ARCHIVAL REPORT

Impaired Prefrontal-Basal Ganglia FunctionalConnectivity and Substantia Nigra Hyperactivityin Schizophrenia

Jong H. Yoon, Michael J. Minzenberg, Sherief Raouf, Mark D’Esposito, and Cameron S. Carter

Background: The theory that prefrontal cortex (PFC) dysfunction in schizophrenia leads to excess subcortical dopamine has generatedwidespread interest because it provides a parsimonious account for two core features of schizophrenia, cognitive deficits and psychosis,respectively. However, there has been limited empirical validation of this model. Moreover, the identity of the specific subcortical brainregions and circuits that may be impaired as a result of PFC dysfunction and mediate its link to psychosis in schizophrenia remainsunclear. We undertook this event-related functional magnetic resonance imaging study to test the hypothesis that PFC dysfunction isassociated with altered function of and connectivity with dopamine regulating regions of the basal ganglia.

Methods: Eighteen individuals with schizophrenia or schizoaffective disorder and 19 healthy control participants completed event-related functional magnetic resonance imaging during working memory. We conducted between-group contrasts of task-evoked,univariate activation maps to identify regions of altered function in schizophrenia. We also compared the groups on the level offunctional connectivity between a priori identified PFC and basal ganglia regions to determine if prefrontal disconnectivity in patientswas present.

Results: We observed task-evoked hyperactivity of the substantia nigra that occurred in association with prefrontal and striatalhypoactivity in the schizophrenia group. The magnitude of prefrontal functional connectivity with these dysfunctional basal gangliaregions was decreased in the schizophrenia group. Additionally, the level of nigrostriatal functional connectivity predicted the level ofpsychosis.

Conclusions: These results suggest that functional impairments of the prefrontal striatonigral circuit may be a common pathwaylinking the pathogenesis of cognitive deficits and psychosis in schizophrenia.

Key Words: Basal ganglia, fMRI, prefrontal cortex, psychosis,schizophrenia, substantia nigra

Two cornerstones of our emerging understanding of schizo-phrenia are the role of excess subcortical dopamine (DA)(1–4) and prefrontal cortex (PFC) dysfunction (5–7) in the

pathogenesis of psychosis and cognitive deficits, respectively.However, since their co-occurrence rather than the presence ofeither core symptom alone is more characteristic of schizophre-nia, the elucidation of how these symptoms are pathophysiolo-gically linked could help to uncover disease mechanisms.

One of the most influential theories of schizophrenia proposesthat PFC dysfunction leads to disinhibited DAergic activity,enhanced subcortical DA neurotransmission, and psychosis (8).Indirect support for this model comes from the demonstrationthat lesioning of the rodent PFC analogue results in increasedsubcortical DA levels (9) and the observation that cognitivedeficits usually predate psychosis onset in schizophrenia (10).A small number of in vivo schizophrenia studies have con-firmed an association between markers of PFC dysfunction orpathology and neurochemical markers of enhanced subcorticalDA function (11,12). However, many aspects of the mechanisms

From the Department of Psychiatry and Imaging Research Center (JHY,

MJM, SR, CSC), University of California Davis, Sacramento; and

Department of Psychology and Helen Wills Neuroscience Institute

(MD), University of California Berkeley, Berkeley, California.

Address correspondence to Jong H. Yoon, M.D., University of California

Davis, Department of Psychiatry and Imaging Research Center, 4701 X

St, Sacramento, CA 95817; E-mail: jhyyoon@ucdavis.edu.

Received Jul 11, 2012; revised Nov 7, 2012; accepted Nov 19, 2012.

0006-3223/$36.00http://dx.doi.org/10.1016/j.biopsych.2012.11.018

by which PFC dysfunction may lead to enhanced DA functionremain unclear.

We undertook this study to increase our understanding of thespecific brain regions and circuits mediating the hypothesizedimpaired PFC regulation of the DA system in schizophrenia. Inparticular, we wanted to determine whether specific basalganglia (BG) structures could be involved in this process. TheBG contains some of the most important DA regulatory regions.They include the midbrain nuclei, ventral tegmental area (VTA),and substantia nigra (SN), which produce and release themajority of brain DA (13), and other structures, such as thestriatum, which sends gamma-aminobutyric acidergic projectionsto, and may exert inhibitory control of, midbrain DA neurons(14,15). While a small number of studies have reported altered BGfunction (including of the midbrain) in schizophrenia (16,17),none, to our knowledge, has specifically examined the hypothesisof an association between dysfunction of DA regulating struc-tures of the BG and the PFC.

We tested this hypothesis with an event-related workingmemory (WM) functional magnetic resonance imaging (fMRI)experiment. We first mapped regions of abnormal activity inschizophrenia to determine if we could detect concomitantdysfunction in the PFC and DA regulating regions of the BG. Tomore directly test our hypothesis, we then measured prefrontalfunctional connectivity with the BG regions showing abnormalactivity. If the BG abnormalities were due to deficits in prefrontalregulation, impaired prefrontal-BG connectivity in schizophreniawould be expected. We employed WM because it is an effectivedriver of not only the PFC but also of BG function and DAsignaling (18,19). Moreover, since WM-associated PFC dysfunctionin schizophrenia is commonly observed (20), WM would be aneffective means of testing the hypothesis of an association

BIOL PSYCHIATRY 2013;74:122–129& 2013 Society of Biological Psychiatry

J.H. Yoon et al. BIOL PSYCHIATRY 2013;74:122–129 123

between PFC and BG dysfunction. We report evidence of task-evoked SN hyperactivity and striatal hypofunction in schizophre-nia occurring in the context of PFC hypofunction, as well asdiminished prefrontal connectivity with these BG regions.

Methods and Materials

SubjectsWe obtained results from 18 subjects with chronic schizo-

phrenia or schizoaffective disorder (SZ) and 19 healthy controlsubjects (C). Groups were well matched on demographic variablesexcept for lower IQ and education in patients (Table 1). TheStructured Clinical Interview for DSM-IV-TR Axis I Disordersconfirmed the diagnosis of schizophrenia or schizoaffectivedisorder in patients and excluded the presence of Axis I condi-tions in control subjects. Negative symptoms and psychosisseverity were quantified using the Scale for the Assessment ofNegative Symptoms and Scale for the Assessment of PositiveSymptoms total scores, respectively. We derived a disorganizationindex from Brief Psychiatric Rating Scale, Scale for the Assessmentof Negative Symptoms, and Scale for the Assessment of PositiveSymptoms subscores (21). Details of diagnostic and symptomquantification procedures can be found in Supplement 1. Exclu-sion criteria for all subjects were IQ �70, drug/alcohol depen-dence history or abuse in the previous 3 months, a positive urinedrug screen on test day, significant head trauma, or any magneticresonance imaging contraindication. Control subjects with psy-chotic first-degree relatives were excluded. We obtained writteninformed consent after providing a complete study description.University of California Davis Institutional Review Board approvedthis study. All subjects were paid for participation.

fMRI ParadigmSubjects completed a delayed-response face WM paradigm. It

entailed encoding a cue face presented for 1 second at thebeginning of each trial, maintaining its mental image throughoutthe 15-second delay period, and making a match discrimination

Table 1. Subject Demographics, Clinical Profile, and Behavioral

Patients (n ¼ 18) C

Mean SD M

Age (Years) 33.1 10.7 2

Gender (% Male) 66.7 5

Education (Years) 12.9 2.1 1

Parental Education (Years) 13.9 3.1 1

IQ 100.6 9.9 1

Handedness (% Right) 94.4 1

GAS 33.1 10.1

Disorganization 6.50 2.64

SANS Total 8.05 4.65

SAPS Total 4.61 3.80

BPRS Total 30.56 9.73

On Antipsychotics 18

Typical 1

Atypical 17

CPZ Equivalents 365 294

Accuracy .75 .11 .8

RT (msec) 1129 228 9

BPRS, Brief Psychiatric Rating Scale; CPZ, chlorpromazine; GAScale for the Assessment of Negative Symptoms; SAPS, Scale fo

with a probe face shown for 1 second, with a right index ormiddle finger button press when a match or nonmatch occurred,respectively. Cue-probe match probability was 50%. The intertrialinterval was 13 seconds. A total of 50 trials were presented,evenly divided across five blocks. A total of 750 fMRI volumeswere obtained for each subject.

fMRIWhole-brain functional scans (T2* echo planar images,

repetition time 2000 msec, echo time 40 msec, flip angle 901,field of view 22 cm, 4.0 mm axial slices with 3.4 mm2 in-planeresolution) were acquired on a 1.5T GE scanner (Milwaukee,Wisconsin). Preprocessing with SPM5 (http://www.fil.ion.ucl.ac.uk/spm5; The Wellcome Trust for Neuroimaging, UniversityCollege London, London, United Kingdom) included temporaland spatial realignment, normalization to the echo planarimages template, .001 Hz high-pass filtering, and spatialsmoothing (8 mm full width at half maximum Gaussian kernelfor whole-brain analysis; 2 mm kernel for subcortical analysis).In Supplement 1, we present data demonstrating the necessityof the smaller smoothing scale for detecting midbrain activity(Figure S1 in Supplement 1).

Controlling for Movement and On-Task Performance. Weexcluded two SZ and one C subject for exhibiting greater than 4mm within run movement. A multivariate analysis of variance ofthe mean scan-to-scan head movement for the six movementparameters (linear movement in x, y, z axes and rotationalmovement of roll, pitch, yaw) indicated no significant groupdifference among the 18 patients and 19 control subjectsincluded in the study, F(6,30) ¼ 1.15, p ¼ .359; Wilks’ l ¼ .813.Summary statistics for these measures along with detailsof additional methods controlling for movement can befound in Supplement 1. To ensure that all subjects includedin this study were engaged with the task, we excludedsubjects whose task performance was below 60% accuracy. Therewas one patient who displayed substandard task performancebut this subject was already excluded due to excessivemovement.

Performance on the Cognitive Task

ontrol Subjects (n ¼ 19)

ean SD t/Chi-Square p Value

8.8 7.3 1.30 .202

7.9 .30 .582

5.8 2.4 3.99 �.001

4.1 3.6 .37 .721

09.1 8.0 2.94 .006

00 1.09 .298

6 .09 3.47 .001

53 207 2.45 .020

S, Global Assessment of Symptoms; RT, reaction time; SANS,r the Assessment of Positive Symptoms.

www.sobp.org/journal

124 BIOL PSYCHIATRY 2013;74:122–129 J.H. Yoon et al.

fMRI Analysis. We conducted a slow event-related analysiswith covariates for encoding, maintenance, and response forcorrect trials and a separate set of covariates for incorrect trials.The three task phases were modeled in the following manner:encoding—a single covariate during cue presentation at thebeginning of the trial (t ¼ 0 sec); maintenance—a singlecovariate in the middle of the delay period (t ¼ 8 sec); andresponse—a single covariate during probe presentation (t ¼ 16sec). We included the first temporal derivative of covariates in theregression matrix to account for potential group differences inblood oxygen level–dependent (BOLD) signal temporal dynamicsresulting from reaction time differences or other variables.A random-effects second-level comparison between groups onthe contrast between WM phases and implicit baseline wasconducted.

BOLD Time Series. We generated time series by trial aver-aging the BOLD signal across all voxels within a functional regionof interest (ROI) for all correct trials for each subject. Signalchange was calculated by normalizing each value in the timeseries by the mean fMRI signal across the entire scan. The percentchange at time zero was then subtracted from every point in thetime series. We then averaged the time series across all subjectswithin a group.

Basal Ganglia Analysis. We localized activations with theWake Forest University Pickatlas (http://www.fmri.wfubmc.edu;Advanced NeuroScience Imaging Resesarch Lab, Wake ForestUniversity School of Medicine, Winston-Salem, North Carolina)(22). This atlas is based on the Talairach Daemon (University ofTexas, Austin, Texas), one of the most commonly useddatabases of anatomic regions (23). We overlaid anatomicmasks to confirm that an activation cluster was locatedwithin a particular region. For midbrain activations, we con-firmed SN localization by verifying that voxels were situatedwithin the anatomic SN mask and excluded from the rednucleus mask.

Functional Connectivity. We utilized the beta series correla-tion method (7,24) to quantify task epoch-specific, e.g., WMresponse, functional connectivity between brain regions.Detailed descriptions of this method can be found inSupplement 1.

Localization of Cortical ROIs from Independent Scans forFunctional Connectivity Analyses. We compared betweengroups the level of prefrontal functional connectivity with thedysfunctional BG regions identified in the univariate analysis(Figure 1). For this analysis, we identified a PFC (LocPFC) ROI froma separate 1-back face WM fMRI experiment completed by all oursubjects. We did this to avoid ROI localization biasing thefunctional connectivity analyses. We refer to these as the LocPFCROIs to distinguish them from the PFC region identified from thedelayed-response WM data and to emphasize their independentlocalization. We also identified the fusiform face area (LocFFA), avisual area specialized for face processing, from the 1-backface WM scan. The LocFFA served as a control region with whichwe tested the specificity of the prefrontal connectivity results.See Supplement 1 for a detailed explanation of our methodsand the rationale for the localization of ROIs fromindependent data.

Statistical Testing. Significance controlling for multiple com-parisons was determined at the cluster level of p � .05, using aninitial threshold of t ¼ 2.5 (25). For analyses of the BG, smallvolume corrected (SVC) statistics for the volumes of interest wereapplied, e.g., SN and head of caudate. Two-tailed statistical testswere used for all analyses.

www.sobp.org/journal

Results

Behavioral ResultsResponses of subjects with schizophrenia or schizoaffective

disorder were slower p ¼ .020 and less accurate p ¼ .001compared with control subjects (Table 1).

fMRI Results: PFC and Caudate Hypofunction inSchizophrenia

Across the three WM task phases, we observed significantC � SZ activity only during response. In the cortex, we found onecluster in the left parahippocampal gyrus and another in the rightinferior frontal gyrus, p � .05, corrected (Figure 1A). In the BG, wefound clusters in the left and right head of caudate, p � .05, SVC(Figure 1B). Given the crossed anatomical connections andfunctional interactions at multiple levels within the prefrontalstriatonigral circuit (26,27) and the more prominent between-group difference in the right caudate, we focused on this regionin the analyses presented below. Additional analyses using datacombined from left and right caudate yielded essentially identicalresults (data not shown).

fMRI Results: SN Hyperactivity in SchizophreniaWe observed significant SZ � C activity only during WM

response. There were bilateral SN clusters just dorsal to thecerebral peduncles and superolateral to the midline interpedun-cular notch. The left 120 mm3 cluster, coordinates (�8 �16 �12),met corrected threshold for significance, p � .05, SVC (Figure 1C).Although a few voxels within the right cluster met voxel-wisefamily-wise error corrected significance, the entire cluster did notmeet the a priori specified cluster-level standard for significance(Table 2). Consequently, all subsequent analyses were restrictedto the left SN cluster. The trial-averaged BOLD time series of theleft SN cluster confirmed that both groups exhibited transientpeaks during scans 10 to 14 (Figure 1G), reflecting response-related activity, and greater SZ � C signal. No cortical regionshowed significant SZ hyperactivity.

fMRI Results: Diminished PFC-Basal Ganglia FunctionalConnectivity

Although the co-occurrence of the BG functional abnormal-ities with PFC hypoactivity is consistent with a pathophysiologiclink between PFC and BG dysfunction in schizophrenia, a strongersupport for this link would be the demonstration of diminishedfunctional connectivity between these regions. To directlyaddress this issue, we tested whether there was diminishedconnectivity between the PFC and (in parallel) those caudate andSN regions that showed between-group differences in activity inthe univariate analyses. As explained above in Methods andMaterials and in Supplement 1, we conducted this analysis usingPFC ROIs identified from independent fMRI data, denoted asLocPFC (Figure S2 in Supplement 1). Furthermore, we tested theprefrontal specificity of our prediction with a control region fromthe visual cortex (LocFFA) that was also localized from indepen-dent data. This region was involved in processing faces (Figure S2in Supplement 1) but was not hypothesized to play a regulatoryrole on BG function. We compared BG functional connectivitywith the LocFFA with that of the LocPFC.

PFC-SN Connectivity. We first examined group differences inPFC-SN functional connectivity. Although the groups did notexhibit significant differences in movement, studies suggest thathead movement may significantly affect estimates of functionalconnectivity (28–30) and confound group comparisons of this

-2

0

2

4

6

1 3 5 7 9 11 13 15

BO

LDScan (TR=2 sec)

R Caudate CSZ

-2

0

2

4

6

1 3 5 7 9 11 13 15

BO

LD

Scan (TR=2 sec)

RIFG CSZ

t-value 0 5

C > SZ

C > SZ

t-value 0 5

R

-2

0

2

4

6

1 3 5 7 9 11 13 15

BO

LD

Scan (TR=2 sec)

L Caudate CSZ

-2

0

2

4

6

1 3 5 7 9 11 13 15

BO

LD

Scan (TR=2 sec)

SN CSZ

t-value 0 5

R

SZ > C

L

R L

L

Figure 1. Substantia nigra (SN) hyperactivity in the context of prefrontal cortex and caudate hypoactivity in schizophrenia. Healthy control subjects(C) � subjects with chronic schizophrenia or schizoaffective disorder (SZ) contrast maps with threshold of t ¼ 2.5 applied showing significant groupactivation differences in (A) a cluster in the right inferior frontal gyrus (RIFG) and (B) left and right head of caudate, p � .05, corrected. (C) SZ � Ccontrast maps with threshold of t ¼ 2.5 applied. A close-up of the midbrain showing bilateral regions of greater SN activity in schizophrenia with theblack arrow pointing to the left cluster meeting corrected statistical significance, p � .05. (D–G) Trial-averaged blood oxygen level–dependent (BOLD)time series of the clusters showing significant group differences in activity depicted to the left. Scans 10 to 14 correspond to activity evoked during theresponse phase of working memory indicated by gray bars. Black arrows indicate timing of probe face presentation. Error bars are SEM. L, left; R, right; TR,repetition time.

J.H. Yoon et al. BIOL PSYCHIATRY 2013;74:122–129 125

measure. Therefore, to control for movement in a manneranalogous to what was done in the univariate analysis, weincluded each subject’s six total scan-to-scan movement para-meters as covariates of noninterest in an analysis of covariance

(ANCOVA) of the functional connectivity measures, along with thefactors of group (SZ and C) and circuit (LocPFC-SN and LocFFA-SN).This analysis revealed that connectivity differed between groups ina circuit-specific manner (Figure 2A), with a significant group �

www.sobp.org/journal

Table 2. List of fMRI Activation Peaks During the Response Phase of the Face WM Task

Region Size (mm3) Peak t p, Corrected Coordinates [x y z]

C � SZ L Parahippocampal Gyrus 29120 5.36 �.001a�20 �50 �4

R Inferior Frontal Gyrus 7976 4.20 .035a 52 42 �8

L Caudate Head 424 3.55 .022 �12 4 4

R Caudate Head 776 4.05 .005 6 12 �2

SZ � C L SN 120 4.33 .048 �8 �16 �12

R SN 48 3.72 .102 8 �16 �14

Coordinates are given in the MNI coordinate system.C, healthy control subjects; fMRI, functional magnetic resonance imaging; L, left; MNI, Montreal Neurological Institute; R, right; SN, substantia nigra; SVC,small volume corrected; SZ, subjects with chronic schizophrenia or schizoaffective disorder; WM, working memory.

aThese p values are corrected for whole-brain multiple comparisons. All others are SVC p values.

126 BIOL PSYCHIATRY 2013;74:122–129 J.H. Yoon et al.

circuit interaction [F(1,29) ¼ 5.01, p ¼ .033]. Follow-up ANCOVAsrevealed a significant effect of group on LocPFC-SN connectivity[F(1,29) ¼ 7.04, p ¼ .013] with C � SZ but a nonsignificanteffect of group in LocFFA-SN connectivity [F(1,29) ¼ .76,p ¼ .389].

PFC-Caudate Connectivity. We then assessed group differ-ences in functional connectivity between the PFC and caudateusing the same approach described above. This analysis revealed asignificant effect of group [F(1,29) ¼ 16.04, p � .001] and a trendtoward a significant group � circuit interaction [F(1,29) ¼ 3.186,p ¼ .083]. There was a significant effect of group for both theLocPFC-caudate [F(1,29) ¼ 15.565, p � .001] and LocFFA-caudate[F(1,29) ¼ 8.692, p ¼ .006] circuits with C � SZ (Figure 2B).

The group difference in connectivity for both the PFC-SN andPFC-caudate circuits was significant after Bonferroni correction formultiple comparisons. The Bonferroni-corrected critical p valuewas .05/2 or .025. Because we excluded incorrect trials, weexamined the effect of lower number of trials in the SZ groupby including this variable as a covariate in an ANCOVA of thesignificant univariate and functional connectivity results reportedabove. In all cases, the number of trials was a nonsignificantcovariate (for all p � .585), and group differences in the fMRImeasures remained significant with the inclusion of this covariate.

Figure 2. Diminished prefrontal functional connectivity with the sub-stantia nigra (SN) and caudate in schizophrenia. (A) Functional connec-tivity between prefrontal cortex (LocPFC) and SN was decreased inpatients, #p ¼ .013, while the connectivity between fusiform face area(LocFFA) (a visual cortical region) and SN was similar between groups.*Group � condition interaction, p ¼ .033. (B) Functional connectivitybetween the caudate and LocPFC in subjects with chronic schizophreniaor schizoaffective disorder (SZ) was significantly decreased, #p � .001, butin a nonspecific manner because its connectivity with the LocFFA wasalso significantly decreased in SZ, **p ¼ .006. *Main effect of group,p � .001, and a nonsignificant group � circuit interaction, p ¼ .083. C,healthy control subjects.

www.sobp.org/journal

Nigrostriatal Connectivity Predicts Psychosis SeverityGiven the well-recognized role of DA in the pathogenesis of

psychosis and the central role of the SN in regulating DA, we testedthe association between the functional properties of this structureor circuits involving this structure (SN univariate activity, SN-caudate, and PFC-SN functional connectivity) with the level ofpsychosis. We did not find a significant association betweenpsychosis and SN univariate activity or PFC-SN functional connec-tivity. However, the magnitude of SN-caudate functional connec-tivity was highly correlated with the level of psychosis, r ¼ .61, p ¼.007 (Figure 3). The correlation with psychotic symptoms wassignificant after correcting for multiple comparisons (critical p valuewas .05/3 ¼ .017). To determine the specificity of this finding withpsychosis, we examined the correlation between SN-caudateconnectivity with the two other symptom domains (negativesymptoms and disorganization) and general psychopathology(Global Assessment Scale), with the prediction that they wouldbe nonsignificant. Substantia nigra-caudate connectivity was notsignificantly correlated with other symptom domains or generalpsychopathology; for all correlations, r � .36 and p � .15.

None of our major fMRI findings (SN activity, PFC-caudate,PFC-SN, and SN-caudate functional connectivity) were signifi-cantly associated with the use of antipsychotics, as quantified bychlorpromazine equivalents, p � .2.

Discussion

We observed functional abnormalities within the prefrontalstriatonigral circuit during WM in schizophrenia. Along with thefrequently reported PFC hypoactivity, patients exhibited striatalhypoactivity and nigral hyperactivity. We examined the possibility

R = 0.38

0.0

5.0

10.0

15.0

-0.3 0.1 0.5

SAPS

SN-Caudate

Figure 3. Nigrostriatal functional connectivity predicts severity of psy-chosis. The magnitude of functional connectivity between the caudateand substantia nigra (SN) demonstrated a highly significant correlationwith psychosis severity, as quantified by the total Scale for the Assess-ment of Positive Symptoms (SAPS) score, p ¼ .007.

J.H. Yoon et al. BIOL PSYCHIATRY 2013;74:122–129 127

that the BG abnormalities could be due to prefrontal disconnec-tivity by measuring prefrontal functional connectivity with theseregions and observed it to be decreased in schizophrenia. Wefound a strong association between the strength of nigrostriatalfunctional connectivity and psychosis severity. Taken together,our results are consistent with the theory that prefrontaldysfunction in schizophrenia leads to disrupted PFC-BG circuits,nigral disinhibition, and psychosis. In the context of the emergingrecognition of the prefrontal striatonigral circuit’s important rolein cognition (18,19), our results suggest that this circuit may be acommon pathway for the expression of cognitive deficits andpsychosis in schizophrenia. Therefore, this circuit may play acentral role in the pathogenesis of this condition.

Our SN findings deserve special consideration because of itscentral role in regulating DA. An obvious limitation of this study isthat, although the nigra is highly enriched with DAergic neurons,its BOLD signal reflects total activity of and inputs into all thedifferent cell types found within this structure. Nonetheless, thefollowing lines of evidence suggest the plausibility that nigralhyperactivity, as measured by fMRI, may reflect enhanced DAergictone in schizophrenia. A recent PET/fMRI study showed that task-evoked ventral midbrain BOLD signal can strongly correlate withthe magnitude of striatal DA release (31). It could be argued thatthe enhanced nigral BOLD signal reflects greater inhibitory inputsto this region. This interpretation, however, is at odds withemerging evidence that synaptic inhibition leads to decreasedBOLD signal (32–34). Thus, increased nigral activity would repre-sent greater net excitatory signaling onto (35) and/or spikingactivity (36) of this region. Either interpretation points to thepossibility of enhanced DAergic drive biasing the system towardincreased DA neurotransmission in schizophrenia.

Whether enhanced nigral activity and DAergic drive translateinto increased postsynaptic striatal DA levels and psychosis willhave to be clarified by future studies. For now, we can point toseveral lines of indirect evidence consistent with such a link.These include the well-documented dense nigrostriatal DAergicconnectivity (37), the demonstration of increased striatal DArelease during responses executed during a cognitive task (38),and, as stated above, the tight correlation between task-evokedmidbrain fMRI signal with striatal DA release magnitude (31). Thepossibility that the nigral hyperactivity associated with prefrontaldysfunction in schizophrenia could lead to enhanced striatal DAis supported by the following evidence. Among the severalschizophrenia PET studies that have reported increased pharma-cologically evoked striatal DA release (1–4), one study showed aninverse relationship between release magnitude and spectro-scopic markers of PFC pathology (39). Another study by Meyer-Lindenberg et al. (12) demonstrated an inverse relationshipbetween task-evoked PFC hypoperfusion and metabolic markersof midbrain DA function in schizophrenia.

Our findings, interpreted in light of the following neuroana-tomical evidence, suggest that the striatum may be involved inmediating the effects of PFC dysfunction on midbrain function inschizophrenia. The PFC has sparse direct connections withDAergic midbrain in primates (40), but it has dense connectionswith the striatum. Excitatory prefrontal efferents synapse ontogamma-aminobutyric acidergic medium spinal neurons, whichproject to and may inhibit the firing of DA neurons (14,15). Seenin this context, the striatal hypoactivity and decreased PFC-striatal functional connectivity in schizophrenia observed in thisstudy support the possibility that PFC dysfunction and its failureto drive inhibitory striatonigral cells may be contributing to nigralhyperactivity. This possibility, however, should be considered in

light of evidence suggesting that direct striatonigral inhibitionmay be weak (41).

We found a strong association between task-evoked nigros-triatal functional connectivity and psychosis severity. This asso-ciation is consistent with the body of literature pointing tothe striatum as a major mediator of DA psychosis (42–46).While psychosis is a cardinal symptom of schizophrenia, it isdifficult to model in animals. Psychosis is subjectively experi-enced and entails complex and seemingly human-specificconstructs such as beliefs. Consequently, task-evoked nigrostria-tal functional connectivity could be useful for reverse tran-slation into preclinical animal models as a brain-based psychosismarker.

It is reasonable to question why ventral midbrain hyperactivityin schizophrenia has not been reported previously among theabundant published fMRI studies employing WM or relatedparadigms. We believe that two methods we employed couldexplain this. First, we minimally smoothed our images. We showin Supplement 1 how smoothing with the commonly appliedkernel of 8 mm would have literally smoothed out the nigralhyperactivity (Figure S1 in Supplement 1). Second, the apparentresponse phase-specificity of nigral hyperactivity suggests thatevent-related analysis may be necessary for its detection. PastfMRI studies of schizophrenia have predominantly utilizedblocked designs (20), which may lack sensitivity to detect taskphase-specific abnormalities.

The fact that our significant fMRI findings were restricted tothe response phase of WM and not found in other phasesthought to reflect essential WM component processes, e.g., delay,raises questions about cognitive mechanisms and specificity toWM. The present study cannot directly address these questionsbecause of design limitations. We utilized a WM paradigm not toaddress issues related to WM specificity or cognitive mechanisms,but rather as a means of testing the hypothesis of an associationbetween PFC and BG dysfunction in SZ. Consequently, ourparadigm did not include additional conditions that would beneeded to address these issues, which are left for future studiesto resolve. While we did not predict that SN hyperactivity wouldbe restricted to the response phase, this finding is not surprising,given the prevailing thought that response or action selection isone of the primary functions of the BG and the well-documentednigral involvement in motor-related processes.

Another limitation of this study is that we studied subjectstaking antipsychotics and other psychiatric medications.Although we did not detect an association with antipsychotics,future studies with unmedicated subjects will be required tomore rigorously examine a medication confound.

Two recent fMRI studies have reported excess midbrainresponses during neutral conditions and blunted responsesduring rewarding or affectively salient conditions in schizophre-nia (16,17). It is not surprising that midbrain hyperactivity is notuniversally present in schizophrenia because simple up or downmodels of dysfunction are likely inadequate to characterize thecomplex regulation and response properties of the midbrain inschizophrenia. Our results suggest that the PFC may be animportant factor determining the directionality of midbraindysfunction in schizophrenia. During WM, PFC deficits couldresult in the failure to dampen task-evoked nigral activity. Inreward processing, recent work suggesting that PFC dysfunctioncontributes to suboptimal action value representations in schizo-phrenia (47) leaves open the possibility of impaired PFC-mediated up-regulation of reward prediction error-relatedDAergic activity in this condition. Future research will be needed

www.sobp.org/journal

128 BIOL PSYCHIATRY 2013;74:122–129 J.H. Yoon et al.

to clarify possible impairments in process-specific PFC modula-tion of midbrain function.

The primacy of PFC dysfunction in our model is consistentwith previously elaborated developmental theories of schizo-phrenia (8,48) and symptom chronology (10): premorbid cogni-tive deficits usually precede positive symptoms. However, thepossibility of a primary BG pathology leading to PFC dysfunction,as suggested by a recent rodent model (49) and findings ofcortical DA deficits in schizophrenia (50–53), cannot be excludedby the present results. This alternative hypothesis should beaddressed in future studies.

It may seem surprising that our results implicate the SN andthe nigrostriatal circuit and not the VTA and mesolimbic circuits inpsychosis. It is now recognized that the previously held view of astrict anatomic segregation between the VTA and SN systemsdoes not hold, particularly in primates. Dopamine neurons thatinnervate the ventral striatum are found in both the VTA andnigra (54), while DA neurons projecting to the cortex and limbicstructures are found not only in the VTA, but also in the SN (37).These findings are consistent with the emerging view that the SNplays an important role in cognition, in addition to its well-recognized role in supporting motor processes (55). Nonetheless,our results should not be interpreted to suggest the unimpor-tance of the VTA and associated limbic systems in psychosis. Wespeculate that our paradigm may have been more sensitive todetecting abnormalities of the cognitive regions of the BG. Itremains to be clarified whether similar findings for the VTA can beuncovered using paradigms that engage limbic system circuitry.

In summary, our demonstration of prefrontal striatonigralabnormalities in schizophrenia supports the proposition that thiscircuit may be a pathway linking cognitive deficits and psychosisin schizophrenia. Furthermore, our work has identified in vivobrain correlates of psychosis that may be useful to futurepathophysiologic studies of schizophrenia.

This study was supported by K08 MH076174 from the NationalInstitute of Mental Health, National Institutes of Health, to JHY. Dr.Yoon, Mr. Raouf, and Dr. D’Esposito report no biomedical financialinterests or potential conflicts of interest. Dr. Minzenberg hasreceived an investigator initiated research grant from Shire.Dr. Carter has received research funding from GlaxoSmithKline.

Supplementary material cited in this article is available online.

1. Abi-Dargham A, Gil R, Krystal J, Baldwin RM, Seibyl JP, Bowers M, et al.(1998): Increased striatal dopamine transmission in schizophrenia:Confirmation in a second cohort. Am J Psychiatry 155:761–767.

2. Breier A, Su TP, Saunders R, Carson RE, Kolachana BS, de BartolomeisA, et al. (1997): Schizophrenia is associated with elevatedamphetamine-induced synaptic dopamine concentrations: Evidencefrom a novel positron emission tomography method. Proc Natl AcadSci U S A 94:2569–2574.

3. Kegeles LS, Abi-Dargham A, Frankle WG, Gil R, Cooper TB, Slifstein M,et al. (2010): Increased synaptic dopamine function in associativeregions of the striatum in schizophrenia. Arch Gen Psychiatry 67:231–239.

4. Laruelle M, Abi-Dargham A, van Dyck CH, Gil R, D’Souza CD, Erdos J,et al. (1996): Single photon emission computerized tomographyimaging of amphetamine-induced dopamine release in drug-freeschizophrenic subjects. Proc Natl Acad Sci U S A 93:9235–9240.

5. Berman KF, Zec RF, Weinberger DR (1986): Physiologic dysfunction ofdorsolateral prefrontal cortex in schizophrenia. II. Role of neuroleptictreatment, attention, and mental effort. Arch Gen Psychiatry 43:126–135.

www.sobp.org/journal

6. Ingvar DH, Franzen G (1974): Abnormalities of cerebral blood flowdistribution in patients with chronic schizophrenia. Acta PsychiatrScand 50:425–462.

7. Yoon JH, Minzenberg MJ, Ursu S, Ryan Walter BS, Wendelken C,Ragland JD, Carter CS (2008): Association of dorsolateral prefrontalcortex dysfunction with disrupted coordinated brain activity inschizophrenia: Relationship with impaired cognition, behavioral dis-organization, and global function. Am J Psychiatry 165:1006–1014.

8. Weinberger DR (1987): Implications of normal brain development forthe pathogenesis of schizophrenia. Arch Gen Psychiatry 44:660–669.

9. Pycock CJ, Kerwin RW, Carter CJ (1980): Effect of lesion of corticaldopamine terminals on subcortical dopamine receptors in rats. Nature286:74–76.

10. Lesh TA, Niendam TA, Minzenberg MJ, Carter CS (2011): Cognitivecontrol deficits in schizophrenia: Mechanisms and meaning. Neurop-sychopharmacology 36:316–338.

11. Bertolino A, Knable MB, Saunders RC, Callicott JH, Kolachana B,Mattay VS, et al. (1999): The relationship between dorsolateralprefrontal N-acetylaspartate measures and striatal dopamine activityin schizophrenia. Biol Psychiatry 45:660–667.

12. Meyer-Lindenberg A, Miletich RS, Kohn PD, Esposito G, Carson RE,Quarantelli M, et al. (2002): Reduced prefrontal activity predictsexaggerated striatal dopaminergic function in schizophrenia. NatNeurosci 5:267–271.

13. Bjorklund A, Dunnett S (2007): Dopamine neuron systems in thebrain: An update. Trends Neurosci 30:194–202.

14. Grace AA, Bunney BS (1985): Opposing effects of striatonigral feed-back pathways on midbrain dopamine cell activity. Brain Res 333:271–284.

15. Paladini CA, Celada P, Tepper JM (1999): Striatal, pallidal, and parsreticulata evoked inhibition of nigrostriatal dopaminergic neurons ismediated by GABA(A) receptors in vivo. Neuroscience 89:799–812.

16. Murray GK, Corlett PR, Clark L, Pessiglione M, Blackwell AD, Honey G,et al. (2008): Substantia nigra/ventral tegmental reward predictionerror disruption in psychosis. Mol Psychiatry 13(239):267–276.

17. Romaniuk L, Honey GD, King JR, Whalley HC, McIntosh AM, Levita L,et al. (2010): Midbrain activation during pavlovian conditioning anddelusional symptoms in schizophrenia. Arch Gen Psychiatry 67:1246–1254.

18. Landau SM, Lal R, O’Neil JP, Baker S, Jagust WJ (2009): Striataldopamine and working memory. Cereb Cortex 19:445–454.

19. Lewis SJ, Dove A, Robbins TW, Barker RA, Owen AM (2004): Striatalcontributions to working memory: A functional magnetic resonanceimaging study in humans. Eur J Neurosci 19:755–760.

20. Minzenberg MJ, Laird AR, Thelen S, Carter CS, Glahn DC (2009): Meta-analysis of 41 functional neuroimaging studies of executive functionin schizophrenia. Arch Gen Psychiatry 66:811–822.

21. Barch DM, Carter CS, MacDonald AW 3rd, Braver TS, Cohen JD (2003):Context-processing deficits in schizophrenia: Diagnostic specificity,4-week course, and relationships to clinical symptoms. J AbnormPsychol 112:132–143.

22. Maldjian JA, Laurienti PJ, Kraft RA, Burdette JH (2003): An automatedmethod for neuroanatomic and cytoarchitectonic atlas-based inter-rogation of fMRI data sets. Neuroimage 19:1233–1239.

23. Lancaster JL, Woldorff MG, Parsons LM, Liotti M, Freitas CS, Rainey L,et al. (2000): Automated Talairach atlas labels for functional brainmapping. Hum Brain Mapp 10:120–131.

24. Rissman J, Gazzaley A, D’Esposito M (2004): Measuring functionalconnectivity during distinct stages of a cognitive task. Neuroimage 23:752–763.

25. Friston KJ, Holmes A, Poline JB, Price CJ, Frith CD (1996): Detectingactivations in PET and fMRI: Levels of inference and power. Neuro-image 4:223–235.

26. Gerfen CR, Staines WA, Arbuthnott GW, Fibiger HC (1982): Crossedconnections of the substantia nigra in the rat. J Comp Neurol 207:283–303.

27. Kelly PA, Graham DI, McCulloch J (1982): Specific alterations in localcerebral glucose utilization following striatal lesions. Brain Res 233:157–172.

28. Power JD, Barnes KA, Snyder AZ, Schlaggar BL, Petersen SE (2012):Spurious but systematic correlations in functional connectivity MRInetworks arise from subject motion. Neuroimage 59:2142–2154.

J.H. Yoon et al. BIOL PSYCHIATRY 2013;74:122–129 129

29. Satterthwaite TD, Wolf DH, Loughead J, Ruparel K, Elliott MA,Hakonarson H, et al. (2012): Impact of in-scanner head motion onmultiple measures of functional connectivity: Relevance for studies ofneurodevelopment in youth. Neuroimage 60:623–632.

30. Van Dijk KR, Sabuncu MR, Buckner RL (2011): The influence of headmotion on intrinsic functional connectivity MRI. Neuroimage 59:431–438.

31. Schott BH, Minuzzi L, Krebs RM, Elmenhorst D, Lang M, Winz OH, et al.(2008): Mesolimbic functional magnetic resonance imaging activa-tions during reward anticipation correlate with reward-related ventralstriatal dopamine release. J Neurosci 28:14311–14319.

32. Devor A, Tian P, Nishimura N, Teng IC, Hillman EM, Narayanan SN,et al. (2007): Suppressed neuronal activity and concurrent arteriolarvasoconstriction may explain negative blood oxygenation level-dependent signal. J Neurosci 27:4452–4459.

33. Schafer K, Blankenburg F, Kupers R, Gruner JM, Law I, Lauritzen M,Larsson HB (2012): Negative BOLD signal changes in ipsilateralprimary somatosensory cortex are associated with perfusiondecreases and behavioral evidence for functional inhibition. Neuro-image 59:3119–3127.

34. Lee JH, Durand R, Gradinaru V, Zhang F, Goshen I, Kim DS, et al.(2010): Global and local fMRI signals driven by neurons definedoptogenetically by type and wiring. Nature 465:788–792.

35. Logothetis NK, Pauls J, Augath M, Trinath T, Oeltermann A (2001):Neurophysiological investigation of the basis of the fMRI signal.Nature 412:150–157.

36. Mukamel R, Gelbard H, Arieli A, Hasson U, Fried I, Malach R (2005):Coupling between neuronal firing, field potentials, and FMRI inhuman auditory cortex. Science 309:951–954.

37. Haber SN, Fudge JL, McFarland NR (2000): Striatonigrostriatal path-ways in primates form an ascending spiral from the shell to thedorsolateral striatum. J Neurosci 20:2369–2382.

38. Lappin JM, Reeves SJ, Mehta MA, Egerton A, Coulson M, Grasby PM(2009): Dopamine release in the human striatum: Motor and cognitivetasks revisited. J Cereb Blood Flow Metab 29:554–564.

39. Bertolino A, Breier A, Callicott JH, Adler C, Mattay VS, Shapiro M, et al.(2000): The relationship between dorsolateral prefrontal neuronal N-acetylaspartate and evoked release of striatal dopamine in schizo-phrenia. Neuropsychopharmacology 22:125–132.

40. Frankle W, Laruelle M, Haber S (2006): Prefrontal cortical projectionsto the midbrain in primates: Evidence for a sparse connection.Neuropsychopharmacology 31:1627–1636.

41. Chuhma N, Tanaka KF, Hen R, Rayport S (2011): Functional connectomeof the striatal medium spiny neuron. J Neurosci 31:1183–1192.

42. Carlsson A, Lindqvist M (1963): Effect of chlorpromazine or haloper-idol on formation of 3methoxytyramine and normetanephrine inmouse brain. Acta Pharmacol Toxicol (Copenh) 20:140–144.

43. Seeman P, Chau-Wong M, Tedesco J, Wong K (1975): Brain receptorsfor antipsychotic drugs and dopamine: Direct binding assays. ProcNatl Acad Sci U S A 72:4376–4380.

44. Farde L, Wiesel FA, Halldin C, Sedvall G (1988): Central D2-dopaminereceptor occupancy in schizophrenic patients treated with antipsy-chotic drugs. Arch Gen Psychiatry 45:71–76.

45. Kapur S, Zipursky R, Jones C, Remington G, Houle S (2000): Relation-ship between dopamine D(2) occupancy, clinical response, and sideeffects: A double-blind PET study of first-episode schizophrenia. Am JPsychiatry 157:514–520.

46. Nordstrom AL, Farde L, Wiesel FA, Forslund K, Pauli S, Halldin C,Uppfeldt G (1993): Central D2-dopamine receptor occupancy inrelation to antipsychotic drug effects: A double-blind PET study ofschizophrenic patients. Biol Psychiatry 33:227–235.

47. Gold JM, Waltz JA, Matveeva TM, Kasanova Z, Strauss GP, HerbenerES, et al. (2012): Negative symptoms and the failure to represent theexpected reward value of actions: Behavioral and computationalmodeling evidence. Arch Gen Psychiatry 69:129–138.

48. Feinberg I (1982): Schizophrenia: Caused by a fault in programmedsynaptic elimination during adolescence? J Psychiatr Res 17:319–334.

49. Kellendonk C, Simpson EH, Polan HJ, Malleret G, Vronskaya S, WinigerV, et al. (2006): Transient and selective overexpression of dopamineD2 receptors in the striatum causes persistent abnormalities inprefrontal cortex functioning. Neuron 49:603–615.

50. Abi-Dargham A, Mawlawi O, Lombardo I, Gil R, Martinez D, Huang Y,et al. (2002): Prefrontal dopamine D1 receptors and working memoryin schizophrenia. J Neurosci 22:3708–3719.

51. Durstewitz D, Seamans JK (2008): The dual-state theory of prefrontalcortex dopamine function with relevance to catechol-o-methyltransferasegenotypes and schizophrenia. Biol Psychiatry 64:739–749.

52. Winterer G, Weinberger DR (2004): Genes, dopamine and corticalsignal-to-noise ratio in schizophrenia. Trends Neurosci 27:683–690.

53. Davis KL, Kahn RS, Ko G, Davidson M (1991): Dopamine in schizo-phrenia: A review and reconceptualization. Am J Psychiatry 148:1474–1486.

54. Haber SN, Fudge JL, McFarland NR (2000): Striatonigrostriatal path-ways in primates form an ascending spiral from the shell to thedorsolateral striatum. J Neurosci 20:2369–2382.

55. Williams SM, Goldman-Rakic PS (1998): Widespread origin of theprimate mesofrontal dopamine system. Cereb Cortex 8:321–345.

www.sobp.org/journal