Neural substrates of cognitive flexibility in cocaine and gambling addictions

Cocaine dependence and pa hological gambling ( elabelled gambling diso de 1 ) ha e ecen ly been join ly classi ied as addic i e diso de s, based on e idence o o e lap in e ms o clinical p esen a ion, neu obiological p o ile and gene ic liabili y. 1 In ea men -seeking samples, bo h diso de s a e cha ac e ised by pe sis en engagemen wi h he ha m ul beha iou despi e i s ad e se consequences in in e pe sonal and occupa ional domains. This pe sis ence e lec s cogni i e in lexibili y in upda ing and in eg a ing he alues o po en ial ac ions wi h e e ence o p e ious and expec ed ou comes. 2 In addi ion o i s ele ance o heo ies o addic ion, cogni i e in lexibili y is also clinically ele an because pe se e a i e esponding is a signi ican p edic o o poo e addic ion ea men ou comes. 3 Pe sis en esponding in he ace o nega i e eedback can be modelled expe imen ally using he e e sal lea ning pa adigm, a disc imina ion ask in which pa icipan s lea n o espond o a ein o ced s imulus, bu mus hen lea n o espond o ano he , p e iously i ele an , s imulus. 4 Two complemen a y aspec s o e e sal lea ning a e ele an o addic i e beha iou s: he s ong mo i a ional endency o espond o p e iously ein o ced s imuli and a di icul y in lexibly shi ing owa ds no el ein o cing s imuli, which a e collec i ely indica ed by ‘pe se e a ion’. 5 Beha iou al s udies indica e ha bo h cocaine use s and pa hological gamble s exhibi e e sal lea ning pe se e a ion compa ed wi h con ols. 6 In cocaine use s, pe se e a ion is mainly obse ed ollowing changes in esponse–ou come con ingencies (i.e. in e e sal lea ning asks in ol ing p obabilis ic wins and losses). 7,8 By con as , lexible esponding is ela i ely p ese ed in non- ewa ded s imulus-disc imina ion asks in ol ing ixed schedules. 9 In pa hological gamble s, he e is educed lexibili y a e e e sal o p e iously ewa ded con ingencies. 10 Re e sal lea ning is unde pinned by neu al ci cui y comp ising do somedial, do sola e al and en ola e al p e on al co ex (dmPFC, dlPFC and lPFC espec i ely) and hei connec ions wi h he s ia um and amygdala. 11 Acco dingly, e e sal lea ning pe se e a ion has been associa ed wi h abno mal dmPFC and dlPFC unc ion in cocaine use s 12,13 and dec eased ac i a ion o igh lPFC in pa hological gamble s. 14 Mo eo e , he e is e idence ha e e sal lea ning is gene ically media ed by dopamine D2 exp ession, as heal hy olun ee s ca ying he DRD2/ANKK1 Taq1A A1 allele (linked o dec eased D2 ecep o a ailabili y) exhibi poo e beha iou al pe o mance, and blun ed lPFC ac i a ion du ing shi ing. 15 The A1 allele has also been associa ed wi h cocaine and gambling addic ions. 16 Accep ing he subs an ial neu obiological o e lap be ween hese diso de s, cocaine addic ion ne e heless in ol es signi ican d ug-induced, ac i e, de imen al e ec s on he dopamine gic sys em. 17–19 P eclinical s udies ha e demons a ed ha p olonged cocaine adminis a ion induces signi ican e e sal lea ning de ici s, 20 and ha hese changes a e media ed by D2-dopamine ansmission. 21 In humans, compa able e ec s a e indica ed by co ela ions agains cocaine ch onici y in neu opsychological s udies 22,23 bu he e is no e idence ega ding he speci ici y o hese de ici s wi h espec o he beha iou al addic ions. We easoned ha cocaine-induced neu oadap i e changes should be absen in pa hological gamble s. 24 As such, pa hological gambling can se e as a con ol g oup o disambigua e he ac i e de imen al e ec s o cocaine dependence on e e sal lea ning while accoun ing o sha ed ulne abili y and ein o cemen his o y. 25,26 The p ima y aim o his s udy was o con as b ain ac i a ion pa e ns associa ed wi h cogni i e shi ing and esponse pe se e a ion in indi iduals diagnosed wi h cocaine dependence . 158 Neu al subs a es o cogni i e lexibili y in cocaine and gambling addic ions An onio Ve dejo-Ga cia, Luke Cla k, Juan Ve dejo-Roma ´n, Na alia Albein-U ios, Jose ´M. Ma inez-Gonzalez, Blanca Gu ie ez and Ca les So iano-Mas Backg ound Indi iduals wi h cocaine and gambling addic ions exhibi cogni i e lexibili y de ici s ha may unde lie pe sis ence o ha m ul beha iou s. Aims We in es iga ed he neu al subs a es o cogni i e in lexibili y in cocaine use s . pa hological gamble s, aiming o disambigua e common mechanisms . cocaine e ec s. Me hod Eigh een cocaine use s, 18 pa hological gamble s and 18 con ols pe o med a p obabilis ic e e sal lea ning ask du ing unc ional magne ic esonance imaging, and we e geno yped o he DRD2/ANKK Taq1A polymo phism. Resul s Cocaine use s and pa hological gamble s exhibi ed educed en ola e al p e on al co ex (PFC) signal du ing e e sal shi ing. Cocaine use s u he showed inc eased do somedial PFC (dmPFC) ac i a ion ela i e o pa hological gamble s du ing pe se e a ion, and dec eased do sola e al PFC ac i a ion ela i e o pa hological gamble s and con ols du ing shi ing. P elimina y gene ic indings indica ed ha cocaine use s ca ying he DRD2/ANKK Taq1A1+ geno ype may de i e unique s imula o y e ec s on shi ing- ela ed en ola e al PFC signal. Conclusions Reduced en ola e al PFC ac i a ion du ing shi ing may cons i u e a common neu al ma ke ac oss gambling and cocaine addic ions. Addi ional cocaine- ela ed e ec s ela e o a wide pa e n o ask- ela ed dys egula ion, e lec ed in signal abno mali ies in do sola e al and dmPFC. Decla a ion o in e es L.C.: The Cen e o Gambling Resea ch a UBC is unded by suppo om he P o ince o B i ish Columbia and he B i ish Columbia Lo e y Co po a ion. The o he au ho s decla e no con lic s o in e es conce ning his s udy. Copy igh and usage BThe Royal College o Psychia is s 2015. The B i ish Jou nal o Psychia y (2015) 207, 158–164. doi: 10.1192/bjp.bp.114.152223 h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess pa hological gambling. We hypo hesised ha cocaine dependence, compa ed o a beha iou al addic ion, would be associa ed wi h dis inc i e al e a ions in p e on al egions ec ui ed by e e sal lea ning. As a seconda y explo a o y aim, we sough o examine whe he his b ain dys egula ion was dopamine gically linked, by s udying associa ions wi h he DRD2/ANKK1 Taq1A gene ic a ian . Me hod Pa icipan s The sample consis ed o 54 pa icipan s: 18 indi iduals mee ing DSM-IV-TR c i e ia o cocaine dependence (cocaine use s) no mee ing c i e ia o any o he Axis I o Axis II diso de , 18 indi iduals mee ing DSM-IV-TR c i e ia o pa hological gambling (gamble s) no mee ing c i e ia o any o he Axis I o Axis II diso de , and 18 heal hy compa ison indi iduals who did no mee DSM-IV-TR c i e ia o Axis I o Axis II diso de s (con ols). Table 1 p esen s sociodemog aphic in o ma ion. The h ee g oups did no di e signi ican ly in age, yea s o educa ion, o IQ measu ed by he Kau man B ie In elligence Tes . 27 Table 1 also p esen s d ug/gambling use cha ac e is ics (mon hly amoun and du a ion o use) as eco ded by he In e iew o Resea ch on Addic i e Beha io s 28 and psychological symp oms as measu ed by he Gene al Heal h Ques ionnai e. 29 Pa icipan s had e y limi ed exposu e o d ugs o he han cocaine, alcohol o obacco; less han 20% o pa icipan s had used cannabis, app oxima ely 5% had used MDMA o hallucinogens, and no pa icipan s had used amphe amines o opia es. Table 1 also displays da a on sel - epo ed abs inence du a ion. In cocaine use s he mean du a ion o abs inence was 2.7 mon hs and in gamble s he mean du a ion o abs inence was 5.7 mon hs. In addi ion, abs inence was moni o ed o 3 weeks du ing he s udy, measu ed wi h u ine es s o alcohol and d ug use and c oss-checked sel - and colla e al epo s o gambling. Table 1 also displays DRD2/ANKK Taq1A geno ype dis ibu ions. In ag eemen wi h popula ion-based da a, 30,31 he A1+ geno ype was ound in be ween 30 and 40% o pa icipan s wi hin each g oup. Sociodemog aphic, d ug/gam- bling use and psychological cha ac e is ics by geno ype subg oups a e p esen ed in online Table DS1. Online Fig. DS1 displays a lowcha o he ec ui men p ocess. Cocaine use s we e ec ui ed as hey commenced ea men in he ou pa ien clinic Cen o P o incial de D ogodependencias in G anada (Spain). Gamble s we e ec ui ed as hey commenced ea men in he ou pa ien clinic Asociacio ´n G anadina de Jugado es en Rehabili acio ´n in G anada (Spain). Bo h clinics p o ide psychological he apies o addic i e diso de s. The inclusion c i e ia we e as ollows: (a) aged be ween 18 and 45 yea s; (b) es ima ed IQ le els abo e 80; (c) mee ing DSM-IV-TR c i e ia o cocaine dependence o pa hological gambling – as assessed by he S uc u ed Clinical In e iew o 159 Neu al subs a es o cogni i e lexibili y Table 1 Demog aphic and clinical cha ac e is ics o he h ee s udy g oups Demog aphic a iables Con ols n=18 Mean (s.d.) Gamble s n=18 Mean (s.d.) Cocaine use s n=18 Mean (s.d.) P Age (yea s) 31.17 (4.74) 33.56 (7.97) 34.27 (6.87) 0.349 Gende (male/ emale) 17/1 16/2 17/1 0.774 La e ali y ( igh -le ) 17/1 17/1 14/4 0.193 Yea s o educa ion 10.56 (1.92) 10.28 (2.11) 9.78 (1.66) 0.468 Ve bal IQ 106.89 (8.98) 102.67 (7.39) 100.94 (7.58) 0.082 DRD2/ANKK Taq1A1+geno ype A1+ 675 A1– 121013 Clinical a iables Cocaine (0 HC/ 0 PG/ 18 CDI) Age a onse cocaine use (yea s) 21.28 (5.83) Mon hly amoun cocaine use s (g) 16.86 (25.49) Du a ion cocaine (mon hs) 43.75 (36.32) Abs inence cocaine (mon hs) 2.73 (5.43) Gambling (0 HC/ 18 PG/ 0 CDI) Age a onse gambling (yea s) 22.17 (8.71) Mon hly amoun gambling (h) 42.53 (41.47) Du a ion gambling (mon hs) 26.12 (24.56) Abs inence gambling (mon hs) 7.69 (6.51) Tobacco (8 HC/ 8 PG/ 14 CDI) Age a onse obacco use (yea s) 17.75 (5.55) 15.50 (3.51) 15.71 (2.58) 0.411 Mon hly obacco use (cig) 286.25 (222.90) 667.50 (278.55) 564.29 (362.21) 0.051 Du a ion obacco (mon hs) 76.37 (104.25) 175.50 (101.00) 137.57 (121.16) 0.219 Alcohol (7 HC/ 14 PG/ 15 CDI) Age a onse alcohol use (yea s) 19.14 (5.53) 16.29 (1.70) 17.87 (4.55) 0.411 Mon hly alcohol use (SDU) 10.07 (9.75) 17.43 (20.09) 31.69 (20.70) 0.038 Du a ion alcohol (mon hs) 83.75 (56.21) 75.43 (63.31) 88.69 (93.90) 0.928 Cannabis (4 HC/ 1 PG/ 6 CDI) Age a onse alcohol use (yea s) 18.75 (3.77) 25 18.71 (7.47) 0.664 Mon hly cannabis use (join s) 0.96 (0.75) 4 107.83 (125.57) 0.267 Du a ion cannabis (mon hs) 19.25 (19.35) 4 110 (137.33) 0.405 GHQ soma ic symp oms 0.39 (0.85) 1.72 (2.33) 1.43 (1.86) 0.079 GHQ anxie y 1.28 (2.19) 1.64 (2.34) 2.12 (2.50) 0.577 GHQ social dys unc ion 0.83 (1.65) 1.00 (1.84) 1.25 (1.91) 0.796 GHQ dep ession 0.44 (0.98) 1.27 (2.15) 1.50 (2.42) 0.248 s.d., s anda d de ia ion; IQ, in elligence quo ien ; g, g ams; h, hou s; cig, ciga e es; SDU, s anda d d inking uni s, GHQ, Gene al Heal h Ques ionnai e. In addi ion o he epo ed equencies o o he d ugs in ake, ou CDI epo ed occasional use o MDMA (mean li e ime use = 6 uni s), while o he wo CDI epo ed occasional use o hallucinogens (mean li e ime use = 14 uni s). No pa icipan s epo ed amphe amines o opia es use. h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess Ve dejo-Ga cia e al DSM-IV Diso de s – Clinician Ve sion (SCID-I-CV); 32 (d) being ea men commence s; and (e) abs inence du a ion >15 days. Ab- s inence in he cocaine use s was con i med by wo u ine es s pe week plus an ad hoc es on he scanning day i sel . Posi i e u ine oxicologies o any o he d ug we e also exclusiona y. Gambling abs inence in he pa hological gamble s was assessed by sel - epo c oss- alida ed by spouses o ela i es. The exclusion c i e ia we e: (a) p esence o any o he Axis I o Axis II diso de s, wi h he excep ions o alcohol misuse and nico ine dependence; (b) his- o y o head inju y o neu ological, in ec ious, sys emic o any o he diseases a ec ing he cen al ne ous sys em; (c) ha ing ol- lowed o he ea men s wi hin he 2 yea s p eceding s udy onse and (d) ha ing en e ed ea men by cou eques . Como bid Axis I diso de s we e assessed wi h he SCID-I-CV. Axis II diso - de s we e assessed using he In e na ional Pe sonali y Diso de s Examina ion (IPDE). 33 We also used he Conne s’ Adul ADHD Diagnos ic In e iew o DSM-IV (CAADID) 34 o assess adul ADHD (which was also exclusiona y). Con ols we e ec ui ed om local employmen agencies. In addi ion o he o me exclusion c i e ia, heal hy con ols could no mee any diagnosis o subs ance use diso de s – wi h he excep ion o nico ine dependence. Axis I and II diso de s we e also assessed in his g oup using he SCID-I-CV, he IPDE and he CAADID. All he diagnoses we e made by a egis e ed clinical psychologis . The s udy was app o ed by he E hics Commi ee o Resea ch in Humans o he Uni e si y o G anada (Spain). All pa icipan s signed an in o med consen o m ce i ying hei olun a y pa icipa ion Func ional MRI ( MRI) ask We used he p obabilis ic e e sal lea ning ask, as desc ibed in Cools e al. 11 In each ial, pa icipan s we e equi ed o choose be ween wo s imuli (abs ac , colou ed pa e ns) p esen ed o he le and igh isual ields (loca ion was andomised). Pa icipan s we e old ha , acco ding o a p ede ined ule, one s imulus was co ec on each ial ( he CS+), and he o he s imulus ( he CS7) was inco ec . A a ious poin s h oughou he ask, he ule deciding he co ec s imulus would change; he pa icipan should change hei esponse when hey we e con iden ha he ule had changed. The ask employed p obabilis ic eedback such ha he CS+ was ewa ded ~85% o imes, and he CS7was ewa ded ~15% o imes. This ga e ise o wo ypes o e o s: p obabilis ic e o s (whe e pa icipan s chose he co ec s imulus bu ecei ed nega i e eedback), and pe se e a i e e o s (whe e pa icipan s keep esponding o he p e iously ein o ced s imuli, despi e nega i e eedback). The ask was ained be o e scanning (using sligh ly di e en s imuli) and hen implemen ed inside he scanne in 3 consecu i e blocks o 11 min each. Each block consis ed o 10 disc imina ion s ages, yielding 9 e e sals. Re e sal o he s imulus– ewa d con ingency occu ed a e 10 o 15 co ec esponses (including p obabilis ic e o s). The numbe o p obabilis ic e o s be ween each e e sal a ied om 0 o 4. S imuli we e p esen ed h ough magne ic- esonance-compa ible liquid-c ys al display goggles (Resonance Technology, No h idge, CA, USA). Beha iou al esponses we e eco ded h ough a i e- bu on box, E oke Response Pad Sys em (Resonance Technology Inc.). On each ial, s imuli we e p esen ed o 2000 ms, wi hin which ime he esponse had o be made (o else a ‘ oo la e’ mes- sage was p esen ed). Pa icipan s esponded using he le o igh bu on on a bu on box posi ioned on pa icipan s’ ches . Feedback was a g een ‘smiley’ ace o co ec esponses, and a ed sad ace o inco ec esponses, and was p esen ed immedia ely a e he pa icipan s’ esponse. The eedback aces we e p esen ed cen ally o 500 ms, du ing which ime he s imuli also emained on he sc een. Following eedback, he e was a a iable in e - ial in e al (a ixa ion c oss) ha was adjus ed so ha he o e all in e s imulus in e al was 3253 ms, enabling p ecise desynch onisa ion om he epe i ion ime (TR) (o 2000 ms) and su icien sampling ac oss he hemodynamic esponse unc ion. Beha iou al measu es The main pe o mance measu es we e hi a es (p opo ion o co ec esponses by o al ials), numbe o pe se e a i e e o s, and pe se e a ion e o a es (numbe o pe se e a i e e o s di ided by numbe o sequences on which he pe se e a ion c i e ion was me ). To mee he pe se e a ion c i e ion, pa icipan s had o make a leas one consecu i e esponse o he p e iously ewa ded s imulus immedia ely ollowing e e sal. Imaging da a acquisi ion and p ep ocessing We used a 3.0 Tesla clinical MRI scanne , equipped wi h an eigh - channel phased-a ay head coil (In e a Achie a, Philips Medical Sys ems, Eindho en, The Ne he lands). Du ing acquisi ion, h ee T2*-weigh ed echo-plana imaging (EPI) was ob ained (TR = 2000 ms, echo ime (TE) = 35 ms, ield o iew (FOV) = 2306230 mm, 96696 ma ix, lip angle = 908,214-mm axial slices, 1-mm gap, 330 scans each). A sagi al h ee-dimensional T1-weigh ed u bo-g adien -echo sequence (160 slices, TR = 8.3 ms, TE = 3.8 ms, lip angle = 88, FOV = 2406240, 1 mm 3 oxels) was ob ained in he same expe imen al session o ana omical localisa- ion o unc ional indings. The b ain images we e analyzed using S a is ical Pa ame ic Mapping so wa e (Wellcome Depa men o Cogni i e Neu ology, Ins i u e o Neu ology, Queen Squa e, London, UK), unning unde Ma lab R2009 (Ma hWo ks, Na ick, MA, USA). P ep ocessing s eps we e slice iming co ec ion, e-slicing o he i s image o he ime se ies, no malisa ion (using a ine and smoo hly non-linea ans o ma ions) o an EPI empla e in he Mon eal Neu ological Ins i u e space, and spa ial smoo hing by con olu ion wi h a 3D Gaussian ke nel ( ull wid h a hal maximum (FWHM) = 8 mm). DRD2/ANKK1 Taq1A geno yping The DRD2/ANKK Taq1A polymo phism ( s1800497) is loca ed in a pu a i e subs a e binding domain o he ANNK1 gene and esul s in a Glu713Lys subs i u ion. In ou s udy, his polymo phism was cha ac e ised using a TaqMan allelic disc imina ion assay om Li e Technologies.Cycling was pe o med on a S epOne Plus he mocycle wi h condi ions ecommended by Li e Technologies. Th ee geno ypes o he dopamine DRD2/ANNK1-TaqIa locus can be di e en ia ed: he A1A1 geno ype, he A1A2 geno ype, and he A2A2 geno ype. Because o he small p e alence o he A1A1 geno ype (3% o he heal hy Whi e popula ion), A1A1 and A1A2 pa icipan s a e commonly g ouped as A1+ pa icipan s, whe eas A2A2 pa icipan s a e e e ed o as A1– pa icipan s. The p e alence o a leas one A1 allele (A1+ g oup) has been associa ed wi h an up o 30% educ ion in D2 ecep o densi y. 35 S a is ical analyses Beha iou al analyses Beha iou al da a we e analysed wi h SPSS e sion 19. We conduc ed one-way ANOVAs ollowed by Tukey es s o compa e he h ee g oups on he e e sal lea ning beha iou al measu es. We also conduc ed wo-way ANOVAs (wi h G oup and DRD2/ ANKK Taq1A geno ype as ac o s) o examine di e ences be ween geno ype subg oups. 160 h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess Neu al subs a es o cogni i e lexibili y Neu oimaging analyses The ime se ies we e high-pass il e ed (128 s), and a canonical hemodynamic esponse unc ion was modeled o a del a unc ion a pa icipan s’ esponse on each ial, which co-occu ed wi h he p esen a ion o he eedback. The ollowing e en s we e modeled: (a) co ec esponses; (b) pe se e a i e e o s (e o s ollowing a ule change whe e pa icipan s ail o swi ch esponse); (c) inal e e sal e o s (las nega i e eedback in he se ies o pe se e a i e e o s ollowed by a esponse swi ch); and (d) p obabilis ic e o s (co ec esponses o which misleading nega i e eedback was gi en). E o ials ha could no be classi ied as p obabilis ic o e e sal e o s we e no included in he model. The main con as o in e es was inal e e sal e o s . pe se e a i e e o s, which e lec s he beha iou al shi componen . We also calcula ed he co ec . inco ec (and he e e sed inco ec . co ec ) con as o map posi i e and nega i e eedback- ela ed ac i a ion, and pe se e a i e minus p obabilis ic e o s o map pe sis en esponding con olling o nega i e eedback. One-sample - es s we e conduc ed on he esul ing i s -le el con as images o assess wi hin-g oup ac i a ions in each o he con as s. These esul s we e co ec ed o mul iple compa isons wi h a combina ion o oxel in ensi y and clus e ex en h esholds. The spa ial ex en h eshold was de e mined by 1000 Mon e Ca lo simula ions using AlphaSim, 36 as implemen ed in he SPM REST oolbox. 37 The inpu pa ame e s included b ain mask o 152 295 oxels, an indi idual oxel h eshold p obabili y o 0.005 and a clus e connec ion adius o 5 mm, a 10.2, 10.4 and 9.1 FWHM smoo hness o he con as s co ec . inco ec , inal e e sal e o s . pe se e a i e e o s and pe se e a i e . p obabilis ic e o s, espec i ely. A minimum clus e ex en o 262, 260 and 212 oxels espec i ely was es ima ed o sa is y a amily-wise e o (FWE) co ec ed P- alue o P FWE 50.05. Nex , we conduc ed a se ies o h ee g oup ANOVAs o assess be ween-g oup di e ences using he same i s -le el con as images. S a is ical signi icance in hese es s was de ined by he same inpu pa ame e s, masking esul s by he ac i a ion maps de i ed om he one-sample - es s. The e o e, o he con as s co ec . inco ec (and inco ec . co ec ), inal e e sal e o s . pe se e a i e e o s and pe se e a i e minus p obabilis ic e o s, espec i ely, a minimum clus e ex en o 104, 14, 47 and 12 oxels (wi hin b ain masks o 41 129, 1812, 14 149 and 1009 oxels), was es ima ed o sa is y a P FWE 50.05. In hose con as s yielding signi ican g oup di e ences, we conduc ed addi ional analyses in SPSS o assess G oup6DRD2/ANKK Taq1A Geno ype in e - ac ions on b ain ac i a ion clus e s di e ing be ween g oups. Speci ically, we conduc ed wo-way ANOVAs (wi h G oup and DRD2/ANKK Taq1A geno ype as ac o s) on peak ac i a ions de i ed om he MRI con as s, ollowed by ele an pai wise compa isons. To exclude a po en ial pe o mance con ound, all analyses we e eplica ed con olling o he beha iou al measu es o inal e e sal e o s and pe se e a i e e o s. Resul s we e equi alen in bo h app oaches, and hence we only epo he o iginal, non-co a ied analyses. Likewise, since bo h obacco and alcohol use ha e been linked o e e sal pe o mance and dopamine gic unc ion, we conduc ed addi ional analyses including he mon hly amoun and du a ion o use o hese subs ances as co a ia es. Fu he , we conduc ed a se ies o wo- way ANOVAs o examine whe he smoking use o cannabis use s a us in e ac ed wi h G oup o DRD2/ANKK Taq1A geno ype e ec s on peak ac i a ions de i ed om he MRI con as s. Co ela ion analyses Co ela ion analyses we e pe o med in SPSS using he peak ac i a ions de i ed om he MRI con as s. The be a eigen alues co esponding o each egion we e ex ac ed o each pa icipan , and hen co ela ed wi h he beha iou al measu es o numbe o pe se e a i e e o s and pe se e a ion e o a es. We also co ela ed he be a eigen alues co esponding o each egion wi h sel - epo es ima es o abs inence du a ion. Resul s Beha iou al measu es Beha iou al measu es a e p esen ed in Table 2. The h ee g oups only di e ed in he a e o pe se e a i e e o s, wi h cocaine use s commi ing mo e pe se e a i e e o s han pa hological gamble s o con ols. In addi ion, in he cocaine g oup, pe se e a ion e o s we e co ela ed posi i ely wi h li e ime du a ion o cocaine use ( = 0.470, P= 0.025). By con as , pe se e a ion e o a es we e nega i ely co ela ed wi h ime since gambling onse in he pa hological gamble s ( =70.409, P= 0.049). G oup x DRD2/ANKK Taq1A geno ype analyses showed no signi ican in e ac ion e ec s on beha iou al measu es. Neu oimaging Co ec . inco ec esponses Collapsing ac oss g oups, co ec ( . inco ec ) esponses we e associa ed wi h inc eased signal in s ia um, supe io and medial on al gy i, la e al o bi o on al co ex, an e io and pos e io cingula e, pos e io insula, amygdala, supe io empo al gy i, angula gy i and occipi al egions. The e e se con as indica ed signal associa ed wi h nega i e eedback in igh dlPFC, igh insula and supplemen a y mo o a ea (see online Table DS2 and Fig. DS2). Be ween-g oup compa isons indica ed no eliable di e ences in hese con as s. Final e e sal e o s . pe se e a i e e o s Shi ing (i.e. inal e e sal e o s) was associa ed wi h signi ican signal inc eases ac oss all g oups in do sal an e io cingula e co ex, bila e al an e io insula /o bi o on al co ex, igh do so- la e al p e on al and en ola e al p e on al co ices, in e io 161 Table 2 Beha io al measu es summa ising pe o mance in he p obabilis ic e e sal lea ning ask in cocaine use s, non-d ug using gamble s and non-d ug using con ols Con ols Mean (s.d.) Gamble s Mean (s.d.) Cocaine use s Mean (s.d.) P Hi a e (% co ec esponses) 64.25 (7.29) 63.39 (8.32) 62.02 (8.53) 0.705 Pe se e a i e e o s 21.72 (9.38) 20.94 (8.53) 27.61 (12.79) 0.120 Sequences on which c i e ion o pe se e a ion was me 14.94 (6.71) 14.72 (6.72) 15.56 (6.56) 0.927 Pe se e a ion e o a e 1.49 (0.26) 1.48 (0.40) 1.87 (0.75) 0.039 To al ials o comple e he ask 508.50 (70.45) 517.50 (77.66) 530.17 (85.64) 0.707 s.d., s anda d de ia ion. h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess Ve dejo-Ga cia e al pa ie al co ex, s ia um, halamus and pos e io isual a eas ex ending o he usi o m gy us. Final e e sal e o s we e also associa ed wi h educed signal in he os al an e io cingula e and medial on al gy i, he pos e io cingula e gy us and he le angula and pa ahippocampal gy i (online Fig. DS3 and Table DS3). Pai wise be ween-g oup compa isons showed ha bo h cocaine use s and pa hological gamble s had signi ican ly dec eased ac i a ion in he igh lPFC (in e io on al gy us) compa ed o con ols. In addi ion, he cocaine use s had signi ican ly dec eased ac i a ion in he igh dlPFC (middle on al gy us) compa ed wi h bo h pa hological gamble s and con ols (online Fig. DS3 and Table DS3). We ound no signi ican co ela ions wi h beha iou al measu es. Pe se e a i e . p obabilis ic e o s Pe se e a ion was associa ed wi h signi ican ac i a ion in supe io and medial on al gy i and os al an e io cingula e gy us. Be ween-g oup compa isons showed no signi ican di e ences be ween cocaine use s o gamble s compa ed o con ols, al hough he cocaine use s did display signi ican ly highe ac i a ion han pa hological gamble s in he medial on al gy us (see online Fig. DS4). Medial on al gy us ac i a ion was also nega i ely co ela ed wi h he numbe o pe se e a i e e o s in cocaine use s ( =70.470, P= 0.025), bu posi i ely co ela ed wi h he numbe o pe se e a i e e o s in pa hological gamble s ( = 0.467, P= 0.025, see online Fig. DS4). G oup6DRD2/ANKK Taq1A geno ype in e ac ions Clus e s showing signi ican be ween-g oup di e ences we e u he examined in ela ion o G oup6DRD2/ANKK Taq1A geno ype analyses. The e was a signi ican g oup6geno ype in e - ac ion (F(2,46) = 4.81, P= 0.013) in he igh lPFC, d i en by opposing e ec s o he dopamine geno ype in he cocaine use s ela i e o he o he wo g oups (see online Fig. DS5). Pai wise analyses showed ha wi hin A17ca ie s, cocaine use s had lowe ac i a ion han bo h pa hological gamble s and con ols. Con e sely, cocaine A1+ ca ie s had signi ican ly highe ac i a ion han pa hological gamble s o he same geno ype. No u he signi ican in e ac ions we e obse ed. Co ela ion be ween pa e ns o b ain ac i a ion and abs inence du a ion We did no ind signi ican co ela ions be ween ask- ela ed b ain ac i a ions and du a ion o abs inence o cocaine o gambling use. Sensi i i y analyses Co a ia e models including mon hly amoun and du a ion o alcohol and obacco use did no change he o e all pa e n o esul s. Fu he , smoking s a us (smoke s . non-smoke s) and cannabis use s a us (cannabis use s . non-use s) showed no signi ican in e ac ions wi h G oup o Geno ype on any o he peak ac i a ions de i ed om MRI analyses. Discussion Ou esul s demons a e ha educed signal in igh lPFC du ing shi ing is common o bo h cocaine use s and pa hological gamble s. This sha ed e ec was supplemen ed by a wide pa e n o ask- ela ed dys egula ion in he cocaine use s, wi h dec eased igh dlPFC ac i a ion du ing shi ing, and inc eased medial p e on al co ex ac i a ion du ing pe se e a ion. These b ain ac i a ion di e ences we e pa alleled by beha iou al esul s, whe e he cocaine use s commi ed mo e pe se e a i e e o s compa ed wi h bo h pa hological gamble s and con ols. P elimina y explo a o y gene ic analyses o he DRD2/ANKK Taq1A geno ype sugges an unde lying dopamine gic con ibu ion o e e sal- ela ed b ain ac i i y: bo h con ols and pa hological gamble s ca ying he (high isk) A1+ geno ype had dec eased swi ch- ela ed lPFC signal, bu his pa e n was e e sed in he cocaine g oup, in which he A1+ ca ie s exhibi ed g ea e shi ing- ela ed ac i a ion. Ou ask ac i a ion esul s eplica e he well-desc ibed pa e n o do sal p e on al/insula ac i a ions in esponse o e o - ela ed nega i e eedback, in conce wi h ec ui men o mo e en al and la e al aspec s o p e on al co ex, an e io cingula e and s ia um du ing shi ing. 11,38 Bo h cocaine use s and pa hological gamble s showed diminished ac i a ion o he igh lPFC du ing shi ing, consis en wi h he pa e n p e iously desc ibed in pa ho- logical gamble s pe o ming a simila e e sal ask. 14 The igh lPFC is a key egion o success ul sel -con ol o beha iou and emo ional egula ion. 39 Fu he , p e ious MRI s udies ha e shown ha igh lPFC ac i a ion is dis inc i ely inc eased in indi iduals wi h high esilience o addic ion, 40 and dec eased in indi iduals wi h amily isk o addic ion. 41 We in e ha dys unc ion o his egion is commonly in ol ed in bo h cocaine and gambling addic ions. The inc eased shi ing- ela ed ac i a ion in cocaine use s ca ying he (high isk) A1 allele compa ed o pa hological gamble s ca ying he same allele migh be explained by he dopamine in e ed-U p inciple, by which A1+ ca ie s, wi h lowe disposi ional dopamine unc ion, may de i e s imula o y ‘bene i s’ om cocaine-induced changes. 42 Fo exam- ple, ea men wi h he dopamine D2- ecep o agonis cabe goline in heal hy olun ee s p o oked opposi e e ec s in A17 . A1+ ca ie s, also mani es ed in he ac i a ion o he igh lPFC. 43 Howe e , because o he small sample size o geno ype subg oups, hese indings should be ea ed as p elimina y, and could al e na i ely be explained by linked in ol emen o o he dopamine gene polymo phisms. Fo example, ecen e idence sugges s ha he dopamine agonis olcapone has unique s imula o y e ec s on p e on al co ex ac i i y in smoke s ca ying he COMT al/ al geno ype, which is also associa ed wi h lowe dopamine unc ion. 44 Fu u e, adequa ely powe ed, molecula gene ic s udies a e wa an ed o es he no ion o whe he indi iduals wi h low disposi ional dopamine ansmission de i e s imula o y b ain e ec s om cocaine consump ion. In addi ion o o e lapping de ici s, cocaine use s showed dec eased igh dlPFC ac i a ion (BA 9) compa ed wi h gamble s and con ols du ing shi ing, and inc eased medial on al gy us ac i a ion (BA 10) compa ed wi h gamble s du ing pe se e a ion. BA 9 has been speci ically in ol ed in he upda ing o he wo king memo y s o es ha se s imulus- esponse con ingencies, 45 whe eas BA10 has been p ima ily in ol ed in he con ol o s imulus-o ien ed a en ion. 46 The e o e, bo h indings a e compa ible wi h he model-based . model- ee sys ems app oach o e e sal lea ning. 2 Acco ding o his model, cocaine-induced changes a ec sys ems in ol ed in he upda ing o s imulus- ou come alues ha se e o adjus p edic ions abou u u e ou comes. P eclinical s udies ha e shown cocaine-induced neu al adap a ions in he a p elimbic co ex ( he unc ional homologue o he human dlPFC), 47,48 which is c i ical o he lea ning o no el s imulus-ou come associa ions. 49 Simila ly, in humans, du a ion o cocaine use is nega i ely associa ed wi h BA 9 g ay ma e olumes. 50 As a consequence o neu oadap a ions in b ain egions in ol ed on model-based p edic ions, cocaine use s may become mo e dependen on sys ems handling model- ee cached 162 h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess Neu al subs a es o cogni i e lexibili y ep esen a ions o s imulus-ou come alues. 2 We p opose ha his compensa o y mechanism would be exempli ied by signi ican ly inc eased medial p e on al ac i a ion du ing pe se e a ion in cocaine use s compa ed o gamble s. The e o e, ou esul s sugges ha cocaine use s equi e addi ional ec ui men o medial p e on al egions o compensa e o comp omised la e al p e- on al egions specialised in he upda ing o s imulus-ou come p edic ions. Con e sely, pa hological gamble s engage his egion o a lesse ex en han bo h cocaine use s and con ols, and his pa e n co ela es wi h ewe pe se e a ions wi hin his g oup (see Domb o ski e al 51 o a simila e ec in dep ession, which migh be explained by hese popula ions being o e ly sensi i e o misleading p obabilis ic eedback, hence less likely o pe se e a e in his ask). This in e p e a ion is also consis en wi h he neu opsychological p o ile we ha e desc ibed p e iously o his coho , whe e he cocaine use s ha e a selec i e impai men in wo king memo y compa ed o he pa hological gamble s (i.e. poo e upda ing o s imulus-ou come alues), whe eas he pa hological gamble s ha e s eepe delay discoun ing (i.e. enhanced ein o cemen sensi i i y). 52 Ou esul s illus a e he neu al unde pinnings o e e sal lea ning in cocaine and gambling addic ions. Since e e sal lea ning is a well- alida ed ansla ional model o in lexibili y/ pe se e a ion, and is linked o addic ion se e i y and clinical p ognosis, hese esul s in o m bo h mechanis ic and clinical esea ch in addic i e diso de s. In ega ds o clinical implica ions, ou indings sugges ha b ain s imula ion and/o cogni i e enhancemen in e en ions a ge ing he dlPFC may con ibu e o he alle ia ion o pe se e a ion in he con ex o cocaine addic ion. 53 We no e ha ou MRI p ocedu e did no de ec signi ican beha iou al al e a ions in e e sal lea ning pe o mance be ween gamble s and con ols. One possible in e ence is ha signs o o e compulsi i y may be less in pa hological gamble s, compa ed wi h cocaine addic ion. Howe e , he p obabilis ic ask in ol ed se ial e e sals on a semi- egula schedule in o de o op imise case-by-case ec ui men o on o-s ia al neu al ci cui y, bu wi h educed beha iou al sensi i i y o g oup di e ences. The obse a ion ha gamble s showed a di e en ial pa e n o e e sal- ela ed b ain ac i i y is compa ible wi h he neu opsychological di e ences ound in p e ious s udies using beha iou -sensi i e e e sal asks. 10 In his ega d, ou indings sugges ha b ain s imula ion and/o cogni i e enhancemen in e en ions a ge ing he lPFC and i s key unc ions (e.g. cogni i e con ol, esponse inhibi ion) could ha e u ili y in he ea men o pa hological gambling. Clinicians may also adap s anda d in e en ions o bu e he impac o en ola e al dys unc ion on eal-li e unc ioning (i.e. ins uc ing clien s o pay a en ion o nega i e eedback and aining hem o gene a e and ehea se al e na i e s a egies). Ou s udy has se e al s eng hs, including he di ec compa ison o cocaine and gambling g oups wi h minimum exposu e o alcohol/o he d ugs and wi hou he con ounding e ec s o psychia ic como bidi ies. Mo eo e , ec ui men was based on consecu i e admissions o public ea men cen es, hence making he sample uly clinically ep esen a i e. Fu he , bo h g oups we e ca e ully supe ised o con inuous abs inence du ing he s udy, hus uling ou con ounding e ec s o acu e d ug use/gambling, wi hd awal o c a ing. Cu en d ug use was objec i ely moni o ed h oughou s udy comple ion using ongoing u ine oxicologies ha uled ou any use o cocaine and o he d ugs. The e is howe e a po en ial limi a ion in he supe ision o gambling abs inence in he gambling g oup, which was based on epo s om bo h pa icipan s and signi ican o he s, bu is s ill suscep ible o epo ing biases. 54,55 Mo eo e , he gene ic analyses in pa icula should be in e p e ed in he con ex o he ela i ely small sample size and he examina ion o a single cogni i e domain and, as such, should be app aised as p elimina y. Fu u e s udies a e wa an ed o examine he addi i e con ibu ion o di e en polymo phisms ac oss he dopamine gene ic pa hway, and he clinical ele ance o his e e sal lea ning- ela ed neu oimaging pheno ype o cocaine and gambling ea men ou comes. An onio Ve dejo-Ga cia, PhD, School o Psychology and Psychia y, Monash Uni e si y, Melbou ne, Aus alia, Ins i u e o Neu oscience F. Olo iz, Uni e sidad de G anada, G anada, Spain and Red de T as o nos Adic i os, Uni e sidad de G anada. G anada, Spain; Luke Cla k, PhD, Depa men o Psychology, Cen e o Gambling Resea ch a UBC, Uni e si y o B i ish Columbia, Canada; Juan Ve dejo-Roma ´n, MSc, Ins i u e o Neu oscience F. Olo iz, Uni e sidad de G anada, G anada, Spain; Na alia Albein-U ios, PhD, Ins i u e o Neu oscience F. Olo iz, Uni e sidad de G anada, G anada, Spain; Jose ´M. Ma inez-Gonzalez, PhD, Red de T as o nos Adic i os, Uni e sidad de G anada, G anada, Spain and Cen o P o incial de D ogodependencias, Dipu acion de G anada, G anada, Spain; Blanca Gu ie ez, PhD, Ins i u e o Neu oscience F. Olo iz, Uni e sidad de G anada, G anada, Spain, Depa men o Psychia y, Uni e sidad de G anada, G anada, Spain and CIBERSAM, Ca los III Heal h Ins i u e, Ba celona, Spain; Ca les So iano-Mas, PhD, Depa men o Psychia y, Bell i ge Biomedical Resea ch Ins i u e-IDIBELL, Ba celona, Spain and CIBERSAM, Ca los III Heal h Ins i u e, Ba celona, Spain, Depa men o Psychobiology and Me hodology o Heal h Sciences, Uni e si a Au o `noma de Ba celona, Spain Co espondence: An onio Ve dejo-Ga cia, School o Psychology and Psychia y, Monash Uni e si y, 3800 Welling on Rd. Clay on Campus, Melbou ne, Aus alia. Email add ess: An onio.Ve [email protected] Fi s ecei ed 1 Jun 2014, inal e ision 18 No 2014, accep ed 23 No 2014 Funding This s udy has been unded by g an s om he Spanish Minis y o Heal h; p ojec g an COPERNICO, D ug Abuse Plan (Plan Nacional sob e D ogas Con oca o ia 2009) and p og am g an RETICS, Ca los III Heal h Ins i u e (Ins i u o de Salud Ca los III, Red de T as o nos Adic i os). CS-M is unded by a ‘Miguel Se e ’ con ac om he Ca los III Heal h Ins i u e (CP10/00604). Re e ences 1Ame ican Psychia ic Associa ion. Diagnos ic and S a is ical Manual o Men al Diso de s, 5 h ed. Ame ican Psychia ic, 2013. 2Lucan onio F, S alnake TA, Shaham Y, Ni Y, Schoenbaum G. The impac o o bi o on al dys unc ion on cocaine addic ion. Na Neu osci 2012; 15: 358–66. 3Tu ne TH, LaRowe S, Ho ne MD, He on J, Malcolm R. Measu es o cogni i e unc ioning as p edic o s o ea men ou come o cocaine dependence. J Subs Abuse T ea 2009; 37: 328–34. 4Cla k L, Cools R, Robbins TW. The neu opsychology o en al p e on al co ex: decision-making and e e sal lea ning. B ain Cogn 2004; 55: 41–53. 5Izquie do A, Jen sch JD. Re e sal lea ning as a measu e o impulsi e and compulsi e beha io in addic ions. Psychopha macology 2012; 219: 607–20. 6Leeman RF, Po enza MN. Simila i ies and di e ences be ween pa hological gambling and subs ance use diso de s: a ocus on impulsi i y and compulsi i y. Psychopha macology 2012; 219: 469–90. 7Fe na ´ndez-Se ano MJ, Pe ales JC, Mo eno-Lo ´pez L, Pe ´ ez-Ga cı ´aM, Ve dejo-Ga cı ´a A. Neu opsychological p o iling o impulsi i y and compulsi i y in cocaine dependen indi iduals. Psychopha macology 2012; 219: 673–83. 8E sche KD, Roise JP, Robbins TW, Sahakian BJ. Ch onic cocaine bu no ch onic amphe amine use is associa ed wi h pe se e a i e esponding in humans. Psychopha macology 2008; 197: 421–31. 9Vonmoos M, Hulka LM, P elle KH, Jenni D, Baumga ne MR, S ohle R, e al. Cogni i e dys unc ions in ec ea ional and dependen cocaine use s: ole o a en ion-de ici hype ac i i y diso de , c a ing and ea ly age a onse . B J Psychia y 2013; 203: 35–43. 10 Vanes LD, an Hols RJ, Jansen JM, an den B ink W, Oos e laan J, Goud iaan AE. Con ingency lea ning in alcohol dependence and pa hological gambling: lea ning and unlea ning ewa d con ingencies. Alcohol Clin Exp Res 2014 (doi:10.1111/ace .12393). 11 Cools R, Cla k L, Owen, AM, Robbins TW. De ining he neu al mechanisms o p obabilis ic e e sal lea ning using e en - ela ed unc ional magne ic esonance imaging. J Neu osci 2002; 22: 4563–7. 163 h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess Ve dejo-Ga cia e al 12 Camchong J, MacDonald III AV, Nelson B, Bell C, Muelle BA, Specke S, e al. F on al hype conneci i y ela ed o discoun ing and e e sal lea ning in cocaine subjec s. Biol Psychia y 2011; 69: 1117–23. 13 E sche KD, Roise JP, Abbo S, C aig KJ, Mulle U, Suckling J, e al. Response pe se e a ion in s imulan dependence is associa ed wi h s ia al dys unc ion and can be amelio a ed by a D(2/3) ecep o agonis . Biol Psychia y 2011; 70: 754–62. 14 de Rui e MB, Vel man DJ, Goud iaan AE, Oos e laan J, Sjoe ds Z, an den B ink W. Response pe se e a ion and en al p e on al sensi i i y o ewa d and punishmen in male p oblem gamble s and smoke s. Neu opsychopha macology 2009; 34: 1027–38. 15 Jocham G, Klein TA, Neumann J, on C amon DY, Reu e M, Ullspe ge M. Dopamine DRD2 polymo phism al e s e e sal lea ning and associa ed neu al ac i i y. J Neu osci 2009; 29: 3695–704. 16 Noble EP. Addic ion and i s ewa d p ocess h ough polymo phisms o he D2 dopamine ecep o gene: a e iew. Eu Psychia y 2000; 15: 79–89. 17 Madha an A, A gilli E, Bonci A, Whis le JL. Loss o D2 dopamine ecep o unc ion modula es cocaine-induced glu ama e gic synap ic po en ia ion in he en al egmen al a ea. J Neu osci 2013; 33: 12329–36. 18 Volkow ND, Fowle JS, Wang GJ, Hi zemann R, Logan J, Schlye DJ, e al. Dec eased dopamine D2 ecep o a ailabili y is associa ed wi h educed on al me abolism in cocaine abuse s. Synapse 1993; 14: 169–77. 19 Volkow ND, Chang L, Wang GL, Fowle JS, Ding YS, Sedle M, e al. Low le el o b ain dopamine D2 ecep o s in me hamphe amine abuse s: associa ion wi h me abolism in he o bi o on al co ex. Am J Psychia y 2011; 158: 2015–21. 20 Po e JN, Olsen AS, Gu nsey K, Dugan BP, Jedema HP, B adbe y CW. Ch onic cocaine sel -adminis a ion in hesus monkeys: impac on associa i e lea ning, cogni i e con ol, and wo king memo y. J Neu osci 2011; 31: 4926–34. 21 Lee B, G oman S, London ED, Jen sch JD. Dopamine D2/D3 ecep o s play a speci ic ole in he e e sal o a lea ned isual disc imina ion in monkeys. Neu opsychopha macology 2007; 32: 2125–34. 22 Fe nandez-Se ano MJ, Pe ez-Ga cia M, Schmid Rio-Valle J, Ve dejo-Ga cia A. Neu opsychological consequences o alcohol and d ug abuse on di e en componen s o execu i e unc ions. J Psychopha macol 2010; 24: 1317–32. 23 Woicik PA, U ban C, Alia-Klein N, Hen y A, Maloney T, Telang F, e al. A pa e n o pe se e a ion in cocaine addic ion may e eal neu ocogni i e p ocesses implici in he Wisconsin Ca d So ing Tes . Neu opsychologia 2011; 49: 1660–9. 24 an Hols RJ, de Rui e MB, an den B ink W, Vel man DJ, Goud iaan AE. A oxel-based mo phome y s udy compa ing p oblem gamble s, alcohol abuse s, and heal hy con ols. D ug Alcohol Depend 2012; 124: 142–8. 25 Cla k L, Limb ick-Old ield EH. Diso de ed gambling: a beha io al addic ion. Cu Opin Neu obiol 2013; 23: 655–9. 26 Limb ick-Old ield EH, an Hols RJ, Cla k L. F on o-s ia al dys egula ion in d ug addic ion and pa hological gambling: consis en inconsis encies? Neu oImage Clin 2013; 2: 385–93. 27 Kau man JC, Kau man AS. Time o he changing o he gua d: a a ewell o sho o ms o in elligence es s. J Psychoeduca ional Assess 2001; 19: 245–67. 28 Ve dejo-Ga cia AJ, Lopez-To ecillas F, Aguila de A cos F, Pe ez-Ga cia M. Di e en ial e ec s o MDMA, cocaine, and cannabis use se e i y on dis inc i e componen s o he execu i e unc ions in polysubs ance use s: a mul iple eg ession analysis. Addic Beha 2005; 30: 89–101. 29 Lobo A, Pe ez-Eche e ia MJ, A al J. Validi y o he scaled e sion o he Gene al Heal h Ques ionnai e (GHQ-28) in a Spanish popula ion. Psychol Med 1986; 16: 135–40. 30 Lo ´pez-Cas oma ´n J, Vaque o-Lo enzo C, Pe ez-Rod iguez MM, Diaz-He nandez M, Fe nandez-Pique as J, Saiz-Ruiz J, e al. Gende e ec on associa ion be ween DRD2 polymo phism and subs ance dependence in a Spanish sample. D ug Alcohol Depend 2009; 101: 210–2. 31 Pe ez de los Cobos J, Baige M, T ujols J, Sinol N, Volpini V, Banuls E, e al. Allelic and geno ypic associa ions o DRD2 TaqI A polymo phism wi h he oin dependence in Spanish subjec s: a case con ol s udy. Beha B ain Func 2007; 3: 25. 32 Fi s MB, Spi ze RL, Gibbon M, Williams JBW. S uc u ed Clinical In e iew o DSM-IV Axis I Diso de s (SCID-I). Ame ican Psychia ic P ess, 1997. 33 Lo ange AW, Sa o ius N, And eoli A, Be ge P, Buchheim P, Channabasa anna SM, e al. The In e na ional Pe sonali y Diso de Examina ion. The Wo ld Heal h O ganiza ion/Alcohol, D ug Abuse, and Men al Heal h Adminis a ion in e na ional pilo s udy o pe sonali y diso de s. A ch Gen Psychia y 1994; 51: 215–24. 34 Conne s CK. Clinical use o a ing scales in diagnosis and ea men o a en ion-de ici /hype ac i i y diso de . Pedia Clin No h Am 1999; 46: 857–70. 35 Ri chie T, Noble EP. Associa ion o se en polymo phisms o he D2 dopamine ecep o gene wi h b ain ecep o -binding cha ac e is ics. Neu ochem Res 2003; 28: 73–82. 36 Wa d BD. AlphaSim P og am Documen a ion o AFNI, Simul aneous In e ence o MRI Da a. Medical College o Wisconsin, Milwaukee. 2000 (h p://a ni.nimh.nih.go /pub/dis /doc/manual/AlphaSim.pd ). 37 Song XW, Dong ZY, Long XY, Li SF, Zuo XN, Zhu CZ, e al. REST: a oolki o es ing-s a e unc ional magne ic esonance imaging da a p ocessing. PLoS ONE 2011; 6: e25031. 38 Mi chell DG, Luo Q, A ny SB, Kasp zycki T, Gup a K, Chen G, e al. Adap ing o dynamic s imulus- esponse alues: di e en ial con ibu ions o in e io on al, do somedial, and do sola e al egions o p e on al co ex o decision making. J Neu osci 2009; 29: 10827–34. 39 Tabibnia G, Mon e osso JR, Baicy K, A on AR, Pold ack RA, Chak apani S, e al. Di e en o ms o sel -con ol sha e a neu ocogni i e subs a e. J Neu osci 2011; 31: 4805–10. 40 Glase YG, Zubie a JK, Hsu DT, Villa ue e S, Mickey BJ, T ucco EM, e al. Indi ec e ec o co ico opin- eleasing ho mone ecep o 1 gene a ia ion on nega i e emo ionali y and alcohol use ia igh en ola e al p e on al co ex. J Neu osci 2014; 34: 4099–107. 41 Smi h DG, Jones PS, Bullmo e ET, Robbins TW, E sche KD. Cogni i e con ol dys unc ion and abno mal on al co ex ac i a ion in s imulan d ug use s and hei biological siblings. T ansl Psychia y 2013; 3: e257. 42 K oene S, Chandle LJ, Phillips PE, Seamans JK. Dopamine modula es pe sis en synap ic ac i i y and enhances he signal- o-noise a io in he p e on al co ex. PLoS ONE 2009; 4: e6507. 43 Cohen MX, K ohn-G imbe ghe A, Elge CE, Webe B. Dopamine gene p edic s he b ain’s esponse o dopamine gic d ug. Eu J Neu osci 2007; 26: 3652–60. 44 Asha e RL, Wiley o EP, Rupa el K, Goelz PM, Hopson RD, Valdez JN, e al. E ec s o olcapone on wo king memo y and b ain ac i i y in abs inen smoke s: a p oo -o -concep s udy. D ug Alcohol Depend 2013; 133: 852–6. 45 Cla k CR, Egan GF, McFa lane AC, Mo is P, Webe D, Sonkkilla C, e al. Upda ing wo king memo y o wo ds: a PET ac i a ion s udy. Hum B ain Mapp 2000; 9: 42–54. 46 Wage TD, Jonides J, Reading S. Neu oimaging s udies o shi ing a en ion: a me a-analysis. Neu oImage 2004; 22: 1679–93. 47 Hea ing M, Ko ecki L, Ma on Fe nandez de Velasco E, Faja do-Se ano A, Chung HJ, Lujan R, e al. Repea ed cocaine weakens GABA(B)-Gi k signaling in laye 5/6 py amidal neu ons in he p elimbic co ex. Neu on 2013; 80: 159–70. 48 Kasane z F, La ou cade M, De oche-Gamone V, Re es JM, Be son N, Balado E, e al. P e on al synap ic ma ke s o cocaine addic ion-like beha io in a s. Mol Psychia y 2013; 18: 729–37. 49 Mihindou C, Guillem K, Na ailles S, Vouillac C, Ahmed SH. Disc imina i e inhibi o y con ol o cocaine seeking in ol es he p elimbic p e on al co ex. Biol Psychia y 2013; 73: 271–9. 50 Ba os-Losce ales A, Ga a an H, Bus aman e JC, Ven u a-Campos N, Llopis JJ, Belloch V, e al. Reduced s ia al olume in cocaine-dependen pa ien s. Neu oimage 2011; 56: 1021–6. 51 Domb o ski AY, Cla k L, Siegle GJ, Bu e s MA, Ichikawa N, Sahakian BJ, e al. Rewa d/punishmen e e sal lea ning in olde suicide a emp e s. Am J Psychia y 2010; 167: 699–707. 52 Albein-U ios N, Ma inez-Gonza ´lez JM, Lozano O, Cla k L, Ve dejo-Ga cı ´a A. Compa ison o impulsi i y and wo king memo y in cocaine addic ion and pa hological gambling: implica ions o cocaine-induced neu o oxici y. D ug Alcohol Depend 2012; 126: 1–6. 53 Wesley MJ, Bickel WK. Remembe he u u e II: me a-analyses and unc ional o e lap o wo king memo y and delay discoun ing. Biol Psychia y 2014; 75: 435–48. doi:10.1016/j.biopsych.2013.08.008. 54 Kuen zel JG, Hende son MJ, Mel ille CL. The impac o social desi abili y biases on sel - epo among college s uden and p oblem gamble s. J Gambl S ud 2008; 24: 307–19. 55 Toce-Ge s ein M, Ge s ein DR, Volbe g RA. A hie a chy o gambling diso de s in he communi y. Addic ion 2003; 98: 1661–72. 164 h ps://doi.o g/10.1192/bjp.bp.114.152223 Published online by Camb idge Uni e si y P ess