Beruflich Dokumente
Kultur Dokumente
a
Departments of Clinical Psychology and Psychiatry, Zhejiang University School of Medicine, Hangzhou, Zhejiang 310058, China
b
Department of Psychology and Behavioral Sciences, Zhejiang University School of Science, Hangzhou, Zhejiang 310058, China
Available online 2 April 2007
KEYWORDS Abstract
Antisocial; Aims. — The mismatch negativity (MMN) to frequency deviant tones has yielded conflicting
Mismatch negativity results in patients with schizophrenia. This might be because Axis I schizophrenia overlaps with
(MMN); Axis II diagnoses such as paranoid or schizotypal personality disorders. This study was designed
Paranoid; to address this issue.
Personality disorder; Methods. — We evaluated the auditory MMN to frequency deviance in 17 patients with paranoid,
Schizophrenia; 15 schizotypal, and 16 antisocial personality disorders. These were compared to 25 healthy
Schizotypal subjects.
Results. — N1 to both deviant and standard tones was shorter in the paranoid group when
compared to healthy controls. MMN latencies were shorter at Fz, Cz, and Pz in the paranoid
group when compared to healthy controls, schizotypal, and antisocial groups. MMN amplitudes
were higher at Fz and Cz in the schizotypal and antisocial groups when compared to healthy
controls and the paranoid group.
Conclusions. — Patients with paranoid personality disorder had faster automatic detection
of auditory stimuli and of their change, but normal inhibition of irrelevant stimuli. By con-
trast, patients with schizotypal and antisocial personality disorders had normal discrimination
of the auditory stimuli, but might have a deficit in inhibition on irrelevant stimuli. Our
results might help differentiate these personality types, and clarify some MMN findings in
schizophrenia.
© 2007 Elsevier Masson SAS. All rights reserved.
∗ Corresponding author.
E-mail addresses: wangmufan@msn.com, DrWang@Doctor.com (W. Wang).
0987-7053/$ — see front matter © 2007 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.neucli.2007.03.001
90 Y. Liu et al.
Introduction
Among human event-related potentials (ERPs), the mis- neurons responsible for the detection of frequency and
match negativity (MMN) is a frontal negativity, which occurs duration changes might be different [23,24], and, indeed,
when a deviant stimulus does not match a sensory trace left frequency deviance could produce better and more reli-
by the preceding standard stimuli. It is described in the audi- able MMN than duration deviance [17]. Some schizophrenic
tory [34] and visual [48] modalities. In the auditory modality, patients who produced a reduced duration MMN displayed
MMN peaks at about 100—250 ms. As an orienting response, an increased frequency MMN instead [29]. In addition, MMN
its presence indicates that the subject is able to discrimi- was larger in the treatment-refractory schizophrenics [30],
nate environmental stimuli in a passive mode. MMN latency while patients with paranoid schizophrenia (some of these
reflects the onset of auditory stimulus change in structure, might be diagnosed as paranoid personality disorder) had
while its amplitude reflects the magnitude of the change and smaller frequency MMN at Fz [41].
saturates with the increment of perceived discrimination According to the DSM system [1], personality disorders
[31,40]. As some investigators showed that its amplitude like the paranoid and schizotypal types (Axis II) share sim-
increases when active attention is involved [35,54], MMN ilar clinical features with schizophrenia (Axis I). Genetic
has sometimes been considered to reflect an attentive acti- and neurochemical studies also suggest that paranoid [2,49]
vation of the human brain [47]. The supratemporal area and or schizotypal [5,20,45] personality disorders might be a
the frontal lobe, which are interconnected with each other, milder variant of schizophrenia. Since patients with per-
are both involved in the generation of the MMN. Morpho- sonality disorders are less affected by such confounding
logically, the MMN is typically largest at the frontocentral factors as hospitalization and psychosis, studies were often
electrodes in adults, and thus overlaps with other ERP com- conducted in some forms of personality disorder in order
ponents such as N1 and N2. In non-passive paradigms, it also to clarify the discrepancy issues reported in schizophre-
overlaps with the N2b component. In order to identify MMN nia. Thus, voluntary cerebral processing, as reflected by
more clearly, many investigators use passive conditions, in the ERP P3 component in an active paradigm, was stud-
which no specific task performance is required during the ied in schizotypal [18,25], antisocial [10], and borderline
test. Therefore, the MMN paradigm is considered as a pow- [6,25] personality disorders. All these studies demonstrated
erful tool for examining the early involuntary or automatic a reduced P3 amplitude, that is, a similar result as that
processing. obtained in schizophrenia [14,46]. Therefore, one might
Since the first MMN trials to duration [43] and fre- wonder whether earlier automatic processing, as reflected
quency [16] deviance, robust results were obtained in by the frequency MMN, could vary among different forms
schizophrenia (reviewed in [28,32]). The reduced MMN of personality disorders that are close to schizophrenia in
(especially in the frontal region) in schizophrenia was sub- terms of the spectrum of disorders. The answer to this ques-
sequently replicated for duration —– (for recent studies tion would help clarify the frequency MMN issues that are
see [3,11]) and partly for frequency deviance (for recent reported in schizophrenia.
studies see [4,44,51,52]). Such reduced MMN amplitude To address this issue, we evaluated the frequency MMN
might reflect deteriorated auditory perception (reviewed in patients with schizotypal and paranoid personality dis-
in [28,32]). However, several studies failed to evidence orders. We also included a group of antisocial personality
any MMN reduction to frequency deviance in schizophre- disorders, as an increased MMN has been demonstrated in
nia [19,29,37,38]. Neurophysiologic studies show that brain individuals at high risk for alcoholism [55], and alcoholism
MMN in personality disorders 91
predisposition is linked to several factors including the anti- reference, in keeping with other studies [27,36]. Bipolar record-
social personality disorder [13]. Considering together the ings of the electrooculogram (EOG) were made with electrodes at
larger MMN in treatment-refractory schizophrenia [30] and the outer canthus and above the right eye. Ground electrode was
the smaller one in paranoid schizophrenia [41], we hypoth- fixed to one arm. Electrode impedance was kept ≤10 K. All record-
esized that the frequency MMN would be larger in both ings were made on a Nihon Kohden Neuropack-sigma device using a
band-pass of .01—30 Hz, and a sampling rate of 500 Hz. The config-
antisocial and schizotypal personality disorders, but smaller
uration baud rate was set at 9600 bit/s, and the data configuration
in paranoid personality disorder. length was set at 8 bits in the hardware of the device. Connections
were set such that a negative activity produced an upward deflec-
Methods and materials tion. Traces with an EOG of amplitude exceeding ±100 V (which
indicates blink or ocular movement) were automatically rejected.
Only artifact-free sweeps were automatically selected for averag-
Subjects ing. About 1100 tones were delivered in each trial but only 1000
sweeps were finally selected for averaging. The sampling epoch was
This study was carried out in 73 subjects: 25 healthy volunteers 100 ms pre-stimulus and 500 ms post-stimulus.
(9 women, mean age: 28.1 years ± 8.4 (S.D); range: 18—51 years) Both latencies and peak-to-baseline amplitudes of the maximal
were recruited among students, hospital staff or paid volunteers negative deflection within a specified latency ranges were mea-
from the general population. Forty-eight outpatients were diag- sured, based on visual inspection. N1 was measured within a latency
nosed as suffering from either paranoid (n = 17, 6 women; mean range of 50—150 ms. Difference waves were obtained by subtracting
age: 32.3 years ± 12.0; range 18—55 years), schizotypal (n = 15, the frequent from the deviant tone ERPs, as proposed in [31]. The
11 women; mean age: 30.5 years ± 12.7; range: 18—62 years), or MMN was measured on the difference waveforms within a latency
antisocial (n = 16, 1 woman; mean age: 29.3 years ± 11.5; range: range of 100—250 ms [31].
18—63 years) personality disorder, based on the DSM-IV-TR criteria
[1]. Since the present study was focused on the Axis II diagnoses, we
did not quantify the anxiety or depression severity of our subjects.
However, it appeared from clinical evaluation that most patients Statistical analyses
had suffered from Axis I disorders such as mild to moderate anx-
iety or depression or both. The Axis II diagnoses were confirmed MANOVA (group×gender×electrode sites) was used to study laten-
using the Personality Measure [39]. This questionnaire designed for cies and amplitudes of N1 to standard tones (N1-standard), N1 to
assessing the disordered personality functioning styles has proven deviant tones (N1-deviant), and MMN at the three midline elec-
reliable in Chinese culture [53]. A semi-structured interview was trodes in the four groups. Whenever a significant main effect was
performed for each healthy control in order to ascertain that they found, post-hoc analyses (Duncan’s multiple new range tests) were
were not suffering or had not suffered from any psychiatric prob- performed to evaluate between-group differences for the corre-
lems. Each patient was diagnosed with one type of personality sponding parameter. Moreover, the Cohen’s d (effect size) was
disorder only; and patients receiving more than one Axis II diag- calculated as a supplement indicator for the magnitude of vari-
noses were excluded from the study. Some patients might have ations. We used Spearman’s rank order correlation to search for
trends of obsession-compulsion, but failed to meet the diagnoses possible relationships between age and latencies or amplitudes. A
of the obsessive-compulsive personality disorder alone. A recent CT p-value < .05 was considered significant.
scan or MRI was available, in order to ensure that all patients were
free from any organic brain lesions. Though about 50% of patients
had received anxiolytics or antidepressants before presenting to our
clinic, all were free from any drug or alcohol for at least 72 h prior Results
to the test. Patients with any drug or alcohol abuse were excluded
from the study. All subjects gave their written informed consents All subjects showed clear N1 components to standard and deviant
to be included in the study. There were no significant age differ- tones at each electrode site. The MMN component in the difference
ences between the different groups (one-way ANOVA, main effect, waves was clearer at Fz and Cz compared to Pz in all subjects. The
F [3,69] = .53, p = .66). grand averages of original and difference waves in each group are
shown in Figure 1.
Stimuli and recording parameters
Figure 1 Superimposed grand averages of ERPs elicited by deviant tones (thin solid lines), standard tones (dashed lines), difference
waveforms (thick solid lines) at the three midline electrodes, and electrooculogram (EOG) in healthy subjects and patients with
paranoid, schizotypal and antisocial personality disorders.
Mismatch negativity tively) groups. Cohen’s d-values indicated that the related effect
sizes were moderate to strong (Table 2).
MANOVA also showed significant between-group differences for Contrary to N1, MANOVA also showed a significant between-
MMN latencies at the three electrode sites (Table 1). Post-hoc group difference in MMN amplitudes. Post-hoc analyses demon-
tests showed that the mean latencies of the paranoid group at strated that the mean amplitudes in both schizotypal and antisocial
Fz (154.7 ms), Cz (154.7 ms) and Pz (152.6 ms) were significantly groups were significantly higher than in the healthy controls at Fz
shorter than those of the healthy controls (174.4, 173.4 and (schizotypal: 5.6 V; antisocial: 4.5 V; controls: 3.0 V) and Cz
173.4 ms, respectively), schizotypal (177.8, 178.5, and 177.5 ms (schizotypal: 4.8 V; antisocial: 3.7 V; controls: 2.4 V), but not
respectively), and antisocial (169.3, 168.9 and 168.6 ms, respec- at Pz. Mean MMN amplitudes in the schizotypal group at Fz (5.6 V)
and Cz, (4.8 V) and those in the antisocial group at Fz (4.5 V)
Figure 2 Scatter-plot of individual MMN amplitudes at Fz in healthy subjects and patients with paranoid, schizotypal and antisocial
personality disorders. Group means and standard deviations are given in Table 2.
MMN in personality disorders 93
Table 1 MANOVA results on latencies and amplitudes of standard and deviant N1, and the mismatch negativity (MMN) in the
four groups of subject
Latency
N1 deviant
Group 3,65 3.19 .03
Gender 1,65 4.80 .03
Electrode 2,130 3.54 .03
Group×Gender×Electrode 6,130 .62 .71
N1 standard
Group 3,65 3.46 .02
Gender 1,65 .21 .65
Electrode 2,130 5.40 .01
Group×Gender×Electrode 6,130 1.90 .09
MMN
Group 3,66 4.29 .01
Gender 1,66 .01 .91
Electrode 2,130 .43 .65
Group×Gender×Electrode 6,130 1.20 .31
Amplitude
N1 deviant
Group 3,65 .78 .51
Gender 1,65 2.02 .16
Electrode 2,130 30.91 .00
Group×Gender×Electrode 6,130 1.85 .09
N1 standard
Group 3,65 2.32 .08
Gender 1,65 .07 .79
Electrode 2,130 19.09 .00
Group×Gender×Electrode 6,130 .71 .64
MMN
Group 3,65 4.11 .01
Gender 1,65 .78 .38
Electrode 2,130 32.29 .00
Group×Gender×Electrode 6,130 .54 .77
were also significantly higher than in the paranoid group (Fz: 3.2 V; ity disorder when compared to healthy controls and patients
Cz: 2.7 V). Cohen’s d-values indicated that these size effects were with schizotypal and antisocial personality disorders. It
moderate to strong at Fz, and weak to moderate at Cz (Table 2). was higher at Fz and Cz in patients with schizotypal and
Since many auditory MMN studies were interested in the frontal antisocial personality disorders when compared to healthy
region [40], the individual MMN amplitudes at Fz in the four groups controls and patients with paranoid personality disorder.
are shown in Figure 2. MMN amplitudes in the patient groups were
Additionally, both deviant and standard tones elicited ear-
more scattered than in controls, but they were no apparent outliers
or sub-clusters in either group.
lier N1s in patients with paranoid personality disorder than
Gender effects were only found for N1-deviant latencies at healthy subjects. The mean MMN latencies in our healthy
Fz, (males: 100.7 ms ± 14.9; females: 108.2 ms ± 20.9, p < .05; subjects were comparable to those previously reported
d = −.36), Cz (males 99.6 ms ± 14.5; females 107.6 ms ± 20.3, [11,12,50]. The mean MMN amplitudes in the four groups
p < .05; d = −.39), and Pz (males 100.2 ms ± 15.0; females 107.3 were maximal at Fz, which is also in line with previous
ms ± 20.4, p < .05, d = −.34), with weak effect sizes. No correlations reports [31,40]. Noteworthy, the reference in this study was
between age and latencies or amplitudes of any ERP parameter were the linked earlobes and not the mastoids, so that there could
found in either group or in all subjects pooled together. be some doubt on whether N2b contributed to the MMN.
Although the distractive task we used should in principle
Discussion warrant inattentiveness and absence of N2b, the compliance
to the task of our subjects, especially patients, was poorly
The MMN to frequency deviant tones was earlier at the monitored. This issue deserves further investigation for a
three midline electrodes in patients with paranoid personal- clearer identification of MMN and N2b in these subjects.
94 Y. Liu et al.
Table 2 Latencies and amplitudes (mean ±S.D.) of N1 (to both standard and deviant stimuli) and mismatch negativity (MMN) at
the three midline electrode sites in healthy subjects and patients with paranoid, schizotypal, and antisocial personality disorders
Latency (ms)
N1 deviant
Fz 110.0 ± 18.3 98.0 ± 16.0 a (−.66) 103.0 ± 17.4 99.5 ± 16.6
Cz 108.6 ± 18.6 96.7 ± 15.2 a (−.63) 103.0 ± 17.7 99.0 ± 15.6
Pz 108.4 ± 18.3 96.8 ± 15.9 a (−.64) 103.5 ± 17.9 99.7 ± 15.7
N1 standard
Fz 105.1 ± 12.1 92.7 ± 11.8 a (−1.02) 99.7 ± 15.9 94.0 ± 12.7 a (−.92)
Cz 105.6 ± 13.5 91.5 ± 12.5 a (−1.04) 98.9 ± 16.7 93.6 ± 12.2
Pz 105.9 ± 13.8 94.1 ± 14.4 a (−.89) 99.9 ± 18.1 94.4 ± 12.6
MMN
Fz 174.4 ± 19.1 154.7 ± 24.6 a (−1.03) 177.8 ± 16.5 b (.94) 169.3 ± 21.8 b (.37)
Cz 173.4 ± 19.1 154.7 ± 25.4 a (−.98) 178.5 ± 16.4 b (.94) 168.9 ± 23.0 b (.56)
Pz 173.4 ± 19.5 152.6 ± 26.7 a (−1.07) 177.5 ± 18.8 b (.93) 168.6 ± 22.7 b (.60)
Amplitude (V)
N1 deviant
Fz 4.9 ± 3.6 4.5 ± 2.3 5.3 ± 2.6 5.3 ± 2.3
Cz 4.7 ± 2.8 4.3 ± 2.3 5.1 ± 2.7 5.4 ± 2.0
Pz 3.4 ± 2.0 2.8 ± 2.1 3.8 ± 2.4 4.0 ± 2.0
N1 standard
Fz 3.5 ± 2.3 2.5 ± 1.5 2.4 ± 1.7 2.8 ± 1.2
Cz 3.3 ± 1.9 2.5 ± 1.2 2.4 ± 1.8 3.0 ± 1.5
Pz 2.5 ± 1.7 1.5 ± .9 1.5 ± 1.3 2.4 ± 1.6
MMN
Fz 3.0 ± 2.1 3.2 ± 2.1 5.6 ± 2.2 a (1.23), b (1.14) 4.5 ± 2.1 a (.71), b (.62)
Cz 2.4 ± 2.5 2.7 ± 2.2 4.8 ± 2.4 a (.96), b (.95) 3.7 ± 1.9 a (.52)
Pz 1.6 ± 2.6 1.6 ± 1.8 3.1 ± 2.1 2.3 ± 1.7
Note: a, p < .05 vs healthy group; b, p < .05 vs paranoid group; numbers in parentheses after a or b indicates the respective effect size
(Cohen’s d).
Our finding of an earlier MMN in patients with paranoid addressed in further studies, all the more as two lines of evi-
personality disorder was also in accordance with a previous dence make it conceivable. Firstly, given that P3 amplitude
study of paranoid schizophrenia [41], which clearly showed represents the inhibition of regional cortical activity and
a frequency MMN at Fz peaking at about 135—140 ms. This reflects the utilization of resources at a more central level of
finding, together with the shorter latencies of N1-deviant information processing [22,42], the P3 reduction in schizo-
and N1-standard, suggest that our patients with paranoid typal and antisocial personality disorders [10,18,25] could
personality disorders would perceive both types of tones imply that the cortical areas were less inhibited in these
faster than healthy subjects or other patients, that is, they patients. Secondly, other neurophysiologic studies showed
would be more readily alerted by environmental acoustic that patients with schizotypal personality disorder have a
stimuli. By contrast, the fact that their amplitudes of N1- decreased prepulse inhibition of the startle reflex [8], and
deviant, N1-standard, and MMN were similar to those of of antisaccade and ocular motor responses [7].
healthy subjects suggests that our patients with schizoty- The higher MMN amplitude, which was observed in
pal and paranoid personality disorders would display normal patients with antisocial personality disorder, is also in keep-
alertness to both type of tones and their contrast. ing with the MMN increment that is observed in subjects
It has been suggested that N1 and MMN would reflect at high risk for alcoholism [55]. This might imply a lack of
sensory gating and its neuronal correlates in temporopari- inhibition on the irrelevant stimuli per se.
etal and prefrontal areas [15,33]. Consequently, higher MMN By contrast, we found normal MMN amplitude, instead
amplitudes would imply that subjects have difficulty to of a reduced one, in patients with paranoid personality
filter out irrelevant stimuli [21,26]. Whether or not the disorder, which is similar to a previous study of paranoid
higher MMN found in our study denotes deficient inhibition schizophrenic patients [41]. This result would indicate that
in patients with schizotypal and antisocial personality dis- patients with paranoid personality disorder do not manifest
orders remains unknown, as our experimental protocol was impaired context processing and have a normal capacity to
not designed to test sensory gating. This hypothesis might be filter our irrelevant stimuli. This group displayed more alert-
MMN in personality disorders 95
cial and national samples of the Danish adoption study of [39] Parker G, Hadzi-Pavlovic D. A questionnaire of style: refining
schizophenia. Arch Gen Psychiatry 1994;51:456—68. the dimensions of personality disorder style. J Personal Disord
[21] Kisley M, Olincy A, Robbins E, Polk SD, Adler LE, Waldo MC, 2001;15:300—18.
et al. Sensory gating impairment associated with schizophre- [40] Picton TW, Alain C, Otten L, Ritter W, Achinm A. Mismatch
nia persists into REM sleep. Psychophysiology 2003;40:29— negativity: Different waters in the same river. Audiol Neurootol
38. 2000;5:111—39.
[22] Kok A. Event-related potential (ERP) reflections of mental [41] Sato Y, Yabe H, Todd J, Michie P, Shinozaki N, Sutoh T, et al.
resources: a review and synthesis. Biol Psychol 1997;45:19—56. Impairment in activation of frontal attention-switch mecha-
[23] Kraus N, McGee T, Carrell T, King C, Littman T, Nicol T. Discrim- nism in schizophrenic patients. Biol Psychol 2002;62:49—63.
ination of speech-like signals in auditory thalamus and cortex. [42] Schupp HT, Lutzenberger W, Rau H, Birbaumer N. Positive shifts
J Acoust Soc Am 1994;96:2758—68. of event-related potentials: a state of cortical disfacilitation as
[24] Kraus N, McGee T, Littman T, Nicol T, King C. Encoding of reflected by the startle reflex probe. Electroencephalogr Clin
acoustic change involves non-primary auditory thalamus. J Neurophysiol 1994;90:135—44.
Neurophysiol 1994;72:1270—7. [43] Shelley AM, Ward PB, Catts SV, Michie PT, Andrews S,
[25] Kutcher SP, Blackwood DH, Gaskell DF, Muir WJ, St Clair McConaghy N. Mismatch negativity: an index of preat-
DM. Auditory P300 does not differentiate borderline person- tentive processing deficit in schizophrenia. Biol Psychiatry
ality from schizotypal personality disorder. Biol Psychiatry 1991;30:1059—62.
1989;26:766—74. [44] Shinozaki N, Yabe H, Sato Y, Hiruma T, Sutoh T, Nashida T,
[26] Light G, Braff D. Do self-reports of perceptual anomalies et al. The difference in mismatch negativity between the
reflect gating deficits in schizophrenia patients? Biol Psychiatry acute and post-acute phase of schizophrenia. Biol Psychol
2000;47:463—7. 2002;59:105—19.
[27] Micheyl C, Carlyon RP, Shtyrov Y, Hauk O, Dodson T, [45] Siever LJ, Amin F, Coccaro EF, Trestman R, Silverman J, Hor-
Pullvermueller F. The neurophysiological basis of auditory con- vath TB, et al. CSF homovanillic acid in schizotypal personality
tinuity illusion: a mismatch negativity study. J Cogn Neurosci disorder. Am J Psychiatry 1993;150:149—51.
2003;15:747—58. [46] Squires-Wheeler E, Friedman D, Skodol AE, Erlenmeyer-Kimling
[28] Michie PT. What has MMN revealed about the auditory system L. A longitudinal study relating P3 amplitude to schizophre-
in schizophrenia? Int J Psychophysiol 2001;42:177—94. nia spectrum disorders and global personality functioning. Biol
[29] Michie PT, Budd TW, Todd J, Rock D, Wichmann H, Box J, et al. Psychiatry 1993;33:774—85.
Duration and frequency mismatch negativity in schizophrenia. [47] Sussman ES. Is MMN truly pre-attentive? In: Shtyrov Y, Pulver-
Clin Neurophysiol 2000;111:1054—65. mullerF F., editors. Programme and Abstract Book of the Fourth
[30] Milovan DL, Baribeau J, Roth RM, Stip E. ERP study of Conference on Mismatch Negativity (MMN) and its Clinical and
pre-attentive auditory processing in treatment-refractory Scientific Applications, Cambridge. April 22—26, 2006. S1.2.
schizophenia. Brain Cogn 2004;55:355—7. [48] Tales A, Newton P, Troscianko T, Butler S. Mismatch negativity
[31] Näätänen R. Attention and Brain Function. Hillsdale, NJ: in the visual modality. Neuroreport 1999;10:3363—7.
Lawrence Erlbaum; 1992. [49] Tienari P, Wynne LC, Lasky K, Moring J, Nieminen P, Sorri A,
[32] Näätänen R. Mismatch negativity: clinical research and possible et al. Genetic boundaries of the schizophrenia spectrum: evi-
applications. Int J Psychophysiol 2003;48:179—88. dence from the Finnish Adoptive Family Study of Schizophrenia.
[33] Näätänen R, Picton TW. The N1 wave of the human electric Am J Psychiatry 2003;160:1587—94.
and magnetic response to sound: a review and an analysis if [50] Todd J, Michie PT, Jablensky AV. Association between
the component structure. Psychophysiology 1987;24:375—425. duration mismatch negativity (MMN) and raised temporal dis-
[34] Näätänen R, Gaillard AWK, Mantysalo S. Early selective- crimination thresholds in schizophrenia. Clin Neurophysiol
attention effect on evoked potential reinterpreted. Acta 2003;114:2061—70.
Psychol (Amst) 1978;42:313—29. [51] Umbricht D, Krljes S. Mismatch negativity in schizophrenia: a
[35] Näätänen R, Paavilainen P, Tiitinen H, Jiang D, Alho K. Attention meta-analysis. Schizophr Res 2005;76:1—23.
and mismatch negativity. Psychophysiology 1993;24:375—425. [52] Umbricht D, Koller R, Schmid L, Skrabo A, Grubel C, Huber T, et
[36] Näätänen R, Pakarinen S, Rinne T, Takegata R. The mismatch al. How specific are deficits in mismatch negativity generation
negativity (MMN): toward the optimal paradigm. Clin Neuro- to schizophrenia? Biol Psychiatry 2003;53:1120—31.
physiol 2004;115:140—4. [53] Wang W, Hu L, Mu L, Chen D, Song Q, Zhou M, et al. Func-
[37] O’Donnell BF, Hokama M, McCarley RW, Smith RS, Salisbury tioning styles of personality disorders and five-factor normal
DF, Mondrow E, et al. Auditory ERPs to non-target stimuli in personality traits: a correlation study in Chinese students. BMC
schizophrenia: relationship to probability, task-demands, and Psychiatry 2003;3:11.
target ERPs. Int J Psychophysiol 1994;17:219—31. [54] Woldorff M, Hackley S, Hillyard SA. The effects of channel-
[38] Oades RD, Dittman-Balcar A, Zerbin D, Grzella I. Impaired selective attention on mismatch negativity wave elicited by
attention-dependent augmentation of MMN in nonparanoid vs deviant tones. Psychophysiology 1991;28:30—42.
paranoid schizophrenic patients: a comparison with obsessive- [55] Zhang XL, Cohen HL, Porjesz B, Begleitner H. Mismatch nega-
compulsive disorder and healthy subjects. Biol Psychiatry tivity in subjects at high risk for alcoholism. Alcohol Clin Exp
1997;41:1121—96. Res 2001;25:330—7.