B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ l oca te /b ra in res

Research Report

Time-dependent effects of amphetamine on feeding in rats

Wesley White⁎, Luke K. Sherrill, Ilsun M. WhitePsychology Department, 601 Ginger Hall, Morehead State University, Morehead, KY 40351, USA

A R T I C L E I N F O

⁎ Corresponding author. Fax: +1 606 783 5077.E-mail address: w.white@moreheadstate.

0006-8993/$ – see front matter © 2007 Elsevidoi:10.1016/j.brainres.2007.08.005

A B S T R A C T

Article history:Accepted 1 August 2007Available online 9 August 2007

Following administration of a moderate dose of amphetamine, rats appear to pass through asequence of physiological/psychological states, including stimulant and depressant states.The present research evaluated whether these states could be inferred from time-dependentchanges in feeding-relatedmeasures. Male rats were housed in individual stations (light–dark12–12 h, free access to water) where, at 3-h intervals, they could respond for food for 1 h. Thework requirement was fixed ratio 1, and each lever press produced six 94-mg food pellets.When the pattern of responding for food stabilized across the light–dark cycle, a series of 6 or 7tests was run. During each test, rats received a saline treatment (1.0 ml/kg, subcutaneously)followed by a 48-h monitoring period, and then they received an amphetamine treatment(2.0 mg/kg, subcutaneously) followed by a 72-h monitoring period. Different groups weretreated at either light onset or light offset. Lever presses and head-in-feeding-bin responsesweremonitored throughout these tests. Administration of amphetamine at light onset and atlight offset produced cumulative food intake functions having four regions: post-treatmenthours 1–6 (hypophagia), 7–12 (normal intake), 13–27 (hypophagia), and 28 and beyond (normalintake). The sequence, duration, and quality of the amphetamine-induced changes in foodintake resembled those formerly seen in cue state and activity, and provided further evidenceof a transient withdrawal state 20–24 h post-amphetamine treatment.

© 2007 Elsevier B.V. All rights reserved.

Keywords:PsychostimulantFood intakeCircadianWithdrawalAnorexia

1. Introduction

Amphetamine, a psychostimulant, produces activating effectsin the short term (the first several or so hours post adminis-tration). These short-term effects and the mechanisms thatmediate them have been extensively studied. Amphetamineand related compounds are used recreationally in part becauseof such effects (reviewed in Berridge, 2006; Robinson andBerridge, 1993; Segal and Kuczenski, 1994).

Amphetamine produces additional time-dependent effectsduring the first day or so following administration. Investiga-tions of the effects of amphetamine on cue state and on activityhave provided good evidence for this. Barrett, Caul andcolleagues have examined the impact of amphetamine on cue

edu (W. White).

er B.V. All rights reserved

state ina series ofdrugdiscrimination studies involvingamphe-tamine and haloperidol (Barrett et al., 1992, 2005; Caul et al.,1996, 1997; Stadler et al., 1999). By “cue state,” Barrett, Caul andcolleagues meant the distinguishable internal sensations pres-ent at a particular time following drug treatment. In one study,rats treated with 10 mg/kg amphetamine responded on anamphetamine-paired lever 4 and 6 h after treatment, onamphetamine- and haloperidol-paired levers equally 8, 12,and 16 h after treatment, on a haloperidol-paired lever 20 and24 h after treatment, and again on each lever equally 32 h aftertreatment (Barrett et al., 1992). White and colleagues haveexamined the impact of amphetamine on activity. Rats treatedwith 2.0 or 4.0 mg/kg amphetamine were hyperactive 1 to 6 hafter treatment, normally active 7 to 18 h after treatment, hypo-

.

76 B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

active 19 to 24 h after treatment, and again normally active 25 hafter treatment (White andWhite, 2006).

Changes in cue state and activity appear correlated acrosstime, and the changes may signify the presence of differentamphetamine-induced states. An amphetamine-like cue stateand hyperactivity from hours 1 to 6 post-treatment indicate thepresence of a stimulant state, whereas a haloperidol-like cuestate and hypo-activity fromhours 19 to 24 post-treatmentmayindicate the presence of a withdrawal state. The withdrawalstate may be preceded by a latent state, when measures onlyappear to normalize, and itmay be followed by a recovery state,when measures actually do normalize. Withdrawal is com-prised of a constellation of symptoms (Barr and Markou, 2005),and oneway to bolster the claim that withdrawal is present at aparticular time is to show that other symptoms indicative ofwithdrawal are also present at that time. One characteristicsymptomofwithdrawal from amphetamine is diminished foodintake (hypophagia).

Thepurpose of this studywas to seewhether amphetaminealtered food intake ina time-dependentmanner comparable tothat observed for cue state and activity. We were particularlyinterested in determining whether amphetamine producedhypophagia during the same interval that it reportedly pro-duced a haloperidol-like cue state and hypo-activity.

Certain amphetamine administration regimes, such asregimes involving chronic escalating doses, have been used toproduce a relatively prolonged condition that has been likenedto depression (Barr and Markou, 2005). In contrast, in this re-search, a moderate dose of amphetamine, 2.0 mg/kg, was re-peatedly administered at intervals of at least 5 days, a regimethat is better suited to produce a transient withdrawal. Short-term effects of amphetamine on food intake have been exten-sively studied, whereas longer-term effects have not been.When longer-termeffects have been studied, investigators havetended tomeasure total intake at the end of a long interval suchas 24 h (Chen et al., 2001). In order to enhance the opportunity toidentify time-dependent changes, we assessed the effects ofamphetamine on intake frequently and over a long interval. Inparticular we allowed rats to lever press for food pellets at mealopportunities that beganevery 3hand thatwere 1h induration,andwemonitored respondingand food intake for 3 days follow-ing amphetamine administration. The effects of a drug dependinpart onwhen it is administered in the light–dark cycle (Daviesand Wellman, 1991; Reinberg, 1999). To evaluate whetherchanges could be observed that were independent of adminis-tration time, we examined the effects of treating differentgroups at light onset and at light offset.

Fig. 1 – Mean ratios completed across acquisitiondays for thegroup treated at light offset. 1-h meal opportunities, duringwhich each lever press produced 6 pellets (fixed ratio 1), werescheduled at 3-h intervals. Arrows indicate training daysfrom which performance on the measure was stable.

2. Results

2.1. Acquisition

Groupsofeight ratswere treatedateither lightonsetor lightoffsetof the12–12h light–dark cycleduring testing.Thegroupswerenottested until they showed stable responding on the feedingschedule. The feeding schedule allowed animals to respond forfood for 1 h every 3 h. During a feeding hour, a lever press couldresult in a “package” of six 94-mg pellets (Fixed ratio 1, FR1, or“ratio”). The lever press produced the first pellet, and an animal

had to place its head in the feeding bin to produce subsequentpellets in the package. Both groups of rats adjusted to the feedingschedule in a similar manner, and so acquisition data will beshown only for the group eventually treated at light offset duringtests.

Fig. 1 shows the mean number of ratios the animals com-pleted across days of exposure to the feeding schedule (Trainingdays). The number of ratios completed increased from days 1 to3, decreased from days 3 to 16, but did not differ from days 16 to20, F(7,19)=29.012, pb .0001 and Fisher's PLSD post hoc tests. Insummary, performance was stable after 15 days of training.

2.2. Testing

Animals received a series of 5-day tests. On day 1, differentgroups were treated, at either light onset or light offset, withsaline (Sal). On day 3, theywere treated, at the same times, with2.0 mg/kg amphetamine (Amp). Feeding opportunities werescheduled as before. Fig. 2 shows the mean number of ratioseach group completed on eachday of each test. The upper panelshows results for the group treated at light onset. An ANOVAproduced a significant effect of Test, F(6,42)=5.035, pb .001, asignificant effect of Test day, F(4,28)=51.448, pb .0001, and asignificant interaction, F(24,168)=1.762, pb .05. Fisher's PLSD forthe main effect of Test day indicated that fewer ratios werecompleted on the day of amphetamine administration than onany other day, p values b .05, but that no other days differed.Fisher's PLSD for the main effect of Test indicated that moreratios were completed during Test 1 than during other tests, pvalues b .05. AnANOVAbasedon the data of the group treatedatlight offset (Fig. 2, lower panel) produced a significant effect ofTest, F(5,35)=7.814, pb .0001, and of Test day, F(4,28)=36.146,pb .0001. Fewer ratios were completed on the day of amphet-amine administration than on other days, andmore ratioswerecompleted during Tests 1 and 2 than during the other tests(Fisher's PLSD, p values b .05). Overall, for treatment at both lightonset and light offset, fewer ratioswere completed on theday ofamphetamine administration, andmore ratios were completedduring the earliest tests.

Fig. 3 shows the number of ratios each group completed ateach meal opportunity during the 2-day period following saline

Fig. 2 – Mean ratios completed during each day of each 5-daytest for the group treated at light onset (upper panel) and forthe group treated at light offset (lower panel). “SalRec” is therecovery or baseline day following the day of salinetreatment (“Sal”), and “AmpRec1” and “AmpRec2” arerecovery or baseline days following the day of amphetamine(“Amp”) treatment. * indicates tests duringwhichmore ratioswere completed, and ** indicates days during which fewerratios were completed.

Fig. 3 – Mean ratios completed at each meal opportunityduring the first 2 days following saline administration oftests. The upper panel is for the group treated at light onset,and the lower panel is for the group treated at light offset.Results from tests 1 and 2, 3 and 4, and 5–7 or 5 and 6 havebeen averaged (Sal1&2, Sal3&4, Sal567 or Sal5&6,respectively). The bar across the top indicates when lightswere on or off. The variability bar (“SEM”) shows the meanstandard error of the mean.

77B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

administration of tests. The time course with which amphet-amine affected performance (see below) was evaluated bymaking comparisons to these baselines. The upper panelshows the data for the group treated at light onset. Three testsets were created by averaging the results of Tests 1 and 2(Sal1&2), 3 and 4 (Sal3&4), and 5, 6 and 7 (Sal567). The circadianpattern of ratios completed appeared to be similar across the2 days post saline and across the three test sets. An ANOVAyielded a significant effect of Test set, F(2,14)=10.740, pb .005, asignificant effect of Meal opportunity, F(15,105)=12.193, pb .0001,and a significant interaction, F(30,210)=1.725,pb .05. The stabilityof the circadian pattern was assessed via post hoc comparisonsbased on the main effect of Meal opportunity. The number ofratios completed was highest at Meal opportunities 1, 5, and 6following light onset, intermediate atMeal opportunities 4 and 8,and lowest at Meal opportunities 2 and 3 (Fisher's PLSD, p valuesb .05). This pattern tended to occur on the 2 days post saline andacross the three test sets.

For the group treated at lights offset (Fig. 3, lower panel), anANOVA produced a significant effect of Test set, F(2,14)=26.713,pb .0001, and of Meal opportunity, F(15,105)=17.255, pb .0001.The number of ratios completed was highest at Meal opportu-nities 1 and 2 after light offset, intermediate at Meal opportunity8, and lowest at Meal opportunities 4, 5, 6, and 7 (Fisher's PLSDbased on themain effect of Meal opportunity, p values b .05). Thepattern was observed on the 2 days post saline and across thethree test sets. On the whole, both groups had highly reproduc-

ible circadian patterns of responding, but the group treated atlight onset had an additional peak of responding at the firstmealopportunity in the light period.

On day 3 of tests, groupswere treated, at light onset or offset,with 2.0 mg/kg amphetamine, after which they could respondfor food on the same schedule of meal opportunities as before:beginning 1 h after treatment, 1-h meal opportunities werescheduled every 3 h, and during these meal opportunities eachlever press produced six 94-mg food pellets. The graphs in Fig. 4show the change in cumulative food intake at the first sixteenmeal opportunities following amphetamine administration,relative to saline administration. The upper and middle panelsshow the results for the groups treated at light onset and offset,respectively. To produce individual functions in these panels,the difference in the number of food pellets consumed at eachmeal opportunity following saline and amphetamine treatmentwas found (amphetamine−saline). These differenceswere thencumulated across the sixteen meal opportunities. The resultsfor sets of tests were then averaged as in Fig. 3. Amphetamineproduced a progressive reduction in the number of pelletsconsumed (“Cumulative Pellet Deficit”).

An ANOVA for the group treated at light onset (upper panel)produced a significant effect of Test set, F(2,14)=6.047, pb .05,and of Meal opportunity, F(15,105)=12.426, pb .0001. A signifi-cant interaction was not obtained. Given that the functions hadthe same general form, in order to identify the major trendsproduced by amphetamine, themain effect of Meal opportunity

Fig. 4 – Cumulative pellet deficit due to amphetamine atthe end of each meal opportunity. The deficit is relative tosaline treatment. The upper panel is for the group treated atlight onset, and the middle panel is for the group treated atlight offset. Tests were averaged as in Fig. 3. The bar acrossthe top indicates when lights were on or off. The variabilitybar (“SEM”) shows the mean standard error of the mean. Thelower panel is the average for both groups and all tests.The error bars represent standard errors of themean. Verticallines show the four regions into which an analysis ofpair-wise comparisons divided the function. Each pellet was94 mg. In the lower panel, * denotes a difference from MealOpportunity 9 but not from Meal Opportunity 8. In theupper and middle panels, * denotes a difference fromTest1&2 to the last test set, and ** denotes a greater differencefor the group treated at light offset.

78 B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

wasanalyzedwith Fisher's PLSD (significant effectswere pb .05).A mean deficit of approximately 40 pellets was obtained at theend of the first meal opportunity. A significant change did notoccur from Meal opportunity 1 to 2. However, the deficitincreased from Meal opportunity 2 to 3, and the deficit wasgreater at Meal opportunity 9 than at Meal opportunities 3, 4, 5,and 7, indicating that the deficit progressed from Mealopportunity 3 to 9. The deficits at Meal opportunities 10–16were less than or equivalent to the deficit atMeal opportunity 9,indicating that thismeal opportunity corresponded to a nadir in

cumulative intake. The analysis suggested that the overallfunction relating cumulative pellet deficit to meal opportunitycould be divided into four regions corresponding to Mealopportunity 1, Meal opportunity 2, Meal opportunities 3–9, andMeal opportunities 10–16.

A parallel ANOVA for the group treated at light offset (middlepanel) alsoproduceda significant effect of Test set,F(2,14)=8.345,pb .005, a significant effect of Meal opportunity, F(15,105)=3.073,pb .0004, and no significant interaction. Fisher's PLSD, done onthe main effect of Meal opportunity, indicated that the deficit,already present at the end of Meal opportunity 1, progressedthrough Meal opportunity 2, diminished at Meal opportunities 3and 4, and progressed again fromMeal opportunities 5 to 8. Theanalysis divided the overall function into four regionscorresponding to Meal opportunities 1–2, 3–4, 5–8, and 9–16.

The lowerpanel shows thecumulative change inpellet deficitat each meal opportunity averaged across groups and tests. AnANOVA and post hoc comparisons divided the function intoregions that includedMealopportunities1and2, 3and4, 5–9, and10–16, F(15,225)=4.501, pb .0001, Fisher's PLSD, p values b .05.Overall, Fig. 4 suggests that amphetamineproduceda short-termphase of hypophagia and a longer-term one (approximately 13–24 h post-treatment) that was preceded and followed by phasesof seemingly normal intake.

Though the functions in Fig. 4 had some similarities, theyalso suggested effects on cumulative intake that appeared todepend upon both test set and treatment time. To assess theseeffects, anANOVAwasdonewithTreatment time (lightonset orlight offset) and Test set (Test1&2 or Test567/Test 5&6) asfactors. Separate analyses were done for Meal opportunities8 and 16. For Meal opportunity 8, only the effect of Test set wassignificant, F(1,14)=7.582, pb .05. Treatment groups had similarcumulative deficits at Meal opportunity 8 during Test1&2, andduring the last Test set this deficit was similarly attenuated inboth groups. For Meal opportunity 16, the effect of Treatmenttime approached significance, F(1,14)=3.616, p=.0780, and theeffect of Test set was significant, F(1,14)=9.866, pb .01. Treat-ment time groups did not differ during Test1&2, but did differduring the last test set, t (14)=2.689, pb .05. For the group treatedat light onset, the change from Test1&2 to the last test set wasnot significant, whereas for the group treated at light offset, thechangewas significant, t (7)=−2.755, pb .05. In summary, repeat-edly treating with amphetamine at light offset resulted in alower deficit in intake 2 days following treatment than repeat-edly treating at light onset.

Duringa feedinghour, ananimalcould retrieveandconsumesixpellets after a leverpress. The time fromthe leverpress to thehead-in-bin response that retrieved the sixth pellet is ameasureof consummatory behavior and will be called consummatorytime. Consummatory times over 65 s were not included inanalyses. The remaining durations constituted over 97% of theobservations. Fig. 5 shows mean consummatory times forgroups treated with saline and amphetamine at light onset(upper panels) or light offset (lower panels).

For the group treated at light onset, the upper left panel showsmean consummatory times during light and dark phases of the2 days following treatments. No effects of treatment (saline oramphetamine), phase (light or dark), or day (day 1 or 2) wereobserved. The upper right figure shows mean consummatorytimes during different regions revealed by the Cumulative Pellet

Fig. 5 – Consummatory times for groups treated at light onset (upper panels) and offset (lower panels). Left panels show meanconsummatory times during phases of the light–dark cycle following saline or amphetamine treatment. The bar across thetop indicates when lights were on or off, and variability is standard error of the mean. Right panels showmean consummatorytimes during different regions revealed by the Cumulative Pellet Deficit analysis (see Fig. 4 lower panel). Variability is standarderror of the mean. Labels above bars indicate meal-opportunity regions having longer consummatory times.

79B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

Deficit analysis (see Fig. 4, lower panel). ANOVA indicated aneffect of region, F(3,21)=6.336, pb .005, but no effect of treatmentand no interaction. During meals 1–2, consummatory timestrended shorter following amphetamine than saline, t (7)=2.286,p=.0562.

For the group treated at light offset, phase in the light–darkcycle did have an effect on consummatory times (lower leftpanel), F(1,7)=61.121, pb .0001, and consummatory times wereshorter during the dark phase than during the light phase,Fisher's PLSD,pb .0001. Saline andamphetaminedidnot producedifferent consummatory times during any region followingtreatment (lower right panel), t (7), pN .05, though a main effectof region, F(3,21)=27.872, pb .0001, and a treatment×region inter-action, F(3,21)=4.015, pb .05, were obtained. The latter effectsweredue tomodulationbyphasesof the light–darkcycle. Insum,consummatory behaviorwas not differentially affected by salineand amphetamine at any time post treatment, and it wasmodu-lated by phase in the light–dark cycle when treatmentwas givenat light offset, but not when treatment was given at light onset.

The mean weight of both groups was around 500 g on theday of the first saline administration, and both groups gainedroughly 20 g over the course of the first six tests. During eachtest, bodyweight increased from the day of saline treatment tothe day of amphetamine treatment, but an increase in bodyweight was attenuated from the day of amphetamine treat-ment to the day of saline treatment.

3. Discussion

The group treated at light onset had a peak of ratios completedat the first meal opportunity in the light phase, whereas the

group treated at light offset did not. Presumably, this peakwasproduced by the treatment and the station maintenance thattook place near light onset. This peak was observed not onlyon the day of saline administration, but also on the day after itand on the two recovery days following amphetamine admin-istration, days when treatment and maintenance were notperformed. Although the pattern was instigated in part byexogenous events, it provided a stable baseline against whichthe effects of amphetamine administration could be assessed.

Administration of amphetamine at light onset and at lightoffset produced cumulative food intake functions having fourregions. Looking across both administration times, these regionscorresponded very roughly to post-treatment hours 1–6 (hypo-phagia), 7–12 (normal intake), 13–27 (hypophagia), and 28 andbeyond (normal intake). The sequence, duration, and quality ofthe amphetamine-induced changes in food intake resembledthose formerly seen in cue state (Barrett et al., 1992) and activity(White and White, 2006). The three measures undergo changesconsistent with stimulant, latent, withdrawal, and recoverystates. An amphetamine cue state, hyperactivity, and hypopha-gia overlap from hours 1 to 5 post-amphetamine treatment,indicating a stimulant state. A haloperidol-like cue state, hypo-activity, and hypophagia appear to overlap from approximatelyhours20 to24post-treatment, suggesting that awithdrawal statemay be present during this time.

The twophasesofhypophagiawere expressed in the contextof different states and were separated by a latent period whenintake appeared to have recovered. Our observations suggestthe possibility that the phases of hypophagia are not a unitaryresponse and may be mediated via different processes. Al-thoughmuch researchhas investigated thedeterminants of thehypophagia occurring shortly after amphetamine treatment,

80 B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

very little research has evaluated whether these determinantsand those of the longer-term hypophagia can be dissociated. Ifthe twophasesofhypophagiaaredue todifferentdeterminants,then amphetamine-induced hypophagia may be better under-stood by looking at both short- and long-term hypophagia.

A basic question raised by the results is why intake declinedfrom hours 13 to 27 post-amphetamine treatment.

The decline cannot be ascribed to a direct effect of amphet-amine, because amphetamine has a half-life of 1–3 h in rats(Fuller et al., 1977; Hutchaleelaha et al., 1994), and negligibleamounts of drug would be in the body during this interval.

Withdrawal from amphetamine is thought to reduce foodintake by interfering with appetitive behavior (Barr and Phillips,1999; Orsini et al., 2001), at least under conditions of highresponse cost (Salamone and Correa, 2002). Whether reducedintake in the present study was due to a change in processesrelated to appetitive behavior is debatable. In the present study,the impact of appetitive behavior on food intake was limited inseveral ways: Travel distance was minimal, the ratio require-ment was one, each response produced six pellets, and ampletime to earn food (1 h) was afforded at each meal opportunity.Nevertheless, an impact of appetitive processes remains pos-sible, because appetitive behavior is expressed via a complexcontrol structure entailing sensory, motor, and integrative pro-cesses (Timberlake and Lucas, 1989).

Consummatory timeswere theduration required to consumefive pellets and to retrieve the sixth. Mean consummatory timesduring thewithdrawal regionwere the same (approximately 30 s)following both saline and amphetamine. This result suggeststhat the decline in intake during the withdrawal region was notdue to a gross impairment of consummatory behavior. A lack ofeffect of withdrawal from amphetamine on consummatorybehavior has been reported previously (Barr and Phillips, 1999;Salamone and Correa, 2002). In the present study, amphetamineaffected consummatory behavior at a higher level of mealorganization, such as the number of packages taken.

To our knowledge, no systematic research has been doneregarding mechanisms that might mediate longer-term hypo-phagia produced by widely spaced administrations of moder-ate amphetamine doses. Whether this hypophagia reflectsmalaise (Cole et al., 1995; Turchan et al., 1998), a specific deficitin feeding-related motivation (Stanek, 2006; Vicentic andJones, 2007), or some other condition is uncertain.

Although intake and weight gain were attenuated byamphetamine administration, animals did not compensate byincreasing intake on the following days. A symptom of with-drawal from amphetamine sometimes reported is a reboundhyperphagia (Srisurapanont et al., 1999). Such a rebound wasprobably not observed in the present study because the drugregimen was mild enough and the feeding schedule was richenough to preclude significant weight loss.

Administering at different time in the light–dark cycle pro-duced some differences. Boundaries between stimulant, latent,withdrawal, and recovery regions differed somewhat. This wasprobably the case for a couple of major reasons. First, for eachgroup, the cumulative change in food intake due to amphet-aminewas relative toadifferent control pattern. Second, for eachgroup, amphetamine-induced effects interacted with a differentsequence of states entrained by the light–dark cycle. Thetreatment–time groups also showed some differences in the

capacity to recover intake following amphetamine. Groupstreated at both times had a smaller deficit in cumulative intakeat Meal opportunity 8 during later tests. Such tolerance to theanorectic effects of amphetamine has been reported frequently(Caul et al., 1988; Kuo and Cheng, 2002). For the group treated atlight offset, a greater attenuation in the cumulative intake deficitwas observed from earlier to later tests at Meal opportunity 16.The group treated at light offset showed a slightly better capacityto recover, even though moderate doses, widely spaced admin-istrations, and only six tests were employed. For the grouptreated at light offset, consummatory times were shorter duringthe active period and longer during the dark period, whereas forthe group treated at light onset, this circadian pattern was lost,and consummatory times tended to be uniformly short.Treatment at light onset may more readily disrupt normalcircadian patterns.

Othermeasures show a biphasic response following amphet-amineadministrationandarepotential indicatorsofwithdrawal,including the frequency of ultrasonic vocalizations (Covingtonand Miczek, 2003; Thompson et al., 2006), the threshold ofintracranial self-stimulation (Cryan et al., 2003), and the propor-tion of EEG activity indicative of REM sleep (Edgar and Seidel,1997). Evaluating whether common conditions produce similartime-dependent effects in different measures would suggestwhether amphetamine-induced states were mediated by aunitary mechanism or by distributed mechanisms.

Each of the successive 5-day tests began with a 2-day re-baseline to which the effects of drug were compared. Thisdesign seemed necessary, because amphetamine administra-tion sometimes appears to shift reward-related set points(Koob and Bloom, 1988). The design did not allow effects due torepeated drug administration to be fully revealed.

Several procedural features were selected to model humanrecreational drug use. A dose of amphetamine was used that arat might be expected to self-administer over a reasonableperiod of time (Deminiere et al., 1989). Drug was administeredrepeatedly, but at intervals widely spaced enough (5 days) toprevent drug accumulation and to allow intake to recover. Inone condition, drug was administered at the start of the ina-ctive period (light onset for the rat), a time at which recrea-tional drug use presumably peaks in humans. Widely spacedadministrations of moderate doses of amphetamine may notproduce dramatic withdrawal symptoms, but this and similarresearch suggests that such regimes significantly impactmotivation.

4. Summary

In the present study, a moderate dose of amphetamine wasadministered every 5 or more days. Two phases of hypophagiawere observed during the first 24 h following amphetamineadministration, a short-term hypophagia (hours 1–6 post-treatment) and a longer-term hypophagia (hours 13–27 post-treatment). The phases of hypophagia could be mediated bydifferent processes and mechanisms, and a complete under-standing of amphetamine-induced hypophagia may not beachievable by investigating only the short-tem phase. Minimalappetitive behavior was required to procure food, and theefficiency of consummatory behavior was not impaired during

81B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

either phase of hypophagia: Consequently, short- and longer-term hypophagia could not be readily ascribed to a gross impair-ment in either appetitive or consummatory processes. Followingrepeated amphetamine treatment, longer-termhypophagiawasdiminished, and the group treated at light offset was able topartially compensate during the recovery phase. The changesacross time in food intake appeared to be indicators of a series ofamphetamine-induced states, because they corresponded tochanges seen in cue state and activity. Together, the measuresindicate that a psycho-stimulant state is present approximately1–5 h following amphetamine treatment, and that a transientwithdrawal is present 20–24 h post-treatment. The changesacross time in the measures appear to be correlated, suggestingthat they may be mediated by a common mechanism. Theresultsmay be relevant to the effects of human recreational druguse.

5. Experimental procedure

5.1. Animals

The study consisted of two conditions that were runsuccessively. Each condition included eight male Wistar rats(Harlan, Indianapolis, IN). Prior to the start of a condition,animalswere pair-housed in an animal colony on a 12-h light/12-h dark cycle and in a temperature of 20–22 °C. They hadfree access to food (Purina 5001 Rodent Diet, Lab Diet) andwater. Just prior to the start of a condition, animals werehandled and were pre-exposed in their home cages to thepellets that they would consume during the study. Animalsweighed between 300 and 400 g at the start of a condition.Animal care and experimental procedures were in accor-dance with NIH guidelines.

5.2. Apparatus

Animals were trained in four standard operant conditioningstations (Med Associates) that contained a retractable lever, afeeder that dispensed 94-mg pellets, and a bin that couldbe illuminated and that was equipped with a head-in-bindetector.

The animals were tested in one of eight “24-h stations” thatwere designed for long-term housing. Each station consisted of asound-attenuating, wooden compartment (58 cm×42 cm×58 cmhigh) that enclosed a plastic cubicle (40 cm×20 cm×40 cm high).Each station contained a response lever, a pellet dispenser, and afood bin similar to those in the operant stations. The lever wassituated just below the bin in the left half of one end wall. Theright half of the end wall contained a drinking tube that wasattached to a water bottle. The floor of each cubicle was a blackmetal pan that contained a thin layer of absorbent micro-wavedtopsoil. Each compartment had a fan (Sunon, sf11580A) thatprovided ventilation and that masked noises and a lightfixture (Lampi-Pico accent light, 4-W) that produced a 12-h light/12-h dark cycle.

Devices in operant conditioning stations and in 24-h stationswere connected to an interface (Med Associates) and a computer.Software (MedAssociates) arrangedcontingenciesandmonitoredbehavior. Stationswere located inwell-isolated temperature- and

humidity-controlled rooms (approximately 1.8 m×2.1 m×2.6 mhigh).

5.3. Drug

Powdered D-amphetamine sulfate (Sigma, St Louis) wasmixedin saline (2.0 mg/ml base). Saline was used as the controltreatment (1.0 ml/kg).

5.4. Procedure

5.4.1. TrainingEach animalwas deprived to 85%of its free-feeding bodyweightand, in an operant station, was trained in a series of conditions(habituation, bin training, auto-shaping, fixed ratio 1) to pressthe lever for 94-mg pellets (Bio Serv, #F0058). The animals werethen placed on free food availability in the colony for 5 days.

5.4.2. AcquisitionAnimals were then transferred to 24-h stations for theremainder of the study. Throughout this time animals were ona 12-h light/12-h dark schedule and had free access to water.During the first 24 h in the stations, the animals could earn foodpellets on an FR1 schedule. Each lever press produced six 94-mgpellets.The leverpressproducedthe first pellet, and thedeliveryof subsequent pellets was contingent on a head-in-bin re-sponse. The lever press turned on the bin light, and the lightremained on until the animal made a head-in-bin response toretrieve the sixth pellet.

On subsequent days, food accesswas restricted to 1-h periodsspaced 3 h apart. One-hour meal opportunities began 1, 4, 7, and10 h after light onset and light offset. The beginning of a mealopportunity was signaled by the delivery of a pellet and theillumination of the feeding bin for 10 s. Animals remained on thisschedule until the number of ratios completed per day and thedistribution of ratios completed across the light–dark cyclestabilized.

5.4.3. TestingAfter intake stabilized, 5-day tests were conducted. The priorfeeding conditions remained in effect. For one group of eightsubjects, at light onset of test day 1, each animal was injectedwith saline (1.0 ml/kg subcutaneously under the loose skin onthe back of the neck). Two days later, at light onset of test day3, each animal was administered 2.0 mg/kg amphetamine inthe samemanner. The other group of eight subjects received asimilar series of tests, but they were treated at light offset. Thegroup treated at light onset was tested seven times, and thegroup treated at light offset was test six times.

Tests and test days were successive, except on a couple ofoccasions when experimenters were unavailable. Animalstreated at light onset received an additional baseline daybetween saline and amphetamine phases of Test 3. Also,between Tests 3 and 4, they were placed in individual plastictubs on free access to food andwater for 5 days, afterwhich theyreceived two additional days of baseline.

Body weights were taken at the time of saline or amphe-tamine treatment. Apparatus maintenance was done at thesame time and involved re-filling the pellet dispensers andwater bottles and changing the pans and top-soil. Otherwise,animals were not disturbed.

82 B R A I N R E S E A R C H 1 1 7 1 ( 2 0 0 7 ) 7 5 – 8 2

5.5. Dependent measures

Throughout acquisition and testing, the number of ratioscompleted at eachmeal opportunitywas recorded. The numberof non-reinforced lever presseswas also recorded every hour, aswell as the number of occasions upon which an animal put itshead in the feeding bin (“checks”).

5.6. Data analysis

Data were analyzed using within subjects ANOVA. Significanteffects were analyzedwith additional within-subjects ANOVAs,followed up by PLSD post hoc comparisons or paired t-tests.

Acknowledgments

The research was supported by grants R15DA015351 andP20RR016481. Tiffany McNabb, Susan Roy, Marcus Hundley,Ian Smith, Richard Cates, and Donald Patton worked on pilotstudies.

R E F E R E N C E S

Barr, A.M., Markou, A., 2005. Psychostimulant withdrawal as aninducing condition in animal models of depression. Neurosci.Biobehav. Rev. 29, 675–706.

Barr, A.M., Phillips, A.G., 1999. Withdrawal following repeatedexposure to D-amphetamine decreases responding for asucrose solution as measured by a progressive ratio scheduleof reinforcement. Psychopharmacology 141, 99–106.

Barrett, R.J., White, D.K., Caul, W.F., 1992. Tolerance, withdrawal,and supersensitivity to dopamine mediated cues in adrug–drug discrimination. Psychopharmacology 109,63–67.

Barrett, R.J., Caul, W.F., Smith, R., 2005. Withdrawal, tolerance, andsensitization to dopamine mediated interoceptive cues in ratstrained on a three-lever drug-discrimination task. Pharmacol.Biochem. Behav. 81, 1–8.

Berridge, C.W., 2006. Neural substrates of psychostimulant-inducedarousal. Neuropsychopharmacology 31, 2332–2340.

Caul, W.F., Jones, J.R., Barrett, R.J., 1988. Amphetamine's effects onfood consumption and body weight: the role of adaptiveprocesses. Behav. Neurosci. 102, 441–450.

Caul, W.F., Barrett, R.J., Huffman, E.M., Stadler, J.R., 1996. Reboundresponding following a single dose of drug using anamphetamine–vehicle–haloperidol drug discrimination.Psychopharmacology 128, 274–279.

Caul, W.F., Stadler, J.R., Barrett, R.J., 1997. Amphetamine-inducedwithdrawal responding: effects of repeated drugadministration. Psychopharmacology 133, 351–355.

Chen, T.Y., Duh, S.L., Huang, C.C., Lin, T.B., Kuo, D.Y., 2001.Evidence for the involvement of dopamine D(1) and D(2)receptors in mediating the decrease of food intake duringrepeated treatment with amphetamine. J. Biomed. Sci. 8,462–466.

Cole, R.L., Konradi, C., Douglass, J., Hyman, S.E., 1995. Neuronaladaptation to amphetamine and dopamine: molecularmechanisms of prodynorphin gene regulation in ratstriatum. Neuron 14, 813–823.

Covington, H.E., Miczek, K.A., 2003. Vocalizations duringwithdrawal from opiates and cocaine: possible expressions ofaffective distress. Eur. J. Pharmacol. 467, 1–13.

Cryan, J.F., Hoyer, D., Markou, A., 2003. Withdrawal from chronicamphetamine induces depressive-like behavioral effects inrodents. Biol. Psychiatry 54, 49–58.

Davies, B.T.,Wellman,P.J., 1991. Effects of circadiancycle and timeoftesting on drug-induced anorexia in rats. Behav. Pharmacol. 2,199–204.

Deminiere, J.M., Piazza, P.V., Le Moal, M., Simon, H., 1989.Experimental approach to individual vulnerability topsychostimulant addiction. Neurosci. Biobehav. Rev. 13,141–147.

Edgar, D.M., Seidel, W.F., 1997. Modafinil induces wakefulnesswithout intensifying motor activity or subsequent reboundhypersomnolence in the rat. J. Pharmacol. Exp. Ther. 283,757–769.

Fuller, R.W., Baker, J.C., Molloy, B.B., 1977. Biological disposition ofrigid analogs of amphetamine. J. Pharm. Sci. 66, 271–272.

Hutchaleelaha, A., Sukbuntherng, J., Chow, H.H., Mayersohn, M.,1994. Disposition kinetics of D- and L-amphetamine followingintravenous administration of racemic amphetamine to rats.Drug Metab. Dispos. 22, 406–411.

Koob, G.F., Bloom, F.E., 1988. Cellular and molecular mechanismsof drug dependence. Science 242, 715–723.

Kuo, D.Y., Cheng, J.T., 2002. Role of cerebral dopamine but notplasma insulin, leptin and glucocorticoid in the developmentof tolerance to the anorectic effect of amphetamine. Neurosci.Res. 44, 63–69.

Orsini, C., Koob, G.F., Pulvirenti, L., 2001. Dopamine partial agonistreverses amphetamine withdrawal in rats.Neuropsychopharmacology 25, 789–792.

Reinberg, A.E., 1999. Concepts in chronopharmacology. Annu. Rev.Pharmacol. Toxicol. 32, 51–66.

Robinson, T.E., Berridge, K.C., 1993. The neural basis of drugcraving—an incentive–sensitization theory of addiction. BrainRes. Rev. 18, 247–291.

Salamone, J.D., Correa,M., 2002.Motivational viewsof reinforcement:implications for understanding the behavioral functions ofnucleus accumbens dopamine. Behav. Brain Res. 137, 3–25.

Segal, D.S., Kuczenski, R., 1994. Behavioral pharmacology ofamphetamine. In: Cho, A.K., Segal, D.S. (Eds.), Amphetamineand its analogs. Academic Press, San Diego, pp. 115–150.

Srisurapanont, M., Jarusuraisin, N., Jittiwutikan, J., 1999.Amphetamine withdrawal: I. Reliability, validity and factorstructure of a measure. Aust. N. Z. J. Psychiatry 33, 89–93.

Stadler, J.R., Caul, W.F., Barrett, R.J., 1999. Characterizingwithdrawal in rats following repeated drug administration usingan amphetamine–vehicle–haloperidol drug discrimination.Psychopharmacology 143, 219–226.

Stanek, L.M., 2006. Cocaine- and amphetamine-related transcript(CART) and anxiety. Peptides 27, 2005–2011.

Thompson, B., Leonard, K.C., Brudzynski, S.M., 2006.Amphetamine-induced 50 kHz calls from rat nucleusaccumbens: a quantitative mapping study and acousticanalysis. Behav. Brain Res. 168, 64–73.

Timberlake, W., Lucas, G.A., 1989. Behavior systems and learning:From misbehavior to general principles. In: Klein, S.B.,Mowrer, R.R. (Eds.), Contemporary Learning Theories:Instrumental Conditioning Theory and theImpact of Biological Constraints on Learning. Erlbaum,Hillsdale, NJ, pp. 237–275.

Turchan, J., Przewlocka, B., Lason, W., 1998. Effects of repeatedpsychostimulant administration on the prodynorphin systemactivity and kappa opioid receptor density in the rat brain.Neuroscience 85, 1051–1059.

Vicentic, A., Jones, D.C., 2007. The CART (cocaine- andamphetamine-regulated transcript) system in appetite anddrug addiction. J. Pharmacol. Exp. Ther. 320, 499–506.

White, W., White, I.M., 2006. The effects of dose onamphetamine-induced acute withdrawal as indicated by anactivity measure in rats. Physiol. Behav. 87, 368–376.