Correlation between P50 suppression and psychometric schizotypy among non-clinical Japanese subjects

International Journal of Psychophysiology 52(2004) 147–157

0167-8760/04/$ - see front matter� 2003 Elsevier B.V. All rights reserved.doi:10.1016/j.ijpsycho.2003.06.001

Correlation between P50 suppression and psychometric schizotypyamong non-clinical Japanese subjects

Jijun Wang *, Hiroshi Miyazato , Hiroto Hokama , Ken-Ichi Hiramatsu , Tsuyoshi Kondoa, a,b a a a

Department of Neuropsychiatry, Faculty of Medicine, University of the Ryukyus, 207 Uehara, Nishihara, Okinawa 903-0215,a

JapanMiyazato Hospital, 1763-2 Umusa, Nago-shi, Okinawa 905-0006, Japanb

Received 11 March 2003; received in revised form 6 June 2003; accepted 10 June 2003

Abstract

The link between P50 suppression and psychometric schizotypy was previously reported in non-clinical English-speaking subjects; however, whether a similar relationship exists within a different ethnic sample is unknown.Furthermore, whether such a relationship can also be accounted for by such basic personality characteristics asextraversion or neuroticism has not yet been reported. In the present study, we investigated the correlations of P50suppression with psychometric schizotypy, and with extraversion or neuroticism among non-clinical Japanese. Subjectswere 34 healthy volunteers. The auditory P50 potential was obtained using a paired stimulus paradigm. Psychometricschizotypy was assessed using schizotypal personality questionnaire(SPQ). Extraversion and neuroticism wereassessed using Maudsley personality inventory(MPI). P50 suppression correlated not only with total SPQ score, butalso with extraversion and with neuroticism. However, the partial correlation analysis revealed a significant partialcorrelation of P50 suppression with SPQ when controlled for extraversion or neuroticism, and a non-significant partialcorrelation of P50 suppression with extraversion or neuroticism when controlled for SPQ. When subjects were dividedinto two subgroups according to the mean SPQ score, the degree of P50 suppression was lower in the high than inthe low SPQ scorers. Our results indicate that P50 suppression is one of the neurobiological substrates underlyingpsychometric schizotypy, and that this relationship cannot be accounted for by measures of extraversion or neuroticism.� 2003 Elsevier B.V. All rights reserved.

Keywords: P50 suppression; Psychometric schizotypy; Extraversion; Neuroticism

1. Introduction

Increased interest in studying psychometric schi-zotypy led to the finding that the schizotypal traitsof non-clinical population resemble the syndromeof schizophrenia in the factorial structure. Three

*Corresponding author. Tel.:q81-98-895-1157; fax:q81-98-895-1419.

E-mail addresses: [email protected](J. Wang),[email protected](J. Wang).

major factors extracted from schizotypal traits ofthe normal population measured using schizotypalpersonality questionnaire(SPQ, Raine, 1991) werewithdrawn, unreality and active(Gruzelier, 1996;Gruzelier and Doig, 1996), or interpersonal, cog-nitive-perceptual and disorganized(Raine et al.,1994). A similar three-dimensional model was alsoconfirmed in psychiatric patients without psychoticsymptoms(Vollema and Hoijtink, 2000). Studies

148 J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

of schizophrenia symptomatology identified simi-lar factors in schizophrenic patients, namely, onenegative and two positive syndromes(Liddle andMorris, 1991; Gur et al., 1991). It is suggestedthat three factors of schizophrenic symptomatologymay reflect an exaggeration of the three analogousfactors in the general population(Raine et al.,1994). The remarkable similarity between schizo-typy and schizophrenia in the factorial structuremight be partly associated with the same neuralsubstrate abnormalities.P50 suppression, which measures the brain func-

tion of sensory gating, might be such a neurobiol-ogical candidate. It has long been hypothesizedthat subjects with schizophrenia spectrum disordershave deficits in inhibitory functioning that lead todifficulty in filtering trivial internal and externalstimuli, perhaps accounting for the observed atten-tion and cognitive abnormalities. Deficiency in theearly sensory gating mechanism has been consis-tently demonstrated in schizophrenic patients(forinstance, Adler et al., 1982; Clementz et al., 1998a;etc) and their first-degree relatives(Clementz etal., 1998b; Siegel et al., 1984; Waldo et al., 1991;Myles-Worsley, 2002). The P50 suppression defi-ciency was also revealed in schizotypal personalitydisorder (SPD) subjects as compared with unaf-fected controls(Cadenhead et al., 2000), suggest-ing an association of P50 suppression deficits withschizotypal traits as well as with familial geneticvulnerability to schizophrenia. The P50 suppres-sion dysfunction was further related to psychomet-ric schizotypy in a non-clinical population by Croftet al. (2001), with the finding of a lesser degreeof P50 suppression in subjects reporting moreperceptual anomalies and magical ideation(anunreality factor). However, whether the relation-ship between P50 suppression and psychometricschizotypy exists within a different ethnic popula-tion remains unknown. It is possible that person-ality traits might be ethnically biased(Reynolds etal., 2000; Chen et al., 1997). The purpose of thepresent study was to replicate the study by Croftet al.(2001) in a Japanese population. In addition,P50 suppression tests were also administered toone group of schizophrenic patients in the presentstudy, in order to investigate whether the highscorers on schizotypal questionnaires are similar

to schizophrenics in terms of P50 suppression(Croft et al., 2001).While the study of schizotypy in the normal

population can be a useful strategy for elucidatingthe schizophrenic process, controversy over thenature of schizotypy remains. There are two broad-ly different perspectives on schizotypy: a psychi-atric perspective, where schizotypy representsattenuated psychotic symptoms, and a psycholog-ical perspective, where schizotypy represents devi-ant personality traits(Gruzelier, 2003). From thepsychological perspective, it is hypothesized thatall functional psychoses share an underlying latenttrait (psychoticism) which extends beyond psy-chotic states and can be measured in normalpopulation(Eysenck and Barrett, 1993). Differentvarieties of psychosis and schizotypy are believedto be manifestations of the other two major per-sonality constructs(extraversion and neuroticism)(Eysenck, 1992). Given that the close relationshipbetween schizotypy and basic personality traits hasbeen reported extensively(for instance, Braun-stein-Bercovitz et al., 2002; Ross et al., 2002), itwould be more advantageous to investigate wheth-er P50 suppression correlate with such basic per-sonality traits as extraversion or neuroticismsimultaneously. This approach might help to clarifywhether the link between P50 suppression andpsychometric schizotypy is specific, or just aneffect secondary to the correlation of P50 suppres-sion with more basic personality traits such asextraversion or neuroticism.

2. Methods

2.1. Subjects

Subjects were 34 paid healthy volunteers, all ofwhom had provided written informed consent afterreceiving an explanation of this study. They wererecruited from the staff and students of Universityof the Ryukyus, Okinawa, Japan. The demographiccharacteristics of the subjects is shown in Table 1.Subjects reported no personal history of neurologicor psychiatric disorders and no familial history ofpsychiatric illness. There were six light smokers,who smoked approximately 10 cigarettes per day.All subjects were right-handed(Edinburgh hand-

149J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

Table 1The demographic characteristics and means of P50 measurements

Non-clinical subjects Schizophrenic

Total subjects Low SPQ High SPQpatients

Cases 34 21 13 18Gender(myf) 16y18 8y13 8y5 13y5Age 33.4"8.2 32.9"8.5 35.2"7.8 35.7"7.8S1 latency(ms) 52.2"10.0 51.8"9.7 53.0"10.8 52.7"10.8S2 latency(ms) 49.5"10.5 50.2"10.6 48.2"10.5 53.6"11.5S1 amplitude(mV) 1.72"0.78 1.90"0.89 1.42"0.44 1.84"0.91S2 amplitude(mV) 0.48"0.52 0.31"0.40 0.76"0.59 1.31"1.07

P50 suppressionS2yS1 ratio 0.30"0.34 0.15"0.22 0.54"0.36 0.87"0.67S1–S2 difference(mV) 1.24"0.86 1.59"0.81 0.67"0.62 0.54"1.45

edness inventory, Oldfield, 1971). Schizotypaltraits were assessed using the Japanese translationof the 74 item Schizotypal personality question-naire(SPQ) (Raine, 1991; translated by Fujiwara,1993). The questionnaire was administered to 258college students(167 males and 91 females) inShiga University of Medical Science, Shiga, Japan.Compared with the study by Raine(1991), theresults showed a lower mean SPQ score of8.14"7.77 but a similar internal reliability(as0.90) and a similar criterion validity(rs0.61,Ns31, P-0.001) (Someya et al., 1994). Basic fea-tures of personality were assessed using the Japa-nese version of Jensen’s 80 item Maudsleypersonality inventory(MPI) (Jensen, 1958; Iwa-waki, 1969), which mainly includes two-dimen-sions: extraversion(24 items) and neuroticism(24items). MPI was administered to subjects at aninterval of 5.9"3.5 months later; however, datafor five of the subjects were unobtainable.Schizophrenic patients were recruited from

Miyazato Hospital, Okinawa, Japan, in the presentstudy. Eighteen medicated patients who met DSM-IV criteria for schizophrenia were included. Theantipsychotic medication was administered at anequivalent chlorpromazine dosage of 657"266mgyday. Clinical symptoms were evaluated usingthe positive and negative syndrome scale(PANSS)(Kay et al., 1987; general means28.6, S.D.s10.6; positive scale means13.8, S.D.s6.7; nega-tive scale means15.7, S.D.s6.1). There was nogroup difference in either age(F s0.66, P)1,50

0.05) or gender composition(x s3.02,Ps0.073)2

between schizophrenic patients and the non-clini-cal subjects.

2.2. Procedures

The recording procedures used in the Depart-ment of Anatomy, University of Arkansas forMedical Science, USA, were followed(Skinner etal., 1999). Subjects were seated on an armchair ina well-lit, sound-attenuating shielded room. Gold-plated surface electrodes were used, and electroderesistance was maintained at-5 KV. The P50potential was recorded at the vertex(Cz) referredto a frontal electrode(Fz). Eye movements(EOG)were detected by using diagonally placed canthalelectrodes, while jaw movements(electromyo-gram, EMG) were detected by using a lead overthe mentalis muscle referred to the chin. A sub-clavicular ground was used instead of mastoid orearlobe since the subjects wore headphones duringthe recording. Prior to the recording, the hearingthreshold was tested for each subject. The teststimulus was a rarefield click of 0.1 ms durationset at least 50 dB above threshold, usually 95–105 dB, as required. Paired clicks of S1–S2 at aninterstimulus interval of 250 ms were deliveredonce every 5 s to the subjects. This interstimulusinterval of the paired clicks was selected in thepresent study, because several previous reportssuggested that a shorter ISI(250 ms) is moresensitive to differences in sensory gating of the

150 J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

P50 potential(Rasco et al., 2000; Skinner et al.,2002). Signals were collected at 1 sampleyms,recorded and processed using a Synax 2100(NECCorporation, Tokyo, 1999). The signal was filteredat 2 Hz–1 kHz for the EEG channel, at 2 Hz–1kHz for the EOG channel and at 20 Hz–1 kHzfor the EMG channel. EEG signals that containedinterference from EOG or EMG were excludedfrom the average. The recording terminated when64 acceptable trials were averaged. If)8 trials(12.5%) were excluded because of this criterionin order to obtain 64 trials for the average, thesubjects were removed from the study. Two sub-jects were removed from the study for failing tomeet this criterion. The subjects were instructed tokeep their eyes open and to count the number oftrials presented as a means of maintaining vigi-lance. Only subjects who reported)90% accuracyin stimulus counts were included in the presentstudy. None of the subjects were excluded becauseof this criterion.

2.3. Data analysis

The P50 potential was identified as the largestpositive amplitude wave occurring between 30 and80 ms after the stimulus. P50 amplitude measureswere performed by the classic peak-to-peak meth-od, measuring the amplitude from the precedingnegativity (Nb), or from the preceding baseline ifNb was absent. P50 suppression was determinedusing the ratio of P50 amplitude in response to thesecond stimulus(S2) divided by the P50 amplitudein response to the first stimulus(S1). It was alsoassessed using the difference between P50 ampli-tudes in response to S1 and S2. In order to preventoutliers from disproportionately affecting the groupmeans, the S2yS1 ratios greater than two wereassigned the value 2, according to the methods ofNagamoto et al.(1991).The correlations of P50 measurements with total

SPQ score and its three factor scores were com-puted using Spearman correlating analysis(Bon-ferroni adjustment used for three factor scores;a9s0.0167). Spearman’s measure of associationwas also used to test for relations between P50and two MPI factors(Bonferroni adjustment used;a9s0.025). Exploratory partial correlating analy-

ses were performed on P50 suppression with SPQwhen adjusting for the effect of extraversion orneuroticism, and also on P50 suppression withextraversion or neuroticism when adjusting for theeffect of SPQ. Before the partial correlating anal-yses, the ratio of S2yS1 amplitude was transformedinto its square root to minimize the skewed distri-bution. No transformation was done with the dif-ference between S1 and S2 amplitudes because ofits normal distribution. Pearson correlation analysiswas also performed between S1–S2 difference andpsychometric measures, and between SPQ andMPI measures.The non-clinical subjects were further divided

into two subgroups of high and low SPQ scorers,according to the mean SPQ score of all subjects.There were 13 subjects with an SPQ score abovethe mean, and 21 subjects with an SPQ scorebelow the mean. The two subgroups did not differin age (F s0.60, Ps0.444) and the gender1,32

composition(x s1.77,Ps0.291). Subgroup dif-2

ferences in P50 measurements were assessed usingMann–Whitney test. The group differences in P50variables between the non-clinical subjects andschizophrenic patients were also evaluated usingMann–Whitney test.

3. Results

3.1. Psychometric assessments

The mean total SPQ score of all subjects was10.5"7.7. The mean SPQ factor score was5.9"5.4 for the withdrawn, 3.0"3.1 for the unre-ality and 2.3"2.1 for the active. MPI assessmentsrevealed a mean score of 28.1"11.7 for theextraversion dimension, and 12.0"8.4 for theneuroticism dimension. The intercorrelationsbetween SPQ and MPI are shown in Table 2. Theextraversion dimension had a correlation with thetotal SPQ score(rsy0.494,Ps0.007) and withthe withdrawn factor(rsy0.694,P-0.001).

3.2. Correlations of between P50 measurementsand psychometric assessments

Spearman’s correlations between P50 measure-ments and psychometric assessments are shown in

151J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

Table 2Pearson’s correlation coefficients amongst psychometric variables

SPQ(ns34) MPI (ns29)

Total Withdrawn Unreality Active Extraversion Neuroticism

SPQTotal r 1.0 0.878 0.703 0.754 y0.494 0.380

P – -0.001 -0.001 -0.001 0.007 0.042Withdrawn r 1.0 0.360 0.466 y0.694 0.238

P – 0.036 0.006 -0.001 0.214Unreality r 1.0 0.563 0.067 0.375

P – 0.001 0.730 0.045Active r 1.0 y0.193 0.375

P – 0.316 0.045

MPIExtraversion r 1.0 y0.172

P – 0.374Neuroticism r 1.0

P –

Table 3Spearman’s correlation coefficients(andP-values) between P50 and psychometric variables

SPQ(ns34) MPI (ns29)

Total Withdrawn Unreality Active Extraversion Neuroticism

S1 latency r 0.142 0.061 0.107 0.284 y0.007 y0.090P 0.425 0.734 0.548 0.104 0.972 0.644

S2 latency r y0.151 y0.120 y0.143 y0.054 y0.001 y0.004P 0.394 0.500 0.420 0.761 0.997 0.985

S1 amplitude r y0.153 y0.103 y0.119 y0.095 0.060 y0.146P 0.388 0.562 0.503 0.592 0.757 0.451

S2 amplitude r 0.405 0.391 0.202 0.199 y0.365 0.402P 0.017 0.022 0.252 0.259 0.051 0.030

S2yS1 ratio r 0.525 0.486 0.299 0.321 y0.436 0.463P 0.001 0.004 0.086 0.064 0.018 0.011

S1–S2 r y0.371 y0.320 y0.245 y0.220 0.280 y0.345difference P 0.031 0.065 0.163 0.210 0.141 0.067

Table 3. The P50 amplitude in response to S1 didnot show any correlation with whether total SPQscore or its factor scores. The P50 amplitude inresponse to S2 correlated with the total SPQ score(rs0.405,Ps0.017), although it showed no sig-nificant correlation with any factor score of SPQ.The ratio of S2yS1 amplitude demonstrated a verysignificant correlation not only with the total SPQscore (rs0.525, Ps0.001) but also with with-drawn factor(rs0.486, Ps0.004). The correla-tion between the S1–S2 amplitude difference andthe total SPQ score was also significant(rs

y0.371,Ps0.031). In addition, Pearson correla-tion analysis revealed that the S1–S2 amplitudedifference correlated significantly with total SPQscore(rsy0.403,Ps0.018; Fig. 1) and with thewithdrawn factor(rsy0.442,Ps0.009).Regarding the relationship between P50 meas-

urements and the two MPI subscales, the ratio ofS2yS1 amplitude showed a correlation with bothextraversion subscale(rsy0.436,Ps0.018) andneuroticism subscale(rs0.463,Ps0.011). How-ever, when controlling for the effect of the totalSPQ score, the square root of the S2yS1 ratio did

152 J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

Fig. 1. The linear correlations of P50 suppression with total SPQ score, and with extraversion or neuroticism. P50 suppression wasassessed using the ratio of S2yS1 amplitude and the difference between S1 and S2 amplitudes. The ratio was transformed to itssquare root to minimize its skewed distribution. Both P50 suppression measures demonstrated a significant correlation with totalSPQ score. Note the non-significant correlation of S1–S2 difference with extraversion or neuroticism.

not correlate significantly either with extraversion(rsy0.216, d.f.s26, Ps0.270) or with neuroti-cism (rs0.256, d.f.s26, Ps0.188). On the con-trary, the square root of the S2yS1 ratio stillshowed a significant partial correlation with SPQwhen controlling for extraversion(rs0.492, d.f.s

26,Ps0.008) or for neuroticism(rs0.529, d.f.s26,Ps0.004). Furthermore, this partial correlationbetween P50 suppression and the total SPQ scorewas still significant when controlling for bothextraversion and neuroticism simultaneously(rs0.414, d.f.s25, Ps0.032). Additionally, the S1–

153J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

Fig. 2. The grand average waveforms of P50 auditory evokedpotentials averaged separately for two non-clinical subgroups.S1: first click; S2: second click. Solid line: low SPQ scorers(ns21); dotted line: high SPQ scorers(ns13). Note: the P50potentials in response to S1, indicated by the black dot, weresimilar between two subgroups; however, the P50 potentials inresponse to S2, also indicated by the black dot, were inhibitedremarkably in the low SPQ scorers but not in the high SPQscorers.

Fig. 3. P50 suppression comparison between non-clinical sub-jects and schizophrenic patients. Note the non-significant dif-ference in P50 suppression between the high SPQ scorers andthe schizophrenics.

S2 amplitude difference did not show a significantlinear correlation with either extraversion(rs0.360,Ps0.055) or neuroticism(rsy0.358,Ps0.057) (Fig. 1). No correlation of age with anyP50 measurement was detected.

3.3. Low SPQ vs. high SPQ scorers

The mean values of P50 variables are shownfor all subjects in Table 1. The P50 evokedpotentials averaged separately for the low and highSPQ scorers are shown in Fig. 2. Visual inspectionfound that, while both subgroups had similar P50responses to the first stimulus(S1), they demon-strated different responses to the second stimulus(S2). Mann–Whitney test revealed that, betweenthe two non-clinical subgroups, there was no dif-ference in P50 latency(S1 latency:Zs0.25,Ps0.807; S2 latency:Zs1.15, Ps0.261) and P50amplitude in response to S1(Zs1.79,Ps0.076).A very significant difference was detected withP50 amplitude in response to S2(Zs2.79, Ps0.005), the ratio of S2yS1 amplitude(Zs3.40,P-0.001) and the S1–S2 amplitude difference(Zs3.10,Ps0.001), demonstrating a poorer sen-sory gating function in the high SPQ scorers than

in their low SPQ counterparts. Mann–Whitney testwas also performed on P50 measurements betweenmale and female subjects, and failed to detect anysignificant differences(S1 latency:Zs0.50, Ps0.621; S2 latency:Zs0.59, Ps0.574; S1 ampli-tude: Zs0.47, Ps0.646; S2 amplitude:Zs1.39,Ps0.175; S2yS1 amplitude ratio:Zs1.35, Ps0.187; S1–S2 difference:Zs0.52,Ps0.621).

3.4. Non-clinical subjects vs. schizophreniapatients

Mean values of P50 measurements of 18 schizo-phrenic patients are also shown in Table 1. Schizo-phrenic patients demonstrated a poorer P50suppression as compared to the non-clinical sub-jects in general(S2yS1 ratio:Zs3.27,Ps0.001).However, separate comparisons of two non-clinicalsubgroups with schizophrenic patients revealed asignificant difference in P50 suppression onlybetween low SPQ scorers and schizophrenicpatients(S2yS1 ratio:Zs3.98,P-0.001), but notbetween high SPQ scorers and schizophrenicpatients(S2yS1 ratio: Zs1.16, Ps0.258) (Fig.3).

4. Discussion

Raine’s SPQ has been applied in our departmentto screen college students at risk for SPD(Mannan

154 J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

et al., 2001). More than 50% of subjects with SPQscores)20 were later verified as SPD using thestructure clinical interview for DSM-III-R. Thiscut-off point was far below that used by Raine(1991), and was even below the mean SPQ scoreof their college students. The mean total SPQscore was only 11 in the present study, whichtogether with other SPQ data from Japanese col-lege students(Someya et al., 1994), suggests thatJapanese subjects possibly understate their schi-zotypal traits. However, this kind of cultural dif-ference did not obscure the relationship betweenpsychometric schizotypy and P50 suppression,since the findings of the present study are basicallyconsistent with those of Croft et al.(2001).In the present study, P50 suppression assessed

using the ratio of S2yS1 amplitude or the differ-ence between S1 and S2 amplitudes showed asignificant correlation with the total SPQ scoresamong the non-clinical Japanese. In addition, P50suppression also correlated with the withdrawnfactor score in the present study. In the study ofCroft et al. (2001), although an associationbetween P50 suppression and the unreality factorof their PSQ(personality syndrome questionnaire)was emphasized, the correlation between P50 sup-pression and the subjects’ total PSQ score wasalso significant(rs0.433, Ps0.009; Croft, per-sonal communication). The PSQ is an 80 itemquestionnaire, which was developed from the SPQand emphasizes the 3 factorial structure(active,withdrawn and unreality) of schizotypy(Gruzelier,1996; Gruzelier and Doig, 1996; Croft et al.,2001). Therefore, a main conclusion, which couldbe drawn from both studies, is that P50 suppressionvery possibly correlates with schizotypy as a wholesyndrome. This could also be supported by thetendency towards a relationship between P50 sup-pression and the active subscale in both studies.The role of thalamo-cortical activation systems isthe key to understanding schizotypy and schizo-phrenia. It is proposed that lateral shifts in theorchestration of hemispheric activation via thethalamocortical activation systems underpins theactivated and withdrawn syndromes, which areshifted to the left with activated syndrome andshifted to the right with withdrawn syndrome(Gruzelier, 2002, 2003). A generalized dysfunction

between the non-specific arousal system and spe-cific thalamocortical input is possibly associatedwith the ‘unreality’ experience(Gruzelier, 2002,2003).P50 suppression possibly measures the GABA

mediated inhibitory processes in the thalamo-lim-bic-cortical circuitry. Freedman and his colleaguesidentified the pyramid neurons of the CA3 regionof the hippocampus as the sources of cerebral-evoked P50 response, and reported that the P50suppression deficit in schizophrenic patients wasrelated to desensitization of the alpha-7 nicotincreceptor on the hippocampal GABA-ergic inter-neurons(Adler et al., 1992; Griffith et al., 1998;Freedman et al., 2002). There is also strong evi-dence suggesting that the P50 potential has neuralorigin in the temporal auditory cortex, and in thereticular activating system(RAS) of the brainstem,especially the ascending projection of PPN(pedun-culopontine nucleus) (Reite et al., 1988; Makelaet al., 1994; Reese et al., 1995; Garcia-Rill et al.,1995). Our work on animal equivalents of P50potential(P13 in rat) lends further support to theprobable brainstem source of this waveform(Miyazato et al., 1999a,b). However, the P50potential, as well as other mid-latency EP compo-nents, may be produced by underlying stimulus-bound synchronous oscillation in the gammafrequency band(40 Hz); and P50 suppression maybe a proxy for gamma band response(GBR)suppression(Basar et al., 1987; Clementz et al.,1997; Muller et al., 2001; Crawford et al., 2002).The neural mechanisms of synchronous gammaactivity mainly include the GABA-ergic interneu-ron network model and the thalamo-cortical arous-al model(Lee et al., 2003).In the present study, the non-clinical subjects

with high SPQ scores not only had a poorer P50suppression than their low SPQ counterparts, butalso demonstrated the deficiency similar in severityto that of schizophrenic patients, which furtherattests to the continuum of subclinical schizotypy-clinical schizophrenia. In the study of Croft et al.(2001), the P50 suppression results of low andhigh ‘unreality’ subjects were also suggested to beconsistent with comparisons of controls with eitherschizophrenia patients or SPD subjects in theliterature. Additionally, subjects who met predeter-

155J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

mined criteria for ‘schizotaxia’(a concept similarto negative schizotypy) showed a positive responseto low doses of risperidone(Tsuang et al., 2002).All these findings are in line with the psychiatricperspective on schizotypy where schizotypy rep-resents attenuated psychotic symptoms(Gruzelier,2003).The present study further revealed that the rela-

tionship between P50 suppression and schizotypyin the non-clinical subjects may be uncontaminatedby basic personality features of extraversion orneuroticism as measured using MPI, suggestingthat a specific relationship exists between thisneurocognitive function and schizotypy. This isconsistent with the findings reported in the familiesof schizophrenics(Freedman et al., 2002). In thesefamilies, most of the psychopathology other thanschizophrenia itself is found in siblings who donot have P50 inhibitory deficits, whereas individ-uals with P50 inhibitory deficits have some symp-toms of schizotypal personality disorder. Thespecific relationship between schizotypy and P50suppression revealed in the present study mightconfound the psychological perspective on schi-zotypy suggested by Eysenck(1992). However,we acknowledge that, for several reasons, it isdifficult practically to disentangle the effect ofbasic personality characteristics on the relationshipbetween schizotypy and sensory gating function.First, the close intercorrelations among schizotypy,extraversion, neuroticism and other personalitycharacteristics were consistently reported(forinstance, Ross et al., 2002; etc). The correlationbetween negative schizotypy and extraversion wasalso very significant in the present study. Second,findings to the contrary were reported. In the highschizotypals, the anxiety component of schizotypy,more than the perceptual-disorganization(schizo-phrenia-like) component, was found to account forthe neurocognitive deficit of latent inhibition(LI),which is used to assess selective attention pro-cesses and distraction by irrelevant stimuli(Braun-stein-Bercovitz, 2000; Braunstein-Bercovitz et al.,2002). The P50 suppression deficits also occurredin the high-anxious individuals such as posttrau-matic stress disorder patients(Gillette et al., 1997;Neylan et al., 1999; Skinner et al., 1999). Whetherthe relationship between P50 suppression and schi-

zotypy could be accounted for by the anxietycomponent needs to be clarified in future studies.Third, the neurobiological basis of personalityoverlaps with the mechanism of P50 suppression.For instance, extraversion has been interpreted asthe autoregulative behavioral response to a lowtonic activity of the non-specific activating system(Eysenck, 1967; Hegerl et al., 1995).The main limitation of the present study was

the small sample size, especially in relation topsychometric assessments. This limitation possiblycontributed to the finding that the correlationsbetween P50 suppression and the unreality or theactive factors of schizotypy were not significant.Confirmation of the findings in additional subjectsis necessary. In conclusion, the present study sug-gests that P50 suppression is one of the neurobiol-ogical substrates underlying psychometric schizo-typy, and that this relationship cannot be accountedfor by basic personality measures of extraversionor neuroticism.

Acknowledgments

This study was supported by Japan Grant-in-Aid for Scientific Research(No. 13671022).

References

Adler, L.E., Pachtman, E., Franks, R.D., Pecevich, M., Waldo,M.C., Freedman, R., 1982. Neurophysiological evidence fora defect in neuronal mechanisms involved in sensory gatingin schizophrenia. Biol. Psychiatry 17, 639–654.

Adler, L.E., Hoffer, L.J., Griffith, J., Waldo, M.C., Freedman,R., 1992. Normalization by nicotine of deficient auditorysensory gating in the relatives of schizophrenics. Biol.Psychiatry 32, 607–616.

Basar, E., Rosen, B., Basar-Eroglu, C., Greitschus, F., 1987.The association between 40-Hz EEG and the middle latencyresponse of the auditory evoked potential. Int. J. Neurosci.33, 103–117.

Braunstein-Bercovitz, H., 2000. Is the attentional dysfunctionin schizotypy related to anxiety? Schizophr. Res. 46,255–267.

Braunstein-Bercovitz, H., Rammsayer, T., Gibbons, H., Lubow,R.E., 2002. Latent inhibition deficits in high-schizotypalnormals: symptom-specific or anxiety-related? Schizophr.Res. 53, 109–121.

Cadenhead, K.S., Light, G.A., Geyer, M.A., Braff, D.L., 2000.Sensory gating deficits assessed by the P50 event-relatedpotential in subjects with schizotypal personality disorder.Am. J. Psychiatry 157, 55–59.

156 J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

Chen, W.J., Hsiao, C.K., Lin, C.C., 1997. Schizotypy incommunity samples: the three-factor structure and correla-tion with sustained attention. J. Abnorm. Psychol. 106,649–654.

Clementz, B.A., Blumenfeld, L.D., Cobb, S., 1997. The gammaband response may account for poor P50 suppression inschizophrenia. Neuroreport 8, 3889–3893.

Clementz, B.A., Geyer, M.A., Braff, D.L., 1998a. Multiple siteevaluation of P50 suppression among schizophrenia andnormal comparison subjects. Schizophr. Res. 30, 71–80.

Clementz, B.A., Geyer, M.A., Braff, D.L., 1998b. Poor P50suppression among schizophrenia patients and their first-degree biological relatives. Am. J. Psychiatry 155,1691–1694.

Crawford, H.J., McClain-Furmanski, D., Castagnoli, N., Cas-tagnoli, K., 2002. Enhancement of auditory sensory gatingand stimulus-bound gamma band(40 Hz) oscillations inheavy tobacco smokers. Neurosci. Lett. 317, 151–155.

Croft, R.J., Lee, A., Bertolot, J., Gruzelier, J.H., 2001. Asso-ciations of P50 suppression and desensitization with percep-tual and cognitive features of ‘unreality’ in schizotypy. Biol.Psychiatry 50, 441–446.

Eysenck, H.J., 1967. The biological basis of personality,Springfield, III: Thomas C.

Eysenck, H.J., 1992. The definition and measurement ofpsychoticism. Pers. Individ. Differ. 13, 757–785.

Eysenck, H.J., Barrett, P., 1993. The nature of schizotypy.Psychol. Rep. 73, 59–63.

Freedman, R., Adler, L.E., Olincy, A., Waldo, M.C., Ross,R.G., Stevens, K.E., et al., 2002. Input dysfunction, schi-zotypy, and genetic models of schizophrenia. Schizophr.Res. 54, 25–32.

Fujiwara, T., 1993. Japanese translation of ‘The SPQ: a scalefor the assessment of schiztypal personality based on DSM-III-R criteria’. Adv. Schizophr. Res. 3, 21–25 Japanese.

Garcia-Rill, E., Biedermann, J.A., Chambers, T., Skinner, R.D.,Mrak, R.E., Husain, M., et al., 1995. Mesopontine neuronsin schizophrenia. Neuroscience 66, 321–335.

Gillette, G.M., Skinner, R.D., Rasco, L.M., Fielstein, E.M.,Davis, D.H., Pawelak, J.E., et al., 1997. Combat veteranswith posttraumatic stress disorder exhibit decreased habitu-ation of the P1 midlatency auditory evoked potential. LifeSci. 61, 1421–1434.

Griffith, J.M., O’Neill, J.E., Petty, F., Garver, D., Young, D.,Freedman, R., 1998. Nicotinic receptor desensitization andsensory gating deficits in schizophrenia. Biol. Psychiatry44, 98–106.

Gruzelier, J.H., 1996. The factorial structure of schizotypy:part I. Affinities with syndromes of schizophrenia. Schi-zophr. Bull. 22, 611–620.

Gruzelier, J.H., 2002. A Janusian perspective on the nature,development and structure of schizophrenia and schizotypy.Schizophr. Res. 54, 95–103.

Gruzelier, J.H., 2003. Theory, methods and new directions inthe psychophysiology of the schizophrenic process andschizotypy. Int. J. Psychophysiol. 48, 221–245.

Gruzelier, J.H., Doig, A., 1996. The factorial structure ofschizotypy: part II. Cognitive asymmetry, arousal, handed-ness, and sex. Schizophr. Bull. 22, 621–634.

Gur, R.E., Mozley, P.D., Resnick, S.M., Levick, S., Erwin, R.,Saykin, A.J., et al., 1991. Relations among clinical scalesin schizophrenia. Am. J. Psychiatry 148, 472–478.

Hegerl, U., Gallinat, J., Mrowinski, D., 1995. Sensory corticalprocessing and the biological basis of personality. Biol.Psychiatry 37, 467–472.

Iwawaki, M., 1969. Sin Seikaku Kensa Hou—Maudsley Per-sonality Inventory. Seisin Syobo, Tokyo, pp. 115–177Japanese.

Jensen, A.R., 1958. The Maudsley personality inventory. ActaPsychol. 14, 314–325.

Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive andnegative syndrome scale(PANSS) for schizophrenia. Schi-zophr. Bull. 13, 261–276.

Lee, K.H., Williams, L.M., Breakspear, M., Gordon, E., 2003.Synchronous gamma activity: a review and contribution toan integrative neuroscience model of schizophrenia. BrainRes. Rev. 41, 57–78.

Liddle, P.F., Morris, D.L., 1991. Schizophrenic syndromes andfrontal lobe performance. Br. J. Psychiatry 158, 340–345.

Makela, J.P., Hamalainen, M., Hari, R., McEvoy, L., 1994.Whole-head mapping of middle-latency auditory evokedmagnetic fields. Electroencephalogr. Clin. Neurophysiol. 92,414–421.

Mannan, M.R., Hiramatsu, K.I., Hokama, H., Ohta, H., 2001.Abnormalities of auditory event-related potentials in stu-dents with schizotypal personality disorder. Psychiatry Clin.Neurosci. 55, 451–457.

Miyazato, H., Skinner, R.D., Cobb, M., Andersen, B., Garcia-Rill, E., 1999a. Midlatency auditory-evoked potentials inthe rat: effects of interventions that modulate arousal. BrainRes. Bull. 48, 545–553.

Miyazato, H., Skinner, R.D., Garcia-Rill, E., 1999b. Sensorygating of the P13 midlatency auditory evoked potential andthe startle response in the rat. Brain Res. 822, 60–71.

Muller, M.M., Keil, A., Kissler, J., Gruber, T., 2001. Suppres-sion of the auditory middle-latency response and evokedgamma-band response in a paired-click paradigm. Exp.Brain Res. 136, 474–479.

Myles-Worsley, M., 2002. P50 sensory gating in multiplexschizophrenia families from a Pacific island isolate. Am. J.Psychiatry 159, 2007–2012.

Nagamoto, H.T., Adler, L.E., Waldo, M.C., Griffith, J.M.,Freedman, R., 1991. Gating of auditory response in schizo-phrenics and normal controls. Effects of recording site andstimulation interval on the P50 wave. Schizophr. Res. 4,31–40.

Neylan, T.C., Fletcher, D.J., Lenoci, M., McCallin, K., Weiss,D.S., Schoenfeld, F.B., et al., 1999. Sensory gating inchronic posttraumatic stress disorder: reduced auditory P50suppression in combat veterans. Biol. Psychiatry 46,1656–1664.

157J. Wang et al. / International Journal of Psychophysiology 52 (2004) 147–157

Oldfield, R.C., 1971. The assessment and analysis of handed-ness: the Edinburgh inventory. Neuropsychologia 9, 97–113.

Raine, A., 1991. The SPQ: a scale for the assessment ofschiztypal personality based on DSM-III-R criteria. Schi-zophr. Bull. 17, 556–564.

Raine, A., Reynolds, C., Lencz, T., Scerbo, A., Triphon, N.,Kim, D., 1994. Cognitive-perceptual, interpersonal, anddisorganized features of schizotypal personality. Schizophr.Bull. 20, 191–201.

Rasco, L., Skinner, R.D., Garcia-Rill, E., 2000. Effect of ageon sensory gating of the sleep state-dependent P1yP50midlatency auditory evoked potential. Sleep Res. Online 3,97–105.

Reese, N.B., Garcia-Rill, E., Skinner, R.D., 1995. The pedun-culopontine nucleus–auditory input, arousal and pathophy-siology. Prog. Neurobiol. 47, 105–133.

Reite, M., Teale, P., Zimmerman, J., Davis, K., Whalen, J.,1988. Source location of a 50 ms latency auditory evokedfield component. Electroencephalogr. Clin. Neurophysiol.70, 490–498.

Reynolds, C.A., Raine, A., Mellingen, K., Venables, P.H.,Mednick, S.A., 2000. Three-factor model of schizotypalpersonality: invariance across culture, gender, religious affil-iation, family adversity, and psychopathology. Schizophr.Bull. 26, 603–618.

Ross, S.R., Lutz, C.J., Bailley, S.E., 2002. Positive andnegative symptoms of schizotypy and the five-factor model:a domain and facet level analysis. J. Pers. Assess. 79,53–72.

Siegel, C., Waldo, M.C., Mizner, G., Adler, L.E., Freedman,R., 1984. Deficits in sensory gating in schizophrenia patientsand their relatives. Arch. Gen. Psychiatry 41, 607–612.

Skinner, R.D., Rasco, L.M., Fitzgerald, J., Karson, C.N.,Matthew, M., Williams, D.K., et al., 1999. Reduced sensorygating of the P1 potential in rape victims and combatveterans with posttraumatic stress disorder. Depress. Anxiety9, 122–130.

Skinner, R.D., Miyazato, H., Garcia-Rill, E., 2002. The sleep-state dependent P50 auditory evoked potential in neuropsy-chiatric diseases. In: Hirata, K., Koga, Y., Nagata, K.,Yamazaki, K. (Eds.), Recent Advances in Human BrainMapping. Elsevier, Amsterdam, pp. 813–825.

Someya, T., Sasaki, T., Orita, S., Takahashi, S., 1994. Aboutreliability and validity of schizotypal personality question-naire(SPQ). Arch. Psychiatr. Clin. Eval. 5, 98 Japanese.

Tsuang, M.T., Stone, W.S., Tarbox, S.I., Faraone, S.V., 1 Mar2002. An integration of schizophrenia with schizotypy:identification of schizotaxia and implications for researchon treatment and prevention. Schizophr. Res. 54 (1–2),169–175.

Vollema, M.G., Hoijtink, H., 2000. The multidimensionalityof self-report schizotypy in a psychiatric population: ananalysis using multidimensional Rasch models. Schizophr.Bull. 26, 565–575.

Waldo, M.C., Carey, G., Myles-Worsley, M., Cawthra, E.,Adler, L.E., Nagamoto, H.T., et al., 1991. Codistribution ofa sensory gating deficit and schizophrenia in multi-affectedfamilies. Psychiatry Res. 39, 257–268.