ORIGINAL INVESTIGATION

Changes in cue reactivity and attentional bias followingexperimental cue exposure and response prevention:a laboratory study of the effects of D-cycloserinein heavy drinkers

Sunjeev K. Kamboj & Rachel Massey-Chase &

Lydia Rodney & Ravi Das & Basil Almahdi &H. Valerie Curran & Celia J. A. Morgan

Received: 15 December 2010 /Accepted: 1 March 2011# Springer-Verlag 2011

AbstractRationale The effects of D-cycloserine (DCS) in animalmodels of anxiety disorders and addiction indicate a role forN-methyl D-aspartate (NMDA) receptors in extinctionlearning. Exposure/response prevention treatments foranxiety disorders in humans are enhanced by DCS,suggesting a promising co-therapy regime, mediated byNMDA receptors. Exposure/response prevention may alsobe effective in problematic drinkers, and DCS mightenhance habituation to cues in these individuals. Sinceheavy drinkers show ostensible conditioned responses toalcohol cues, habituation following exposure/responseprevention should be evident in these drinkers, with DCSenhancing this effect.Objectives The objective of this study is to investigate theeffect of DCS on exposure/response prevention in heavydrinkers.Methods In a randomised, double-blind, placebo-controlled study, heavy social drinkers recruited from

the community received either DCS (125 mg; n=19) orplacebo (n=17) 1 h prior to each of two sessions ofexposure/response prevention. Cue reactivity and atten-tional bias were assessed during these two sessions and ata third follow-up session. Between-session drinkingbehaviour was recorded.Results Robust cue reactivity and attentional bias to alcoholcues was evident, as expected of heavy drinkers. Within-and between-session habituation of cue reactivity, as well asa reduction in attentional bias to alcohol cues over time wasfound. However, there was no evidence of greater habituationin the DCS group. Subtle stimulant effects (increasedsubjective contentedness and euphoria) which wereunrelated to exposure/response prevention were foundfollowing DCS.Conclusions DCS does not appear to enhance habituationof alcohol cue reactivity in heavy non-dependent drinkers.Its utility in enhancing treatments based on exposure/response prevention in dependent drinkers or drug usersremains open.

Keywords Cue exposure . Response prevention .

D-Cycloserine . Cue reactivity . Attentional bias . Alcohol .

Dependence . Hazardous drinking . Habituation . Extinction

Introduction

Psychological and pharmacological treatments for addic-tions often aim to reduce reactivity to the drug cues thatmotivate drug-seeking behaviour. For example, treatmentsbased on a classical conditioning model of alcohol use,such as ‘exposure and response prevention’ are expected to

S. K. Kamboj (*) :R. Massey-Chase : L. Rodney : R. Das :H. V. Curran : C. J. A. MorganClinical Psychopharmacology Unit, Research Departmentof Clinical, Educational and Health Psychology,University College London,Gower Street,London WC1E 6BT, UKe-mail: sunjeev.kamboj@ucl.ac.uk

B. AlmahdiCentre of Anaesthesia, University College Hospital,University College London,Gower Street,London WC1E 6BT, UK

PsychopharmacologyDOI 10.1007/s00213-011-2254-z

cause extinction or habituation1 of cue reactivity (Drum-mond et al. 1995).

The process of habituation to alcohol cues encompassesreductions in subjective craving, physiological arousal andattentional bias which—on the basis of a classical condi-tioning model of addiction (Drummond et al. 1995)—should decrease over repeated unreinforced trials ofexposure to cues. While a number of studies with alcohol-dependent patients have shown promising results usingexposure/response prevention (e.g. Monti et al. 1993;Drummond and Glautier 1994; Sitharthan et al. 1997), theefficacy of cue exposure treatments more generally has yetto be convincingly demonstrated (Conklin and Tiffany2002). In particular, relapse rates are often high asindividuals remain susceptible to the effects of renewal,spontaneous recovery and reinstatement following re-exposure to alcohol-related cues (Bouton 2002). Refinementof cue exposure therapies to enhance consolidation ofextinction memories, or to strengthen memory tracesassociated with conditioned response-opposed behaviour(response ‘prevention’) have been suggested as means ofovercoming the limited efficacy of current cue exposureprocedures in addictions (Conklin and Tiffany 2002). Theserefinements might involve modification of exposure/re-sponse prevention procedures themselves (for example,incorporating ‘extinction reminders’ during exposure/re-sponse prevention), although another possibility is nootropicenhancement of exposure/response prevention (Hoffman2007).

Extinction learning (including extinction of conditioneddrug-cue reactivity) is mediated by central N-methyl D-aspartate receptors (NMDA-Rs; e.g. Lee and Kim 1998).Augmenting NMDA-R activity—for example using D-cycloserine (DCS), a partial agonist at the glycine site ofthe NMDA-R—is one potential strategy for enhancingexposure/response prevention. Indeed, there have been anumber of demonstrations of enhanced consolidation of

fear extinction following DCS treatment in rodents (e.g.Walker et al. 2002; Ledgerwood et al. 2003).

Given the low toxicity and established safety of DCS,animal learning studies of fear extinction have quickly beentranslated into studies on anxiety disorders in humans. Anumber of clinical trials (Norberg et al. 2008) have nowshown improved outcomes in anxiety disorder patientstreated with combined exposure/response prevention-DCS(however, see Guastella et al. 2007 and Storch et al. 2007for exceptions). On the basis of these findings a number of‘critical parameters’ for DCS have been proposed (Roth-baum 2008). For example, doses of 50–125 mg adminis-tered 1 h before learning may be optimal for exposure-based interventions in anxiety disorders. These relativelylow doses are less likely to produce antagonist-like effectswhich are expected of partial agonists at higher doses,especially when concentrations of the natural agonist are atsaturating levels. Furthermore, studies examining the time-course of enhancement of exposure treatments suggest thatthe early phases are most affected by DCS, with the largestdifferential treatment effects evident within the first twosessions (Chasson et al. 2010).

In animal models of addiction (e.g. cocaine dependence)DCS enhances extinction of conditioned place preference(CPP) (Botreau et al. 2006; Nic Dhonnchadha et al. 2010).On this basis—and considering the effects reported inhuman trials of DCS in exposure treatments for anxietydisorders—it is timely to examine the effects of DCS inexposure/response prevention for substance dependencedisorders.

Given the potential for iatrogenic effects of exposure/response prevention in some vulnerable individuals (Marissenet al. 2007) a cautious approach is warranted in applyingDCS as a co-therapy in substance dependence. In a recentdescriptive study, Price et al. (2009) suggest that com-bined exposure/response prevention-DCS treatment mayacutely enhance ‘response activation’ in cocaine-dependent individuals. This effect is difficult to reconcilewith the extant literature on extinction of CPP with DCSin animal models of cocaine dependence but may point tosubjective effects of DCS on craving which are indepen-dent of exposure/response prevention and have yet to befully explored. In a pilot study of combined exposure/response prevention-DCS treatment, Santa Ana et al.(2009) examined relatively light, non-abstaining, non-treatment seeking tobacco smokers. Greater within-sessionreduction in skin conductance responding during thesecond session, and lower carbon monoxide levels at oneweek follow-up were found in the DCS group. Overallhowever, there was no clear evidence of enhancedbetween-session habituation, which would be the pre-dicted effect of combined exposure/response prevention-DCS, reflecting greater consolidation of extinction learning

1 We have tended to use the term ‘habituation’ instead of ‘extinction’throughout this paper to describe reductions in cue reactivity. Webelieve this distinction is important: although these terms are usedinterchangeably in the literature, ‘extinction’ is perhaps more clearlydefined as a neurobiological process involving reduction in theconditioned response following repeated exposure to the conditionedstimulus, with concurrent reduction in neural activity in centralnervous system structures in which the CS–CR and/or CS-CSrelationships are represented. Habituation on the other hand, alsoinvolves reduced responsiveness with repeated stimulation, but inaddition to the neurobiological and behavioural changes seen inextinction, may also encompass confounding factors such as fatigueand ‘stimulus satiation,’ especially in human studies (Groves andThompson 1970). While we hypothesise that the observed reductionsin responses to cues described in this paper are due to extinction perse, we are unable to exclude the influence of these other factors. Wetherefore propose that ‘habituation’ is a more apt phrase to describethe phenomena described in this study.

Psychopharmacology

over the course of several hours, or days. Indeed a greaterreduction in between-session outcomes—especially early ones—rather than faster or larger within-session reductions in fearhas been found in animal models of extinction (Richardson etal. 2004) and treatment trials of exposure/response preventionplus DCS for anxiety disorders (Guastella et al. 2008; Ressleret al. 2004). Thus, studies to date provide an unclear pictureas to whether DCS can enhance exposure/response preven-tion in problematic drug users.

The current study

Here, we examine the effects of two isolated oral doses ofDCS (125 mg), compared with placebo, on cue reactivityfollowing repeated unreinforced exposure to in vivo alcoholcues in a laboratory analogue of exposure/response preven-tion. Specifically, we examine the effects of two sessions ofexperimental exposure/response prevention plus DCS/Placebo on ‘cue reactivity’ (subjective craving, galvanicskin response—GSR) and attentional bias. The latter isof particular interest as attentional bias has a potentiallycausal role in maintaining drug/alcohol dependence(Robinson and Berridge 2003) and its modification hasbecome a recent goal in novel psychological treatments(Schoenmakers et al. 2010).

Since the effects of DCS on cue reactivity are stillunknown, our sample consisted of a carefully selectedgroup of non-treatment-seeking heavy drinkers who did notmeet criteria for alcohol dependence. An advantage ofexamining this population experimentally is that in contrastto dependent individuals and those seeking abstinence, theyare more likely to show consistent cue reactivity (Rohsenow etal. 2001) and attentional bias effects2 (Townshend and Duka2007), thus allowing relatively uncomplicated measurementof habituation.

Relatively little is known about the subjective effects ofDCS in recreational drinkers and drug users (see however,Nesic et al. 2008; Trevisan et al. 2008). This is of particularrelevance for problematic drinkers as they may be moresensitive to drugs affecting the NMDA-R system. As suchthere may be modulatory effects on the NMDA-R indepen-dently of effects produced by exposure/response prevention (e.g. Krupitsky et al. 2007). Furthermore, drugs which are co-administered with a psychological treatment (such as expo-sure/response prevention) should ideally have minimal sub-jective effects (e.g. sedation, stimulation, etc.) to avoid state-dependent learning (Hoffman 2007). Therefore, examiningthe subjective effects of DCS in heavy drinkers (independently

of exposure/response prevention) was itself a (secondary) aimof this study. Furthermore, given the description of ‘responseactivation’ to cues noted by Price et al. (2009) in cocaine-dependent individuals receiving DCS, we also explored thepossibility that this occurs in heavy drinkers receiving DCS inthe presence of alcohol cues.

We hypothesised that heavy drinkers would show cuereactivity (subjective craving; GSR); that repeated unrein-forced exposure to alcohol cues would result in a reductionin cue reactivity following exposure/response preventionand this reduction would be enhanced by DCS. In addition,since attentional processes have a putative causal role indrug-seeking and taking behaviour, we hypothesised thatattentional biases would also reduce following exposure/response prevention (Marissen et al. 2007), again, with agreater reduction evident in the DCS group.

Method

Participants

Participants were recruited from the local community(inner London) using poster and internet advertisement.Respondents are aged 18–65 years who reported atscreening to be regularly drinking more than 42 ‘units’of alcohol a week3 (a ‘unit’ corresponds to 10 ml of pureethanol) and drinking alcohol on most days (≥5 days/week). Screening involved health and psychiatric interviewsby trained clinical psychologists.

Importantly, potential participants were excluded andsignposted to their family doctor if they met the criteria foralcohol dependence, assessed using the Structured ClinicalInterview for DSM-IV (Spitzer et al. 1995) or if they reportedserious concerns about their mental health. No other sign-posting was offered. Those with an allergy to DCS or historyof epilepsy, renal failure, and porphyria were also excluded,as were participants who were pregnant or likely to becomepregnant. Participants received monetary compensation.

Eighty-two participants were initially telephone screenedfor eligibility. Of these, 33 did not meet inclusion criteriaand nine refused to participate. Forty were randomised(using a Latin squares procedure) to receive placebo (n=20)or DCS (n=20). One participant (placebo group; male)could not complete the second phase of the study and wasreplaced. Four participants (one in the DCS group and threein the placebo) misreported their alcohol intake at screeningand their AUDIT scores were ≤8. They were not replaced,

2 Despite the putative causal role of attentional bias towards alcohol/drug cues, treatment seekers may exhibit an approach-avoidanceconflict which means that some may direct their attention away fromsuch cues

3 i.e. regularly drinking more than twice the recommended maximumamount of alcohol for men according to UK Department of HealthGuidelines

Psychopharmacology

and their data were excluded from the analysis below. Thefinal sample therefore consisted of n=36.

All participants gave written, informed consent. Ethicalapproval was granted by an institutional (UCL/UCLH)Ethics Committee.

Design

Prior to each of two exposure/response prevention sessions,participants received either an oral dose of 125 mg DCS (n=19; six females) or matched placebo (n=17; five females)in a randomised, placebo-controlled design. DCS (KingPharmaceuticals, Leicester, UK) and placebo (milk powder)were formulated in identical opaque capsules. Participants andexperimenters were blind to treatment group.

Procedure

Participants attended the study centre for three experimentalsessions separated by at least 2 days. The mean gapbetween sessions 1 and 2 and between sessions 2 and 3was 4 days. Verbal confirmation of abstinence over the past12 h was obtained from all participants. Recent abstinencewas also established using a breathalyser test.

Each participant was tested under the same contextualconditions on the three occasions (time of day andlaboratory). Sessions commenced between 9 a.m. and 6p.m.4 according to participant availability on the threesessions.

Session 1: baseline assessment and first cue exposuresession

On the first testing day, participants completed baselinequestionnaires and a dot probe task with simultaneous eyetracking (which was the baseline attentional bias measure).

Immediately after the dot probe task, participantsreceived DCS/placebo treatment; 60 min after treatment,they completed Visual Analogue Scale (VAS) ratings forsubjective craving, mood and bodily symptoms. Cueexposure was then performed with continuous skin con-ductance measurement and VAS craving ratings obtainedafter each of four trials of in vivo cue exposure.

Session 2: second cue exposure session

Following questionnaire completion (ACQ, MRS and BSS;see below), participants received the same treatment(placebo/DCS) as session 1. After 60 min, the exposure/response prevention procedure was repeated.

Session 3: follow-up

On the third session, participants received no treatment orexposure/response prevention but completed questionnairesand the dot probe task.

Questionnaires

Questionnaires were selected to record baseline participantcharacteristics and to monitor changes in mood, bodilysymptoms and craving following placebo/DCS treatment, aswell as changes induced by exposure/response prevention.

Baseline questionnaire measures: anxiety, depression,impulsivity and drug/alcohol use

On session 1, baseline trait anxiety and depressivesymptoms were assessed using the Spielberger TraitAnxiety Inventory (Spielberger et al. 1970) and BeckDepression Inventory (Beck et al. 1961), respectively.Trait impulsivity was assessed using the BehaviourImpulsivity Scale (Gest 1997). Participants also completedthe Alcohol Use Disorders Identification Test (AUDIT;Saunders et al. 1993). This ten-item questionnaire reliablyidentifies hazardous and harmful drinking behaviour(those scoring >8/40).

A drug-use questionnaire (unpublished) was used toassess history of use of other recreational drugs (includingtobacco, stimulants, sedatives, opiates and hallucinogens).Participants indicated whether they used a particular drug:‘ever,’ ‘regularly,’ ‘in the last month,’ and ‘when used last?’

Repeated measures: subjective craving and mood

The Alcohol Craving Questionnaire (ACQ; Tiffany et al.2000) was used as a between-session measure of cravingand administered prior to drug (placebo or DCS) treatmentand cue exposure on all three sessions. In addition,subjective craving state was assessed using a single item100-mm VAS measure: ‘At this moment in time, how muchdo you want a drink?’ (Drummond and Glautier 1994). TheVAS craving measure was used within sessions to assessmomentary subjective craving (pre- and post-drug treatmentas well as pre-, during and post-cue exposure) and served asthe main measure of within- and between-session subjectivecraving.

To monitor subjective changes related to drug (placeboor DCS) condition, the 16-item Mood Rating Scale (MRS;Bond and Lader 1974) was used to assess subjective mood(pre- and 1 h post-placebo/DCS). Principle componentsanalysis of this measure yields factors for alertness,contentedness and calmness. In addition, a 14-item BodilySymptoms Scale (BSS) was adapted from our previous

4 A similar number of sessions occurred in mornings and afternoonsfor the two groups (p>0.1).

Psychopharmacology

psychopharmacological studies for self report on a numberof potential subjective state effects that may be associatedwith DCS: anxiety, depression, memory problems, palpita-tions, nausea, numbness, euphoria, drowsiness, tension,headache, concentration, shaking, vertigo and confusion.All of these ratings were made on a series of 100-mm VASfrom ‘not at all’ (0) to ‘very much’ (100). The MRS, BSSand momentary VAS craving item were used pre- and 1 hafter DCS/placebo treatment.

Repeated measures: alcohol use, expectancy and readinessto change

At the end of sessions 1 and 2, participants were instructedto record daily units of alcohol consumed on a credit-card-sized diary of alcohol intake. Quantities of alcohol whichconstituted ‘a unit’ (equivalent to 8 g pure alcohol) wereexplained using examples.

Alcohol intake in the 2 days prior to sessions two andthree was used as a measure of between-session drinkingbehaviour.

Participants also completed self-report measures assess-ing cognitions about the effects of alcohol (AlcoholExpectancy Questionnaire, Townshend and Duka 2001)and the Readiness to Change Questionnaire (Heather andHonekopp 2008), based on the stage model of change inaddiction which classified participants as being at either the‘pre-contemplative’, ‘contemplative’ or ‘action’ stage ofchange. These two questionnaires were administered on thefirst testing day and repeated during the third, follow-upsession.

Attentional bias tasks

Reaction times and eye movements towards probes wereused as indices of attentional bias towards alcohol-relatedstimuli. These measures were recorded at the first and thirdtesting sessions.

Dot probe task

We adapted our previous dot probe experiment by usingalcohol stimuli. The task is described fully in Morgan etal. (2008). Briefly, participants were seated with theirheads in a chin/forehead rest 70 cm in front of a 19-in.monitor, on which all stimuli were presented. Stimuliconsisted of two photographs (matched for composition)simultaneously presented side by side for either 500 or2,000 ms. The photographs consisted of ten wine-neutral,ten beer-neutral and 20 neutral-neutral pairs.5 Following

stimulus presentation, the pictures were replaced by aprobe (a dot) that appeared on either the left or right of thescreen. Participants were instructed to respond to the dotprobe as quickly and accurately as possible by pressingeither the right or left trigger button on a game pad.Reaction times were recorded and attentional bias scoreswere calculated by subtracting the reaction time for whenthe probe replaced the alcohol pictures (alcohol congruent)from when the probe replaced the matched neutral pairs(alcohol incongruent) trials. Reaction times below 100 msand over 1,000 ms following off-set of stimuli wereconsidered to be outliers and therefore excluded fromanalysis. In total, 5.8% of trials were excluded as outliers.

Only the neutral-neutral and beer-neutral stimuli wereused in the analysis presented below as the majority ofparticipants (24 of 36) preferred beer and beer stimuliproduced a more robust attentional bias effect. Each pairwas presented a total of four times: at both durations (500and 2,000 ms). There were therefore 120 trials in total usedin the analysis. The task was preceded by five practice trialsof neutral-neutral pairs.

Eye movements in response to the stimuli were recordedat a 1 KHz sampling rate, using a desktop mountedEyeLink 1000 eye tracking device (SR Research, Ontario,Canada). Prior to recording, participants’ eye movementswere calibrated by fixations on a 3×3 point grid. Driftcorrection was performed between each trial to ensureparticipants were attending to the centre of the screenbefore stimulus onset. Attentional bias measures werecomputed by calculating percentage dwell time. This wasa measurement of the amount of time the participant spentlooking at the target picture in each trial. The target pictureconsisted of beer stimuli in the beer-neutral pairs and forthe neutral-neutral pairs it was arbitrarily assigned.

Experimental cue exposure and response prevention

On the basis of previous reports suggesting relatively lowcue reactivity to alcohol cues in problematic drinkers(Carter and Tiffany 1999), we used a combination ofimaginal–contextual cues plus in vivo cue exposure toenhance responsivity to in vivo cues and overcome theartificiality of the laboratory setting. This was modelled onLitt et al. (2000) approach for eliciting cue reactivity. Thuseach of the four in vivo trials began following imaginalrevivification of contextual and sensory cues (time, place,social context, sights, smells, etc.) based on a craving script(available from the corresponding author).

In vivo cues consisted of the participant’s drink of choicepresented in a pint (beer) or wine glass (red/white wine). Ofthe 36 participants, 12 preferred wine and 24 preferred beer.Instructions included asking the participant to imagineinteracting with the drink in their ordinary context while

5 Alcohol stimuli were rated as significantly more pleasant thanneutral ones (p<0.001).

Psychopharmacology

handling and smelling the drink. This imaginal rehearsaland in vivo interaction lasted for 5 min and constituted onetrial of experimental exposure/response prevention. Aftereach of the four exposure trials, a 0–100-mm craving VASwas completed which asked ‘At this moment in time howmuch do you want a drink?’ (Drummond and Glautier1994). The first VAS response (subtracting baselinecraving) served as a subjective measure of ‘initial activationof craving.’

The procedure was ‘fixed’ in that it ended followingcompletion of the four exposure trials rather than after apre-determined percentage reduction in cue reactivity. Thisapproach maximised experimental control and overcamepotential difficulties related to participants giving artificial-ly low subjective responses in order to terminate theprocedure prematurely (since prolonged unreinforced crav-ing can be experienced as unpleasant).

Within-session habituation of subjective craving wasindexed as the difference between craving at the end of thefourth exposure/response prevention trial and maximumcue-induced craving on sessions one and two. Similarly,within-session physiological habituation was indexed as thedifference between the average accumulated skin conduc-tance responses (SCR) of the last exposure/responseprevention trial and the trial causing maximum SCR.

Skin conductance response

SCRs were acquired via a 24-bit digital amplifier (ContactPrecision Instruments), using silver chloride electrodes withisotonic electrode gel placed on the medial phalanges of theindex and middle fingers of the non-dominant hand of theparticipant. Recording was at a sampling rate of 500 Hz.Waveforms were smoothed and visually inspected forartefacts before being processed using PSYLAB analysissoftware (PSYLAB 7, Contact Precision Instruments).Average accumulated amplitude data were obtained forthe duration of in vivo exposure.

Ten participants’ data were excluded from skin conduc-tance analysis due to technical failures or non-response.

Statistical analysis

All data were analysed using the Statistical Package forSocial Sciences version 14.

One-way analyses of variance (ANOVAs) were used tocompare baseline characteristics between groups (placebo,DCS). For repeated measurements analysis of variance(RMANOVA), the between-subjects factor was treatmentgroup (DCS, placebo) and within-subjects factors weresession (1, 2 and 3), time (pre- and post-drug treatment)and/or exposure point (max and end point). For the dot probeanalysis, additional within-subjects factors were stimulus

duration (500 and 2,000 ms), stimulus type (neutral andbeer). Post hoc t tests were Bonferroni corrected.

Results

Baseline measures: demographics, anxiety, depression,impulsivity and alcohol use

As displayed in Table 1, the placebo and DCS groups weresimilar in terms of age, education level and alcohol-relatedcharacteristics. The groups had similar AUDIT scores, agesof first drink, age of regular drinking and readiness tochange scores (all p values>0.1). Groups also hadequivalent ‘trait’ anxiety, depression and impulsivity scores(all p values>0.1; Table 2) and previous and currentrecreational drug-use characteristics.

Subjective effects of DCS: mood, bodily symptomsand craving

Further subjective data are summarized in Table 2. Therewere no drug effects on alertness or calmness factors of theMRS. However there was a time (pre, post) x groupinteraction (F(1, 34)=4.673; p=0.038) on the contented-ness factor. Bonferroni corrected (α=0.025) post hoc, paired-samples t tests showed that these effects were due to anincrease in contentedness following administration of DCS [t(18)=3.817; p=0.001] but no change following placebo.This was consistent with effects on the BSS euphoria item,where there was a session×time×group interaction (F(1, 34)=5.852; p=0.021). Bonferroni corrected (α=0.0125) posthoc, paired-samples t tests suggested that this was due to atrend towards increased euphoria in the DCS group onsession 1 (t(18)=−2.27; p=0.035). There were no otherbetween-group differences on BSS items.

Treatment effects on subjective momentary craving werealso assessed. There was no time (pre–post)×group ortime×group×session (1 and 2) interactions (p values>0.5)

Table 1 Participant characteristics

Placebo DCS

Demographic details

N (male/female) 17 (12/5) 19 (13/6)

Age 26.82 (10.49) 25.58 (8.76)

Alcohol use details

AUDIT score 19.11 (6.50) 18.30 (5.11)

Age of 1st drink 11.06 (3.77) 11.42 (3.45)

Age of regular drinking 17.53 (7.25) 15.92 (1.97)

Family history of alcoholism 6 6

Except for N and family history, values are means (±SD)

Psychopharmacology

indicating that increased contentedness and euphoria on theMRS and BSS were not accompanied by changes incraving following DCS in the absence of alcohol cues.

Cue exposure

Subjective craving: within- and between-session changes

As shown in Fig. 1a, significant within-session reductionwas found for subjective craving using the momentarymeasure of craving (main effect of exposure time point; F(1,34)=15.62; p<.001). Between-session reduction in (momen-tary) craving was also significant (main effect of session: F(1, 34)=12.58; p=0.001) although there was no effect ofgroup and no interactions between group and exposure timepoint or session (all F values<3.2; p values>0.08).

Changes in subjective craving were also assessedusing the 12-item ACQ (Tiffany et al. 2000) across thetwo exposure/response prevention sessions and the thirdfollow-up session. There was a main effect of session (F(2, 68)=5.25; p=0.008). Bonferroni corrected (α=0.025)

post hoc paired-sample t tests showed that this main effectwas due to a reduction in craving between sessions 1 and2 (t(35)=3.27; p=0.002) with no significant changebetween sessions 2 and 3 (t(35)=−0.046; p=0.96). Again,there was no interaction between group and session (F(2,68)=0.62; p>0.5).

Skin conductance responses: within- and between-sessionchanges

Both treatment groups showed significant within-sessionhabituation to alcohol cues (F(1, 24)=13.73; p=0.001;Fig 1b). There was no main effect of group nor anysignificant interactions between group, exposure time point(max and endpoint) or session (all Fs<1.44; p values>0.24).

Initial activation of craving: does DCS facilitate responseactivation?

Since Price et al. (2009) suggest that DCS may ‘facilitateresponse activation’ (i.e. enhance craving)—at least to

Placebo DCS

Pre Post Pre Post

Mood/trait variables

Depression 9.47 (5.79) 9.92 (5.97)

Anxiety 43.76 (11.25) 45.00 (12.03)

Impulsivity Motor 25.35 (3.94) 25.16 (3.64)

Non-planning 31.06 (6.22) 30.00 (5.01)

Cognitive 19.24 (3.00) 19.37 (3.34)

Mood rating scale

Alertness 43.54 (16.75) 41.96 (13.68) 56.57 (13.52) 44.35 (13.01)

Contentedness 33.35 (15.54) 34.74 (15.65) 45.03 (15.40) 38.59 (13.21)

Calmness 31.14 (13.84) 31.91 (19.23) 30.79 (13.95) 36.21 (14.87)

Bodily symptom scale

Anxiety 28.05 (22.09) 24.42 (22.38) 47.89 ( 22.26) 29.72 (24.02)

Depression 25.95 (21.62) 18.89 (18.87) 29.33 (20.59) 25.61 (20.75)

Memory 21.53 (19.66) 20.37 (18.35) 29.61 (21.48) 25.22 (19.79)

Palpitations 15.37 (19.21) 18.74 (19.71) 26.44 (27.12) 25.39 (20.15)

Nausea 16.21 (16.61) 15.21 (14.59) 23.61 (26.26) 20.39 (21.83)

Numbness 30.00 (25.16) 25.58 (23.38) 37.67 (27.94) 30.78 (27.62)

Euphoria 29.11 (20.93) 27.32 (22.24) 26.56 (19.07) 34.67 (24.52)

Drowsiness 38.53 (26.96) 32.42 (24.38) 54.06 (24.48) 37.89 (26.10)

Tension 28.05 (26.85) 21.58 (19.60) 29.67 (26.72) 31.11 (28.19)

Headache 21.58 (24.81) 16.63 (18.11) 21.56 (27.02) 22.33 (24.79)

Concentration 30.00 (24.26) 23.58 (19.28) 46.44 (25.24) 33.44 (24.47)

Shaking 11.79 (16.74) 11.32 (14.69) 18.72 (24.43) 19.17 (18.68)

Vertigo 11.79 (19.87) 11.21 (17.39) 13.5 (21.79) 12.00 (21.24)

Confusion 17.63 (22.40) 15.26 (19.07) 24.61 (27.73) 19.61 (22.10)

Table 2 Trait characteristicsand subjective scores pre- andpost-treatment means (±SD)

Psychopharmacology

cocaine cues—it was of interest to determine if there was asimilar effect in heavy drinkers in response to alcohol cues.

‘Initial activation of craving’ responses—the subjectiveresponses to the first in vivo presentation of alcoholcompared with baseline craving—are shown in Fig. 2a.As can be seen, on both days, the first presentation ofalcohol causes an increase in subjective craving relative tobaseline but the increase is smaller on day two in bothgroups (main effect of day (F(1, 33)=8.88; p=0.005). Thisis another index of between-session habituation. However,there was no main effect of group (F(1, 33)=1.47; p=0.23),and no group×day interaction (F(1, 33)=0.59; p=0.45).

In the same way as for subjective reactivity, physiolog-ical between-session habituation of cue reactivity wasassessed by comparing accumulated baseline SCRs to firstin vivo (initial reactivity) exposures across days (Fig. 2b).A RMANOVA with session (1 and 2) and exposure point(baseline and first exposure) as within-subjects factors andtreatment group as a between-subjects factor was con-ducted. Unlike subjective responding, SCR did not showsignificant between-session habituation for either treatmentgroup (F(1, 24)=0.63; p=0.43).

Attentional bias: dot probe task

There was a significant stimulus duration×stimulus interaction(F(1, 31)=15.02; p=0.001) based on a bias towards beerstimuli at 500 ms (t33=−4.45; p<0.001) but not at 2,000 ms(t32=−0.58; p=0.57). There was also a significant session×stimulus interaction (F(1, 31)=6.98; p=0.013, see Fig. 3),which post hoc paired sample t tests showed was due to asignificant bias to beer stimuli on session 1 (t34=−5.42; p<0.001) but not session 3 (t33=−1.13; p=0.27), indicatingreduction in attentional bias over time. There were nosignificant interactions involving group (all F values<2.2;p values>0.15).

Eye tracking measures

A RMANOVAwas conducted on the percentage dwell timedata, using within-subjects factors of session, stimulusduration (500 and 2,000 ms) and stimulus (neutral andbeer) and a between-subjects factor of group (placebo,DCS). A significant interaction was found for duration×stimulus×group (F(1, 32)=4.59; p=0.04). Exploration of

0

10

20

30

40

50

60

70

80

90

Placebo DCS Placebo DCS

Session 1 Session 2

Maximum

Endpoint

Sel

f-R

ated

VA

S C

ravi

ng

(%)

a

b

Fig. 1 a Within- and between-session changes in subjectivecraving to in vivo cues. Filledbars indicate maximum subjec-tive craving recorded after oneof the four cue exposure trials(dark bars, placebo; light bars,DCS). The filled hatched barsindicate subjective cravingrecorded on the fourth exposuretrial (endpoint) in thecorresponding treatment groupsacross the two sessions. bWithin- and between-sessionchanges in physiological reac-tivity to in vivo cues. Filled barsindicate the average SCR am-plitude recorded during the trialin which participants showedmaximum reactivity relative tobaseline SCR amplitude (darkbars, placebo; light bars, DCS).Hatched bars indicate SCRduring the fourth trial relative tobaseline (endpoint) in thecorresponding treatment groupsacross the two sessions

Psychopharmacology

this interaction using paired-samples t tests revealedsignificantly greater percentage dwell times for beer stimuliin both the DCS and placebo groups compared with neutralat each of the stimulus durations (500 and 2,000 ms); all pvalues<0.001. The pattern of results was the same for boththe DCS and placebo groups and there were no other effectsof group.

There were also main effects of stimulus duration (F(1,32)=257.48; p<0.001) and stimulus type (F(1, 32)=500.20; p<0.001) with longer percentage dwell times forthe 2,000 ms stimulus duration and beer stimuli.

Drinking behaviour, alcohol expectancy and reasonsfor change

In the 2 days prior to session 2, the placebo and DCS grouprecorded drinking 6.52±3.95 and 5.59±3.92 units respec-tively. Prior to session 3 the placebo and DCS groups,respectively, recorded 5.05±4.62 and 6.74±4.43 units. Atwo-way repeated measures ANOVA indicated no maineffect of group (F(1, 33)=0.12; p=0.73) and no group×interval interaction (F(1, 33)=2.13; p=0.154).

As might be expected there were no incidental attitudinalchange towards alcohol as a result of experimentalexposure/response prevention: scores on the readiness tochange questionnaire did not change significantly over thecourse of the sessions, and there were no group effects.Similarly no effects of session or group emerged on theAlcohol Expectancy Questionnaire (all p values>0.1).

Discussion

This is the first study that we are aware of which comparesthe effects of DCS and placebo treatment on cue reactivityand attentional bias following experimental cue exposure/response prevention in heavy drinkers. The main positivefinding was that the DCS group showed mild subjectiveeffects (increased euphoria/contentedness), consistent withpreviously reported NMDA-R modulating effects of DCS(Krystal et al. 2003; see also Nesic et al. 2008). However,there was no evidence of greater habituation on subjective,physiological or attentional measures in the DCS group.Although we found no between-group difference in

Fig. 2 a Initial subjective cuereactivity to in vivo alcoholcues. Filled bars indicate base-line VAS craving in the absenceof alcohol cues (dark bars,placebo; light bars, DCS).Hatched bars are VAS ratingsfollowing the first presentationof in vivo alcohol cues. b Initialphysiological cue reactivity.Filled bars indicate baselineSCR in the absence of alcoholcues (dark bars, placebo; lightbars, DCS). Hatched bars areaverage SCR ratings over thecourse of the first presentationof in vivo alcohol cues

Psychopharmacology

between-session drinking behaviour, an absence of baselineor follow-up measures of drinking precludes any firmstatement on drinking behaviour from being made.

The rationale for the current experiment came fromstudies on animals using the standard CPP model foraddiction which showed that extinction is enhanced in thepresence of DCS, at least in models of cocaine dependence(e.g. Botreau et al. 2006). Furthermore, conditioningmodels of anxiety indicated that DCS enhanced extinctionof fear in both humans and animals. In light of thesepositive findings how can the absence of effect of DCS inthe current study be understood? There are a number ofpossibilities.

Firstly, considering studies on conjunctive DCS treat-ment in anxiety, it has been suggested (Guastella et al.2007) that significant effects of DCS may only be evidentin ‘clinical populations,’ i.e. those with more severedifficulties (in this case, alcohol dependence). However,while Guastella et al. (2007)) found that DCS failed topotentiate habituation of subclinical anxiety in humans, anabsence of statistically significant effects of DCS has, sincethen, also been found in two clinical studies of OCD(Storch et al. 2007; 2010, although the latter showed a non-significant effect favouring those in the DCS group). In thecurrent study we focused on a carefully selected group ofnon-dependent problematic drinkers since a cautiousapproach is warranted at this early stage of knowledgeabout the effects of DCS on drug and alcohol craving.However, despite their non-dependent status, the participantsin the current study had clinically relevant characteristics.Their AUDIT scores clearly indicated their hazardousdrinking patterns and therefore they represented a group forwhom moderation drinking, for example, might be anappropriate therapeutic goal. Furthermore, they showed robustsigns of acquired incentive salience (indexed through dotprobe performance) and ostensible conditioned responding

(subjective and physiological cue reactivity), which arecharacteristics seen in dependent drinkers.

Alternatively, it is possible that our ‘short-term’outcome measures (subjective/physiological cue reactivityand attentional bias) are insufficiently sensitive to detect groupdifferences. For instance, in clinical trials of exposure/response prevention for substance dependence, theprinciple outcome measures are generally long-term ones(e.g. latency to relapse) and behavioural, rather thansubjective, cognitive and physiological, as used here. Ina recent study of modelled cocaine dependence insquirrel monkeys, it was suggested that the main impactof DCS following extinction training was to reducereacquisition (Nic Dhonnchadha et al. 2010). Clearly, inhuman studies, the equivalent outcome (reduced relapserates) could only be examined over a longer time-course inindividuals who intended to remain abstinent followingexposure/response prevention.

Another explanation for an absence of an effect is that adose of DCS of 125 mg might be suboptimal. Most extantstudies of DCS-exposure/response prevention in anxietydisorders have used a lower dose of 50 mg (Norberg et al.2008), although similar effects seem to be produced by500 mg (Ressler et al. 2004). Our use of 125 mg was basedon the study by Kushner et al. (2007) which examinedexposure/response prevention in obsessive compulsivedisorder. That study used a human-equivalent dose ofDCS (2.08 mg/kg) that closely matched the dose used in themajority of studies on animal fear conditioning (2.42 mg/kg; Norberg et al. 2008). In fact, studies showing enhancedextinction effects of DCS on cocaine CPP have usedsubstantially higher doses (15 mg/kg: Botreau et al. 2006;Paolone et al. 2009; 30 mg/kg Nic Dhonnchadha et al.2010). While this might suggest that 125 mg is too low adose to produce facilitative effects on drug-cue extinction, arecent study of alcohol CPP in mice using the same high

20

10

0

10

20

30

40

50

60

Neut Beer Neut Beer

Session 1

Att

enti

on

al B

ias

Placebo

DCS

Session 3

Fig 3 Attentional bias in reac-tion times to stimuli across druggroups and days. Dot probereaction times in ms to neutraland beer stimuli collapsedacross both SOAs (500 and2,000 ms) for placebo (darkbars) and DCS (light bars)

Psychopharmacology

dose that was used in cocaine-CPP studies also found nofacilitative effects of DCS (Groblewski et al. 2009). Itshould be noted however that it is challenging to determineeffective oral doses in humans from animal studies whichuse intraperitoneal routes of administration. So while ourresults are consistent with the only relevant study in theanimal literature using alcohol CPP (Groblewski et al.2009), it is still unclear what the effective dose of DCS forenhancing extinction of appetitively conditioned stimulimight be in humans.

It is also possible that DCS’s effects on extinction ofappetitive responding—at least in humans—are generallysmaller than those seen in anxiety. However, if this were thecase, the absence of trends favouring DCS on any of theoutcomes we explored suggests that its facilitative effects, ifthey exist, would be very small indeed. Such relativelysmall effects of DCS may particularly be found in thoseprone to heavy drinking or alcohol dependence (comparedwith other drugs) due to alterations in NMDA-R function-ing resulting from prolonged drinking (Krystal et al. 2003).

It is noteworthy that we found clear evidence ofhabituation using our exposure protocol which aimed toenhance in vivo exposure using imaginal contextual cues.Obviously in the absence of robust habituation, it would nothave been possible to draw any conclusions about thepresence or absence of DCS effects on this process. Indeed,it is possible that our exposure/response prevention proto-col had ‘saturating’ effects and resulted in ceiling levels ofcue reactivity reduction. As such subtle effects of DCS mayhave been masked. Indeed, results from studies on anxietydisorders suggest that ‘under-dosing’ of exposure-basedtreatments is required for DCS effects to be observed. Suchceiling effects may be particularly evident in subclinicalpopulations (Hoffman 2007).

Apart from an absence of DCS effects on exposure/response prevention, another finding of interest was thereduction in attention bias, which has potential clinicalsignificance. Attentional bias towards drug stimuli is notonly a sign of problematic drug/alcohol use, but is thoughtto have a causal role in dependence (Robinson andBerridge 2003). Research examining the use of attentionalbias modification training for alcohol dependence is arelatively recent development, although preliminary resultssuggest participants who undergo such training (to attendaway from alcohol stimuli in a modified dot probe trainingtask) show reduced attentional bias to alcohol comparedwith baseline (Schoenmakers et al. 2010). The controlgroup, who do not undergo training, show no change inattentional bias over time (Schoenmakers et al. 2010).Other studies in which repeated measures of attentional biasin heavy drinkers have been performed also show thatspontaneous reductions in attentional bias do not occur incontrol groups (Field et al. 2007; Schoenmakers et al.

2010). While we did not have a control condition in whichexposure/response prevention was not performed, theprevious cited studies suggest that such a group would notshow a spontaneous reduction in attentional bias over time.While we are unable to comment on whether this reductionin attentional bias had a generalised effect (i.e. to stimuliother than those presented in the dot probe) or effects onsubsequent craving or drinking behaviour, these data dosuggest that attentional bias can be modified throughexperimental exposure/response prevention and that furtherwork may be warranted to examine its role in the efficacyof exposure/response prevention in dependent populations.

In summary, the predicted effects of combined DCS-exposure/response prevention were not found in the currentstudy. Further research is needed to elucidate whatpharmacological and behavioural manipulations, if any,will create an effective protocol for nootropic enhancementof exposure/response prevention in the treatment ofaddictions. The finding that DCS produced mild stimulanteffects might suggest that it could interfere with exposure-based therapies due to state-dependent learning (Hoffman2007). Furthermore, subjective stimulation may also in-crease craving and prime alcohol use, although thestimulant effect found here was not associated withenhanced response activation in our sample, in contrast toan apparent cue-activation effect in cocaine users (Price etal. 2009).

Our finding of within- and between-session habituationusing in vivo and imaginal–contextual cues is noteworthyas it clearly shows that our main ‘treatment’ was effica-cious. Such within- and between-session outcomes are notcommonly reported in cue exposure studies and we suggestthat these outcomes are valid and important in their ownregard. Furthermore, the effect of exposure/response pre-vention on attentional bias warrants further investigation asit may have implications for studies which aim to producebehavioural changes in drug taking through manipulatingattentional bias (Schoenmakers et al. 2010).

Acknowledgements The research was supported by an MRC grantawarded to SK, HVM and CJAM. R M-C and LR received fundingfrom the Central Research Fund of the University of London and theUCL Graduate School. The authors have no conflicts of interest todeclare. Preliminary results were presented at the 2010 InternationalCollege of Neuropsychopharmacology congress in Hong Kong andthe 2010 British Association of Psychopharmacology conference. Wethank the anonymous peer reviewers for their constructive commentson our paper.

References

Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) Aninventory for measuring depression. Arch Gen Psychiatry 4:561–571

Psychopharmacology

Bond AJ, Lader MH (1974) The use of analogue scales in ratingsubjective feelings. Br J Med Psychol 47:211–218

Botreau F, Paolone G, Stewart J (2006) d-Cycloserine facilitatesextinction of a cocaine-induced conditioned place preference.Behav Brain Res 172:173–178

Bouton ME (2002) Context, ambiguity, and unlearning: sources ofrelapse after behavioral extinction. Biol Psychiatry 52:976–986

Carter BL, Tiffany ST (1999) Meta-analysis of cue-reactivity inaddiction research. Addiction 94:327–340

Chasson GS, Buhlmann U, Tolin DF, Rao SR, Reese HE, Rowley T,Welsh KS, Wilhelm S (2010) Need for speed: evaluating slopesof OCD recovery in behavior therapy enhanced with D-cycloserine. Behav Res Ther 48:675–679

Conklin CA, Tiffany ST (2002) Applying extinction research and theoryto cue-exposure addiction treatments. Addiction 97:155–167

Drummond DC, Glautier S (1994) A controlled trial of cue exposuretreatment in alcohol dependence. J Consult Clin Psychol 41:809–817

Drummond DC, Tiffany ST, Glautier S, Remington B (1995)Addictive behaviour: cue exposure theory and practice. Wiley,Sussex

Field M, Duka T, Eastwood B, Child R, Santarcangelo M, Gayton M(2007) Experimental manipulation of attentional biases in heavydrinkers: do the effects generalise? Psychopharmacology192:593–608

Gest S (1997) Behavioral inhibition: stability and associations withadaptation from childhood to early adulthood. J Pers Soc Psychol72:467–475

Groblewski PA, Lattal KM, Cunningham CL (2009) Effects of d-cycloserine on extinction and reconditioning of ethanol-seekingbehavior in mice. Alcohol Clin Exp Res 33:772–782

Groves PM, Thompson RF (1970) Habituation: a dual process theory.Psychol Rev 77:419–450

Guastella AJ, Dadds MR, Lovibond PF, Mitchell P, Richardson R(2007) A randomized controlled trial of the effect of D-cycloserine on exposure therapy for spider fear. J Psychiatr Res41:466–471

Guastella AJ, Richardson R, Lovibond PF, Rapee RM, Gaston JE,Mitchell P, Dadds MR (2008) A randomized controlled trial of D-cycloserine enhancement of exposure therapy for social anxietydisorder. Biol Psychiatry 63:544–549

Heather N, Honekopp J (2008) A revised edition of the readiness tochange questionnaire (treatment version). Addiction Researchand Therapy 16(5):421–433

Hoffman SG (2007) Enhancing exposure-based therapy from atranslational research perspective. Behav Res Ther 45:1987–2001

Krupitsky EM et al (2007) Effect of memantine on cue-inducedalcohol craving in recovering alcohol-dependent patients. Am JPsychiatry 164:519–523

Krystal JH, Petrakis IL, Mason G, Trevisan L, D’Souza DC (2003) N-methyl-D-aspartate glutamate receptors and alcoholism: reward,dependence, treatment, and vulnerability. Pharmacol Ther 99:79–94

Kushner MG, Kim SW, Donahue C, Thuras P, Adson D, Kotlyar M,McCabe J, Peterson J, Foa EB (2007) D-Cycloserine augmentedexposure therapy for obsessive–compulsive disorder. Biol Psy-chiatry 62:835–838

Ledgerwood L, Richardson R, Cranney J (2003) Effects of D-cycloserine on extinction of conditioned freezing. Behav Neuro-sci 117:341–349

Lee H, Kim JJ (1998) Amygdalar NMDA receptors are critical fornew fear learning in previously fear-conditioned rats. J Neurosci18:8444–8454

Litt MD, Cooney NL, Morse P (2000) Reactivity to alcohol-relatedstimuli in the laboratory and in the field: predictors of craving intreated alcoholics. Addiction 95:889–900

Marissen MAE, Franken IHA, Blanken P, Van den Brink W, HendriksVM (2007) Cue exposure therapy for the treatment of opiateaddiction: results of a randomized controlled clinical trial.Psychother Psychosom 76:97–105

Monti PM, Rohsenow DJ, Rubonis AV, Niayra RS, Sirota AD, ColbySM, Goddard P, Abrams DB (1993) Cue exposure coping skillstreatment for male alcoholics: a preliminary investigation. JConsult Clin Psychol 61:1011–1019

Morgan CJA, Rees H, Curran HV (2008) Attentional bias to incentivestimuli in frequent ketamine users. Psychol Med 38:1331–1340

Nesic J, Rusted J, Duka T, Jackson A (2008) Effects of the glycine-site partial agonist d-cycloserine on the cognitive and subjectiveeffects of smoking. J Psychopharmacol 22:A69

Nic Dhonnchadha BA, Szalay JJ, Achat-Mendes C, Platt DM, OttoMW, Spealman RD, Kantak KM (2010) D-Cycloserine detersreacquisition of cocaine self-administration by augmentingextinction learning. Neuropsychopharmacology 35:357–367

Norberg MM, Krystal JH, Tolin DF (2008) A meta-analysis of D-cycloserine and the facilitation of hear extinction and exposuretherapy. Biol Psychiatry 63:1118–1126

Paolone G, Botreau F, Stewart J (2009) The facilitative effects of d-cycloserine on extinction of a cocaine-induced conditioned placepreference can be long lasting and resistant to reinstatement.Psychopharmacology 202:403–409

Price KL, McRae-Clark AL, Saladin ME, Moran-Santa Maria MM,DeSantis SM, Back SE, Brady KT (2009) D-Cycloserine andcocaine cue reactivity: preliminary findings. Am J Drug AlcoholAbuse 35:434–438

Ressler KJ, Rothbaum BO, Tannenbaum B, Anderson P, Graap K,Zimand E, Hodges L, Davis M (2004) Use of D-cycloserine inphobic individuals to facilitate extinction of fear. Arch GenPsychiatry 61:1130–1144

Richardson R, Ledgerwood L, Cranney J (2004) Facilitation of fearextinction by D-cycloserine: theoretical and clinical implications.Learn Mem 11:510–516

Robinson TE, Berridge KC (2003) Addiction. Annu Rev Psychol54:25–53

Rohsenow DJ, Monti PM, Rubonis AV, Gulliver SB, Colby SM,Binkoff JA, Abrams DB (2001) Cue exposure with coping skillstraining and communication skills training for alcohol depen-dence: 6- and 12-month outcomes. Addiction 96:1161–1174

Rothbaum BO (2008) Critical parameters for D-cycloserine enhance-ment of cognitive-behavioral therapy for obsessive-compulsivedisorder. Am J Psychiatry 165:293–296

Santa Ana EJ, Rounsaville BJ, Frankforter TL, Nich C, Babuscio T,Poling J, Gonsai K, Hill K, Carroll KM (2009) D-Cycloserineattenuates reactivity to smoking cues in nicotine dependentsmokers: a pilot investigation. Drug Alcohol Depend 104:220–227

Saunders JB, Aasland OG, Babor TF, De la Fuente JR, Grant ME(1993) Development of the alcohol use disorders identificationtest (AUDIT): WHO collaborative project on early detection ofpersons with harmful alcohol consumption-II. Addiction 88:791–804

Schoenmakers TM, De Bruin M, Lux IFM, Goertz AG, Van KerkhofDHAT, Wiers RW (2010) Clinical effectiveness of attentionalbias modification training in abstinent alcoholic patients. DrugAlcohol Depend 109:30–36

Sitharthan T, Sitharthan G, Hough MJ, Kavanagh DJ (1997) Cueexposure in moderation drinking: a comparison with cognitive-behavior therapy. J Consult Clin Psychol 65:878–882

Spielberger CD, Gorsuch RL, Lushene RD (1970) STAI manual.Consulting Psychologist Press, Palo Alto

Spitzer RL, Williams JBW, Gibbon M, First MB (1995) Structuredclinical interview for DSM-IV (SCID). New York State PsychiatricInstitute, Biometrics Research, New York

Psychopharmacology

Storch EA, Merlo LJ, Bengtson M, Murphy TK, Lewis MH, YangMC, Jacob ML, Larson M, Hirsh A, Fernandez M, Geffken GR,Goodman WK (2007) D-Cycloserine does not enhance exposure-response prevention therapy in obsessive–compulsive disorder.Int Clin Psychopharmacol 22:230–237

Tiffany ST, Carter BL, Singleton EG (2000) Challenges in themanipulation, assessment and interpretation of craving relevantvariables. Addiction 95(S2):S177–S187

Townshend JM, Duka T (2001) Attentional bias associated withalcohol cues: differences between heavy and occasional socialdrinkers. Psychopharmacology 157:67–74

Townshend JM, Duka T (2007) Avoidance of alcohol-related stimuli inalcohol-dependent inpatients. Alcohol Clin Exp Res 31:1349–1357

Trevisan L, Petrakis IL, Pittman B, Gueorguieva R, D’Souza DC,Perry E et al (2008) Absence of significant interactive effects ofhigh dose D-Cycloserine and ethanol in healthy human subjects:preliminary insights into ethanol actions at the glycine (b) site ofNMDA glutamate receptors. Alcohol Clin Exp Res 32:36–42

Walker DL, Ressler KJ, Lu KT, Davis M (2002) Facilitation ofconditioned fear extinction by systemic administration or intra-amygdala infusions of D-cycloserine as assessed with fear-potentiated startle in rats. J Neurosci 22:2343–2351

Psychopharmacology