Available online at www.sciencedirect.com

01 (2008) 111–123www.elsevier.com/locate/schres

Schizophrenia Research 1

Cortical gray and white matter volume in unmedicated schizotypaland schizophrenia patients

Erin A. Hazlett a,⁎, Monte S. Buchsbaum a, M. Mehmet Haznedar a, Randall Newmark a,Kim E. Goldstein a, Yuliya Zelmanova a, Cathryn F. Glanton a, Yuliya Torosjan a,Antonia S. New a,b, Jennifer N. Lo a, Vivian Mitropoulou a, Larry J. Siever a,b

a Department of Psychiatry, Mount Sinai School of Medicine, New York, NY, United Statesb Bronx Veterans Affairs Medical Center, NY and Mental Illness Research, Education and Clinical Center (MIRECC) and VISN 3, United States

Received 13 July 2007; received in revised form 5 December 2007; accepted 13 December 2007Available online 13 February 2008

Abstract

Magnetic resonance imaging (MRI) studies have revealed fronto-temporal cortical gray matter volume reductions in schizophrenia.However, to date studies have not examined whether age- and sex-matched unmedicated schizotypal personality disorder (SPD) patientsshare some or all of the structural brain-imaging characteristics of schizophrenia patients. We examined cortical gray/white matter volumesin a large sample of unmedicated schizophrenia-spectrum patients (n=79 SPD,n=57 schizophrenia) and 148 healthy controls.MRI imageswere reoriented to standard position parallel to the anterior–posterior commissure line, segmented into gray and white matter tissue types,and assigned to Brodmann areas (BAs) using a postmortem-histological atlas. Group differences in regional volume of gray and whitematter in the BAs were examined withMANOVA. Schizophrenia patients had significantly reduced gray matter volume widely across thecortex but more marked in frontal and temporal lobes. SPD patients had reductions in the same regions but only about half that observed inschizophrenia and sparing in key regions including BA10. In schizophrenia, greater fronto-temporal volume loss was associated withgreater negative symptom severity and in SPD, greater interpersonal and cognitive impairment. Overall, our findings suggest that increasedprefrontal volume in BA10 and sparing of volume loss in temporal cortex (BAs 22 and 20)may be a protective factor in SPDwhich reducesvulnerability to psychosis.Published by Elsevier B.V.

Keywords: MRI; Schizophrenia; Schizotypal personality disorder; Frontal lobe volume; Temporal lobe volume; Cingulate gyrus; Negativesymptoms; Gray matter volume; White matter volume

1. Introduction

Numerous studies have reported structural brain-imaging abnormalities in schizophrenia-spectrum disor-

⁎ Corresponding author. Department of Psychiatry, Box 1505,Mount Sinai School of Medicine, New York, NY 10029, UnitedStates. Tel.: +1 212 241 2779.

E-mail address: erin.hazlett@mssm.edu (E.A. Hazlett).

0920-9964/$ - see front matter. Published by Elsevier B.V.doi:10.1016/j.schres.2007.12.472

ders. Similar to schizophrenia, cerebrospinal fluid (CSF) isincreased and cortical volumes are decreased in schizo-typal personality disorder (SPD) (Buchsbaum et al., 1997;Dickey et al., 2000; Koo et al., 2006). Temporal cortexvolume reductions, particularly in the superior temporalgyrus (STG) have been consistently reported in schizo-phrenia (Davidson and Heinrichs, 2003; Downhill et al.,2001; Gur et al., 2000; Hirayasu et al., 2000; Shentonet al., 2001;Wright et al., 2000). Reduced temporal cortex

112 E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

volume has also been observed in SPD in STG (Dickeyet al., 1999; Kawasaki et al., 2004; Takahashi et al.,2006a), Heschl's gyrus (Dickey et al., 2002b), and middleand inferior temporal gyrus (Downhill et al., 2001) insome but not all studies (Takahashi et al., 2006b). Studieswith schizophrenia patients and their relatives also suggestreduced medial temporal volume including the amygdalaand/or hippocampal complex, but these findings have notbeen consistently observed in SPD (Dickey et al., 2002a,1999; Seidman et al., 2002, 2003). One recent studyreported reduced hippocampal volume in SPD (Dickeyet al., 2007). Taken together, these findings are consistentwith a model of common temporal lobe abnormalitiesacross the schizophrenia spectrum.

While SPD and schizophrenia patients appear to showsome common temporal lobe abnormalities, SPD patientsdo not appear to show volumetric decreases in frontalcortex to the same extent as schizophrenia (Kawasaki et al.,2004; Siever et al., 1990, 2002; Siever and Davis, 2004;Suzuki et al., 2005). Reductions in frontal volume areassociated with the deficit-like symptoms of SPD, im-plying that patients with smaller frontal lobe volume aremore likely to display traits such as a sociality (Diwadkaret al., 2006; Raine et al., 1992; Siever et al., 1993). We(Haznedar et al., 2004) and others (Takahashi et al., 2004)have reported that cingulate gray matter volume is rela-tively preserved in SPD compared with schizophrenia.Volume reductions in the parietal lobe have also beenreported to show a spectrum pattern with more widespreadgray matter reductions in schizophrenia than SPD (Zhouet al., 2007). Although relatively fewer anatomical MRIstudies have been conducted in SPD, they suggest thatSPD patients show abnormalities in some, but not all of thebrain regions implicated in schizophrenia.

Individuals with SPD share a broad range of similaritieswith schizophrenia patients in terms of genetics, neuro-biology, and phenomenology (Siever and Davis, 2004).These similarities may be essential for the pathogenesis ofschizophrenia and related to the vulnerability to schizo-phrenia. However, additional pathological changes maybe required for the development of overt and sustainedpsychosis. Siever and Davis (2004) suggest a model ofschizophrenia-spectrum disorders which hypothesizes thatshared deficits may reflect a common neurodevelopmen-tally based cortical pathology (e.g., temporal cortex).However, in contrast to patients with schizophrenia, SPDpatients are hypothesized to have reduced vulnerability topsychosis due to protective factors which include greaterprefrontal reserves. Factors such as capacity to recruitalternate prefrontal areas (e.g., BA10) to compensatefor other dysfunctional areas during cognitive demands(Buchsbaum et al., 2002) and less frontal lobe volume

reduction when compared to schizophrenia (Suzuki et al.,2005) may protect the SPD patients from chronic schi-zophrenia (Siever and Davis, 2004).

Prior magnetic resonance imaging (MRI) work compar-ing SPD and schizophrenia patients, (e.g. Kawasaki et al.,2004; Zhou et al., 2007) has involvedmedicated patients orvoxel-based morphometry which did not provide com-bined three-group contrasts, a comparison of themagnitudeof volume change in SPD and schizophrenia, or statisticalcontrasts of different BAs or hemispheric differences. Themain goal of this study was to use MRI to determine theextent to which cortical gray and white matter volumediffers from normal in a large, unmedicated schizophrenia-spectrum sample. We examined gray and white mattervolume within BAs of the entire cortex and the cingulategyrus in three age- and sex-matched groups: healthycontrols and unmedicated patients with SPD or schizo-phrenia. We hypothesized that there would be a distinctionbetween schizotypy and schizophrenia in terms of under-lying regional cortical volume; SPD patients would showless severe prefrontal and temporal cortex volume reduc-tions than schizophrenia patients. We also explored thesymptomatic correlates of fronto-temporal cortex volumein our large sample and hypothesized that in both groups,fronto-temporal volume reductions would be associatedwith greater symptom severity.

2. Methods

2.1. Participants

We studied three age- and sex-matched groups: 79individuals with schizotypal personality disorder, 57patients with schizophrenia, and 148 healthy controls.The three groups did not significantly differ in age or sex.Demographic and clinical data are presented in Table 1.

All patients met DSM-IV criteria and were offmedication at the time of their MRI scans for a minimumof two weeks. All participants were screened for severemedical or neurological illness and head injury by com-prehensive medical history, and laboratory tests taken by aphysician. Healthy controls and SPD patients received aninterview with a psychologist using the Structured ClinicalInterview for DSM-III-R Axis I disorders (Spitzer et al.,1992) and the Structured Interview for DSM-III-R Perso-nality Disorders (Pfohl and Blum, 1989). Schizophreniapatients were diagnosed using the Comprehensive Assess-ment for Symptoms and History (Andreasen et al., 1992).Participants were excluded if they met lifetime criteria forsubstance dependence or abuse in the past six months, hadtaken any psychotropic medications in the last two weeks,or had a positive urine toxicology screen for drugs of abuse.

Table 1Demographic and clinical characteristics of schizophrenia-spectrum groups and healthy controls

Characteristic Healthy controls SPD patients Sz patients

(n=148) (n=79) (n=57)

Mean SD Range Mean SD Range Mean SD Range

Age (years) 35.6 13.0 17–65 37.9 a 10.1 19–62 34.8 b 14.6 17–65Male/female 99/49 62/17 c 42/15 d

Psychoactive meds N % N % N %Never medicated – – 61 77% 29 51%Previously medicated – – 18 e 23% 28 49%

Clinical symptoms Mean SD Range Mean SD Range Mean SD Range

BPRS total score 48.5 11.3 26–83Positive symptom factor 21.4 6.4 10–37Negative symptom factor 8.5 3.2 3–14

Total number of DSM-III-R symptoms (of 9 possible) 6.8 1.1 5–9Cognitive impairment factor 1.5 .8 0–5Interpersonal factor 2.9 .9 1–5Paranoid factor 2.4 .9 1–5a post-hoc t-test for HC vs. SPD group, t(225)=1.36, p=0.17.b HC vs. Sz group, t(203)=0.38, p=0.70; SPD vs. Sz group, t(134)=1.46, p=0.15.c HC vs. SPD group, t(225)=1.84, p=0.07. Note: MANOVAs reported in results section were also analyzed with analysis of covariance with sex

as a covariate and they yielded same pattern of results in terms of significant interactions with group.d HC vs. Sz group, t(203)=0.94, p=0.35; SPD vs Sz group, t(134)=0.65, p=0.52.e 15 of the previously-medicated SPD patients received a non-neuroleptic medication and 3 received neuroleptics. All SPD and schizophrenia

patients were off psychoactive medications for a minimum of two weeks prior to their MRI scan.

113E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

Healthy controls had no personal history, nor a first degreerelative with an Axis I or Axis II diagnosis.

The healthy controls and majority of the SPD patients(90%) were recruited through advertisement in local news-papers. The remaining SPD patients were recruited throughreferrals from psychiatric clinics at the Bronx VeteransAffairs Medical Center and Mount Sinai. Schizophreniapatients were recruited from inpatient and outpatient unitsat both of these hospitals. Participants provided writteninformed consent in accordance with the InstitutionalReview Board guidelines. A subgroup of the healthycontrols (n=24), SPD patients (n=12), and schizophreniapatients (n=26) were previously included in studies exa-mining Brodmann areas using FDG-PET (Buchsbaumet al., 2002), cingulate volume (Haznedar et al., 2004) andwhole temporal lobe volume (Downhill et al., 2001).

All of the schizophrenia patients received the BriefPsychiatric Scale (BPRS) (Overall and Gorham, 1962)within one week of their MRI scan.We report the total 18-item BPRS score, Positive and Negative Symptom Factorscores (Opler et al., 1984). Since BPRS was not obtainedfor the SPD patients, we grouped their DSM symptomsinto three composite scores: interpersonal, paranoid, andcognitive/perceptual based on our prior work (Bergmanet al., 1996; Mitropoulou et al., 2002). These factors havebeen argued to be more sensitive than the BPRS in SPD

patients and the interpersonal factor resembles the negativesymptom score on the BPRS.

2.2. MRI image collection

T1-weighted axial MRI scans were acquired witha 1.5-T Signa-5x system. The acquisition parameterswere: repetition time=24 msec, echo time=5 msec, flipangle=40°, slice thickness=1.2 mm, pixel matrix=256×256, field of view=23 cm, and total slices=128.MRI scans were re-sectioned to standard Talairach–Tournoux (Talairach and Tournoux, 1988) position.

2.3. Brodmann area measurement and tissue typequantification

Gray and white matter volumes within Brodmannareas and cingulate gyrus were analyzed on coronalMRI slices using our standard methods briefly describedbelow and detailed elsewhere (Buchsbaum et al., 2002;Hazlett et al., 1998; Mitelman et al., 2005a).

A digitized version of a histologically-based atlas(Perry et al., 1991) which includes 33 coronal slice mapsof Brodmann areas defined by microscopic examinationof one entire postmortem brain was used to delineate theBrodmann areas. Coronal slices perpendicular to the

114 E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

anterior–posterior commissure line were reconstructed ina 256×256 pixel matrix. First, we determined the front(first slice containing the cortical ribbon) and back of thebrain (last slice containing the cortical ribbon) and iden-tified 33 evenly spaced slices such that the first slice began1/34th of the distance from front to back. For each tem-poral lobe, we identified the slice with the anterior tem-poral pole and the most posterior extent of the Sylvianfissure and divided the interval into 13 equally spacedslices. The posterior extent was determined by followingthe main fissure posteriorly until it had no sulcal depth.The brain edgewas obtained on the approximately circular33 non-temporal slices and 26 (13 in each hemisphere)temporal slices by depositing points visually on the tips ofthe gyri and then fitting a spline curve to the points. Eachslice was then divided into 20 radial sectors on eachhemisphere surface and 10 midline sectors. Brodmannareas were then assessed for the gray matter, white matter,and CSF pixels within each sector (see later discussion ofsegmentationmethod); meanswereweighted according to

Fig. 1. Delineation of cortical Brodmann areas. Upper left: Perry brain atlas fodrawing and with sectors added. Upper right: Coronal brain edges are tracalgorithm, 20 sectors on the lateral and orbital surface (white numbers) andleft: Perry atlas in three-dimensional perspective within MRI surface. Gray andright: Diagram of frontal Brodmann areas as colored stick diagram: area 10

the number of sectors in each region of interest and pro-portionally combined to obtain a single measure. FourBrodmann areas were combined (1-2-3-5) as delineatedby Perry.

For gray matter and white matter quantification, thecoronal images for each of the lobes were segmented intomatter type by using cut-off values individually determinedin each subject by examining the within-brain-edge histo-gram of axial MRI signal intensity values. Every subjecthad a clear point of rarity between gray and white matter.Validation of our gray/white matter segmentation approachhas been reported (Mitelman et al., 2003; Mitelman et al.,2005a).

2.4. Statistical methods

We obtained volume data from 39 Brodmann areas(BAs: 1-2-3-5, 4,6,7a, 7b,8,9,10,11,12,17,18,19,20,21,22,23,24,25,27,28,29,30,31,32,34,35,36,37,38,39,40,41,42,43,44,45,46, and 47) identified by the Perry atlas (Fig. 1).

r slice #5 showing frontal pole Brodmann areas (e.g., 9, 46, 45…) in rawed on individual MRI slices and set of sectors applied by computer10 in the medial portion (green numbers) for a total of 30. Bottomwhite matter volume within each Brodmann area is calculated. Bottomis magenta, area 9 is blue, area 46 yellow, as example.

115E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

Subgroups of these variables developed on a theoretical andanatomical basis were entered into a series of multivariateanalysis of variance (MANOVA) analyses with diagnosticgroup (Healthy controls vs. SPD vs. Schizophrenia) as thebetween-group factor. Repeated measures included: region(e.g., orbital, dorsolateral), selected sets of Brodmann areas(nested within regions, e.g., orbital: BA11, BA12, BA47and dorsolateral: BA44, BA45, BA46), hemisphere, andtissue type (gray, whitematter). Group ×Region and higherorder interactions were examined to establish regionaldifferences. Follow-up simple interactions were performedto identify the strongest sources of group interactions. Wecomputed relative size as a ratio of area of the region-of-interest volume/total brain volume multiplied by 100.While ratiomeasuresmay not adequately correct for normalvariation in intracranial volume, they are more commonlyreported. Note that in reporting repeated-measures MAN-OVA, means for the main effect of diagnostic group andGroup × Region interactions are collapsed across the BAs.

For the MANOVAs, we report the multivariate (Wilkslambda) F from Statistica (StatSoft, 2003) to adjustprobabilities for repeated-measure effects with more thantwo levels. Fisher's Least Significant Difference (LSD)tests were used to follow-up significant interaction effectswith diagnostic group. We examine only the frontal lobe,cingulate, and lateral temporal lobe with nested MAN-OVA. This approach, which provided tests of hypothe-sized group differences, helps minimize Type I statisticalerror involved with t-tests for each area, group contrast,

Fig. 2. Differences from the healthy control group in cingulate gyrus BrodmCompared with healthy controls, the schizophrenia patients showed the gintermediate (Group×Brodmann area interaction, F[8,556]=2.92, p=0.00volume) patients minus healthy controls.

and hemisphere. In a further effort, we only examinesignificant main effects for diagnostic group and interac-tions between diagnostic group and gray/white andhemisphere or within-region Brodmann area effects. Toexamine symptom correlates of gray/white matter volumein prefrontal and temporal cortex,we used Pearson product-moment correlations. Lastly, because some studies haveidentified additional brain regions-of-interest, we reportexploratory t-tests on every Brodmann area as a supple-mentary table only (Appendix A).

3. Results

We present the mean volumes and standard deviationsfor each of the Brodmann areas by group in a Supple-mentary table (Appendix A). Our hypothesis-drivenMANOVA results are presented below.

3.1. Cingulate gyrus arch volume

A Group×Cingulate Brodmann area (BA25,24,31,23,29)×Matter type×Hemisphere MANOVA examinedwhether there were schizophrenia-spectrum differen-ces in relative cingulate gyrus volume. Compared withhealthy controls, both SPD and schizophrenia patients hadsignificantly less gray matter volume (averaged across allcingulate BAs and hemispheres; mean±SD: healthycontrols=0.226±0.024, SPD=0.215±0.018, schizophre-nia=0.218±0.023), SPDs had significantly greater white

ann area volumes are shown for the schizophrenia-spectrum groups.reatest cingulate volume reduction in BA24 and SPD patients were3, Wilks). Y-axis shows 100× (region-of-interest volume/total brain

Fig. 3. Differences from the healthy control group in gray/white matter volume of the cingulate gyrus showed significantly decreased gray and increasedwhite matter volume in the schizophrenia-spectrum groups (Group×Matter type×Brodmann area interaction for the cingulate arch, F[8,556]=3.73,p=.0003, WilksWilks). ⁎all pb0.04, Fisher's LSD test). “+” denotes SPDNSz patients, p=0.002, Fisher's LSD test.

116 E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

matter volume and the two patient groups did not differfrom each other (healthy controls=0.225±0.024, SPD=0.233±0.018, schizophrenia=0.230±0.023;Group×Mat-ter type interaction, F[2,281]=7.44, p=0.0007 significantFisher's LSD tests were pb0.002).

Compared with healthy controls, the schizophreniapatients displayed the greatest reduction in cingulatevolume in BA24 and SPD patients were intermediate,although none of the follow-up tests reached significance(Fig. 2). The gray/white matter effect was largest forBA24, 31, and 23 (Fig. 3). Follow-up tests revealed thatboth patient groups had significantly less gray and morewhite matter in BA31. SPD patients also showed reducedgray but increased white matter in BA24 and BA23 (allpb0.04, Fisher's LSD). White matter volume in BA24was significantly greater in SPD than schizophreniapatients and this was the only cingulate region where thepatient groups differed (p=0.002, Fisher's LSD). Themain effect of group was not significant (p=0.46), norwere any of the other interactions with group.

3.2. Anterior, orbital, and dorsolateral prefrontalcortex volume

A Group×Region (Anterior: BA 8,9,10; Orbital: BA11,12,47; Dorsolateral: BA 44,45,46) ×Brodmannarea×Matter type×Hemisphere MANOVA examinedwhether schizophrenia-spectrumdifferences existedwithinprefrontal cortex. The schizophrenia patients had signifi-

cantly reduced overall gray but not white matter volume inprefrontal cortex (averaged across hemisphere and all BAs)compared with the controls and SPD patients (Fig. 4).Compared with healthy controls, schizophrenia patientshad reduced volume in the anterior region of the prefrontalcortex, especially BA10, Fisher's while SPD patientsshowed greater-than-normal BA10 volume (Fig. 5). Inthe right hemisphere, both patient groups showed re-duced prefrontal volume while in the left hemisphere,schizophrenia but not SPD patients showed reducedvolume, predominantly in the anterior prefrontal region(BA8, 9, and 10; Group×Region×Hemisphere interac-tion, F [4,560]=5.41, p=.0003, Wilks; Group×Hemi-sphere interaction, F [2,281]=3.96, p=.02). Neither themain effect of group (p=0.15), nor any of the otherinteractions with group were significant.

3.3. Temporal lobe (BA22, 21, and 20) volume

A Group×Temporal Lobe BA Region (superior BA22,middle: BA21, and inferior temporal gyrus:BA20)×Mattertype×Hemisphere MANOVA was conducted to examinegroup differences in the temporal cortex. There was ahighly significant main effect of group which indicated aspectrum pattern with Healthy controlsNSPDNSchizo-phrenia patients for average relative temporal lobe vol-ume (F[2,281]=5.14, p=0.006; mean±SD: Healthycontrols=0.659±0.042, SPD patients=0.649±0.044,Schizophrenia patients=0.638±0.048). This spectrum

Fig. 4. Gray and white matter prefrontal cortex volume differences from normal are shown for the schizophrenia-spectrum groups. The schizophreniapatients showed significantly reduced gray matter volume in prefrontal cortex (averaged across hemisphere and all 9 prefrontal cortex BAs whichinclude BA 8, 9, 10, 11, 12, 47, 44, 45, 46) compared with the healthy controls (*NNSz, p=0.00001, Fisher's LSD test) and SPD patients (“+”denotes SPDNSz, p=0.009, Fisher's LSD test). Group×Matter type interaction, F[2,281]=4.09, p=0.018).

117E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

pattern was found in the left hemisphere while in the righthemisphere, the schizophrenia patients showed reducedvolume and the SPD patients were similar to the healthy

Fig. 5. Significant Group×Region×Brodmann area interaction (F[8,556]=7patients, p=0.003, Fisher's LSD test. +SPDNSz patients, p=0.0001, Fisher'seach of the three regions: anterior, orbital, and dorsolateral) was also signifi

controls (Group×Hemisphere interaction,F[2,281]=10.60,p=0.00004). A Group×Matter type interaction indicated aspectrum pattern for temporal lobe gray matter with

.48, p=0.000001, Wilks) is shown for the prefrontal cortex. ⁎CNSzLSD test. The Group×Region interaction (collapsed across BAs withincant, F[4,560]=3.28, p=0.01, Wilks.

118 E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

CNSPDNSz and the opposite SzNSPDNC for whitematter (F[2,281]=12.67, p=0.000005). Compared withcontrols, both patient groups showed significantly

Fig. 6. (TOP). A significant Group×Brodmann area×Matter type interact*significantly different from healthy controls, all p values between 0.04 and 0(BOTTOM). Group×Brodmann area×Hemisphere interaction, F [4,560]=2.p valuesb0.04, Fisher's LSD test. A Group×Brodmann area interaction (colindicated that the schizophrenia patients had smaller superior and middle tepb0.025, respectively, Fisher's LSD) while SPD patients only had reduced

reduced gray matter volume in the middle temporalgyrus, whereas only the schizophrenia patients showedsignificantly reduced gray matter volume in superior and

ion (F[4,560]=5.89, p=0.0001) was observed in the temporal lobe..000001, +SPDNSz patients, both pb0.0005, Fisher's LSD test. Fig. 679, p=0.026, Wilks. *significantly different from healthy controls, alllapsed across gray/white matter; F [4,560]=8.09, p=0.000002, Wilks)mporal gyrus volume compared with healthy controls (pb0.031 andmiddle temporal gyrus volume (pb0.002).

119E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

inferior temporal gyrus, and increased white mattervolume in the inferior temporal gyrus (Fig. 6-top).

In the superior temporal gyrus, the schizophreniapatients (p=0.03, Fisher's LSD test) but not the SPDpatients showed significantly reduced volume in the lefthemisphere compared with healthy controls (Fig. 6-bottom). In the middle temporal gyrus, both patient groupsshowed significantly reduced left (both pb0.013) but notright volume. In the inferior temporal gyrus there were nogroup differences in laterality (Fig. 6-bottom).

3.4. Absolute brain volume

The three groups did not differ on total brain volume(Healthy controls: mean=1216 cm3±SD=125; SPDpatients: mean =1205 cm3±125, schizophrenia patients:mean=1214 fscm3 ±100, F[2,281]=0.20, p=0.82).Total brain volume was determined by adding the volumeof gray and white matter for all Brodmann regions acrossthe 33 coronal slices.

In order to examine whether the three groups differedin absolute volume for the four lobes, we conducted aGroup×Lobe (frontal, parietal, temporal, occipital)×Matter type×Hemisphere MANOVA. Brodmann areasincluded for each lobe were as follows: frontal lobe: BA8,9,10,11,12,32,44,45,46, and 47; parietal lobe: BA7,39,and 40; temporal lobe: BA20,21,22,27,28,30,35,36,37,38,41, and 42; and occipital lobe: BA17,18, and 19. Thegroups did not differ in overall absolute volume (collapsed

Fig. 7. Group×Lobe×Matter type interaction, F[6, 558]=4.1

across all lobes, matter type, and hemisphere), F[2, 281]=1.27, p=0.28. Compared with the healthy controls, theschizophrenia patients showed reduced volume in all fourlobes, however, gray matter reduction in the temporal lobewas most striking (p=0.0036, Fisher's LSD test) andmoreprominent in the left hemisphere (Group×Lobe×Mattertype interaction, F[6,558]=4.12, p=0.00047, Wilks;Fig. 7; Group×Lobe interaction (F[6,558] = 2.41,p=0.026, Wilks; Group×Lobe×Hemisphere interaction,F[6,558]=2.49, p=0.02). Schizophrenia patients showednonsignificant increases in white matter in all four lobescompared with healthy controls. In contrast, SPD patientsshowed more mild gray matter reductions in frontal,parietal, and temporal cortex, and none in occipital cortex.The Group×Matter type interaction confirmed a spectrumpattern showing controls had greater overall gray mattervolume averaged across hemisphere and the four lobes,schizophrenia patients had the least, and SPD patientswere intermediate, while the pattern was reversed forwhite matter volume with schizophrenia patients hav-ing the greatest white matter volume (F[2,281]=9.74,p=0.00008). A Group×Hemisphere interaction indicatedthat overall volume reduction was marked in both hemi-spheres in schizophrenia while SPD patients were inter-mediate between healthy controls and schizophreniapatients in the left hemisphere but similar to controls inthe right (F[2,281]=8.37, pb0.0003). A MANOVA onrelative volume for the four lobes produced the same pat-tern of significant group differences (e.g., Group×Lobe×

2, p=0.00047, Wilks. *p=0.0036, Fisher's LSD test.

Table 2BPRS symptom correlates of gray and white matter volume inschizophrenia patients

Positive symptoms Negative symptoms

Matter: Matter:

G W G W

BA 24 ns ns ns nsBA 8 ns −0.30* −0.32* 0.41**BA 9 ns ns −0.30* nsBA 10 ns ns −0.37* nsBA 22 ns ns −0.47** 0.37*BA 21 ns ns −0.51** 0.43**BA 20 ns ns −0.41** 0.35*

*pb0.05, Pearson correlation, **pb0.0018, Bonferroni corrected forpb0.05.

120 E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

Matter type interaction, F[6,558]=4.20, pb0.0004,Wilks).

3.5. Symptom correlates

To minimize the number of tests, we conducted volumevs. clinical symptom correlations with only the seven BAsin prefrontal and temporal cortex where we found sig-nificant between-group volume differences (i.e. significantinteractions with Group: BA24, 8, 9, 10, 22, 21, and 20).

Among the schizophrenia patients, smaller gray andlarger white matter volume in fronto-temporal BAs wasassociated with greater severity of BPRS negative symp-toms (Table 2). In contrast, individual differences in posi-tive symptom severity were generally not associated withcortical volume.

Only two of the seven BAs examined showed signi-ficant correlations in the SPD group. Smaller gray mattervolume in BA24was associatedwith greater interpersonalimpairment (r=−0.31, p=.007). Smaller white mattervolume in BA22 (STG) was associated with greater cog-nitive impairment (r=−0.25, p=.026).

4. Discussion

The present study indicates schizophrenia patientsshowed reduced gray matter volume widely across thecortex but more profoundly in the frontal and temporallobes. The SPD patients similarly had reduced volume inthe frontal and temporal lobes only and the decrease wasabout half that observed in schizophrenia. Thus, for thefronto-temporal deficit, the schizophrenia-spectrumconcept is supported with SPD patients intermediatebetween healthy controls and schizophrenia patients.This spectrum finding suggests that SPD probably re-presents a milder form of disease along the schizophreniacontinuum which is consistent with the review of Dickey

et al. (2002a) and other morphometric studies involvingSPD and schizophrenia patients (Kawasaki et al., 2004;Suzuki et al., 2005; Takahashi et al., 2006b).

However, detailed analyses of each Brodmann areahave suggested additional possible variations (Buchsbaumet al., 2002). Four potential possibilities of normal-SPD-schizophrenia differences may be evaluated: a spectrumfor the brain region with SPD intermediate between nor-mal and schizophrenia, a sparing of the region with SPDthe same as normal, a protective factor with SPD largerthan normal, and a commonality with SPD and schizo-phrenia showing identical values. The whole prefrontalcortex, anterior cingulate (BA24) and left temporal lobe(BA21) are all examples of this spectrum relationship.Left temporal lobe BA22 shows very much smaller-than-normal volumes in schizophrenia patients than SPDpatients, indicating the sparing pattern. Polar frontallobe, BA10 is an example of a possibly protective orcompensatory area with total volume (averaged over grayandwhitematter) decreases in schizophrenia and increasesin SPD patients. Yet, the middle temporal gyrus (BA21)and posterior cingulate (BA31) are areas where commongray matter deficits in SPD and schizophrenia are seen.Taken together, this variety of patterns is consistent with amultiple gene model in which several deficits produceschizophrenia, fewer deficits produce SPD, and someprotective or modulatory factors ameliorate full develop-ment of schizophrenia in SPD.Thismodel originated fromthe hypothesis that schizophrenia results from the cu-mulative impact of multiple common small-effect, geneticvariants, interacting with environmental exposures to ex-ceed a biological threshold (Gottesman et al., 1982).

In the temporal lobe, patients with schizophrenia showreduced volume in both left hemisphere BA22 and 21,while SPD patients show small reductions in BA22 but fullreduction in BA21. Dickey (Dickey et al., 1999) foundabout a 10% decrease (0.03 for relative data) for BA22 inSPD but did not report schizophrenia data; our SPD de-crease in the current study was 0.02 for BA22 but 0.08for BA21. In our earlier study with a much smaller SPDsample (n=13 overlapping with current sample), largerdecreases were actually seen for SPD than patients withschizophrenia for bothBA22and 21.Our overall pattern oftemporal lobe volume decreases being less widespread inSPD than schizophrenia patients, in comparison withhealthy controls is consistent with other studies examiningthese three groups, (e.g.Kawasaki et al., 2004; Takahashiet al., 2006b). However, our finding of a shared volumereduction in SPD and schizophrenia patients in middletemporal gyrus is inconsistent with a voxel-based study thatreported shared volume reduction in the left superior but notmiddle temporal gyrus (Kawasaki et al., 2004). Differences

121E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

in tracing (Dickey and Downhill studies visually tracedonly the mid-portion of the temporal lobe and identified thesuperior temporal gyrus), Brodman area identificationmethods (we used a stereotaxic method for separatingareas 22 and 21, somewhatmore analogous to voxel-based-morphometry) of gray–white segmentation (thresholdmethod used here), and symptom severity may be im-portant. Both the Kawasaki and Takahashi studies used aclinic-based SPD sample taking low-dose antipsychoticswhile we studied an unmedicated community samplesuggesting these samples may differ in illness severity.

In the prefrontal cortex, the schizophrenia patientsshowed overall volume loss in gray but not white matterwhile the SPD patients did not differ from normal. Thiseffect was most marked in BA10 where the schizophreniapatients had smaller volume compared with the healthycontrols while SPD patients had greater-than-normalvolume (although not significantly more). This may resultfrom some variation sparing SPD patients from the fulldevelopment of schizophrenia, an adaptive response tosymptoms of the schizophrenia spectrum, or an associateddevelopmental anomaly among other possibilities. Con-sistent with our findings, Suzuki et al. (2005) demonstratedthat the prefrontal volumes were largely preserved in SPDin contrast to widespread prefrontal involvement in schi-zophrenia. It is of interest that in our prior FDG-PET study(Buchsbaum et al., 2002) using a task involving verbalworking memory in a subgroup of subjects from this study,metabolic rates in BA10 were distinctly higher in SPDpatients compared with healthy controls while schizophre-nia patients were lowest and this was determined in thesame way as in our present study.

In contrast to the prefrontal and temporal cortex findingswhere the schizophrenia patients showed more gray mattervolume loss, the SPD patients showed greater differencesfrom normal in cingulate gyrus gray matter. SPD patientshad reduced gray matter volume primarily in BA24, 31,and 23 compared to normal while schizophrenia patientsshowed a reduction in only BA31. However, because SPDpatients showed significantly greater white matter volumein these same Brodmann areas, the net volume loss in theseareas was greater in schizophrenia than SPD. In otherwords, averaged across all the cingulate BAs, both the SPDand schizophrenia patients had reduced gray mattervolume, but the SPD patients had greater-than-normalwhite matter volume. In a separate sample of chronicpatients with schizophrenia, we also observed the largestgray matter reduction in BA24 and 31 (Mitelman et al.,2005b) and others have also found posterior cingulatevolume reduction, (Zhou et al., 2005) consistent with thecurrent results. We previously reported no normal-SPDgroup differences (Haznedar et al., 1997) in the cingulate in

a subgroup (n=12) of the SPD patients from the currentstudy which suggests that large sample sizes are importantfor detecting subtle group differences. We have discussedelsewhere in detail (Mitelman et al., 2005b) that abnorm-alities in the posterior cingulate and BA31 may involvethought disorder and verbal memory. Social anhedonia,important in schizophrenia, is also prominent in SPDpatients (Horan et al., 2007). A factor-analytic study of theBeck Depression Inventory and FDG-PET linked BA31with anhedonia in unipolar patients (Dunn et al., 2002).

The schizophrenia-spectrum fronto-temporal volumeabnormalities we observed were related to symptom se-verity. In patients with schizophrenia, negative symptomseverity of the illness as assessed with the BPRS wasassociated with reduced frontopolar (BA8,9,10) and tem-poral lobe gray matter volume and increased white matter.Among the SPD patients, smaller BA24 gray mattervolume was associated with greater interpersonal impair-ment, while smaller STG white matter volume was asso-ciated with greater cognitive impairment. Previous MRIwork reported smaller STG volume was associated withmore positive symptoms in schizophrenia but that nosignificant correlations emerged among the SPD patients(Takahashi et al., 2006a). Other work indicates prefrontaland temporal gray matter volume is unrelated to variousclinical symptoms in schizophrenia while reduced whitematter volume in these regions is associated with negativesymptoms (Sanfilipo et al., 2000). Differences in patientcharacteristics may explain these differing findings.

This study has a large size with attendant power todetect small differences, an important aspect when com-paring SPD and normal groups. However, due to the rangeof diagnoses, the same symptom severity estimation scaleswere not available for both SPD and schizophreniapatients. Family history data would have been potentiallyinformative, but was not collected in a formal structuredgenetic interview, and so was not included in the analyses.Past treatment with neuroleptics may have had somemorphological effect. All of the patients in our study wereoff medication at the time of theirMRI scan.While 77% ofthe SPD patients had never previously been medicated,51% of the schizophrenia patients met this criterion. Thus,chronic effects of medication could have been more pro-nounced in the schizophrenia group, although the pro-portion difference is not large. Smaller middle temporalgyrus gray matter volumes in first-episode patients withschizophrenia but not affective psychosis (Kuroki et al.,2006) suggest that our finding of a reduction in this sameregion in both SPD and schizophrenia is not due to priormedication treatment, instead it is a schizophrenia-spectrum specific abnormality. Future work comparingnever-medicated andpreviously-medicated schizophrenia-

122 E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

spectrum patients is needed to fully address the issue ofmedication effects. Recently, a lack of correlation betweenduration of untreated psychosis and cingulate or frontalsubregion volume was found (Takahashi et al., 2007) butthe left superior planum temporale volume was negativelycorrelated.Hospitalization is presumablymore common inthe schizophrenia group, so the effects of the inpatientenvironment on cortical size cannot be ruled out.

Our data analysis tends to reveal gray matter deficitsassociated with white matter increases. Since each Brod-mann area is expressed as relative to whole brain, wide-spread mm3 decreases in gray matter might be reflected inrelative white matter increases and since we found graymatter decrease in schizophrenia across all four lobeswhenmm3 datawas analyzed, thismust contribute to the relativewhite matter increase. However, the white matter increasein patients was also widespread when absolute mm3 datawere analyzed. Thus, there are global tissue findings, aswell as, regional ones: bothBA8 andBA10 break the gray-down/white-up rule. A detailed analysis with adjustmentof regional volumes for brain size estimated from bodysize may help to address this question.

In conclusion, our findings support the concept thatSPD is in the schizophrenia spectrum and the frontal andtemporal volume decreases are generallymidway betweennormal and schizophrenia values. Our results suggest thatincreased volume in prefrontal cortex (BA10) and sparingof volume loss in temporal cortex (BA22 and BA20) maybe a protective factor in SPD which reduces vulnerabilityto psychosis.

Role of funding sourceThe formulation of this manuscript has had full support of the

study sponsors recognized in the acknowledgments. No interferencehas been made by them in the study design; collection, analysis, andinterpretation of data; and our sponsors served no role in the writing ofthe report; or in the decision to submit the paper for publication.

ContributorsDrs. Hazlett, Buchsbaum and Siever conceived the idea of the study.

Dr. Buchsbaum conceived the methods for the cortical volume analysis.Dr. Hazlett conducted the statistical analyses and wrote the paper. Drs.Hazlett, Haznedar and New helped recruit and diagnose the studyparticipants. R. Newmark, K.E. Goldstein, Y. Zelmanova, C.F. Glanton,Y. Torosjan, J.N. Lo, and V. Mitropoulou provided technical support(data acquisition and processing). Drs. Buchsbaum and Siever alsocontributed to the write up. All authors approved the final manuscript.

Conflict of interestThere are no conflicts of interest for any of the authors.

AcknowledgmentThis work was supported in part by an Independent Investigator

Award from NARSAD (Dr. Hazlett), NIMH grants (Dr. Hazlett:MH073911; Dr. Buchsbaum: MH40071, MH56489, and MH60023;

Dr. Siever: MH56606), a VA Merit Award (Dr. Siever), a grant fromthe National Center for Research Resources (M01-RR00071) awarded

to the General Clinical Research Center, Mount Sinai School ofMedicine, and a MIRECC VISN3 award.

Appendix A. Supplementary data

Supplementary data associated with this articlecan be found, in the online version, at doi:10.1016/j.schres.2007.12.472.

References

Andreasen, N., Flaum, M., Arndt, S., 1992. The Comprehensive Assess-ment of Symptoms and History (CASH): an instrument for assessingdiagnosis and psychopathology. Arch. Gen. Psychiatry 49, 615–623.

Bergman, A.J., Harvey, P.D., Mitropoulou, V., Aronson, A., Marder, D.,Silverman, J., et al., 1996. The factor structure of schizotypal symp-toms in a clinical population. Schizophr. Bull. 22, 501–509.

Buchsbaum, M., Yang, S., Hazlett, E., Siegel, B., Germans, M.,Haznedar, M., et al., 1997. Ventricular volume and asymmetry inschizotypal personality disorder and schizophrenia assessed withmagnetic resonance imaging. Schizophr. Res. 27, 45–53.

Buchsbaum, M.S., Nenadic, I., Hazlett, E.A., Spiegel-Cohen, J.,Fleischman, M.B., Akhavan, A., et al., 2002. Differential metabolicrates in prefrontal and temporal Brodmann areas in schizophrenia andschizotypal personality disorder. Schizophr. Res. 54, 141–150.

Davidson, L.L., Heinrichs, R.W., 2003. Quantification of frontal andtemporal lobe brain-imaging findings in schizophrenia: a meta-analysis. Psychiatry Res. 122, 69–87.

Dickey, C.C., McCarley, R.W., Voglmaier, M.M., Niznikiewicz, M.A.,Seidman, L.J., Hirayasu, Y., et al., 1999. Schizotypal personalitydisorder and MRI abnormalities of temporal lobe gray matter. Biol.Psychiatry 45, 1393–1402.

Dickey, C.C., Shenton, M.E., Hirayasu, Y., Fischer, I., Voglmaier, M.M.,Niznikiewicz, M.A., et al., 2000. Large CSF volume not attributableto ventricular volume in schizotypal personality disorder. Am. J.Psychiatry 157, 48–54.

Dickey, C.C., McCarley, R.W., Shenton, M.E., 2002a. The brain inschizotypal personality disorder: a review of structural MRI andCT findings. Harv. Rev. Psychiatry 10, 1–15.

Dickey, C.C., McCarley, R.W., Voglmaier, M.M., Frumin, M.,Niznikiewicz, M.A., Hirayasu, Y., et al., 2002b. Smaller leftHeschl's gyrus volume in patients with schizotypal personalitydisorder. Am. J. Psychiatry 159, 1521–1527.

Dickey, C.C., McCarley, R.W., Xu, M.L., Seidman, L.J., Voglmaier,M.M., Niznikiewicz, M.A., et al., 2007. MRI abnormalities of thehippocampus and cavum septi pellucidi in females with schizo-typal personality disorder. Schizophr. Res. 89, 49–58.

Diwadkar, V.A., Montrose, D.M., Dworakowski, D., Sweeney, J.A.,Keshavan, M.S., 2006. Genetically predisposed offspring withschizotypal features: an ultra high-risk group for schizophrenia?Prog. Neuro-psychopharmacol. Biol. Psychiatry 30, 230–238.

Downhill, J.E., Buchsbaum, M.S., Hazlett, E.A., Barth, S., LeesRoitman, S., Nunn, M., et al., 2001. Temporal lobe volume deter-mined by magnetic resonance imaging in schizotypal personalitydisorder and schizophrenia. Schizophr. Res. 48, 187–199.

Dunn, R.T., Kimbrell, T.A., Ketter, T.A., Frye, M.A., Willis, M.W.,Luckenbaugh, D.A., et al., 2002. Principal components of the BeckDepression Inventory and regional cerebral metabolism in unipolarand bipolar depression. Biol. Psychiatry 51, 387–399.

Gottesman, I., Shields, J., Hanson, D.R., 1982. Schizophrenia, TheEpigenetic Puzzle. Cambridge University Press, New York.

123E.A. Hazlett et al. / Schizophrenia Research 101 (2008) 111–123

Gur, R.E., Turetsky, B.I., Cowell, P.E., Finkelman, C., Maany, V.,Grossman, R.I., et al., 2000. Temporolimbic volume reductions inschizophrenia. Arch. Gen. Psychiatry 57, 769–775.

Hazlett, E.A., Buchsbaum, M.S., Haznedar, M.M., Singer, M.B., Schnur,D.B., Jimenez, E.A., et al., 1998. Prefrontal cortex glucose me-tabolism and startle eyeblink modification abnormalities in unmedi-cated schizophrenia patients. Psychophysiology 35, 186–198.

Haznedar, M.M., Buchsbaum,M.S., Metzger, M., Solimando, A., Spiegel-Cohen, J., Hollander, E., 1997. Anterior cingulate gyrus volumeand glucose metabolism in autistic disorder. Am. J. Psychiatry 154,1047–1050.

Haznedar,M.M., Buchsbaum,M.S., Hazlett, E.A., Shihabuddin, L., New,A., Siever, L.J., 2004. Cingulate gyrus volume and metabolism in theschizophrenia spectrum. Schizophr. Res. 71, 249–262.

Hirayasu, Y., McCarley, R.W., Salisbury, D.F., Tanaka, S., Kwon, J.S.,Frumin, M., et al., 2000. Planum temporale and Heschl gyrusvolume reduction in schizophrenia: a magnetic resonance imagingstudy of first-episode patients. Arch. Gen. Psychiatry 57, 692–699.

Horan, W.P., Brown, S.A., Blanchard, J.J., 2007. Social anhedonia andschizotypy: the contribution of individual differences in affectivetraits, stress, and coping. Psychiatry Res. 149, 147–156.

Kawasaki, Y., Suzuki, M., Nohara, S., Hagino, H., Takahashi, T., Matsui,M., et al., 2004. Structural brain differences in patients with schi-zophrenia and schizotypal disorder demonstrated by voxel-basedmorphometry. Eur. Arch. Psychiatry Clin. Neurosci. 254, 406–414.

Koo,M.S., Dickey, C.C., Park, H.J., Kubicki,M., Ji, N.Y., Bouix, S., et al.,2006. Smaller neocortical gray matter and larger sulcal cerebrospinalfluid volumes in neuroleptic-naive women with schizotypal person-ality disorder. Arch. Gen. Psychiatry 63, 1090–1100.

Kuroki, N., Shenton, M.E., Salisbury, D.F., Hirayasu, Y., Onitsuka, T.,Ersner-Hershfield, H., et al., 2006. Middle and inferior temporalgyrus gray matter volume abnormalities in first-episode schizo-phrenia: an MRI study. Am. J. Psychiatry 163, 2103–2110.

Mitelman, S., Shihabuddin, L., Brickman, A., Hazlett, E., Buchsbaum,M.S., 2003. MRI assessment of gray and white matter distributionin Brodmann areas of the cortex in patients with schizophreniawith good and poor outcomes. Am. J. Psychiatry 160, 2154–2168.

Mitelman, S.A., Buchsbaum, M.S., Brickman, A.M., Shihabuddin, L.,2005a. Cortical intercorrelations of frontal area volumes in schizo-phrenia. Neuroimage 27, 753–770.

Mitelman, S.A., Shihabuddin, L., Brickman, A.M., Hazlett, E.A.,Buchsbaum, M.S., 2005b. Volume of the cingulate and outcome inschizophrenia. Schizophr. Res. 72, 91–108.

Mitropoulou, V., Harvey, P.D., Maldari, L.A.,Moriarty, P.J., New, A.S.,Silverman, J.M., et al., 2002. Neuropsychological performance inschizotypal personality disorder: evidence regarding diagnosticspecificity. Biol. Psychiatry 52, 1175–1182.

Opler, L.A., Kay, S.R., Rosado, V., Lindenmayer, J., 1984. Positiveand negative syndromes in chronic schizophrenic inpatients. J. ofNerv. Ment. Dis. 172, 317–325.

Overall, J., Gorham, D., 1962. The brief psychiatric rating scale. Psychol.Rep. 10, 799–812.

Perry, R., Oakley, A., Perry, E., 1991. Coronal Brain Map and DissectionGuide: Localization of Brodman Areas in Coronal Sections.

Pfohl, B., Blum, N., M.Z., 1989. Structured Interview for DSM-III-RPersonality (SIDP-R). University of Iowa, Iowa City.

Raine, A., Sheard, C., Reynolds, G.P., Lencz, T., 1992. Prefrontalstructural and functional deficits associated with individualdifferences in schizotypal personality. Schizophr. Res. 7, 237–247.

Sanfilipo, M., Lafargue, T., Rusinek, H., Arena, L., Loneragan, C.,Lautin, A., et al., 2000. Volumetric measure of the frontal and

temporal lobe regions in schizophrenia: relationship to negativesymptoms. Arch. Gen. Psychiatry 57, 471–480.

Seidman, L.J., Faraone, S.V., Goldstein, J.M., Kremen, W.S., Horton,N.J., Makris, N., et al., 2002. Left hippocampal volume as avulnerability indicator for schizophrenia: a magnetic resonanceimaging morphometric study of nonpsychotic first-degree rela-tives. Arch. Gen. Psychiatry 59, 839–849.

Seidman, L.J., Pantelis, C., Keshavan, M.S., Faraone, S.V., Goldstein,J.M., Horton, N.J., et al., 2003. A review and new report of medialtemporal lobe dysfunction as a vulnerability indicator for schizo-phrenia: a magnetic resonance imaging morphometric family studyof the parahippocampal gyrus. Schizophr. Bull. 29, 803–830.

Shenton, M.E., Dickey, C.C., Frumin, M., McCarley, R.W., 2001. Areview ofMRI findings in schizophrenia. Schizophr. Res. 49, 1–52.

Siever, L.J., Davis, K.L., 2004. The pathophysiology of schizophreniadisorders: perspectives from the spectrum. Am. J. Psychiatry 161,398–413.

Siever, L., Silverman, J., Horvath, T., Klar, H., Coccaro, E., Keefe, R.,et al., 1990. Increased morbid risk for schizophrenia-related dis-orders in relatives of schizotypal personality disordered patients.Arch. Gen. Psychiatry 47, 634–640.

Siever, L., Kalus, O., Keefe, R., 1993. The boundaries of schizo-phrenia. Psychiatry Clin. N. Am. 15, 217–244.

Siever, L.J., Koenigsberg, H.W., Harvey, P., Mitropoulou, V., Laruelle,M., Abi-Dargham, A., et al., 2002. Cognitive and brain function inschizotypal personality disorder. Schizophr. Res. 54, 157–167.

Spitzer, R.L., Williams, J.B., Gibbon, M., First, M.B., 1992. TheStructured Clinical Interview for DSM-III-R (SCID). I: history,rationale, and description. Arch. Gen. Psychiatry 49, 624–629.

StatSoft, I., 2003. Statistica, 6.0 ed. OK, Tulsa. www.statsoft.com.Suzuki, M., Zhou, S.Y., Takahashi, T., Hagino, H., Kawasaki, Y., Niu, L.,

et al., 2005. Differential contributions of prefrontal and temporolimbicpathology to mechanisms of psychosis. Brain 128, 2109–2122.

Takahashi, T., Suzuki, M., Zhou, S.Y., Hagino, H., Kawasaki, Y.,Yamashita, I., et al., 2004. Lack of normal gender differences of theperigenual cingulate gyrus in schizophrenia spectrum disorders.A magnetic resonance imaging study. Eur. Arch. Psychiatry Clin.Neurosci. 254, 273–280.

Takahashi, T., Suzuki,M.,Zhou, S.Y., Tanino,R.,Hagino,H.,Kawasaki,Y.,et al., 2006a. Morphologic alterations of the parcellated superior tem-poral gyrus in schizophrenia spectrum. Schizophr. Res. 83, 131–143.

Takahashi, T., Suzuki,M.,Zhou, S.Y., Tanino,R.,Hagino,H.,Niu, L., et al.,2006b. Temporal lobe gray matter in schizophrenia spectrum: avolumetric MRI study of the fusiform gyrus, parahippocampal gyrus,and middle and inferior temporal gyri. Schizophr. Res. 87, 116–126.

Takahashi, T., Suzuki, M., Tanino, R., Zhou, S.Y., Hagino, H., Niu, L.,et al., 2007. Volume reduction of the left planum temporale graymatter associated with long duration of untreated psychosis inschizophrenia: a preliminary report. Psychiatry Res. 154, 209–219.

Talairach, J., Tournoux, P., 1988. Co-planar Stereotaxic Atlas of theHuman Brain. Thieme, Stuttgart.

Wright, I.C., Rabe-Hesketh, S., Woodruff, P.W., David, A.S., Murray,R.M., Bullmore, E.T., 2000. Meta-analysis of regional brainvolumes in schizophrenia. Am. J. Psychiatry 157, 16–25.

Zhou, S.Y., Suzuki, M., Hagino, H., Takahashi, T., Kawasaki, Y., Matsui,M., et al., 2005. Volumetric analysis of sulci/gyri-defined in vivofrontal lobe regions in schizophrenia: precentral gyrus, cingulategyrus, and prefrontal region. Psychiatry Res. 139, 127–139.

Zhou, S.Y., Suzuki, M., Takahashi, T., Hagino, H., Kawasaki, Y.,Matsui, M., et al., 2007. Parietal lobe volume deficits inschizophrenia spectrum disorders. Schizophr. Res. 89, 35–48.