes mirrors and sometimes opposes the observed drug effect, with the latter instances termed paradoxical conditioning or conditioned compensatory responses.These findings raise critical questions about the nature of the UR and, similarly, the US-specifically, whether the drug itself should be considered the US.Ramsay and Woods (1997) argued that it is not the substance (e.g., cocaine) that should be seen as the US but rather the disturbance it causes in normal physiological functions.Thus, the UCR is the body's reaction to this disturbance, and the CR is thought to be identical to the UCR.When conditioned outcomes are assessed under saline conditions, the CR persists because it is not obscured by the drug, i.e., US actions.Such theories offer explanations for why specific behaviors induced by drugs, like cocaine-induced locomotion, may not occur during conditioning, as the crucial elements, such as the US and UCR, must be identified within the organism, particularly its brain.However, this perspective introduces several significant challenges, notably that the UCR cannot always be defined behaviorally but rather physiologically, through changes in neural activity within the brain.Unfortunately, these neural mechanisms are not yet fully understood.
Neuronal mechanisms: Given the critical role of DA for learning (Wise, 2004), it is unsurprising that much of the neurobiological and pharmacological research on the neural mechanisms behind the acquisition and expression of conditioned drug effects has focused on this neurotransmitter.Although numerous studies support significant involvement of DA, others have reported negative results, particularly regarding the expression of conditioned effects, and occasionally during acquisition.These discrepancies may not necessarily challenge the importance of DA, owing to various methodological factors.For instance, older studies often depended on post-mortem measures that only reflect the state shortly before death and fail to capture actual neurotransmitter release.In vivo studies, such as microdialysis or voltammetry, are better suited but have been underutilized, preventing a clear understanding.Methodologically, neurotoxic lesions and traditional pharmacological interventions (like receptor antagonists) have their limitations in specificity and selectivity, particularly regarding receptor types and anatomical specificity.However, recent advancements, including optogenetic and chemogenetic tools like designer receptors exclusively activated by designer drugs (DREADDs) (reviewed by Runegaard et al., 2019), offer more precise approaches to investigate the role of DA.For example, it is now feasible to temporally manipulate neurochemical effects in specific brain areas, such as inhibiting DA release in the nucleus accumbens only during CS presentation.
However, DA is not the sole neurotransmitter involved in conditioned drug effects.Evidence also points to the role of glutamate and specific glutamate receptors, which should be further explored.Additionally, the changes in intracellular substrates-essentially, the engrams formed during conditioning-remain poorly understood.Future research could benefit from the extensive body of knowledge on the neurochemistry of learning and memory (e.g., Dunn, 1980) to develop hypotheses for identifying these engrams.
Presenting methods in publications: Regardless of the specific focus of a publication on conditioned drug effects, it is crucial for authors to recognize that the success of such research heavily depends on the conditioning procedure and the corresponding controls.Although it might seem obvious, more attention and detail should be devoted to describing the materials and methods used for conditioning.This includes not only factors such as the drugs used, animal strain, age, sex, and housing conditions (refer to the ARRIVE guidelines for more details) but also, and particularly, the detailed processes involved in conditioning and testing.This encompasses specifics of handling and injections, the duration of stays before and after exposure to a conditioning environment, the setup for conditioning, features of the conditioned stimuli, among others.A thorough description of these aspects will aid others in their research endeavors and, if replication is desired, contribute to reinforcing the evidence presented in the original study.
The term "placebo": An interesting facet of the literature on opioids and their conditioned effects on pain is the increasing mention and focus on the term "placebo" in animal studies.Although the term is not new to this field (Siegel, 1976 serves as an earlier example), much of the past animal research preferred to use phrases like "conditioned drug effects," likely because "placebo" encompasses broader concepts beyond conditioning and associative learning, such as expectancy.The recent surge in placebo studies involving human subjects may have encouraged animal researchers to contextualize their findings within the broader scope of placebo research.Unfortunately, however, the term "placebo" is sometimes used in a misleading way.In several publications, certain control groups are labeled as "placebo", although they refer to a simple drug vehicle condition, which neither comprised a conditioning procedure (usually in animal work) nor a placebo-relevant instruction (as in human studies).We suggest that future authors refrain from doing so, because the term "placebo" would otherwise loose its specific meaning.
Final Conclusions: Although research on conditioned drug effects now reflects approximately a century of intensive and seminal work, this scientific field is far from being saturated.As highlighted earlier, numerous critical questions still require further and new experimental scrutiny.Additionally, the growing interest in placebo mechanisms in humans-a topic to which conditioned drug effects undeniably contribute-should motivate continued relevant animal research.This could include the use of preclinical models of human neuropsychiatric disorders, such as depression.This comprehensive review aims to assist current and future researchers in formulating pertinent research questions and in designing their studies accordingly.Box 1

Fig. 2 .
Fig. 2. Within-subjects conditioning design.Two distinct environments are utilized in the experiment, where animals receive injections of the drug (indicated by a red syringe) in one environment and saline (indicated by a blue syringe) in the other.The test for conditioning occurs after saline injections, first in the environment previously paired with the drug (upper right) and then in the compartment previously paired with saline (lower right).
(continued on next page) R.K.W.Schwarting et al.