ORIGINAL INVESTIGATION

Prefrontal dopamine D1 receptors and working memoryin schizotypal personality disorder: a PET studywith [11C]NNC112

Judy L. Thompson & Daniel R. Rosell & Mark Slifstein & Ragy R. Girgis & Xiaoyan Xu &

Yosefa Ehrlich & Lawrence S. Kegeles & Erin A. Hazlett & Anissa Abi-Dargham &

Larry J. Siever

Received: 2 October 2013 /Accepted: 31 March 2014# Springer-Verlag Berlin Heidelberg 2014

AbstractRationale Schizotypal personality disorder (SPD) is associat-ed with working memory (WM) impairments that are similarto those observed in schizophrenia. Imaging studies havesuggested that schizophrenia is associated with alterations indopamine D1 receptor availability in the prefrontal cortex(PFC) that may be related to the WM impairments that char-acterize this disorder.Objectives The aim of this study was to characterize prefron-tal D1 receptor availability and its relation to WM perfor-mance in SPD.Methods We used positron emission tomography (PET) andthe radiotracer [11C]NNC112 with 18 unmedicated SPD and

21 healthy control participants; as an index of D1 receptoravailability, binding potential (BP) measures (BPF, BPND, andBPP) were calculated for prefrontal and striatal subregions. ToassessWM, SPD participants completed the 2-back and PacedAuditory Serial Addition Test (PASAT).Results There were no significant group differences in PFCBP. BPF and BPP in the medial PFC were significantly nega-tively related to PASAT performance (rs=−0.551, p=.022 andrs=−0.488, p=.047, respectively), but BP was not related to 2-back performance.Conclusions In contrast to what has been found in schizo-phrenia, SPD was not associated with significant alterationsin prefrontal D1 receptor availability. Similar to previous schizo-phrenia findings, however, higher prefrontal D1 receptor avail-ability was associated with poorer WM performance (as mea-sured by the PASAT) in SPD. These findings suggest thatschizophrenia and SPDmay share a common pathophysiologicalfeature related to prefrontal dopamine functioning that contrib-utes to WM dysfunction, but that in SPD, alterations in D1 mayoccur only in a subset of individuals and/or to an extent that isminor relative to what occurs in schizophrenia.

Keywords Schizotypal personality disorder . Schizophrenia .

PET . Dopamine . Prefrontal cortex .Workingmemory

Introduction

Cognitive dysfunction is a key feature of schizophrenia thathas been shown to predict poorer functional outcome and isrelatively intractable to current treatments (Green 2006).Schizotypal personality disorder (SPD) is considered the pro-totypic schizophrenia spectrum disorder and shares commonsymptomatic, neuropsychological, neuroanatomical, and

Electronic supplementary material The online version of this article(doi:10.1007/s00213-014-3566-6) contains supplementary material,which is available to authorized users.

J. L. Thompson :M. Slifstein : R. R. Girgis :X. Xu : L. S. Kegeles :A. Abi-DarghamDepartment of Psychiatry, Columbia University College ofPhysicians and Surgeons, New York, NY 10032, USA

D. R. Rosell :Y. Ehrlich : E. A. Hazlett : L. J. SieverDepartment of Psychiatry, Mount Sinai School of Medicine, NewYork, NY 10029, USA

D. R. Rosell :Y. Ehrlich : E. A. Hazlett : L. J. SieverJames J. Peters Veterans Affairs Medical Center, Bronx, NY 10468,USA

L. S. Kegeles :A. Abi-DarghamDepartment of Radiology, Columbia University College ofPhysicians and Surgeons, New York, NY 10032, USA

J. L. Thompson (*)School of Health Related Professions, Rutgers, The State Universityof New Jersey, 1776 Raritan Road, Scotch Plains, NJ 07076, USAe-mail: judy.thompson@rutgers.edu

PsychopharmacologyDOI 10.1007/s00213-014-3566-6

genetic risk features with schizophrenia (Siever and Davis2004). For example, SPD is associated with an array ofclinically relevant working memory (WM) and executivefunctioning deficits that are qualitatively similar to thoseobserved in schizophrenia (McClure et al. 2008;Mitropoulou et al. 2002). Thus, studying cognitive impair-ments and their neurobiological correlates in SPD can be apowerful strategy for advancing knowledge regarding thepathophysiology of and potential therapeutic targets forschizophrenia, as doing so largely obviates the potentiallyobscuring factors of antipsychotic medication exposure andsecondary effects of chronic psychosis (Siever and Davis2004).

Preclinical research has indicated that intact WM perfor-mance is critically dependent on cortical dopamine (DA)functioning, and in particular, stimulation of D1 receptors(D1R) by DA in the prefrontal cortex (PFC). In a pivotalstudy, it was shown that local infusion of D1R antagonistsinto the PFC of nonhuman primates resulted in a degradationof WM performance (Sawaguchi and Goldman-Rakic 1991).Subsequent studies demonstrated that D1R agonists improvedWM performance both in aged primates and in DA-depletedyoung adult primates (Arnsten et al. 1994). Such findingssuggested that the cognitive dysfunction associated withschizophrenia may in part be due to deficits in prefrontal D1stimulation. In humans, it was recently shown that adminis-tration of a D1R antagonist to young adults attenuated WMperformance to a level similar to that of an older adult group(Fischer et al. 2010). Furthermore, we recently found that inSPD participants, performance in several cognitive domains,including WM and executive functioning, improved after4 weeks of treatment with the mixed D1-D2R agonistpergolide (McClure et al. 2010).

To more directly assess D1R and related WM performancein schizophrenia, we have used positron emission tomography(PET) and the D1R radioligand [11C]NNC112. We previouslyfound that patients with schizophrenia were characterized bysignificantly higher D1R availability in the dorsolateral PFC(DLPFC) compared to controls (by 28 %); we also observed asignificant negative association between DLPFC D1R avail-ability and WM performance in the patients with schizophre-nia (Abi-Dargham et al. 2002). A possible explanation forthese findings is that in schizophrenia, there is a compensatory(but inefficient) upregulation of D1R in response to chronicunderstimulation of D1 in this brain region (Abi-Darghamet al. 2002). This interpretation is consistent with convergingevidence from preclinical (Guo et al. 2003; Tsukada et al.2005) and human (Narendran et al. 2005; Slifstein et al.2008) research that reduced DA levels are associated with anupregulation of D1 as measured with [11C]NNC112.

More recently, we found that compared to controls, D1Ravailability was increased in regions of the PFC in drug-naïvepatients with schizophrenia, whereas patients with a history of

antipsychotic treatment did not differ significantly from con-trols; furthermore, among patients with prior antipsychotictreatment, the duration of their medication-free interval waspositively associated with DLPFC D1R availability (Abi-Dargham et al. 2012). These findings suggest that antipsy-chotic exposure may normalize illness-related upregulation ofD1. Results from recent studies comparing D1R availability inschizophrenia as measured with the radiotracers[11C]NNC112 and [11C]SCH23390 (Kosaka et al. 2010;Poels et al. 2013) are consistent in suggesting that D1R levelsin schizophrenia are influenced by factors such as antipsy-chotic exposure and possibly stage of illness, and that suchpatient characteristics have contributed to the mixed findingsyielded by studies of D1R availability in schizophrenia (seeAbi-Dargham et al. 2012; Kosaka et al. 2010; Poels et al.2013).

Assessing D1R and its relation toWMperformance in SPDcan complement the above studies of schizophrenia by exam-ining such phenomena in a related condition while controllingfor factors such as antipsychotic exposure. This approachshould also contribute to knowledge regarding the potentialcortical DA pathology of SPD, a spectrum condition thatshares many features of schizophrenia, including cognitiveand social impairment, yet is not characterized by frank psy-chosis. To our knowledge, there are no published studies ofD1R availability in SPD. Thus, we sought to characterize thisindex of DA functioning and its relation to WM performancein a group of healthy, unmedicated patients with SPD usingPETand [11C]NNC112. Based on previous findings in schizo-phrenia (Abi-Dargham et al. 2002, 2012), our primary hy-potheses were the following: (1) SPD participants would havehigher D1R availability in the PFC compared to matchedcontrols, and (2) prefrontal D1R availability would be nega-tively related toWM performance in SPD. Given the previousfindings suggesting that SPD may be characterized by alter-ations in striatal DA functioning (Abi-Dargham et al. 2004),we also conducted exploratory analyses of D1R availability instriatal subregions.

Methods

Participants

This study was approved by the Institutional Review Boards ofNew York State Psychiatric Institute of Columbia UniversityMedical Center, Mount Sinai Hospital, and the BronxVeterans Affairs Medical Center and performed in accordancewith the ethical standards laid down in the 1964Declaration ofHelsinki. All participants provided written informed consentafter the procedures were fully explained to them, and SPDparticipants were independently assessed for capacity to pro-vide consent. Participants were recruited from the community

Psychopharmacology

through advertisements in local newspapers and Internet, andfor some SPD participants, through clinician referral. A subsetof the SPD participants of this study was included in previ-ously published reports of WM in SPD (Barch et al. 2004;McClure et al. 2008; Mitropoulou et al. 2002, 2005), and asubset of the controls was included in prior [11C]NNC112PET publications (Abi-Dargham et al. 2002, 2012; Poelset al. 2013; Slifstein et al. 2008). Screening included a phys-ical examination, basic blood tests, and electrocardiogram. Allparticipants were between the ages of 18 and 60, medicallyhealthy, not pregnant or nursing, and had no recent recreation-al drug use (confirmed by urine toxicology tests). All SPDparticipants met criteria for SPD according to the Diagnosticand Statistical Manual of Mental Disorders, 4th edition(DSM-IV) (APA 1994), were not currently taking psychotro-pic medications (see below), and were excluded if they metcriteria for a lifetime diagnosis of schizophrenia or bipolardisorder (I or II), current major depressive disorder, or current/recent (past 6 months) substance abuse or dependence.Controls were free of any current or past psychiatric condi-tions, including substance abuse (nicotine dependence waspermitted in both groups). Diagnoses for the SPD participantswere assessed with the Structured Clinical Interview forDSM-IV Axis I Disorders (SCID-IV; First et al. 1994)and Structured Interview for DSM-IV PersonalityDisorders (SIDP-IV; Pfohl et al. 1997); for controls, the ab-sence of psychiatric conditions was confirmed with an abbre-viated version of the SCID-IVor the Diagnostic Interview forGenetic Studies (DIGS; Nurnberger et al. 1994).

Cognitive assessments

The SPD participants were administered the N-back (Cohenet al. 1997) and Paced Auditory Serial Addition Test (PASAT;Gronwall 1977), both measures of WM. The N-back requiresparticipants to monitor a series of letters presented sequential-ly on a computer screen and to respond whenever apredetermined target letter is presented (0-back), or when theletter presented is identical to either the letter that immediatelypreceded it (1-back) or to the letter 2 trials back (2-back); 3blocks of 25 trials each were administered for each condition.The outcome used was hit rate. Here, we report results fromthe 2-back. The PASAT requires participants to listen to arecorded voice present a series of single-digit numbers(50 at a rate of 1 per 2 s) and to add together each adjacent pairand to state the sum aloud; the outcome used was the numberof correct responses. Of note, the SPD participantswere administered the WM measures through a separate studythat partially overlapped in time with the PET study. As aconsequence, the time period between the WM assessmentsand PET was variable across participants; in some cases, morethan a year separated these assessments and PET (see below).

PET acquisition and input function measurement

Scans were acquired in 3D mode with an ECAT EXACTHR+scanner (Siemens/CTI, Knoxville, TN). After a 10-mintransmission scan for attenuation correction, [11C]NNC112was injected intravenously as a bolus over 45 s. Emission datawere collected for 90 min. See the Online Resource text foradditional information regarding radiochemistry and PET ac-quisition. Arterial plasma input function measurements werecollected for kinetic analysis and estimation of the plasma freefraction (fp, fraction of [11C]NNC112 in arterial plasma notbound to protein) as previously described (Abi-Dargham et al.2000, 2012).

Image analysis

All participants received a high-resolution T1-weighted MRIscan. PET reconstruction and coregistration of PET to MRIimages and delineation of regions of interest (ROIs) wereperformed as previously described (Abi-Dargham et al.2012) to generate time activity curves in each ROI. Theprimary ROIs examined were regions of the PFC: DLPFC,medial PFC (MPFC), and orbitofrontal cortex (OFC), as de-fined in Abi-Dargham et al. (2000). We also examined thestriatal subregions, which included five ROIs (Table 2) andwere drawn as defined inMawlawi et al. (2001). A graymattermask was applied to cortical ROIs so that only the gray matteractivity was measured in those regions (Abi-Dargham et al.2002). The cerebellum (CER) was included as a referenceregion.

Receptor parameter estimation

PET data were analyzed with kinetic modeling and arterialplasma input using 2-tissue compartment modeling (2TC) inROIs and 1TC in the CER to derive the total distributionvolume (VT) for each region (Online Resource text). VT isequivalent to the equilibrium ratio of the total concentration ofradioligand in that region to the arterial plasma concentration.Because there is a negligible concentration of D1R in CER(Hall et al. 1994), CER VT was used to estimate VND, whichreflects the concentration of free and nonspecifically-boundtracer, and is assumed to be equal across the brain regions. Asin our previous studies with patients with schizophrenia (Abi-Dargham et al. 2002, 2012), we aimed to report the bindingpotential measure BPP as our primary PET outcome measure,estimated as: BPP=VT (ROI)−VT (CER). Due to significantgroup differences observed in fp (see below), however, we alsopresent results for BPF and BPND. BPF is defined as: BPF=BPP / fp; thus, its calculation accounts for fp. Given the groupdifference in fp, we consider BPF the primary PET outcomemeasure for group comparisons. BPND is defined as: BPND=BPP / VT (CER); this measure was included with the aim of

Psychopharmacology

corroborating the results obtained with BPF. Due to significantgroup differences in some of the PFC ROI volumes, PET datafor cortical ROIs were adjusted to correct for partial volumeeffects (PVE). The striatal ROI data were not PVE corrected,as there were no significant group differences in striatal sub-region volumes. See Online Resource text for additional in-formation regarding the PET outcome measures and PVEcorrection.

Statistical analysis

Group comparisons of the PEToutcome measures BPF, BPND,and BPP were performed using linear mixed modeling forwhich diagnosis (SPD vs control) and ROI were treated asfixed effects, and ROI was treated as a repeated measure; thediagnosis×ROI term was included in all models. For primaryanalyses, a model using “PFC” (comprising DLPFC, MPFC,and OFC) as the ROI was analyzed for each BP outcomemeasure. Between-group planned comparisons were also con-ducted for each ROI, and Cohen’s d (using the observed groupmeans and pooled SD) was used as an index of effect size. Forthe SPD participants, correlations of BP outcome measures(BPF, BPND, and BPP) with the two WM measures wereexamined for each of the three regions of PFC usingSpearman’s rank correlations, due to non-normal distributionsobserved for several variables (18 bivariate correlations intotal). Analyses for primary hypotheses (i.e., group

comparisons of BP in the PFC and the correlations betweenPFC BP and WM) were not corrected for multiple compari-sons. For exploratory analyses, we examined group differ-ences in the “striatal” ROI (comprising the five striatal subre-gions) with mixed modeling as described above and indepen-dent sample t tests for the composite “whole striatum”; thefalse discovery rate procedure (Benjamini and Hochberg1995) was used to correct for multiple comparisons for ex-ploratory analyses. Because age was significantly related toBPF in the DLPFC in the SPD participants (see below), weconducted secondary analyses including age as a covariatewhen examining BP group differences and BP-WM relations(using partial Spearman’s rank correlations, prs, for the latter).Mixed models were also run when including age×diagnosis;because this term was nonsignificant in all cases, it wasdropped from the analyses. All statistical tests were two tailed.

Results

Demographic and clinical characteristics

Eighteen unmedicated SPD and 21 control participants wereincluded in this study. There were no significant group differ-ences in age, sex, ethnicity, smoking status, or participant orparental years of education (Table 1). Twelve of the 18 SPDparticipants were naive to psychotropic medication (and 16 to

Table 1 Demographic characteristics and scan parameters by diagnostic group

Parameter Controls (n=21) SPD participants (n=18) pa

Demographic (mean±SD unless otherwise noted)

Age 40.76±5.80 42.30±7.42 0.473

Sex (% male) 67 % 67 % 1.00

Ethnicity (%) 38 % C, 48 % AA, 14 % H 39 % C, 22 % AA, 39 % H 0.959b

Smoking status (% yes) 29 % 29 %c 0.955

Parental education (years) 14.65±3.22d 13.40±4.07d 0.317

Participant education (years) 15.20±2.09e 14.33±2.91 0.295

Scan parameters (mean±SD)

Injected dose (mCi) 13.24±3.70 11.01±2.93 0.046

Injected mass (μg) 4.91±1.08 4.32±1.48 0.166

Specific activity (Ci/mmol) 921.86±285.72 985.00±553.65 0.650

Plasma free fraction (fp, %) 0.89±0.26f 0.73±0.15 0.024

VND (mL/cm3) 2.25±0.48 2.17±0.46 0.614

AAAfrican American,CCaucasian,HHispanic, SD standard deviation, SPD schizotypal personality disorder, VND nondisplaceable distribution volume(measured in the cerebellum to provide a measure of nonspecific radioligand binding)a For continuous variables, independent sample t test; for dichotomous variables, chi-square tests usedbGroups were compared on a dichotomized ethnicity variable (Caucasian vs ethnic minority)c Smoking status was missing for one SPD participantd Parental education was missing for three SPD participants and one controle Participant education was missing for one controlf Plasma free fraction was missing for one control

Psychopharmacology

antipsychotics), and the remaining had not taken psychotropicmedication for at least 3 weeks prior to PET (with mostreporting no medication use for the last several years). Thetwo SPD participants with previous antipsychotic use reportedlast using antipsychotics over 20 years prior to study partici-pation. Comorbid Axis-I and Axis-II diagnoses of the SPDparticipants are provided in the Online Resource text; note thatsome of the SPD participants met criteria for past substance-use disorders (see Online Resource text).

Scan parameters and regional volumes

Scan parameter data are provided in Table 1. There were nosignificant group differences in injected mass or specific ac-tivity of [11C]NNC112 or in VND. The SPD participants re-ceived significantly less injected radioactivity of[11C]NNC112; however, this was not problematic for theanalyses because injected mass was well within tracer-doselimits for all scans. The mean plasma free fraction (fp) of[11C]NNC112 was significantly smaller in the SPD than inthe control group. The source of this fp group difference is notclear. For example, fp was not significantly related to partici-pant body mass index (in the total sample or either groupseparately) as was observed in one of our previous samples(Kegeles et al. 2010). As detailed above, this fp group differ-ence posed a problem for relying on BPP as our primary BPoutcome measure; therefore, we also report BPF and BPND.Group comparison results of ROI volumes are presented inTable 2.

Timing of cognitive assessments relative to PET scans

As noted above, the timing of WM assessments relative toPET was variable across the SPD participants. The mediannumber of weeks separating PET and administration of thePASAT and N-back, respectively, was 45.86 (range=2.29–144) and 52 (range=4.43–204.57).

Associations of age with binding potential and WMmeasures

When examining associations of age with BP in the PFCROIsand with WM performance in the SPD participants, age wassignificantly related only to DLPFC BPF in the SPD partici-pants (rs=0.494, p=0.037). In contrast, age was significantlynegatively related to BPND in several striatal subregions in thetotal sample and in both groups; see Online Resource text.

Group differences in [11C]NNC112 VT, BPF, BPND, and BPP

PFC There were no significant group differences in VT for thePFC ROIs (Online Resource Table 1). In mixed model anal-yses with the PFC ROI, there was not a significant effect ofdiagnosis in predicting BPF [F(1, 36.01)=1.03, p=0.316, n=20 for controls] or BPND [F(1, 37.05)=0.59, p=0.446].Scatterplots of the BPF values by group are provided inOnline Resource Fig. 1. Between-group planned comparisonsof the predicted marginal means derived from the mixedmodels confirmed the lack of significant group differencesin BPF (Table 3) and BPND (Table 4). Mixed modelingresults with the PFC ROI and BPP were consistent in indicat-ing no significant effect of diagnosis in the PFC [F(1, 37.02)=0.46, p=0.504]. Results were essentially unchanged whenincluding age as a covariate in these PFC models.

Striatal subregions There were no significant group differ-ences in striatal VT (Online Resource Table 1). Mixed model-ing with the striatal ROI indicated a trend effect for diagnosisin predicting BPF [F(1, 36)=3.08, p=0.088, n=20 for con-trols]. Group comparisons indicated that the SPD participantshad higher BPF in the ventral striatum (VST) compared tocontrols (p=0.025; Table 3, Online Resource Fig. 1d); theeffect size for this difference was in the medium to large range(0.763), but the difference was not considered significant aftermultiple comparisons correction. There were no significantdifferences in BPF in the other striatal subregions or wholestriatum. For BPND, there was not a significant effect ofdiagnosis [F(1, 37)=0.29, p=0.595]. Group comparisons in-dicated that the SPD participants had higher BPND in the VSTcompared to controls at trend level (p=0.064; Table 4); theeffect size for this difference was in the medium range (0.612).There were no significant differences in other striatal subre-gions or whole striatum. For BPP, the diagnosis term was notsignificant [F(1, 37)=0.03, p=0.861], nor were the group

Table 2 Volumes (mm3) of the regions of interest by diagnostic group

ROI Controls (n=21) SPD participants (n=18) pa

PFC

DLPFC 25,037±5,708 20,116±5,531 0.010

MPFC 5,546±2,290 4,204±2,059 0.064

OFC 11,821±5,356 10,709±4,104 0.477

Striatal

VST 1,638±428 1,600±427 0.781

pre-DCA 4,598±749 4,245±704 0.140

pre-DPU 3,831±591 3,781±535 0.785

post-CA 997±301 846±266 0.109

post-PU 4,102±689 3,911±512 0.337

Mean±standard deviation

DLPFC dorsolateral prefrontal cortex, MPFC medial prefrontal cortex,OFC orbitofrontal cortex, PFC prefrontal cortex, post-CApostcommissural caudate, post-PU postcommissural putamen, pre-DCAprecommissural dorsal caudate, pre-DPU precommissural dorsal puta-men, ROI region of interest, SPD schizotypal personality disorder, VSTventral striatuma Independent sample t test

Psychopharmacology

differences in the striatal subregions or whole striatum. SeeOnline Resource text for effect sizes related to striatal groupcomparisons. Results were nearly identical when includingage in these striatal models (Online Resource text).

Correlations between WM performance and [11C]NNC112BP in the PFC

The Online Resource text provides the descriptive data of thePASAT and 2-back performance of the SPD participants.PASAT performance was significantly negatively related to[11C]NNC112 BPF and BPP in theMPFC (rs=−0.551, p=.022and rs=−0.488, p=.047, respectively; n=17; Fig. 1), and wasnegatively related at trend level to BPF and BPP in the DLPFC(rs=−0.460, p=.055 and rs=−0.421, p=.082, respectively;Online Resource Fig. 2). These results indicate that SPDparticipants with poorer PASAT performance tended to havehigher BPF and BPP in these prefrontal regions. There were noother significant associations between the PASAT and BP inthe DLPFC, MPFC, or OFC. When controlling for age,PASAT performance was related at trend level to BPF in theMPFC (prs=−0.474, p=.064) and remained significantly

associated with BPP in the MPFC (prs=−0.498, p=.050); thetrend level associations with BP in the DLPFC became non-significant, but their strength remained moderate (prs=−0.354, p=.164 and prs=−0.357, p=.160 for BPF and BPP,respectively). In contrast to the PASAT, performance on the 2-back (n=16) was not significantly related to BP in any regionof the PFC; these results remained unchanged when control-ling for age. See Online Resource Tables 2 and 3 for acomplete list of the BP-WM correlational results.

Discussion

To our knowledge, this is the first report of D1R availabilityand related WM performance in SPD. In contrast to ourhypothesis and previous findings in schizophrenia (Abi-Dargham et al. 2002, 2012), we found no evidence for signif-icant group differences between SPD and control participantsin D1R availability in the PFC as measured with[11C]NNC112 PET. As predicted, however, prefrontal D1Ravailability, particularly in the MPFC, was significantly neg-atively related toWMperformance in SPD asmeasured by thePASAT, indicating that the SPD participants who performedmore poorly on the PASAT tended to have higher levels of

Table 3 [11C]NNC112 BPF by diagnostic group for PFCa and striatalROIs

ROI Controls (n=20b) SPD (n=18) pc

PFC

DLPFC 357.61±109.66 408.58±187.08 0.306

MPFC 373.28±129.22 414.40±179.27d 0.466

OFC 377.23±160.76 447.50±207.56 0.248

Striatal

VST 663.41±219.96 877.40±335.61 0.025

pre-DCA 762.28±237.42 928.88±359.64 0.097

pre-DPU 813.18±266.70 990.80±380.12 0.101

post-CA 584.59±196.56 698.07±309.58 0.181

post-PU 781.91±265.39 939.80±363.83 0.132

whole STR 755.49±241.89 927.82±358.57 0.088

Mean±standard deviation (observed)

DLPFC dorsolateral prefrontal cortex, MPFC medial prefrontal cortex,OFC orbitofrontal cortex, PFC prefrontal cortex, post-CApostcommissural caudate, post-PU postcommissural putamen, pre-DCAprecommissural dorsal caudate, pre-DPU precommissural dorsal puta-men, ROI region of interest, SPD schizotypal personality disorder,STR striatum, VST ventral striatumaBPF values for all PFC (but not striatal) ROIs were adjusted to correctfor partial volume effects (PVE); see Methodsb n=20: BPF values were missing for one control due to unavailability offp termcGroup comparisons of the estimatedmarginal means derived from linearmixed modeling (see text) with the exception of the composite wholeSTR; for this region, an independent sample t test was usedd n=17: The value for one SPD participant was excluded due to excessivenoise in the VT measurement

Table 4 [11C]NNC112 BPND by diagnostic group for PFCa and striatalROIs

ROI Controls (n=21) SPD (n=18) pb

PFC

DLPFC 1.35±0.22 1.28±0.27 0.369

MPFC 1.40±0.24 1.28±0.25c 0.172

OFC 1.40±0.34 1.39±0.30 0.960

Striatal

VST 2.50±0.49 2.78±0.44 0.064

pre-DCA 2.90±0.59 2.94±0.45 0.792

pre-DPU 3.08±0.65 3.14±0.53 0.772

post-CA 2.19±0.41 2.20±0.48 0.969

post-PU 2.96±0.60 2.98±0.50 0.885

whole STR 2.86±0.54 2.94±0.46 0.618

Mean±standard deviation (observed)

DLPFC dorsolateral prefrontal cortex, MPFC medial prefrontal cortex,OFC orbitofrontal cortex, PFC prefrontal cortex, post-CApostcommissural caudate, post-PU postcommissural putamen, pre-DCAprecommissural dorsal caudate, pre-DPU precommissural dorsal puta-men, ROI region of interest, SPD schizotypal personality disorder,STR striatum, VST ventral striatumaBPND values for all PFC (but not striatal) ROIs were adjusted to correctfor partial volume effects (PVE); see MethodsbGroup comparisons of the estimatedmarginal means derived from linearmixed modeling (see text) with the exception of the composite wholeSTR; for this region, an independent sample t test was usedc n=17: The value for one SPD participant was excluded due to excessivenoise in the VT measurement

Psychopharmacology

prefrontal D1. However, associations between prefrontal D1and 2-back performance were not significant; thus, our secondhypothesis was only partially supported. Results from explor-atory analyses suggest that SPD may be characterized bygreater D1R availability in the VST compared to that ofcontrols. These group differences did not survive multiplecomparisons correction but were of medium to large effect.

The significant negative associations we observed betweenprefrontal D1R availability and PASAT performance are sim-ilar to one of our previous findings in schizophrenia (Abi-Dargham et al. 2002); such results may reflect that individualswith SPDwho have impairment in the processes tapped by thePASAT (i.e., WM, attention) are characterized by prefrontalDA dysfunction that, similar to what has been proposed tooccur in schizophrenia (Abi-Dargham et al. 2002, 2012),results in understimulation of D1R and a compensatory butinefficient upregulation of these receptors. However, this in-terpretation is tentative given that we did not find evidence ofsignificantly higher prefrontal D1R availability in SPD. It maybe that compensatory alterations in D1 occur in a subset ofSPD participants with particular WM deficits but to a relative-ly minor degree that does not result in detectable elevations ofD1 at the group level, and/or that there are other processesoccurring in SPD that contribute to variance in D1 levels. Ofnote, preliminary findings from our group (Rosell et al. 2013)indicate that in SPD (N=16), 3 days of treatment with thepotent D1R agonist DAR-0100A significantly improvedPASAT performance compared to placebo; these results areconsistent with the current findings in suggesting a link be-tween impaired WM performance and alterations in D1Rstimulation by DA in SPD.

The inconsistent results we observed between the PASATand 2-back with regard to associations with BP are worthnoting. The lack of a significant association between prefron-tal D1 and 2-back performance contrasts with the findingsfrom one of our previous studies of schizophrenia (Abi-Dargham et al. 2002) in which higher DLPFC D1R availabil-ity was associated with worse performance on both the 2- and

3-back. Research on cognitive functioning in SPD has indi-cated that this spectrum condition is associated with cognitiveimpairment across various domains that is similar to buttypically less severe than what is found in schizophrenia(McClure et al. 2008; Siever and Davis 2004). Thus, althoughspeculative, it is possible that in SPD, associations betweencognitive performance and potential neurobiological corre-lates such as D1R are only clearly evident when using taskswith heavier cognitive demands than those of the 2-back. Ofnote, we did not obtain data with the more difficult 3-back forthis study. Relatedly, we previously found that compared toseveral other cognitive measures, including a visuospatialWM task, the PASATwas the most effective at discriminatingSPD and control participants (Mitropoulou et al. 2005). Thesefindings, along with our current results, suggest that thePASAT may be particularly useful in tapping the cognitiveprocesses that are impaired in the schizophrenia spectrum andthat are especially dependent on the appropriate level ofprefrontal D1R stimulation in SPD.

Exploratory analyses indicated that SPD may be character-ized by greater D1R availability in the VST compared tocontrols. Of interest is a similar trend level finding fromAbi-Dargham et al. (2012), suggesting that antipsychotic-naive patients with schizophrenia may have elevated VSTD1R availability. Thus, it may be worth investigating suchpotential alterations in D1R with preplanned comparisons infuture studies, particularly given this brain region’s putativerole in motivational processes and negative symptoms.

The limitations of this study must be considered wheninterpreting its findings. One issue relates to the lack ofselectivity in vivo of [11C]NNC112 for cortical D1R versusserotonin-2A receptors (5-HT2AR), such that ~25–30 % ofthe binding in the cortex of healthy controls is to 5-HT2AR(Slifstein et al. 2007). This lack of selectivity for cortical D1also characterizes the other currently used D1 tracer,[11C]SCH23390 (Ekelund et al. 2007). In the striatum, how-ever, where 5-HT2A density is negligible compared to D1,most of the binding of [11C]NNC112 is attributable to D1

Fig. 1 a The relation between[11C]NNC112 BPF in the medialprefrontal cortex (MPFC) andperformance on the PacedAuditory Serial Addition Test(PASAT) in participants withschizotypal personality disorder(SPD), n=17; rs=−0.551,p=.022. b The relation betweenBPP in the MPFC and PASATperformance in SPD participants,n=17; rs=−0.488, p=.047

Psychopharmacology

(Slifstein et al. 2007; see Abi-Dargham et al. 2012). Thus, ifSPD is associated with alterations in cortical 5-HT2AR avail-ability, our characterization of SPD-control group differencesin prefrontal D1 levels may have been compromised. We arenot aware of any studies of 5-HT2AR in SPD per se. PETstudies of cortical 5-HT2A in schizophrenia have yieldedmixed results; however, recent findings with antipsychotic-naive first-episode patients provide evidence that this disordermay be associated with lower cortical 5-HT2AR levels com-pared to controls (Rasmussen et al. 2010). In addition, recentwork of Hurlemann et al. (2008) suggests that individuals atincreased clinical risk for psychosis may have lower cortical5-HT2AR availability relative to controls. Given that it isunclear how this limitation of [11C]NNC112 influenced thecurrent results, future investigations using a more selectivetracer, when available, would provide a more definitive char-acterization of prefrontal D1R and its relation to WM perfor-mance in SPD.

Another limitation is that we did not include group com-parisons on catechol-O-methyltransferase (COMT) genotypein this report. Although efforts to collect genetic samples fromparticipants were initiated prior to the completion of this study,due to the lack of samples among those who completed PETprior to the initiation of these efforts, as well as issues offeasibility in collecting samples among some of the latersubjects, ultimately we did not have genotyping data on anadequate number of participants to allow for this. Given thefindings that healthy controls with the COMT genotype Val/Val have significantly higher D1R availability in cortical (butnot striatal) regions as measured by [11C]NNC112 comparedto Met carriers (Slifstein et al. 2008), characterizing the cur-rent sample on COMT genotype would have provided valu-able information regarding potential factors influencing pre-frontal D1 in both groups. Although the question of potentialgroup differences in COMT genotype cannot be addressedwith the current sample, Minzenberg et al. (2006) found thatdistribution in COMT Val/Met genotype did not differ signif-icantly between SPD and control participants.

In addition, we did not systematically collect informationregarding the quantity of tobacco used by the participants whoreported current regular tobacco use; thus, although groupswere matched on smoking status, we could not examinewhether the quantity of tobacco used differed by group. Wealso did not collect an estimate of quantity of past recreationaldrug use of participants; this is particularly relevant for theSPD participants, as several met criteria for a past substance-use disorder (see Online Resource text). As a result, we couldnot examine potential associations between quantity of tobac-co or past drug use and indices of D1R availability or WMperformance. Relatedly, given that the SPD and controlgroups were not matched on past drug use, the potentialinfluence of this factor on the current group comparisonresults is unclear. Thus, additional research that accounts for

past drug use will advance attempts to characterize DA func-tioning and its correlates in SPD.

A limitation with regard to our correlational analyses be-tween D1 and WM performance is that the time period be-tween the WM assessments and PETwas variable across SPDparticipants and, in some cases, spanned more than a year (seeResults). Given that presumably both state and trait factorsinfluence our PETand behavioral measurements, it is possiblethat these variable time intervals influenced our correlationalresults, in particular weakening our ability to detect true cross-sectional associations. We acknowledge this limitation butbelieve it is unlikely that this issue led to significant associa-tions that would not have been detected if standardized andshorter time intervals separated theWMassessments and PET.Of note, previous research has indicated high test-retest reli-ability for the PASAT, including long-term reliability (e.g., rs=0.96, Sjogren et al. 2000; for review, Tombaugh 2006). Thetest-retest reliability of the 2-back is generally considered fairto good (Nuechterlein et al. 2008), although we are not awareof any reports of the long-term reliability of the 2-back per se.It is possible that this limitation contributed to the inconsistentresults obtained between the PASATand 2-back with regard toassociations with D1R.

We conducted multiple correlational analyses to test ourhypothesis that prefrontal D1R availability was related toWMperformance in SPD (see Methods) without correcting therelated p values for this multiple testing. Our rationale fornot using such a correction was that these tests were hypoth-esis driven and that the relatively small size of our SPD samplewould require larger statistical test values to achieve signifi-cance. It should be noted that the magnitude of the significantcorrelations observed was in the moderate to large range. Onthe other hand, due to the small sample size and lack ofcorrection for multiple tests, these results should be consid-ered preliminary and require replication.

In addition, we assessedWM performance among the SPDparticipants of this study, but not among the controls. Thus,although in prior studies we have reported impaired WMperformance in individuals with SPD compared to healthycontrols on both the PASAT (Mitropoulou et al. 2002, 2005)and 2-back (McClure et al. 2008), we cannot draw conclu-sions regarding WM dysfunction in this SPD sample relativeto their matched controls. Relatedly, we were not able toexamine the same associations between prefrontal D1R andWM performance among controls that were examined inSPD; thus, we cannot draw conclusions regarding the poten-tial specificity of the observed associations to SPD.

Another issue relates to the PET outcome measures wereport here. In our prior [11C]NNC112 studies, we have in-cluded BPP and BPND, with BPP as our primary outcome.However, given that in the present study, there was a signif-icant group difference in the mean plasma free fraction ofradiotracer (fp), we also included BPF, as its calculation

Psychopharmacology

accounts for fp (by dividing BPP by fp).We consider BPF as theprimary PET outcome for group comparisons. Given this,there is increased confidence in the group difference wedetected in BP in the VST because it was found with BPFand corroborated at trend level by BPND. In contrast, correla-tional analyses with BPP were not compromised by the fpgroup difference, given that these only included the SPDparticipants. This lends credence to the correlational resultswe obtained with BPP, which were consistent with those foundwith BPF.

In summary, we found only limited support for our hypoth-eses positing that the same prefrontal D1 alterations andassociations with WM performance previously found inschizophrenia (Abi-Dargham et al. 2002, 2012) would char-acterize individuals with SPD. In contrast to what has beenobserved in schizophrenia, we found no evidence for prefron-tal D1R upregulation in SPD. Similar to previous schizophre-nia findings, however, the individuals with SPD whodisplayed poorer WM performance as measured by thePASAT tended to have higher D1R availability in the PFC.These findings combine to suggest that schizophrenia andSPD may share a common pathophysiological feature relatedto prefrontal DA functioning that contributes to the WMdysfunction that is present in both conditions, but that inSPD, compensatory alterations in D1 may occur only in thesubset of individuals who are the most severely impaired withregard to WM and/or occur to an extent that is minor relativeto the alterations that characterize schizophrenia. Such differ-ences may reflect that SPD is characterized by a less severeform of cortical DA pathology compared to schizophrenia, orthat there are other system-level adaptations in SPD thatcompensate for cortical DA dysfunction that do not occur inthe full syndrome of schizophrenia. Future well-powered in-vestigations including both schizophrenia and SPD samples,as well as those selected based on particularWM impairments,would be valuable in further elucidating the cortical DAalterations and their relations to functioning across the schizo-phrenia spectrum. In addition, studies more directly assessingpresynaptic cortical DA functioning, such as with the amphet-amine challenge paradigm and the high-affinity D2Rradioligand [11C]FLB 457, should provide valuable insightwith regard to cortical DA functioning in this condition.

Acknowledgments We thank the research participants of this study andexpress gratitude for the expert assistance of Rawad Ayoub, Jennifer Bae,Felipe Castillo, John Castrillon, Elizabeth Hackett, Elisabeth Iskander,Olga Kambalov, and Ethan Rothstein. This work was supported by aMerit Review grant (7609-028) from the US Department of VeteransAffairs to Larry J. Siever, by grant MH56140 from the National Instituteof Mental Health to Larry J. Siever, and by the Veterans Affairs VISN 3Mental Illness Research, Education and Clinical Center (MIRECC).

Conflict of interest The authors declare no conflicts of interest; how-ever, the following coauthors have appointments at the James J. PetersVeterans Affairs Medical Center, Bronx, NY (in addition to the Mount

Sinai School of Medicine, New York, NY): Daniel R. Rosell, YosefaEhrlich, Erin A. Hazlett, and Larry J. Siever. In addition, some coauthorshave financial relationships with pharmaceutical companies.Specifically, Mark Slifstein serves as a consultant for Amgen and receivesresearch support from Pierre Fabre; Ragy R. Girgis receives researchsupport from Otsuka; Lawrence S. Kegeles receives research supportfrom Pfizer and Amgen; and Anissa Abi-Dargham is a consultant forAmgen and Roche, and receives research support from Otsuka, Takeda,Forest, and Pierre Fabre.

References

Abi-Dargham A, Martinez D, Mawlawi O, Simpson N, Hwang DR,Slifstein M, Anjilvel S, Pidcock J, Guo NN, Lombardo I, Mann JJ,Van Heertum R, Foged C, Halldin C, Laruelle M (2000)Measurement of striatal and extrastriatal dopamine D1 receptorbinding potential with [11C]NNC 112 in humans: validation andreproducibility. J Cereb Blood Flow Metab 20:225–243

Abi-Dargham A, Mawlawi O, Lombardo I, Gil R, Martinez D, Huang Y,Hwang DR, Keilp J, Kochan L, Van Heertum R, Gorman JM,Laruelle M (2002) Prefrontal dopamine D1 receptors and workingmemory in schizophrenia. J Neurosci 22:3708–3719

Abi-Dargham A, Kegeles LS, Zea-Ponce Y, Mawlawi O, Martinez D,Mitropoulou V, O’Flynn K, Koenigsberg HW, Van Heertum R,Cooper T, Laruelle M, Siever LJ (2004) Striatal amphetamine-induced dopamine release in patients with schizotypal personalitydisorder studied with single photon emission computed tomographyand [123I]iodobenzamide. Biol Psychiatry 55:1001–1006

Abi-Dargham A, Xu X, Thompson JL, Gil R, Kegeles LS, Urban N,Narendran R, Hwang DR, Laruelle M, Slifstein M (2012) Increasedprefrontal cortical D(1) receptors in drug naive patients with schizo-phrenia: a PET study with [(11)C]NNC112. J Psychopharmacol 26:794–805

APA (1994) Diagnostic and Statistical Manual of Mental Disorders, 4thedn. American Psychiatric Association, Washington, DC

Arnsten AF, Cai JX, Murphy BL, Goldman-Rakic PS (1994) DopamineD1 receptor mechanisms in the cognitive performance of youngadult and aged monkeys. Psychopharmacology (Berl) 116:143–151

Barch DM, Mitropoulou V, Harvey PD, New AS, Silverman JM, SieverLJ (2004) Context-processing deficits in schizotypal personalitydisorder. J Abnorm Psychol 113:556–568

Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: apractical and powerful approach to multiple testing. J Royal Stat SocSer B 57:289–300

Cohen JD, Perlstein WM, Braver TS, Nystrom LE, Noll DC, Jonides J,Smith EE (1997) Temporal dynamics of brain activation during aworking memory task. Nature 386:604–608

Ekelund J, Slifstein M, Narendran R, Guillin O, Belani H, Guo NN,Hwang Y, Hwang DR, Abi-Dargham A, Laruelle M (2007) In vivoDA D(1) receptor selectivity of NNC 112 and SCH 23390. MolImaging Biol 9:117–125

First MB, Spitzer RL, Gibbon M, Williams JBW (1994) StructuredClinical Interview for DSM-IV Axis-I Disorders, Version 2.0.Biometrics Research New York State Psychiatric Institute, NewYork

Fischer H, Nyberg L, Karlsson S, Karlsson P, Brehmer Y, Rieckmann A,MacDonald SW, Farde L, Backman L (2010) Simulatingneurocognitive aging: effects of a dopaminergic antagonist on brainactivity during working memory. Biol Psychiatry 67:575–580

Green MF (2006) Cognitive impairment and functional outcome inschizophrenia and bipolar disorder. J Clin Psychiatry 67(Suppl 9):3–8

Psychopharmacology

Gronwall DM (1977) Paced auditory serial-addition task: a measure ofrecovery from concussion. Percept Mot Skills 44:367–373

Guo N, Hwang DR, Lo ES, Huang YY, Laruelle M, Abi-Dargham A(2003) Dopamine depletion and in vivo binding of PET D1 receptorradioligands: implications for imaging studies in schizophrenia.Neuropsychopharmacology 28:1703–1711

Hall H, Sedvall G, Magnusson O, Kopp J, Halldin C, Farde L (1994)Distribution of D1- and D2-dopamine receptors, and dopamine andits metabolites in the human brain. Neuropsychopharmacology 11:245–256

Hurlemann R, Matusch A, Kuhn KU, Berning J, Elmenhorst D, Winz O,Kolsch H, Zilles K, Wagner M, Maier W, Bauer A (2008) 5-HT2Areceptor density is decreased in the at-risk mental state.Psychopharmacology (Berl) 195:579–590

Kegeles LS, Slifstein M, Xu X, Urban N, Thompson JL, Moadel T,Harkavy-Friedman JM, Gil R, Laruelle M, Abi-Dargham A (2010)Striatal and extrastriatal dopamine D2/D3 receptors in schizophreniaevaluated with [18 F]fallypride positron emission tomography. BiolPsychiatry 68:634–641

Kosaka J, Takahashi H, Ito H, Takano A, Fujimura Y, Matsumoto R,Nozaki S, Yasuno F, Okubo Y, Kishimoto T, Suhara T (2010)Decreased binding of [11C]NNC112 and [11C]SCH23390 in pa-tients with chronic schizophrenia. Life Sci 86:814–818

Mawlawi O, Martinez D, Slifstein M, Broft A, Chatterjee R, Hwang DR,Huang Y, Simpson N, Ngo K, Van Heertum R, Laruelle M (2001)Imaging human mesolimbic dopamine transmission with positronemission tomography: I. Accuracy and precision of D2 receptorparameter measurements in ventral striatum. J Cereb Blood FlowMetab 21:1034–1057

McClure MM, Barch DM, Flory JD, Harvey PD, Siever LJ (2008)Context processing in schizotypal personality disorder: evidence ofspecificity of impairment to the schizophrenia spectrum. J AbnormPsychol 117:342–354

McClure MM, Harvey PD, Goodman M, Triebwasser J, New A,Koenigsberg HW, Sprung LJ, Flory JD, Siever LJ (2010)Pergolide treatment of cognitive deficits associated with schizotypalpersonality disorder: continued evidence of the importance of thedopamine sy s t em in t h e s ch i zoph r en i a s pe c t r um .Neuropsychopharmacology 35:1356–1362

Minzenberg MJ, Xu K, Mitropoulou V, Harvey PD, Finch T, Flory JD,New AS, Goldman D, Siever LJ (2006) Catechol-O-methyltransferase Val158Met genotype variation is associated withprefrontal-dependent task performance in schizotypal personalitydisorder patients and comparison groups. Psychiatr Genet 16:117–124

Mitropoulou V, Harvey PD, Maldari LA, Moriarty PJ, New AS,Silverman JM, Siever LJ (2002) Neuropsychological performancein schizotypal personality disorder: evidence regarding diagnosticspecificity. Biol Psychiatry 52:1175–1182

Mitropoulou V, Harvey PD, Zegarelli G, New AS, Silverman JM, SieverLJ (2005) Neuropsychological performance in schizotypal person-ality disorder: importance of working memory. Am J Psychiatry162:1896–1903

Narendran R, Frankle WG, Keefe R, Gil R, Martinez D, Slifstein M,Kegeles LS, Talbot PS, Huang Y, Hwang DR, Khenissi L, CooperTB, Laruelle M, Abi-Dargham A (2005) Altered prefrontal

dopaminergic function in chronic recreational ketamine users. AmJ Psychiatry 162:2352–2359

Nuechterlein KH, GreenMF, Kern RS, Baade LE, Barch DM, Cohen JD,Essock S, Fenton WS, Frese FJ 3rd, Gold JM, Goldberg T, HeatonRK, Keefe RS, Kraemer H, Mesholam-Gately R, Seidman LJ,Stover E, Weinberger DR, Young AS, Zalcman S, Marder SR(2008) The MATRICS Consensus Cognitive Battery, part 1: testselection, reliability, and validity. Am J Psychiatry 165:203–213

Nurnberger JI Jr, Blehar MC, Kaufmann CA, York-Cooler C, SimpsonSG, Harkavy-Friedman J, Severe JB, Malaspina D, Reich T (1994)Diagnostic Interview for Genetic Studies. Rationale, unique fea-tures, and training. NIMH Genetics Initiative. Arch GenPsychiatry 51:849–859

Pfohl B, Blum N, Zimmerman M (1997) Structured Interview for DSM-IV Personality: SIDP-IV. American Psychiatric Publishing,Washington, DC

Poels EM, Girgis RR, Thompson JL, SlifsteinM, Abi-DarghamA (2013)In vivo binding of the dopamine-1 receptor PET tracers[(11)C]NNC112 and [(11)C]SCH23390: a comparison study inindividuals with schizophrenia. Psychopharmacology (Berl) 228:167–174

Rasmussen H, Erritzoe D, Andersen R, Ebdrup BH, Aggernaes B, OranjeB, Kalbitzer J, Madsen J, Pinborg LH, BaareW, Svarer C, Lublin H,Knudsen GM, Glenthoj B (2010) Decreased frontal serotonin2Areceptor binding in antipsychotic-naive patients with first-episodeschizophrenia. Arch Gen Psychiatry 67:9–16

Rosell DR, Zaluda LC,McClureMM, Strike KS, Barch DM, Harvey PD,Girgis RR, Lieberman JA, Siever LJ (2013) Clinical testing of a D1agonist for cognitive enhancement in the schizophrenia spectrumSociety of Biological Psychiatry 68th Annual Scientific Meeting.CA, USA, San Francisco

Sawaguchi T, Goldman-Rakic PS (1991) D1 dopamine receptors inprefrontal cortex: involvement in working memory. Science 251:947–950

Siever LJ, Davis KL (2004) The pathophysiology of schizophreniadisorders: perspectives from the spectrum. Am J Psychiatry 161:398–413

Sjogren P, Thomsen AB, Olsen AK (2000) Impaired neuropsychologicalperformance in chronic nonmalignant pain patients receiving long-term oral opioid therapy. J Pain Symptom Manag 19:100–108

Slifstein M, Kegeles LS, Gonzales R, Frankle WG, Xu X, Laruelle M,Abi-Dargham A (2007) [11C]NNC 112 selectivity for dopamine D1and serotonin 5-HT(2A) receptors: a PET study in healthy humansubjects. J Cereb Blood Flow Metab 27:1733–1741

Slifstein M, Kolachana B, Simpson EH, Tabares P, Cheng B, Duvall M,Gordon Frankle W, Weinberger DR, Laruelle M, Abi-Dargham A(2008) COMT genotype predicts cortical-limbic D1 receptor avail-ability measured with [(11)C]NNC112 and PET. Mol Psychiatry 13:821–827

Tombaugh TN (2006) A comprehensive review of the Paced AuditorySerial Addition Test (PASAT). Arch Clin Neuropsychol 21:53–76

Tsukada H, Nishiyama S, Fukumoto D, Sato K, Kakiuchi T, Domino EF(2005) Chronic NMDA antagonism impairs working memory, de-creases extracellular dopamine, and increases D1 receptor binding inprefrontal cortex of conscious monkeys. Neuropsychopharmacology30:1861–1869

Psychopharmacology