ORIGINAL INVESTIGATION

Blockade of NMDA GluN2B receptors selectively impairsbehavioral flexibility but not initial discrimination learning

Gemma L. Dalton & Liya M. Ma & Anthony G. Phillips &

Stan B. Floresco

Received: 15 September 2010 /Accepted: 21 February 2011 /Published online: 8 March 2011# Springer-Verlag 2011

AbstractRationale Behavioral flexibility is the ability to adjustbehavior when environmental contingencies change and iscompromised in disease states such as schizophrenia,attention deficit hyperactivity disorder, and followingdamage to the prefrontal cortex.Methods The present study investigated the contribution ofN-methyl-D-aspartate GluN2B receptor subunits in theinitial learning and in shifting between or within discrim-ination strategies (i.e., strategy set-shifting and reversallearning), using tasks conducted in operant chambers.Strategy set-shifting required rats initially to learn avisual-cue discrimination (day 1) and on day 2, shift tousing an egocentric spatial response strategy to obtainreward. For reversal learning, rats were trained on aresponse discrimination on day 1 and then required toselect the opposite lever on day 2.Results Blockade of GluN2B receptors with systemic admin-istration of Ro25-6981 on day 1 did not impair initialacquisition of either a response or visual-cue discriminationnor did these treatments affect performance of strategy orreversal shifts on day 2. However, administration of Ro25-6981 prior to a set-shift or reversal on day 2 significantlyimpaired performance on both tasks, inducing a selective

increase in perseverative errors, indicative of a disruption theability to suppress a previously acquired prepotent response.Conclusions These data suggest that systemic blockade ofGluN2B receptors Ro25-6981 does not appear to interferewith the initial acquisition of simple visual or responsediscriminations. However, these receptors do appear to playa central and selective role in the modification of previouslyacquired strategies or stimulus–reward associations, facili-tating behavioral inhibition so that novel patterns ofbehavior may emerge.

Keywords NMDA . GluN2B . Set-shifting . Reversallearning . Discrimination learning . Rats

Introduction

The ability to behave in a flexible manner and adjustongoing activity in the face of changing environmentalcontingencies is an essential survival skill. It is wellestablished that different forms of behavioral flexibility,from relatively simple reversal learning to more complexset-shifting strategies, are critically dependent on neuralnetworks that incorporate different regions of the prefrontalcortex (PFC) and its striatal outputs (see Floresco et al.2009 for review). Impairments in these forms of behavioralflexibility are observed in disease states such as schizo-phrenia and ADHD and resemble those observed inindividuals with damage to the PFC who show normalacquisition of a task but selective impairments in shiftingfrom that strategy once it is no longer relevant (Goldberg etal. 1987; Lombardi et al. 1999; Mesulam 2002; Pantelis etal. 1999; Rubia et al. 2007a,b; Willcutt et al. 2005).Understanding the neurochemical underpinnings of theseexecutive functions promises new insights that may aid in

G. L. Dalton : S. B. Floresco (*)Department of Psychology and Brain Research Centre,University of British Columbia,2136 West Mall,Vancouver, BC V6T 1Z4, Canadae-mail: floresco@psych.ubc.ca

G. L. Dalton : L. M. Ma :A. G. PhillipsDepartment of Psychiatry and Brain Research Centre,University of British Columbia,2136 West Mall,Vancouver, BC V6T 1Z4, Canada

Psychopharmacology (2011) 216:525–535DOI 10.1007/s00213-011-2246-z

the development of pharmacological treatments for cogni-tive dysfunction associated with these disorders.

Ionotrophic glutamate receptors, particularly the NMDAsubtype, have been implicated in behavioral flexibility.Systemic administration of non-competitive NMDA antag-onists, such as PCP, ketamine, or MK-801, impairs reversallearning and set-shifting assessed by a variety of procedures(Egerton et al. 2005; Idris et al. 2005; Murray et al. 1995;Stefani and Moghaddam 2005; van der Meulen et al. 2003).However, in many instances, these treatments also impairinitial discrimination learning, in keeping with the long-held notion that these receptors play a key role in theformation of different types of memories (Morris et al.1986; Morris 1989; Murray et al. 1995; Stefani andMoghaddam 2005). Accordingly, interpretation of thesedata is problematic, as it is unclear whether impairments inflexibility induced by systemic NMDA receptor blockadeare attributable specifically to disruptions in switchingbetween different cognitive strategies or merely reflectimpaired learning of novel associations or strategies. Notethat regional specific manipulations of NMDA transmissionwithin the frontal lobes can impart selective impairments inbehavioral flexibility without affecting initial learning. Forexample, Stefani and Moghaddam (2005) reported thatsystemic administration of MK-801 impaired performanceboth during the initial rule acquisition phase and during set-shift. In contrast, infusion of MK801 into the medial PFC(mPFC) selectively increased perseverative respondingduring the shift phase of the experiment, while leavinginitial-rule learning unaffected. Thus, a parsimoniousexplanation for the dissociable effects of systemic versusintra-mPFC NMDA blockade on behavioral flexibility maybe NMDA receptor activity within the PFC is involvedspecifically in the shifting between strategies, whereasNMDA receptors outside of the PFC contribute to theinitial learning of a particular strategy.

NMDA receptors can be categorized according tosubunit composition, with the predominant NR2 subunitsin the forebrain being GluN2A and GluN2B (IUPHARnomenclature; Collingridge et al. 2009; Laurie et al. 1997;Mutel et al. 1998; Wenzel et al. 1997). Previous studiesinvestigating the effect of systemic NMDA receptorblockade on behavioral flexibility have used relativelynon-selective antagonists that block both the GluN2A andGluN2B NMDA receptor subtypes (Egerton et al. 2005;Stefani and Moghaddam 2005, 2010). Therefore, thepossibility remains that distinct NMDA receptor subtypesmay contribute differentially to either the initial learning orshifting between strategies or stimulus–reward associations.In this regard, an emerging literature suggests that theGluN2B containing-receptor may play a key role inresponse inhibition and behavioral flexibility. Higgins etal. (2003) initially describe impaired response inhibition in

rats following systemic pre-treatment with the selectiveGluN2B receptor subunit antagonist Ro63-1908. Morerecently, work from our laboratory reveals that blockadeof GluN2B receptors selectively impairs extinction, but notacquisition of conditioned fear (Dalton et al. 2007, 2008).Similarly, Duffy et al. (2008) report increased perseverationduring a reversal task performed in a water maze followingsystemic administration of the GluN2B antagonist Ro25-6981 in mice. In contrast, deletion of GluN2A receptorsinduces a significant impairment in both discrimination andreversal learning (Brigman et al. 2008). Interestingly,perseverative responding was unaffected in GluN2Aknockout mice (Brigman et al. 2008). Viewed collectively,these findings suggest that although the contribution ofGluN2B receptors to initial discrimination learning may bedependent on a variety of circumstances, these receptorsmay play a more prominent role in the adaptation of learnedresponses in ambiguous circumstances. Specifically,GluN2B receptors may play a central role in the suppres-sion of a prepotent response during tasks demandingbehavioral/cognitive flexibility.

The present study was designed to clarify further thepotentially selective contribution of GluN2B receptors tobehavioral flexibility using the selective antagonist Ro25-6981. Specifically, we tested the effects of systemicblockade of GluN2B receptors on the acquisition of simplevisual-cue or response-based discrimination and, morepertinently, examined how these treatments affected shiftsbetween (strategy-shifting) or within (reversal learning)these different discrimination strategies. In so doing, weemployed operant-based procedures used previously toidentify the contribution of the PFC and ventral striataldopamine to these forms of flexibility (Floresco et al. 2008;Haluk and Floresco 2009). These procedures permitted adetailed analysis of the types of errors committed during ashift, to delineate whether GluN2B receptor blockade mayimpair suppression of previously acquired behaviors, or theacquisition and/or maintenance of new strategies orstimulus–reward associations.

Materials and methods

Subjects

Sixty-two male Sprague Dawley rats (280–350 g) were used.Rats were housed singly and maintained on a 12-h light/darkcycle with free access to standard laboratory chow and water.The colony was maintained at 21°C with a 12:12-h light darkcycle (lights on at 07:00 hours). All experiments were carriedout during the light phase of the cycle. Rats were given 7–8 days to acclimatise to the colony before behavioralprocedures began. Rats were handled and weighed daily during

526 Psychopharmacology (2011) 216:525–535

this acclimatisation period and throughout the course of theexperiment. During behavioral procedures, rats were main-tained on a restricted lab chow diet at 85% of their age-matchedad libitum weight and had free access to water. All experimentswere conducted in accordance with the standards of theCanadian Council on Animal Care and were approved by theCommittee on Animal Care, University of British Columbia.

Apparatus

All testing was conducted in four operant chambers (30.5×24 cm×21 cm; Med-Associates, St. Albans, VT, USA)enclosed in sound attenuating boxes. Each box contained afan to mask outside noises and to provide ventilation. Tworetractable levers were in each chamber, one located oneither side of a central food hopper where sugar pelletreinforcement (45 mg; BioServ, Frenchtown, NJ, USA) wasdelivered by a pellet dispenser. Above each lever, a lightemitting diode stimulus light was positioned centrally andserved as a stimulus for visual-cue discrimination learning.Each chamber was illuminated by a 100-mA house lightlocated at the top-center of the wall opposite the levers.Four infrared photobeams were mounted on the sides ofeach chamber 3 cm above the grid floor, and anotherphotobeam was located in the food receptacle. Locomotoractivity was measured as the number of photobeam breaksthat occurred during a session. All experimental data wererecorded by an IBM personal computer connected to thechambers via an interface.

Initial lever pressing training

These procedures have been described previously (Floresco etal. 2008). Rats were given ~20 reward pellets in their homecage the day before initial exposure to the operant chamber.On the first day of training, two to three crushed pellets wereplaced both in the food hopper and on the active lever. Ratswere trained under a fixed-ratio 1 schedule to a criterion of50 presses in 30 min, first for one lever, then the other(counterbalanced left/right between subjects). On subsequentdays, rats were familiarized with the insertion of the leversinto the chambers and were trained to press them within 10 sof their insertion. These sessions consisted of 90 trainingtrials and began with the levers retracted and the chamber indarkness. A new trial began every 20 s with the illuminationof the house light and insertion of one of the two levers intothe chamber. If the rat failed to respond on the lever within10 s, the lever was retracted, the chamber was darkened, andthe trial was scored as an omission. If the rat respondedwithin 10 s, the lever was retracted, a single pellet wasdelivered immediately, and the house light remained illumi-nated for another 4 s. In every pair of trials, the left or rightlever was presented once in a randomised order. The

stimulus lights above each of the levers were neverilluminated during these training sessions. Rats were trainedon this schedule for 5 days. By the end of this trainingschedule, all rats had achieved a criterion of less than fiveomissions in a 30-min session.

Immediately after the last retractable lever training session,the presence of a side bias was determined for each rat. Thisprocedure was similar to those used with maze-based strategy-shifting procedures that we have employed previously(Floresco et al. 2006a, 2008). This session resembled thepretraining session, with the exception that both levers wereinserted into the chamber simultaneously. Again, thestimulus lights above the levers were not illuminated duringthese trials. Rats were required to complete successfullyseven pairs of trials for the side preference test. Each pair oftrials required the rat to alternate responding on the twolevers. Thus, a reward pellet was delivered following aresponse on either lever during the first trial. For the secondtrial, the rat was required to press the opposite lever. If thesame lever was chosen, both levers were retracted, the houselight was extinguished, and no food was delivered. A newtrial-pair began with both levers being active, where a presson either lever again resulted in the delivery of a rewardpellet. With this procedure, we were able to record not onlythe number of responses made on each lever but also whichlever the rat chose first for each pair of trials. If a rat made adisproportionate number of responses on one lever over theentire session (i.e., >2:1 ratio), that lever was considered toreflect a ‘side bias.’ However, if the total number ofresponses on the left and right lever were comparable (<2:1ratio), the lever that a rat chose first most often wasconsidered its side bias.

Strategy shifting

Day 1: visual-cue discrimination On the first day oftesting, rats were trained on a visual-cue discrimination. Asession began in darkness with the levers retracted. Eachtrial was initiated with the illumination of one of thestimulus lights located above a lever. Three seconds later,the house light was turned on, and both levers wereextended. Rats had 10 s to make a response. A press on alever below the illuminated stimulus light resulted in thedelivery of one reward pellet and retraction of both levers.The house light was extinguished 4 s later. A press on theinactive lever resulted in the retraction of both levers andthe immediate extinction of the house light. Failure tochoose either lever within 10 s resulted in the retraction ofboth levers, and the trial was recorded as an omission. Inevery pair of trials, the left or right stimulus light wasilluminated once, and the order within the pair of trials wasrandomized. Trials continued every 20 s until a rat hadreceived a minimum of 30 trials and achieved criterion

Psychopharmacology (2011) 216:525–535 527

performance of 10 consecutive correct responses. Omissiontrials were not included in the trials to criterion measure.Upon completion of a session, the rat was removed fromthe chamber and placed back in its home cage. For eachtrial, lever choice and position of the stimulus light wererecorded, as were the latencies to respond on a lever afterits insertion and locomotor activity.

Day 2: shift to response discrimination Twenty-four hoursafter visual-cue discrimination training, rats entered thestrategy shift phase of the experiment. During this shift, ratswere required to disengage from the previously relevantvisual-cue strategy and shift to using an egocentric responsediscrimination strategy, requiring them to press the leveropposite their side bias (left or right lever) to obtain reward,regardless of the position of the visual cue. As with theinitial visual-cue discrimination, a stimulus light wasilluminated above one of the two levers for 3 s before theinsertion of the levers. Trials were given in a manneridentical to visual-cue discrimination trials, and for eachtrial, the lever that the animal chose and the location of thestimulus light were recorded. Trials continued every 20 suntil a rat achieved criterion performance of 10 consecutivecorrect choices, within a maximum of 150 trials. If a rat didnot achieve criterion within this allotted number of trials, itsdata was included in the analysis and given a score of 150trials for this measure. We did not test rats on the response-to-cue shift because this procedure typically yields highlyvariable data in control animals and is not as sensitive todisruption following PFC inactivation (Floresco et al.2008). The response-to-cue shift appears to be a relativelyeasier shift for rats to perform, likely attributable to the factthat many rats appear to have an innate bias to learning avisual-cue discrimination when compared to a responsestrategy (Floresco et al. 2008).

Response reversal learning

Day 1: initial response discrimination For this experiment,a separate group of rats were initially trained on theresponse discrimination task described above. On eachtrial, rats were required to press the lever opposite their sidebias, regardless of the position of the visual-cue stimuluslight, which for this experiment, served as a distracter.Trials continued until a rat achieved criterion performanceof 10 correct consecutive choices. There was no limit to thenumber of trials rats received to achieve this criterion. Uponcompletion of a training session, the rat was removed fromthe chamber and placed back in its home cage.

Day 2: response reversal Twenty-four hours after responsediscrimination training, rats were trained on a reversal of this

discrimination. Hence, a correct choice required a press of thelever opposite to that which was reinforced on day 1. All otheraspects of training were identical to those used on day 1 ofresponse discrimination training. Trials continued every 20 suntil a rat achieved criterion performance of 10 correctconsecutive choices, within a maximum of 150 trials.

Error analysis

A key advantage to this type of strategy shifting task is theunambiguous and detailed analysis of the type of errorsmade during the shift (Floresco et al. 2008; Haluk andFloresco 2009). Errors committed during the shift weresubdivided into three error subtypes to determine whetherdrug treatment impaired the ability to either shift away fromthe previously learned strategy (perseverative errors) or toacquire and maintain the new strategy after perseverationhad ceased (regressive or never-reinforced errors). Aperseverative error was scored when a rat responded on alever with the stimulus light illuminated above it on trialsthat required the rat to press the opposite lever during theinitial phase of the shift. For example, following the shift,the rat may now be rewarded for always pressing the leftlever. Accordingly, perseverative error was scored when therat pressed the right lever when the stimulus light wasilluminated above it. Eight out of every 16 consecutivetrials required the rat to respond in this manner (i.e. pressthe lever opposite to that indicated by the previouslylearned light cue rule). As in previous studies (Block etal. 2007; Floresco et al. 2006a,b, 2008; Ragozzino 2002),these types of trials were separated into consecutive blocksof eight trials each. Perseverative errors were scored when arat pressed the incorrect lever on six or more out of amaximum of eight trials per block which required the rat topress the lever that did not have the stimulus lightilluminated above it. Once a rat made fewer than fiveperseverative errors in a block for the first time, allsubsequent errors of this type were no longer counted asperseverative errors because at this point, the rat was usingthe original strategy <75% of the time. Instead, these errorswere now scored as regressive errors. The third type oferror, termed never-reinforced errors, was scored when arat pressed the incorrect lever on trials where the visual-cuelight was illuminated above the same lever that the rat wasrequired to press during the shift. Regressive and never-reinforced errors were used as an index of the animals’ability to maintain and acquire a new strategy, respectively.

Reversal learning errors were also subdivided intoperseverative and regressive subtypes and analyzed overblocks of 16 trials (Jones and Mishkin 1972; Chudasamaand Robbins 2003). Perseverative errors were scored whenrats made an incorrect response and pressed the lever that

528 Psychopharmacology (2011) 216:525–535

was reinforced during initial response discriminationtraining on day 1. Once a rat made fewer than 10perseverative errors within a block of 16 trials for the firsttime, all subsequent errors were scored as regressive.

Drugs and group allocation

The selective NMDA GluN2B antagonist (+/−)-(R*,S*)-alpha-(4-hydroxyphenyl)-beta-methyl-4-(phenylmethyl)-1-piperidine propanol (Ro25-6981) was obtained from TocrisBiosciences (Ellisville, MO, USA). This compound hasbeen shown to be 5,000 times more selective for NMDAreceptors containing GluN2B compared to those containingGluN2A subunits (Fischer et al. 1997). Ro25-6981 wasdissolved in 1 part dimethyl sulfoxide (DMSO)/2 partsphysiological saline and was injected i.p. at a dose of 6 mg/kg (injection volume, 1 ml/kg) 30-min before testing began.The vehicle for all experiments was made up of 1 partDMSO/2 parts physiological saline. Dose and route ofadministration of Ro25-6981 was chosen with reference toFox et al. (2006), Wong et al. (2007), Dalton et al. (2008),and Ge et al. (2010). This specific dose of Ro25-6981 wasalso selected because treatment with this dose has beenshown previously to selectively disrupt simpler forms ofbehavioral flexibility while leaving initial learning intact(Dalton et al. 2007, 2008). Our initial findings revealed thattreatment with 6 mg/kg of Ro25-6981 did not affect theinitial acquisition of either a visual cue or responsediscrimination on day 1. To confirm this null effect, wetested the effects of a higher, 12 mg/kg dose of Ro25-6981on initial discrimination learning (along with a separatecohort of vehicle treated rats whose data were merged withthe original control group).

Each of our experiments included the four groups: group1 (controls) received vehicle on both days 1 and 2, groups 2and 3 received Ro25-6981(6 or 12 mg/kg) on day 1 andvehicle on day 2, group 4 received vehicle on day 1 andRo25-6981 (6 mg/kg) on day 2. Prior to day 1 rats wereassigned to either a vehicle or a drug treatment group forthat day. Approximately half of the rats that receivedvehicle treatment on day 1 went on to receive drugtreatment on day 2. On day 2, these rats were matched forperformance in terms of the number of errors to achievecriterion on day 1 and allocated to either a drug or vehiclegroup. Rats that received Ro25-6981 on day 1 received avehicle injection on day 2.

Data analysis

For each phase of the experiments (initial discriminationlearning, strategy shift, or reversal), the primary dependentvariable of interest was errors to achieve criterion perfor-mance of 10 consecutive correct choices. We also analyzed

the number of trials to criterion data for each experiment.Although treatment with Ro25-6981 prior to testing onday 2 tended to increase trials to criterion relative tocontrols on both the strategy and reversal shifts, analysis ofthese data did not reveal significant effects of treatment (allFs<2.5, ns). Likewise, treatment with Ro25-6981 on day 1also did not significantly alter trials to criterion on either avisual cue or response discrimination (all Fs<1.5, n.s.). Incontrast, analyses of the number of errors [which is one ofthe most sensitive measures of animals’ performance ondiscrimination and strategy shift/reversal tasks (Castañé etal. 2010)] did reveal significant effects of treatment. Foreach experiment, error data were analyzed using two-wayANOVAs, with treatment (vehicle or drug) as the between-subjects factor and error type as a within-subjects factor.Significant effects of treatment or treatment × error typeinteractions were followed up with multiple comparisonsusing Dunnett’s test. Response latencies and locomotoractivity (indexed by the number of photobeam breaksdivided by the time required to complete the session, i.e.,beam breaks per minute), were analyzed with one-wayANOVAs. Trial omissions, which occurred infrequentlyacross both experiments (<2 per treatment group) were alsoanalyzed with one-way ANOVAs.

Results

Experiment 1: NMDA GluN2B receptor blockadeand strategy shifting from a visual-cue to a response strategy

Day 1: visual-cue discrimination learning Analysis of thenumber of errors to criterion by rats receiving injections ofRo25-6981 or vehicle revealed no significant main effect oftreatment [F(3,37)=0.68, ns]. Rats receiving either 6 mg/kg(n=8) or 12 mg/kg (n=7) of Ro25-6981 on day 1 were notimpaired in learning a visual-cue discrimination, making acomparable number of errors to control rats (n=15) andrats designated to receive Ro25-6981 on day 2 (n=11)(Fig. 1a). Similarly, there were no differences betweengroups in average response latency, locomotor activity, ornumber of omissions (all Fs<1.2, ns). Thus, systemicblockade of NMDA GluN2B receptors with Ro25-6981does not disrupt the initial learning of a simple visual cuediscrimination strategy.

Day 2: shift to response strategy Analysis of the error dataobtained during the shift to the response strategy from ratsreceiving injections of Ro25-6981 or vehicle revealed asignificant treatment-by-error-type interaction [F(6,74)=2.65; p<0.01; Fig. 1b, c]. Simple main effects analysesrevealed that rats receiving Ro25-6981 during the shiftmade significantly more perseverative errors (p<0.005)

Psychopharmacology (2011) 216:525–535 529

relative to control rats (Fig. 1c). However, the number ofregressive and never reinforced errors were unaffected bytreatment with Ro25-6981. In contrast, rats that receivedRo25-6981 during visual-cue discrimination learning onday 1 did not differ from controls on these measures. Therewere no differences between groups in average responselatency, number of omissions or locomotor activity (all Fs<1.5, ns). Thus, pretreatment with Ro25-6981 impairedshifting from a visual-cue to a response-based discrimina-tion strategy. The selective increase in perseverative errorsinduced by this treatment suggests that this impairment isattributable to a disruption in the ability to cease attendingto a previously relevant strategy, as opposed to theacquisition or maintenance of a novel strategy.

Experiment 2: NMDA GluN2B receptor blockadeand response reversal learning

Day 1: response discrimination training In a separate groupof rats, treatment with Ro25-6981 had no effect on acquisitionof an egocentric response strategy. Analysis of the number oferrors made revealed no significant main effect of treatment[F(3,45)=1.55, ns]. Rats receiving either 6 mg/kg (n=10) or12 mg/kg (n=8) Ro25-6981 on day 1 were not impaired inlearning a simple response discrimination relative to controlrats (n=19) or rats that would receive Ro25-6981 on day 2(n=12) (Fig. 2a). In a similar manner, there were nodifferences between groups in average response latency ornumber of omissions (all Fs<1.7, ns). Thus, blockade ofNMDA GluN2B receptors does not impair the initialacquisition of a response discrimination.

Day 2: response reversal Similar to the results of experi-ment 1, blockade of NMDA GluN2B receptors prior totesting on day 2 increased the number of errors committedduring the reversal. Analysis of the number of errors tocriterion revealed both a significant main effect of treatment[F(3,45)=5.42; p<0.01; Fig. 2b] and a significant treatment-by-error-type interaction [F(3,45)=3.05; p<0.05]. Dunnett’stests revealed that rats receiving Ro25-6981 during thereversal made significantly more errors (p<0.05) and, inparticular, more perseverative errors (p<0.01) than controlrats or those given Ro25-6981 during initial responsediscrimination training on day 1 (Fig. 2c). Rats that hadreceived Ro25-6981 on day 1 tended to make fewer overallerrors than controls during the reversal, but this effect wasnot statistically significant (p>0.085). Moreover, rats inthese groups did not differ from controls in the number ofperseverative errors during the reversal. There were nodifferences between groups in average response latency,number of omissions or locomotor activity (all Fs<1.5, ns).Thus, pretreatment with Ro25-6981 impairs the ability to

Fig. 1 Experiment 1: systemic blockade of NMDA GluN2B receptorsimpairs strategy shifting but does not affect initial visual-cuediscrimination learning. Data are expressed as mean ± SEM. A boxsurrounding X-axis labels denotes that group received drug on thatday. a Rats given either a 6 mg/kg (hatched bar) or 12 mg/kg dose(grey bar) of Ro25-6981 on day 1 did not differ from controls (whitebars) in the number of errors to achieve criterion on a visual-cuediscrimination task. b Number of errors made on day 2, when ratswere required to shift from a visual cue to a response based strategy. cAnalysis of the type of errors made during the strategy shift. Ratstreated with Ro25-6981 during the shift (6 mg/kg, black bars) showeda selective increase in perseverative errors compared to vehicle-treatedcounterparts. Stars denote p<0.05 significant difference vs vehicle-treated counterparts

530 Psychopharmacology (2011) 216:525–535

shift between different stimulus–reward associations,assessed with a spatial reversal task. As was observed inexperiment 1, this impairment manifested itself as aselective increase in perseverative responding.

Discussion

Here, we report that systemic blockade of GluN2B-containing NMDA receptors with the selective antagonistRo25-6981 induces selective impairments in higher orderforms of behavioral flexibility. Strategy shifting requiresthat an animal learn to respond according to a specificlearned rule and then suppress responding according to thatrule and learn a new strategy after changes in reinforcementcontingencies. In comparison, reversal learning entails useof a previously acquired strategy (e.g., always choose aspecific lever location), but to shift between differentstimulus–reward associations within that strategy. Blockadeof GluN2B receptors prior to strategy shifting or reversallearning impaired the ability to shift between or within aparticular strategy. Importantly, these treatments did notimpair the initial learning of either a response or visual-cuediscrimination, indicating that activity at these particularNMDA receptor sites does not play a critical role indiscrimination learning. In addition, GluN2B blockade hadno effect on response latencies, locomotion, or trialomissions, ruling out the possibility that these effects weredue to general disruptions in motoric or motivationalprocesses. Thus, the present findings suggest that NMDAGluN2B receptors make a selective contribution to learningunder conditions that require modifications of previouslyacquired associative memories.

Ro25-6981, the GluN2B receptor antagonist used here, is ahighly selective non-competitive GluN2B receptor antagonist,blocking these receptors in an activity-dependent manner(Fischer et al. 1997; Lynch et al. 2001; Mutel et al. 1998). Itis 5,000 times more selective for GluN2B over GluN2Asubunits and is approximately 25-fold more potent thanifenprodil as an antagonist at GluN2B subunit-containingNMDA receptors (Fischer et al. 1997). In addition, Ro25-6981 shows significantly lower affinity for σ and 5-HT1Areceptors (Mutel et al. 1998) as well as α1 adrenoceptors(Pinard et al. 2001) compared to ifenprodil, another GluN2Bantagonist, which actually displays a higher affinity for α1

adrenoceptors than it does for GluN2B receptors (Chenard etal. 1991; Chenard and Menniti 1999). Thus, it is highlyprobable that the effects of Ro25-6981 on strategy shiftingand reversal learning reported here can be attributed primarilyto blockade of GluN2B containing NMDA receptors.

Although strategy shifting and reversal learning may appearto be related cognitive strategies on a superficial level, they

Fig. 2 Experiment 2: systemic blockade of NDMA GluN2B receptorsimpairs reversal learning without affecting initial acquisition of anegocentric response rule. Data are expressed as mean ± SEM. A boxsurrounding X-axis labels denotes that group received drug on thatday. a Either a 6 mg/kg (hatched bar) or 12 mg/kg dose (grey bar) ofRo25-6981 on day 1 did not differ from controls (white bars) in thenumber of errors to achieve criterion on an egocentric responsediscrimination task. b Rats given Ro25-6981 (6 mg/kg) on day 2(black bars) made significantly more errors compared to controlsduring reversal learning. c Analysis of the type of errors made duringthe set shift. Rats treated with Ro25-6981 during the reversal showedincreased perseveration toward the previously reinforced levercompared to vehicle-treated counterparts. Stars denote p<0.05significant difference vs vehicle-treated counterparts

Psychopharmacology (2011) 216:525–535 531

display substantial dissociations in terms of the neural circuitryand neurochemistry underlying their execution. In rodents, ithas been shown repeatedly that lesions or inactivation of themPFC impair strategy or attentional set-shifting, but notreversal learning, whereas the orbitofrontal cortex plays a keyrole in reversal learning, but not set-shifting (Bissonette et al.2008; Boulougouris et al. 2007; Floresco et al. 2008; Ghods-Sharifi et al. 2008; McAlonan and Brown 2003; Ragozzino etal. 1999). Furthermore, reversal learning is sensitive todepletions of orbitofrontal 5-HT, but not dopamine, whereasprefrontal dopamine is essential for acquiring an attentionalset and shifting between strategies (Clarke et al. 2004, 2007;Crofts et al. 2001; Floresco et al. 2006a). Reversible lesions orblockade of dopamine D1 receptors in the ventral striatumselectively impair strategy shifting, whereas dopamine trans-mission in the dorsal striatum may be more critical forreversal performance (Castañé et al. 2010; Floresco et al.2006b; Haluk and Floresco 2009; O’Neill and Brown 2007).In light of the clear dissociations in the neural andneurochemical mechanisms underlying set-shifting and rever-sal learning the present data provide the novel insight thatNMDA GluN2B receptor activity may serve as a commonmechanism for the facilitation of both forms of behavioralflexibility. Moreover, the fact that treatment with thisantagonist selectively increased perseverative errors duringboth types of shifts suggests that these receptors play a keyrole in disengaging from older redundant strategies and theinhibition of a prepotent response.

It is notable that the selective increase in perseverativeresponding induced by Ro25-6981 is similar to that observedafter lesions, inactivations, or non-selective blockade ofNMDA receptors in the mPFC or orbital PFC (Boulougouriset al. 2007; Floresco et al. 2008; Ragozzino et al. 1999;Stefani and Moghaddam 2005). This suggests that the effectsof systemic administration of Ro25-6981 may be attributableto blockade of GluN2B receptors within the frontal lobes.GluN2B-containing NMDA receptors show high levels ofexpression not only within the PFC but also in other regionsof the forebrain, including the olfactory bulb, hippocampus,thalamus, and, in particular, the striatum (Jin et al. 1997;Monyer et al. 1994; Rigby et al. 1996; Wang et al. 1995;Watanabe et al. 1993; Wenzel et al. 1997). Learning of simplevisual or response discriminations is likely to be mediated byregions of the dorsal striatum (Palencia and Ragozzino 2005),and alterations in behavior required during shifts may befacilitated by top–down control by frontal regions over striatalsystems that process these rules (Block et al. 2007; Kehagia etal. 2010). Thus, it is possible that descending glutamatergiccortico-striatal pathways involved in behavioral modificationmay act on GluN2B receptors to enable these forms ofbehavioral flexibility. It follows that the enhanced persevera-tion induced by Ro25-6981 blockade may also be attributableto perturbations of NMDA GluN2B receptor-mediated trans-

mission at these cortico-striatal synapses, which may play amore selective role in modifying previously acquired memo-ries rather than the acquisition of novel ones.

Another key finding of the present study is that even thoughRo25-6981 impaired shifting between response and visual cue-based strategies, these treatments did not impair learning of aninitial discrimination that was acquired within one trainingsession. Treatment with either a dose that was effective atdisrupting strategy shifting and reversal learning (6 mg/kg) or ahigher dose (12mg/kg) was ineffective at inducing a significantincrease in errors during training on day 1. This latter findingdiffers from numerous psychopharmacological studies demon-strating that systemic blockade of NMDA receptors usingrelatively non-selective antagonists induce deleterious effectson multiple forms of learning. These include Pavlovianaversive conditioning (Goosens and Maren 2004; Gould etal. 2002), spatial learning (Riekkinen et al. 1996; Ylinen et al.1995), and discrimination learning (Murray et al. 1995;Stefani and Moghaddam 2005). More selective blockade ofGluN2B receptors with Ro25-6981 does not interfere withspatial or fear learning (Dalton et al. 2007; Duffy et al. 2008).Likewise, we observed that blockade of GluN2B on day 1 oftraining did not impair initial visual or response discriminationlearning. This indicates that our observed impairments inflexibility cannot be attributed to a disruption in learning of anovel strategy or stimulus–reward association during the shift.Note also that rats treated with Ro25-6981 on day 1 displayedlevels of perseveration on day 2 that were comparable tocontrol animals, indicating that they were responding inaccordance with the previously acquired rule. This furtherindicates that treatment with Ro25-6981 did not impair theconsolidation of the initially acquired discrimination. This isnot to say that GluN2B-containing NMDA receptors do notcontribute at all to certain forms of discrimination learning.Indeed, more complex discriminations that require multipledays of training to achieve asymptotic performance and havelonger consolidation periods may be more sensitive toGluN2B receptor blockade (Dix et al. 2010). Indeed, whenlearning takes place over several days, mice lacking GluN2Breceptors in the forebrain show significant impairments incortico-hippocampal-dependent tasks (von Engelhardt et al.2008; Brigman et al. 2010). A similar discrepancy has beenobserved in genetic studies using mice lacking GluN2Areceptors that show impairments in spatial learning whentested in an acute setting, while GluN2A deletion has noeffect in spatial learning when learning takes place overmultiple days (Bannerman et al. 2008). Nevertheless, thepresent data, in addition to the abovementioned findingssuggest that unlike non-selective NMDA receptor blockade,selective antagonism of GluN2B-containing NMDA receptorsdoes not appear to interfere with the relatively rapidacquisition of new memories or rules. Rather, these receptorsappear to play a more prominent role in the modification of

532 Psychopharmacology (2011) 216:525–535

already-acquired memories/associations/rules thereby facilitat-ing suppression of patterns of behavior that are no longerrelevant so that more adaptive ones may emerge.

It is important to emphasize that behavioral flexibility is nota unitary phenomenon but, rather, a hierarchical set ofprocesses. Reversal learning and set-shifts between differentstimulus–reward associations or strategies represent two morecomplex forms of flexibility. In comparison, simpler formsinclude response inhibition, requiring executive control torapidly suppress pre-potent actions triggered by internal/external cues (Eagle et al. 2008). Extinction entails a moregradual suppression of conditioned responses upon repeatednon-reinforced presentations of an appetitive or aversiveconditioned stimuli. These relatively simpler forms ofbehavioral flexibility also display considerable dissociationsin terms of their neurochemical underpinnings (see Florescoand Jentsch 2011, for a review). Emerging evidence impli-cates GluN2B receptors in these functions as well. Thus,GluN2B receptor antagonism with Ro63-1908 or CP101-606increases premature responding on a serial reaction time task,indicating that these receptors may be particularly importantin mediating “waiting,” acting as a brake in the control ofimpulsive responses (Higgins et al. 2003). Similarly, Ro25-6981 disrupts within-session extinction of a learned fearresponse (Dalton et al. 2008). In keeping with the presentstudy, treatment with this compound also impairs reversallearning assessed with a water maze task (Duffy et al. 2008).When viewed collectively, these studies, along with thepresent data suggest that activity at NMDA GluN2B-containing receptors may make a fundamental contributionto multiple forms of behavioral flexibility, facilitating thesuppression of innate or previously learned pre-potentresponses so that more adaptive forms of behaviors may beexpressed. As such, development of compounds that mayselectively enhance activity at these receptors may lead tonovel treatments for numerous psychiatric disorders in whichimpaired inhibitory control and aberrant perseverative tenden-cies are prominent symptoms.

Acknowledgments This work was supported by grants from theNatural Sciences and Engineering Research Council of Canada toAGP and SBF. SBF is a Michael Smith Foundation for HealthResearch Senior Scholar.

References

Bannerman DM, Niewoehner B, Lyon L, Romberg C, Schmitt WB,Taylor A, Sanderson DJ, Cottam J, Sprengel R, Seeburg PH,Köhr G, Rawlins JN (2008) NMDA receptor subunit NR2A isrequired for rapidly acquired spatial working memory but notincremental spatial reference memory. J Neurosci 28:3623–3630

Bissonette GB, Martins GJ, Franz TM, Harper ES, Schoenbaum G,Powell EM (2008) Double dissociation of the effects of medial

and orbital prefrontal cortical lesions on attentional and affectiveshifts in mice. J Neurosci 28:11124–11130

Block AE, Dhanji H, Thompson-Tardif SF, Floresco SB (2007)Thalamic-prefrontal cortical-ventral striatal circuitry mediatesdissociable components of strategy set shifting. Cereb Cortex17:1625–1636

Boulougouris V, Dalley JW, Robbins TW (2007) Effects of orbito-frontal, infralimbic and prelimbic cortical lesions on serial spatialreversal learning in the rat. Behav Brain Res 179:219–228

Brigman JL, Feyder M, Saksida LM, Bussey TJ, Mishina M, HolmesA (2008) Impaired discrimination learning in mice lacking theNMDA receptor NR2A subunit. Learn Mem 15:50–54

Brigman JL, Wright T, Talani G, Prasad-Mulcare S, Jinde S, SeaboldGK, Mathur P, Davis MI, Bock R, Gustin RM, Colbran RJ,Alvarez VA, Nakazawa K, Delpire E, Lovinger DM, Holmes A(2010) Loss of GluN2B-containing NMDA receptors in CA1hippocampus and cortex impairs long-term depression, reducesdendritic spine density, and disrupts learning. J Neurosci30:4590–4600

Castañé A, Theobald DE, Robbins TW (2010) Selective lesions of thedorsomedial striatum impair serial spatial reversal learning inrats. Behav Brain Res 210:74–83

Chenard BL, Menniti FS (1999) Antagonists selective for NMDAreceptors containing the NR2B subunit. Curr Pharm Des 5:381–404

Chenard BL, Shalaby IA, Koe BK, Ronau RT, Butler TW, ProchniakMA, Schmidt AW, Fox CB (1991) Separation of alpha 1adrenergic and N-methyl-D-aspartate antagonist activity in aseries of ifenprodil compounds. J Med Chem 34:3085–3090

Chudasama Y, Robbins TW (2003) Dissociable contributions of theorbitofrontal and infralimbic cortex to pavlovian autoshaping anddiscrimination reversal learning: further evidence for the func-tional heterogeneity of the rodent frontal cortex. J Neurosci23:8771–8780

Clarke HF, Dalley JW, Crofts HS, Robbins TW, Roberts AC (2004)Cognitive inflexibility after prefrontal serotonin depletion. Sci-ence 304:878–880

Clarke HF, Walker SC, Dalley JW, Robbins TW, Roberts AC (2007)Cognitive inflexibility after prefrontal serotonin depletion isbehaviorally and neurochemically specific. Cereb Cortex 17:18–27

Collingridge GL, Olsen RW, Peters J, Spedding M (2009) A nomencla-ture for ligand-gated ion channels. Neuropharmacology 56:2–5

Crofts HS, Dalley JW, Collins P, Van Denderen JC, Everitt BJ,Robbins TW, Roberts AC (2001) Differential effects of 6-OHDAlesions of the frontal cortex and caudate nucleus on the ability toacquire an attentional set. Cereb Cortex 11:1015–1026

Dalton, GL, Floresco SB, Phillips AG (2007) Blockade of NMDANR2Breceptors disrupts the extinction, but not acquisition, of auditory fearconditioning: dependence on context. Soc Neurosci Abst 91.21

Dalton GL, Wang YT, Floresco SB, Phillips AG (2008) Disruption ofAMPA receptor endocytosis impairs the extinction, but not acquisi-tion of learned fear. Neuropsychopharmacology 33:2416–2426

Dix S, Gilmour G, Potts S, Smith JW, Tricklebank M (2010) Awithin-subject cognitive battery in the rat: differential effects of NMDAreceptor antagonists. Psychopharmacology 212:227–242

Duffy S, Labrie V, Roder JC (2008) D-serine augments NMDA-NR2Breceptor-dependent hippocampal long-term depression and spa-tial reversal learning. Neuropsychopharmacology 33:1004–1018

Eagle DM, Bari A, Robbins TW (2008) The neuropsychopharmacol-ogy of action inhibition: cross-species translation of the stop-signal and go/no-go tasks. Psychopharmacology 199:439–456

Egerton A, Reid L, McKerchar CE, Morris BJ, Pratt JA (2005)Impairment in perceptual attentional set-shifting following PCPadministration: a rodent model of set-shifting deficits inschizophrenia. Psychopharmacology (Berl) 179:77–84

Fischer G, Mutel V, Trube G, Malherbe P, Kew JN, Mohacsi E, HeitzMP, Kemp JA (1997) Ro 25-6981, a highly potent and selective

Psychopharmacology (2011) 216:525–535 533

blocker of N-methyl-D-aspartate receptors containing the NR2Bsubunit. Characterization in vitro. J Pharmacol Exp Ther283:1285–1292

Floresco SB, Jentsch JD (2011) Pharmacological enhancement ofmemory and executive functioning in laboratory animals. Neuro-psychopharmacology 36:227–250

Floresco SB, Magyar O, Ghods-Sharifi S, Vexelman C, Tse MT(2006a) Multiple dopamine receptor subtypes in the medialprefrontal cortex of the rat regulate set-shifting. Neuropsycho-pharmacology 31:297–309

Floresco SB, Ghods-Sharifi S, Vexelman C, Magyar O (2006b)Dissociable roles for the nucleus accumbens core and shell inregulating set shifting. J Neurosci 26:2449–2457

Floresco SB, Block AE, Tse MT (2008) Inactivation of the medialprefrontal cortex of the rat impairs strategy set-shifting, but notreversal learning, using a novel, automated procedure. BehavBrain Res 190:85–96

Floresco SB, Zhang Y, Enomoto T (2009) Neural circuits subservingbehavioral flexibility and their relevance to schizophrenia. BehavBrain Res 204:396–409

Fox CJ, Russell KI, Wang YT, Christie BR (2006) Contribution ofNR2A and NR2B NMDA subunits to bidirectional synapticplasticity in the hippocampus in vivo. Hippocampus 16:907–915

Ge Y, Dong Z, Bagot RC, Howland JG, Phillips AG, Wong TP, WangYT (2010) Hippocampal long-term depression is required for theconsolidation of spatial memory. Proc Natl Acad Sci USA 107(38):16697–16702

Ghods-Sharifi S, Haluk DM, Floresco SB (2008) Differential effectsof inactivation of the orbitofrontal cortex on strategy set-shiftingand reversal learning. Neurobiol Learn Mem 89:567–573

Goldberg TE, Weinberger DR, Berman KF, Pliskin NH, Podd MH(1987) Further evidence for dementia of the prefrontal type inschizophrenia? A controlled study of teaching the WisconsinCard Sorting Test. Arch Gen Psychiatry 44:1008–1014

Goosens KA, Maren S (2004) NMDA receptors are essential for theacquisition, but not expression, of conditional fear and associativespike firing in the lateral amygdala. Eur J Neurosci 20:537–548

Gould TJ, McCarthy MM, Keith RA (2002) MK-801 disruptsacquisition of contextual fear conditioning but enhances memoryconsolidation of cued fear conditioning. Behav Pharmacol13:287–294

Haluk DM, Floresco SB (2009) Ventral striatal dopamine modulationof different forms of behavioral flexibility. Neuropsychopharma-cology 34:2041–2052

Higgins GA, Ballard TM, Huwyler J, Kemp JA, Gill R (2003) Evaluationof the NR2B-selective NMDA receptor antagonist Ro 63–1908 onrodent behaviour: evidence for an involvement of NR2B NMDAreceptors in response inhibition. Neuropharmacology 44:324–341

Idris NF, Repeto P, Neill JC, Large CH (2005) Investigation of the effectsof lamotrigine and clozapine in improving reversal-learning impair-ments induced by acute phencyclidine and D-amphetamine in therat. Psychopharmacology (Berl) 179:336–348

Jin DH, Jung YW, Ko BH, Moon IS (1997) Immunoblot analyses onthe differential distribution of NR2A and NR2B subunits in theadult rat brain. Mol Cells 7:749–754

Jones B, Mishkin M (1972) Limbic lesions and the problem ofstimulus-reinforcement associations. Exp Neurol 36:362–377

Kehagia AA, Murray GK, Robbins TW (2010) Learning and cognitiveflexibility: frontostriatal function and monoaminergic modula-tion. Curr Opin Neurobiol 20:199–204

Laurie DJ, Bartke I, Schoepfer R, Naujoks K, Seeburg PH (1997)Regional, developmental and interspecies expression of the fourNMDAR2 subunits, examined using monoclonal antibodies.Brain Res Mol Brain Res 51:23–32

Lombardi WJ, Andreason PJ, Sirocco KY, Rio DE, Gross RE, UmhauJC, Hommer DW (1999) Wisconsin Card Sorting Test perfor-

mance following head injury: dorsolateral fronto-striatal circuitactivity predicts perseveration. J Clin Exp Neuropsychol 21:2–16

Lynch DR, Shim SS, Seifert KM, Kurapathi S, Mutel V, GallagherMJ, Guttmann RP (2001) Pharmacological characterization ofinteractions of RO 25-6981 with the NR2B (epsilon2) subunit.Eur J Pharmacol 416:185–195

McAlonan K, Brown VJ (2003) Orbital prefrontal cortex mediatesreversal learning and not attentional set shifting in the rat. BehavBrain Res 146:97–103

Mesulam M-M (2002) The human frontal lobes: transcending thedefault mode through contingent encoding. In: Stuss DT, KnightRT (eds) Principles of frontal lobe function. Oxford UniversityPress, Oxford, pp 8–30

Monyer H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH (1994)Developmental and regional expression in the rat brain andfunctional properties of four NMDA receptors. Neuron 12(3):529–540

Morris RG (1989) Synaptic plasticity and learning: selective impair-ment of learning rats and blockade of long-term potentiation invivo by the N-methyl-D-aspartate receptor antagonist AP5. JNeurosci 9:3040–3057

Morris RG, Anderson E, Lynch GS, Baudry M (1986) Selectiveimpairment of learning and blockade of long-term potentiation byan N-methyl-D-aspartate receptor antagonist, AP5. Nature319:774–776

Murray TK, Ridley RM, Snape MF, Cross AJ (1995) The effect ofdizocilpine (MK-801) on spatial and visual discrimination tasksin the rat. Behav Pharmacol 6:540–549

Mutel V, Buchy D, Klingelschmidt A, Messer J, Bleuel Z, Kemp JA,Richards JG (1998) In vitro binding properties in rat brain of [3H]Ro25-6981, a potent and selective antagonist of NMDA receptorscontaining NR2B subunits. J Neurochem 70:2147–2155

O"Neill M, Brown VJ (2007) The effect of striatal dopamine depletionand the adenosine A2A antagonist KW-6002 on reversal learningin rats. Neurobiol Learn Mem 88:75–81

Palencia CA, Ragozzino ME (2005) The contribution of NMDAreceptors in the dorsolateral striatum to egocentric responselearning. Behav Neurosci 119:953–960

Pantelis C, Barber FZ, Barnes TR, Nelson HE, Owen AM, RobbinsTW (1999) Comparison of set-shifting ability in patients withchronic schizophrenia and frontal lobe damage. Schizophr Res37:251–270

Pinard E, Alanine A, Bourson A, Büttelmann B, Gill R, Heitz M,Jaeschke G, Mutel V, Trube G, Wyler R (2001) Discovery of (R)-1-[2-hydroxy-3-(4-hydroxy-phenyl)-propyl]-4-(4-methyl-ben-zyl)-piperidin-4-ol: a novel NR1/2B subtype selective NMDAreceptor antagonist. Bioorg Med Chem Lett 11:2173–2176

Ragozzino ME (2002) The effects of dopamine D(1) receptorblockade in the prelimbic-infralimbic areas on behavioralflexibility. Learn Mem 9:18–28

Ragozzino ME, Detrick S, Kesner RP (1999) Involvement of theprelimbic-infralimbic areas of the rodent prefrontal cortex inbehavioral flexibility for place and response learning. J Neurosci19:4585–4594

Riekkinen M, Stefanski R, Kuitunen J, Riekkinen P Jr (1996) Effectsof combined block of alpha 1-adrenoceptors and NMDAreceptors on spatial and passive avoidance behavior in rats. EurJ Pharmacol 300(1–2):9–16

Rigby M, Le Bourdellès B, Heavens RP, Kelly S, Smith D, Butler A,Hammans R, Hills R, Xuereb JH, Hill RG, Whiting PJ,Sirinathsinghji DJ (1996) The messenger RNAs for the N-methyl-D-aspartate receptor subunits show region-specific ex-pression of different subunit composition in the human brain.Neuroscience 73:429–447

Rubia K, Smith AB, Brammer MJ, Taylor E (2007a) Temporal lobedysfunction in medication-naive boys with attention-deficit/

534 Psychopharmacology (2011) 216:525–535

hyperactivity disorder during attention allocation and its relationto response variability. Biol Psychiatry 62:999–1006

Rubia K, Smith AB, Brammer M, Taylor E (2007b) Performance ofchildren with attention deficit hyperactivity disorder (ADHD) ona test battery for impulsiveness. Child Neuropsychol 30:659–695

Stefani MR, Moghaddam B (2005) Systemic and prefrontal corticalNMDA receptor blockade differentially affect discriminationlearning and set-shift ability in rats. Behav Neurosci 119:420–428

Stefani MR, Moghaddam B (2010) Activation of type 5 metabotropicglutamate receptors attenuates deficits in cognitive flexibilityinduced by NMDA receptor blockade. Eur J Pharmacol 639:26–32

van der Meulen JA, Bilbija L, Joosten RN, de Bruin JP, Feenstra MG(2003) The NMDA-receptor antagonist MK-801 selectivelydisrupts reversal learning in rats. NeuroReport 14:2225–2228

von Engelhardt J, Doganci B, Jensen V, Hvalby Ø, Göngrich C,Taylor A, Barkus C, Sanderson DJ, Rawlins JN, Seeburg PH,Bannerman DM, Monyer H (2008) Contribution of hippocampaland extra-hippocampal NR2B-containing NMDA receptors toperformance on spatial learning tasks. Neuron 60:846–860

Wang YH, Bosy TZ, Yasuda RP, Grayson DR, Vicini S, Pizzorusso T,Wolfe BB (1995) Characterization of NMDA receptor subunit-specific antibodies: distribution of NR2A and NR2B receptor

subunits in rat brain and ontogenic profile in the cerebellum. JNeurochem 65:176–183

Watanabe M, Inoue Y, Sakimura K, Mishina M (1993) Distinctdistributions of five N-methyl-D-aspartate receptor channelsubunit mRNAs in the forebrain. J Comp Neurol 338:377–390

Wenzel A, Fritschy JM, Mohler H, Benke D (1997) NMDA receptorheterogeneity during postnatal development of the rat brain:differential expression of the NR2A, NR2B, and NR2C subunitproteins. J Neurochem 68:469–478

Willcutt EG, Doyle AE, Nigg JT, Faraone SV, Pennington BF (2005)Validity of the executive function theory of attention-deficit/hyperactivity disorder: a meta-analytic review. Biol Psychiatry57:1336–1346

Wong TP, Howland JG, Robillard JM, Ge Y, Yu W, Titterness AK,Brebner K, Liu L, Weinberg J, Christie BR, Phillips AG, WangYT (2007) Hippocampal long-term depression mediates acutestress-induced spatial memory retrieval impairment. Proc NatlAcad Sci USA 104(27):11471–11476

Ylinen A, Pitkänen M, Sirviö J, Hartikainen T, Sivenius J, Koivisto E,Riekkinen PJ Sr (1995) The effects of NMDA receptorantagonists at anticonvulsive doses on the performance of ratsin the water maze task. Eur J Pharmacol 274:159–165

Psychopharmacology (2011) 216:525–535 535