Aberrant temporal behavior of mismatch negativity generators in schizophrenia patients and subjects at clinical high risk for psychosis

Accepted Manuscript Aberrant temporal behavior of mismatch negativity generators in schizophrenia patients and subjects at clinical high risk for psychosis Minah Kim, Kang Ik Kevin Cho, Youngwoo Bryan Yoon, Tae Young Lee, Jun Soo Kwon PII: DOI: Reference:

S1388-2457(16)31007-0 http://dx.doi.org/10.1016/j.clinph.2016.11.027 CLINPH 2008001

To appear in:

Clinical Neurophysiology

Received Date: Revised Date: Accepted Date:

26 July 2016 23 November 2016 26 November 2016

Please cite this article as: Kim, M., Ik Kevin Cho, K., Bryan Yoon, Y., Young Lee, T., Soo Kwon, J., Aberrant temporal behavior of mismatch negativity generators in schizophrenia patients and subjects at clinical high risk for psychosis, Clinical Neurophysiology (2016), doi: http://dx.doi.org/10.1016/j.clinph.2016.11.027

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Aberrant temporal behavior of mismatch negativity generators in schizophrenia patients and subjects at clinical high risk for psychosis Minah Kim1, Kang Ik Kevin Cho2, Youngwoo Bryan Yoon2, Tae Young Lee3, Jun Soo Kwon1,2,4* 1

Department of Psychiatry, Seoul National University College of Medicine, Seoul, Republic of Korea

2

Department of Brain and Cognitive Science, Seoul National University College of Natural Science,

Seoul, Republic of Korea 3

Department of Neuropsychiatry, Seoul National University Hospital, Seoul, Republic of Korea

4

Institute of Human Behavioral Medicine, SNU-MRC, Seoul, Republic of Korea

Corresponding author: Jun Soo Kwon, M.D., Ph.D. Department of Psychiatry, Seoul National University College of Medicine 101 Daehak-no, Chongno-gu, Seoul, Republic of Korea. 03080. Tel: +82 2 2072 2972 Fax: +82 2 747 9063 E-mail address: [email protected]

Highlights 

Schizophrenic patients and clinical high risk subjects show decreased mismatch negativity (MMN) strength in frontal and temporal cortices.



Disconnection between MMN-related brain regions exists even prior to psychosis onset.



Aberrant MMN generator activity might be related to schizophrenia pathophysiology. 1

Abstract Objective: Although disconnection syndrome has been considered a core pathophysiologic mechanism of schizophrenia, little is known about the temporal behavior of mismatch negativity (MMN) generators in individuals with schizophrenia or clinical high risk (CHR) for psychosis. Methods: MMN was assessed in 29 schizophrenia patients, 40 CHR subjects, and 47 healthy controls (HCs). Individual realistic head models and the minimum L2 norm algorithm were used to generate a current source density (CSD) model of MMN. The strength and time course of MMN CSD activity were calculated separately for the frontal and temporal cortices and were compared across brain regions and groups. Results: Schizophrenia patients and CHR subjects displayed lower MMN CSD strength than HCs in both the temporal and frontal cortices. We found a significant time delay in MMN generator activity in the frontal cortex relative to that in the temporal cortex in HCs. However, the sequential temporofrontal activities of MMN generators were disrupted in both the schizophrenia and CHR groups. Conclusions: Impairments and altered temporal behavior of MMN multiple generators were observed even in individuals at risk for psychosis. Significance: These findings suggest that aberrant MMN generator activity might be helpful in revealing the pathophysiology of schizophrenia.

Keywords: clinical high risk for psychosis, current source density, disconnection syndrome, mismatch negativity, schizophrenia.

2

1. Introduction Mismatch negativity (MMN) is an event-related potential (ERP) component that represents preattentive auditory information processing (Naatanen et al., 2001). Since MMN was thought to be produced by two sequential processes (auditory sensory change detection and involuntary switching of attention), multiple MMN generators were suggested primarily in the temporal and frontal cortices (Naatanen and Michie, 1979). This suggestion was supported by MMN findings from patients with frontal lesions and intracranial recording (Alho et al., 1994, Liasis et al., 2001, Rosburg et al., 2005). Studies using distributed source analysis also reported that MMN generators were located in both the frontal and temporal cortices (Giard et al., 1990, Jemel et al., 2002) and that frontal generator activity peaked after temporal generator activity in healthy subjects (Fulham et al., 2014, Rinne et al., 2000). Successive functional magnetic resonance imaging (fMRI) findings provided several suggestions of frontal cortex contributions to the MMN mechanism, which indicated temporo-frontal sequential activity of MMN generators in healthy volunteers (Doeller et al., 2003, Opitz et al., 2002, Rinne et al., 2005). These findings suggest that the separate temporal behavior of discrete MMN generators may serve as an indicator of functional connectivity between related brain regions at the pre-attentive level. Structural and functional disconnection between brain regions has been suggested as a core pathophysiologic mechanism of schizophrenia; the resulting disorder has been termed disconnection syndrome (Friston, 1999, Friston and Frith, 1995). Reduced white matter volume and integrity in schizophrenia patients were found in widespread regions (Bora et al., 2011, Lee et al., 2013, Wheeler and Voineskos, 2014). In subjects at clinical high risk (CHR) for psychosis, structural connectivity was disrupted, although less prominently than in schizophrenia patients, and was associated with the transition to an overt psychotic disorder (Carletti et al., 2012, von Hohenberg et al., 2014, Ziermans et al., 2012). In addition, previous fMRI studies revealed functional disconnection between the temporal and frontal cortices during the performance of executive function tasks among schizophrenia patients and CHR individuals (Benetti et al., 2009, Crossley et al., 2009). Recently, Gaebler et al. reported fMRI 3

results showing that the sensory processing deficit of schizophrenia patients during an auditory mismatch task was related to disrupted connectivity between the temporal and prefrontal cortices (Gaebler et al., 2015). However, the temporal behavior of discrete MMN generators assessed via electroencephalography (EEG), which has the great advantage of millisecond-order time resolution, has not yet been sufficiently studied in patients with schizophrenia. An impaired MMN response has been consistently reported in schizophrenia patients, suggesting its relationship with the pathophysiology of the disorder (Umbricht and Krljes, 2005). A reduced MMN amplitude in CHR individuals has also been reported, although this finding is not as consistent as in patients with schizophrenia (Bodatsch et al., 2015, Shin et al., 2009). Associations of reduced MMN amplitude with higher-order cognitive deficits and poor functional status were suggested in schizophrenia patients (Kim et al., 2014, Wynn et al., 2010). In CHR subjects, it has been shown that MMN amplitude effectively predicts the transition from a CHR status to overt psychosis and the time to psychosis onset (Bodatsch et al., 2015, Bodatsch et al., 2011, Perez et al., 2014). Recent current source density (CSD) analyses showed that schizophrenia patients displayed MMN generator impairments in widespread brain regions (Miyanishi et al., 2013, Takahashi et al., 2013), but CSD analysis of CHR individuals has not yet been presented. Regarding the separate temporal behavior of discrete MMN generators, only one study reported delayed MMN CSD activation between primary and secondary auditory cortices in schizophrenia (Fulham et al., 2014). However, due to the high variability of CSD onset latencies in the frontal lobe, the delay in frontal cortex CSD activation relative to temporal cortex CSD activation was excluded from analysis in that study. In this study, we aimed to demonstrate disruptions in functional connectivity at the pre-attentive level in patients with schizophrenia and individuals at CHR for psychosis using MMN CSD analysis. We first hypothesized that MMN CSD strength was reduced in both the schizophrenia and CHR groups compared to the healthy control (HC) group not only in the temporal cortex but also in the frontal cortex. Second, we sought to confirm the findings of previous studies that HC subjects would 4

show sequential temporo-frontal activation of discrete MMN generators, but aberrant temporal behavior of frontal and temporal MMN generators was expected in schizophrenia patients and CHR individuals.

2. Methods 2.1. Participants and clinical assessments Twenty-nine patients with schizophrenia, 40 individuals at CHR for psychosis, and 47 HC subjects participated in this study. Study participants were recruited via the Seoul Youth Clinic (www.youthclinic.org), a center for early detection and intervention of people at high risk for psychosis (Kwon et al., 2012). The diagnosis of schizophrenia was confirmed using the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition Axis I Disorders (SCID-I), and the severity of psychotic symptoms was measured using the Positive and Negative Syndrome Scale (PANSS). The CHR subjects met at least 1-of the following 3 criteria of the Structured Interview for Prodromal Symptoms (SIPS) (Miller et al., 2002): attenuated positive symptoms (APS), brief intermittent psychotic symptoms (BIPS), and genetic risk with deterioration (GRD). Prodromal psychotic symptoms were assessed in CHR individuals using the validated Korean version of the SIPS (Jung et al., 2010). In both the schizophrenia and CHR groups, the Global Assessment of Functioning (GAF), Hamilton rating scale for anxiety (HAM-A) and depression (HAMD) were used to assess general functioning, anxiety, and depressive symptoms. Intelligence quotient (IQ) was measured in all participants using the abbreviated version of the Korean-Wechsler Adult Intelligence Scale (Kim et al., 1994). The exclusion criteria included a lifetime diagnosis of substance abuse or dependence, neurological disease, significant head injury accompanied by loss of consciousness, medical illness with documented cognitive sequelae, sensory impairments, or intellectual disability (IQ<70). 5

All of the participants fully understood the study procedure and provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and with the approval of the Institutional Review Board of Seoul National University Hospital. 2.2. EEG recordings and MRI acquisition Subjects were instructed to find Wally from a “Where’s Wally?” picture book while ignoring acoustic stimuli. While subjects concentrated on the game, pseudorandom series of 1000-Hz (80-dB, 10-ms rise/fall) auditory stimuli were binaurally presented using a STIM2 sound generator (Compumedics, Charlotte, NC). The auditory oddball stimuli consisted of 982 (81.8%) standard stimuli lasting 50 ms and 218 (18.2%) deviant stimuli lasting 100 ms. The deviant stimuli were preceded by at least one standard stimuli and the stimulus onset asynchrony was 300 ms. Before the data acquisition session, each subject’s anatomical landmarks and the scalp locations of each electrode were recorded with an Isotrak 3D digitizer (Polhemus, Colchester, VT). Continuous EEG recordings were acquired using a Neuroscan 128 Channel SynAmps system equipped with a 128-channel Quick-Cap based on the modified 10-20 international system (Compumedics, Charlotte, NC). The electrodes at the mastoid sites served as reference electrodes. The EEG data were digitized at a 1000-Hz sampling rate with an online filter of 0.05-100 Hz. Eye-movement artifacts were monitored by recording the vertical and horizontal electro-oculogram using electrodes below and on the outer canthus of the left eye. The resistance at all electrode sites was below 5 kΩ. MRI scans were obtained with a 3-T scanner (Siemens Magnetom Trio, Erlangen, Germany) using a 12-channel head coil. The T1-weighted (T1) images were acquired using a magnetization-prepared rapid gradient echo sequence (TR 1670 ms, RE 1.89 ms, voxel size 1 × 1 × 1, FOV 250 mm, flip angle 9°, and 208 slices). 2.3. MMN CSD analysis ERP data pre-processing and source reconstruction were performed using Curry version 7 software 6

(Compumedics, Charlotte, NC). Bad channels were replaced via linear interpolation of adjacent channels (up to 7% per participant). EEG recordings were re-referenced to the common average reference data, and eye movement artifacts were reduced using the artifact reduction algorithm implemented in Curry 7 software (Semlitsch et al., 1986). Continuous EEG data were band-pass filtered between 0.1 and 30 Hz, epoched to a 100-ms pre-stimulus and a 300-ms post-stimulus and baseline-corrected using the averaged pre-stimulus interval voltage. Epochs containing EEG amplitudes that exceeded ±75 V were rejected automatically, and the number of remaining epochs exceeded 100 in all participants. Importantly, no group differences were detected in the number of remaining epochs after the artifact rejection procedure (F=0.369, p=0.692). The means and standard deviations of the numbers of remaining epochs were as follows: schizophrenia (182.7±28.5), CHR (186.5±19.3), and HC (187.5±25.5). MMN response activity was obtained by subtracting the ERPs elicited by standard stimuli from those elicited by deviant stimuli. A peak detection method was used to determine the peak MMN response latency and amplitude, which was defined as the most negative deflection between 130 and 250 ms post-stimulus onset at the Fz, FCz, and Cz electrode sites. For each subject, the EEG channel locations were co-registered to the structural MRI map using three anatomical landmarks (nasion and left and right preauricular points) and overlaid with Talairach coordinates (Fuchs et al., 1995). Individual realistic head models were constructed using the boundary element method (Fuchs et al., 1998), and minimum norm least squares (MNLS) estimation was then performed to obtain the cortical MMN CSD distribution as a function of time (Fuchs et al., 1999). The dipole orientations were constrained perpendicular to the cortical surfaces. The time series of average CSD activity within the frontal and temporal cortices of the left and right hemispheres were calculated separately for each individual (Fulham et al., 2014, Rinne et al., 2000). Those cortical regions within each subject were differentiated from one another based on the anatomical labeling of the Talairach atlas (Lancaster et al., 2000). Peak CSD strength and latency was defined as the point in which the greatest CSD strength was observed, between 130 and 250 ms post-stimulus onset. For the spatial comparison of CSD strength, Statistical Parametric Mapping (SPM) version 8 software 7

(Wellcome Department of Cognitive Neurology, London, UK) was used to project each individual’s structural MRI map onto the standardized Montreal Neurological Institute (MNI) space. The CSD distribution was interpolated to every voxel in MNI space within 5 mm of the source surface, and the data were smoothed using a 10-mm Gaussian kernel. The volume maps were thresholded at an uncorrected p<0.001 (Fulham et al., 2014). 2.4. Statistical analysis The demographic and clinical characteristics of the subjects were compared across groups using one-way analysis of variance (ANOVA), independent sample t-tests, or Welch’s test if the variances were not equal. The X2 test or Fisher’s exact test was used for analysis of categorical data. A repeatedmeasures ANOVA with age and sex as covariates was performed to reveal intergroup differences in surface MMN response amplitudes and latencies at the Fz, FCz, and Cz electrode sites. The intergroup differences in the MMN peak amplitudes and latencies at each electrode site and the MMN CSD peak strengths and latencies were tested via analysis of covariance (ANCOVA) with age and sex as covariates. Simple contrast tests were used to identify specific intergroup differences for each variable. The delay in CSD peak activation in the frontal cortex relative to the temporal cortex was compared within each group using paired t-tests. The number of subjects who showed an aberrant time sequence of MMN generator activity, detected as earlier CSD peak activation in the frontal lobe than in the temporal lobe or a time delay below 1 standard deviation (SD) of the time delay observed in HC subjects, was compared between groups using the X2 test. IBM SPSS version 22 (IBM, Armonk, NY) was used for statistical analysis. Statistical significance was set at P<0.05.

3. Results 3.1. Characteristics of the subjects No significant intergroup differences were detected in sex, handedness, GAF or the HAM-D score. 8

However, significant differences in age, IQ, years of education, and the HAM-A score were found between groups (Table 1). A post-hoc Tukey’s honest significance test revealed that CHR subjects were younger (CHR vs HC, p<0.001; CHR vs Schizophrenia, p<0.001) and less educated (CHR vs HC, p<0.001; CHR vs Schizophrenia, p=0.001) than HC subjects and schizophrenia patients. The HC group had higher IQ scores than both the CHR (p=0.042) and schizophrenia (p<0.001) groups, and CHR subjects had higher IQ scores than schizophrenia patients (p=0.003). The mean (standard deviation) duration of illness in patients with schizophrenia was 21.9 (36.8) months. 3.2. MMN and CSD results Figure 1 (a) displays the grand-average MMN waveforms at the Fz and FCz electrode sites. A repeated-measures ANOVA considering age and sex as covariates, group (HC, CHR, Schizophrenia) as a between-subject factor, and the fronto-central electrode site (Fz, FCz, Cz) as a within-subject factor revealed that the peak amplitudes of MMN responses at surface electrodes were lower in CHR subjects (p=0.041) and schizophrenia patients (p=0.001) than in HC subjects (F2,110=5.795, p=0.004). The results of ANCOVA investigating the group effect on MMN response amplitudes and latencies at each electrode site are summarized in Table 2. A simple contrast test revealed that the MMN amplitude was reduced in the patients with schizophrenia compared to the HCs at all electrode sites (Fz, p=0.001; FCz, p=0.001; Cz, p=0.032). The CHR subjects showed smaller MMN amplitudes than the HC subjects only at the FCz electrode site (p=0.029); the difference was not significant at Fz (p=0.077) and Cz (p=0.084). Age showed a significant effect at the Fz (F4,222=9.311, p=0.003) and FCz (F4,222=7.028, p=0.009) electrode sites but not at the Cz electrode site (F4,222=2.260, p=0.136). Sex did not have a significant effect. Figure 1 (b) presents the MMN CSD distribution in the brain using the standard MNI space for the three groups. In the comparison between the CSD peak strengths measured in the frontal and temporal cortices, a significant intergroup difference was found in the CSD strength in the right frontal cortex (F2,72=5.332, p=0.006); specifically, the HC group showed a greater CSD strength than both the CHR (p=0.009) and schizophrenia groups (p=0.007). No difference 9

in the CSD peak strengths measured in the temporal cortex nor the CSD peak latencies was found between groups. The SPM results revealed that HC subjects showed higher MMN CSD activity in the bilateral superior temporal gyrus and inferior frontal gyrus and in the right cingulate gyrus, middle frontal gyrus, middle temporal gyrus, precentral gyrus, and supramarginal gyrus than schizophrenia patients (Figure 2). The MMN CSD strengths in the bilateral cingulate gyrus, in the left superior temporal gyrus, inferior temporal gyrus, middle temporal gyrus, and in the right precentral gyrus were weaker in CHR individuals than in HC subjects. 3.3. Temporal behavior of discrete MMN generators Figure 3 (a) shows the time course of the MMN CSD activation in the temporal and frontal lobes across the three groups. In HC subjects, MMN CSD activation in the frontal cortex peaked after MMN CSD activation in the temporal cortex in both the left (t46=3.999, p<0.001) and right (t46=3.116, p=0.003) hemispheres. The mean latency was approximately 20.7-ms (±35.5) and 14.1-ms (±30.9), respectively. CHR individuals showed a similar temporo-frontal MMN CSD activation latency in the right hemisphere (t39=2.397, p=0.021) but not in the left hemisphere (Table 2). Normal temporo-frontal sequential activity of the MMN CSD was not observed in the schizophrenia patients, and no intergroup difference in the latency of the temporo-frontal CSD activation peaks was detected. When the normal range of latency in the temporo-frontal CSD activation was defined as within 1 SD of the average temporo-frontal CSD activation in the HC subjects (left hemisphere, delay>-14.8 ms; right hemisphere, delay>-16.9 ms), a significant intergroup difference in the percentage of participants who showed an aberrant or normal sequence of temporo-frontal CSD activity was found using the X2 test. Figure 3 (b) displays the CSD peak latencies in the temporal and frontal cortices as well as participants who showed an aberrant sequence of CSD activation. The percentage of subjects who showed aberrant temporal behavior of the MMN generators in the left hemisphere was significantly higher in the schizophrenia patients (p=0.013) than in the HCs, and a trend level difference was found between the CHR (p=0.087) and HC groups (schizophrenia patients, 24.1%; CHR individuals, 15.0%; 10

and HC subjects, 4.3%). In the right hemisphere, a greater percentage of schizophrenia patients (37.9%) showed aberrant temporal behavior of MMN CSD activity, reflecting that CSD activity in the temporal cortex peaked after the activity in the frontal cortex in the CHR (10.0%) and HC (10.6%) subjects.

4. Discussion The present study investigated automatic auditory processing and functional connectivity among activated brain regions as indexed by MMN CSD in patients with schizophrenia and subjects at CHR for psychosis. At the surface electrode level, the MMN response amplitudes were lower in the schizophrenia patients and to a lesser degree in the CHR subjects than in the HCs. At the source level, schizophrenia patients showed impaired MMN CSD strengths compared to HCs in widespread brain regions. In CHR individuals, a reduction in MMN CSD strengths compared to HCs was also found, but the extent of these impairments in the involved cortical regions were less than those in schizophrenia patients. In addition, we determined the sequential pattern of the temporo-frontal activities of discrete MMN generators in HCs; this pattern was disrupted in both schizophrenia patients and, to a lesser extent, CHR individuals. In other words, in contrast to HCs, patients with schizophrenia and subjects at CHR for psychosis did not show peak activation of MMN generators in the frontal cortex after that in the temporal cortex. These results indicate altered functional connectivity among MMN-related brain regions between these three groups. To our knowledge, this study is the first to report reduced MMN CSD strengths in the frontal cortex of CHR subjects, although a reduced magnetic counterpart of MMN dipole moment in the superior temporal gyrus of this group was reported in our previous study (Shin et al., 2009). CSD analysis has advantages in localizing multiple distributed sources and in determining the time courses of discrete source activities at the whole-brain level compared to the equivalent dipole model, 11

which is calculated using a fixed dipole location (Luck, 2005). In this study, similar to the findings in schizophrenia patients, the CHR individuals showed reduced MMN CSD strengths not only in the temporal cortices but also in the frontal cortices, including the cingulate cortex. The MMN CSD reduction of the medial frontal regions in the schizophrenia patients was in convergence with findings from previous studies (Fulham et al., 2014, Miyanishi et al., 2013, Takahashi et al., 2013). Attentional switching, augmentation of discrepancies between auditory stimuli, and inhibition of irrelevant activation of MMN generators in the temporal cortex have been suggested as roles for MMN generators in the frontal cortex (Doeller et al., 2003, Opitz et al., 2002, Rinne et al., 2005) and are closely associated with the working memory and executive functions mediated by the cingulate cortex (Carter et al., 1999, Lenartowicz and McIntosh, 2005). For example, Miyanishi reported a significant association between the CSD strength in the medial frontal lobe and the working memory performance in schizophrenia patients. (Miyanishi et al., 2013). Considering that CHR subjects share broad domains of cognitive dysfunction with schizophrenia patients (Bora et al., 2014, de Paula et al., 2015, Piskulic et al., 2016), the results of the current study indicating that abnormality of the MMN CSD in the frontal cortex is apparent among those in the at risk state of psychosis may have a significant impact on our understanding of the pathophysiology of schizophrenia. In line with the results of previous studies, HCs displayed a significant delay in MMN CSD peak activity in the frontal cortex relative to the temporal cortex in the current study (Fulham et al., 2014, Rinne et al., 2000). However, we could not find normal temporo-frontal sequential activation of discrete MMN generators in schizophrenia patients in either hemisphere or in the CHR individuals on the left hemisphere. In the study by Fulham et al., the temporo-frontal delay in MMN CSD onset was reported only in HCs because schizophrenia patients displayed high variability in MMN CSD onset latencies in the frontal cortex (Fulham et al., 2014). However, the observation of high variability in MMN generator activation latencies (i.e., aberrant temporal behavior of discrete MMN generators) might reflect a core pathophysiology of schizophrenia: disconnectivity among MMN-related brain regions. This result is in line with the findings of previous studies that reported increased trial to trial 12

variability and impaired harmonic oscillation in schizophrenia patients, suggesting that schizophrenia patients had deficits in the synchronized functional activity of different brain regions (Ford et al., 1994, Kwon et al., 1999). Structural brain imaging studies have reported altered thalamocortical and cortico-cortical connectivity in both patients with schizophrenia and subjects at CHR for psychosis (Cho et al., 2016, von Hohenberg et al., 2014, Wheeler and Voineskos, 2014), and this disconnectivity might be a cause of abnormal auditory signal transmission and the consecutive alteration in the time course of MMN generator activities. Moreover, previous studies demonstrated a less prominent alteration of white matter structure in CHR subjects than in schizophrenia patients and showed the progressive change in white matter structure during the development of psychosis (Peters et al., 2010, von Hohenberg et al., 2014, Ziermans et al., 2012). Those findings might support the findings of the present study that the normal temporo-frontal sequence of activity of discrete MMN generators was relatively preserved in the right hemisphere in the CHR group but that this normal pattern was disrupted in both hemispheres in the schizophrenia group. N-methyl-D-aspartate (NMDA) receptor-mediated glutamatergic activity has been suggested as a cellular mechanism of MMN response generation, and the close relationship of NMDA receptor hypofunction with schizophrenia pathophysiology has been studied (Goff and Wine, 1997, Javitt et al., 2012, Korostenskaja et al., 2007). Recently, NMDA receptor and glutamate functions in the prodromal state have drawn increasing attention because these functions may reflect the developmental changes in cellular pathways during the transition to the psychosis state. Human CHR studies reported inconsistent results to date; however, a portion of those studies demonstrated an association between abnormalities in glutamatergic activity and the progression of psychotic disorder (Treen et al., 2016). In addition, animal studies showed that repeated blockage of NMDA receptors produced schizophrenia-like structural brain changes and frontal lobe dysfunction in adolescent rats (Li et al., 2016, Wu et al., 2016). Those findings suggest that gradual and repeated hypofunction of NMDA receptors might produce longitudinal disruptions in brain structures and functions that ultimately lead to the onset of psychotic disorder. In this study, we found that the CHR subjects showed 13

intermediate alterations in the MMN amplitude at the surface level as well as the strength and temporal behavior of MMN generator activities at the source level between the HCs and schizophrenia patients. These results support the progressive nature of the NMDA receptor and glutamate system alterations during the development of psychotic disorder. This study had several limitations. First, most of the schizophrenia patients were taking antipsychotic medication at the time of EEG recording. Although it has been reported that certain antipsychotic medications do not influence MMN response amplitudes or latencies in patients with schizophrenia (Rissling et al., 2012, Umbricht et al., 1998, 1999), our results should be interpreted cautiously because little is known about the effect of these medications on MMN source strengths and temporal behaviors. Second, the definition of the normal range of latency in temporo-frontal MMN CSD peak activation (i.e., within 1 SD of that in HCs) is somewhat arbitrary because this study is the first to investigate the temporal behavior of MMN generators in the frontal and temporal cortices in a cohort of schizophrenia patients. However, the sequential temporo-frontal activity of MMN generators in HCs has been consistently found in many previous studies; thus, the delay of MMN generator activation in the frontal lobe relative to the temporal lobe activation in HCs might be used as a reference for the normal range of MMN generator activities. Third, the recently reported transition rate to psychosis in the Seoul Youth Clinic was approximately 13 percent and declined yearly (Lim et al., 2016). Regarding the fact that not all of the CHR subjects converted to psychosis, the observations of intermediate MMN CSD results in the CHR subjects in this study should be interpreted in consideration with the CHR heterogeneity. Fourth, there is a lack of direct intracranial evidence confirming the functional connectivity between the fontal and temporal cortices during the MMN process. Thus, functional connectivity can be analogized only from the previous indirect fMRI and ERP evidence and interpretations should be made with caution.

5. Conclusions 14

We aimed to reveal dysfunction and disconnections among brain regions that are responsible for pre-attentive auditory processing in patients with schizophrenia and in subjects at CHR for psychosis. Reduced MMN CSD strengths and aberrant time behavior of discrete MMN generators were observed in both patients with established schizophrenia and individuals at CHR for psychosis. These findings suggest that there are MMN generator impairments and functional disconnectivity in MMNrelated brain regions even prior to the onset of psychosis. These results might correspond to the findings of previous studies that progressive white matter changes and NMDA receptor hypofunction occur during the development of psychotic disorder. Therefore, we suggest that MMN generator activity can be used to help not only identify individuals with emerging psychosis but also reveal the pathophysiologic mechanism underlying schizophrenia.

Conflict of interest All authors have no conflicts of interest to declare.

Acknowledgments The Basic Science Research Program through the National Research Foundation of Korea (NRF), which is funded by the Ministry of Science, ICT and Future Planning (Grant no. 2016R1E1A1A02921618), supported this research.

References Alho K, Woods DL, Algazi A, Knight RT, Naatanen R. Lesions of frontal cortex diminish the auditory mismatch negativity. Electroencephalogr Clin Neurophysiol 1994;91:353-362. Benetti S, Mechelli A, Picchioni M, Broome M, Williams S, McGuire P. Functional integration between 15

the posterior hippocampus and prefrontal cortex is impaired in both first episode schizophrenia and the at risk mental state. Brain 2009;132:2426-2436. Bodatsch M, Brockhaus-Dumke A, Klosterkotter J, Ruhrmann S. Forecasting psychosis by eventrelated potentials-systematic review and specific meta-analysis. Biol Psychiatry 2015;77:951958. Bodatsch M, Ruhrmann S, Wagner M, Muller R, Schultze-Lutter F, Frommann I, et al. Prediction of psychosis by mismatch negativity. Biol Psychiatry 2011;69:959-966. Bora E, Fornito A, Radua J, Walterfang M, Seal M, Wood SJ, et al. Neuroanatomical abnormalities in schizophrenia: a multimodal voxelwise meta-analysis and meta-regression analysis. Schizophr Res 2011;127:46-57. Bora E, Lin A, Wood SJ, Yung AR, McGorry PD, Pantelis C. Cognitive deficits in youth with familial and clinical high risk to psychosis: a systematic review and meta-analysis. Acta Psychiatr Scand 2014;130:1-15. Carletti F, Woolley JB, Bhattacharyya S, Perez-Iglesias R, Fusar Poli P, Valmaggia L, et al. Alterations in white matter evident before the onset of psychosis. Schizophr Bull 2012;38:1170-1179. Carter CS, Botvinick MM, Cohen JD. The contribution of the anterior cingulate cortex to executive processes in cognition. Rev Neurosci 1999;10:49-57. Cho KI, Shenton ME, Kubicki M, Jung WH, Lee TY, Yun JY, et al. Altered thalamo-cortical white matter connectivity: probabilistic tractography study in clinical-high risk for psychosis and first-episode psychosis. Schizophr Bull 2016;42:723-731. Crossley NA, Mechelli A, Fusar-Poli P, Broome MR, Matthiasson P, Johns LC, et al. Superior temporal lobe dysfunction and frontotemporal dysconnectivity in subjects at risk of psychosis and in first-episode psychosis. Hum Brain Mapp 2009;30:4129-4137. de Paula AL, Hallak JE, Maia-de-Oliveira JP, Bressan RA, Machado-de-Sousa JP. Cognition in at-risk mental states for psychosis. Neurosci Biobehav Rev 2015;57:199-208. Doeller CF, Opitz B, Mecklinger A, Krick C, Reith W, Schroger E. Prefrontal cortex involvement in 16

preattentive auditory deviance detection: neuroimaging and electrophysiological evidence. Neuroimage 2003;20:1270-1282. Ford JM, White P, Lim KO, Pfefferbaum A. Schizophrenics have fewer and smaller P300s: a singletrial analysis. Biol Psychiatry 1994;35:96-103. Friston KJ. Schizophrenia and the disconnection hypothesis. Acta Psychiatr Scand Suppl 1999;395:6879. Friston KJ, Frith CD. Schizophrenia: a disconnection syndrome? Clin Neurosci 1995;3:89-97. Fuchs M, Drenckhahn R, Wischmann HA, Wagner M. An improved boundary element method for realistic volume-conductor modeling. IEEE Trans Biomed Eng 1998;45:980-997. Fuchs M, Wagner M, Kohler T, Wischmann HA. Linear and nonlinear current density reconstructions. J Clin Neurophysiol 1999;16:267-295. Fuchs M, Wischmann HA, Wagner M, Kruger J. Coordinate system matching for neuromagnetic and morphological reconstruction overlay. IEEE Trans Biomed Eng 1995;42:416-420. Fulham WR, Michie PT, Ward PB, Rasser PE, Todd J, Johnston PJ, et al. Mismatch negativity in recentonset and chronic schizophrenia: a current source density analysis. PLoS One 2014;9:e100221. Gaebler AJ, Mathiak K, Koten JW, Jr., Konig AA, Koush Y, Weyer D, et al. Auditory mismatch impairments are characterized by core neural dysfunctions in schizophrenia. Brain 2015;138:1410-1423. Giard MH, Perrin F, Pernier J, Bouchet P. Brain generators implicated in the processing of auditory stimulus deviance: a topographic event-related potential study. Psychophysiology 1990;27:627-640. Goff DC, Wine L. Glutamate in schizophrenia: clinical and research implications. Schizophr Res 1997;27:157-168. Javitt DC, Zukin SR, Heresco-Levy U, Umbricht D. Has an angel shown the way? Etiological and therapeutic implications of the PCP/NMDA model of schizophrenia. Schizophr Bull 2012;38:958-966. 17

Jemel B, Achenbach C, Muller BW, Ropcke B, Oades RD. Mismatch negativity results from bilateral asymmetric dipole sources in the frontal and temporal lobes. Brain Topogr 2002;15:13-27. Jung MH, Jang JH, Kang DH, Choi JS, Shin NY, Kim HS, et al. The reliability and validity of the korean version of the structured interview for prodromal syndrome. Psychiatry Investig 2010;7:257-263. Kim M, Kim SN, Lee S, Byun MS, Shin KS, Park HY, et al. Impaired mismatch negativity is associated with current functional status rather than genetic vulnerability to schizophrenia. Psychiatry Res 2014;222:100-106. Kim Z, Lee Y, Lee M. Tow-and four-subtest short forms of the Korean-Wechsler Adult Intelligence Scale. Seoul J Psychiatry 1994;19:121-126. Korostenskaja M, Nikulin VV, Kicic D, Nikulina AV, Kahkonen S. Effects of NMDA receptor antagonist memantine on mismatch negativity. Brain Res Bull 2007;72:275-283. Kwon JS, Byun MS, Lee TY, An SK. Early intervention in psychosis: Insights from Korea. Asian J Psychiatr 2012;5:98-105. Kwon JS, O'Donnell BF, Wallenstein GV, Greene RW, Hirayasu Y, Nestor PG, et al. Gamma frequency-range abnormalities to auditory stimulation in schizophrenia. Arch Gen Psychiatry 1999;56:1001-1005. Lancaster JL, Woldorff MG, Parsons LM, Liotti M, Freitas CS, Rainey L, et al. Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp 2000;10:120-131. Lee SH, Kubicki M, Asami T, Seidman LJ, Goldstein JM, Mesholam-Gately RI, et al. Extensive white matter abnormalities in patients with first-episode schizophrenia: a Diffusion Tensor Iimaging (DTI) study. Schizophr Res 2013;143:231-238. Lenartowicz A, McIntosh AR. The role of anterior cingulate cortex in working memory is shaped by functional connectivity. J Cogn Neurosci 2005;17:1026-1042. Li JT, Su YA, Wang HL, Zhao YY, Liao XM, Wang XD, et al. Repeated blockade of NMDA receptors during adolescence impairs reversal learning and disrupts GABAergic interneurons in rat 18

medial prefrontal cortex. Front Mol Neurosci 2016;9:17. Liasis A, Towell A, Alho K, Boyd S. Intracranial identification of an electric frontal-cortex response to auditory stimulus change: a case study. Brain Res Cogn Brain Res 2001;11:227-233. Lim KO, Lee TY, Kim M, Chon MW, Yun JY, Kim SN, et al. Early referral and comorbidity as possible causes of the declining transition rate in subjects at clinical high risk for psychosis. Early Interv Psychiatry 2016;doi:10.1111/eip.12363. Luck SJ. An introduction to the event-related potential technique. Massachusetts Institute of Technology, Cambridge, Massachusetts 02142: The MIT Press; 2005. Miller TJ, McGlashan TH, Rosen JL, Somjee L, Markovich PJ, Stein K, et al. Prospective diagnosis of the initial prodrome for schizophrenia based on the Structured Interview for Prodromal Syndromes: preliminary evidence of interrater reliability and predictive validity. Am J Psychiatry 2002;159:863-865. Miyanishi T, Sumiyoshi T, Higuchi Y, Seo T, Suzuki M. LORETA current source density for duration mismatch negativity and neuropsychological assessment in early schizophrenia. PLoS One 2013;8:e61152. Naatanen R, Michie PT. Early selective-attention effects on the evoked potential: a critical review and reinterpretation. Biol Psychol 1979;8:81-136. Naatanen R, Tervaniemi M, Sussman E, Paavilainen P, Winkler I. "Primitive intelligence" in the auditory cortex. Trends Neurosci 2001;24:283-288. Opitz B, Rinne T, Mecklinger A, von Cramon DY, Schroger E. Differential contribution of frontal and temporal cortices to auditory change detection: fMRI and ERP results. Neuroimage 2002;15:167-174. Perez VB, Woods SW, Roach BJ, Ford JM, McGlashan TH, Srihari VH, et al. Automatic auditory processing deficits in schizophrenia and clinical high-risk patients: forecasting psychosis risk with mismatch negativity. Biol Psychiatry 2014;75:459-469. Peters BD, Dingemans PM, Dekker N, Blaas J, Akkerman E, van Amelsvoort TA, et al. White matter 19

connectivity and psychosis in ultra-high-risk subjects: a diffusion tensor fiber tracking study. Psychiatry Res 2010;181:44-50. Piskulic D, Liu L, Cadenhead KS, Cannon TD, Cornblatt BA, McGlashan TH, et al. Social cognition over time in individuals at fclinical high risk for psychosis: findings from the NAPLS-2 cohort. Schizophr Res 2016;171:176-181. Rinne T, Alho K, Ilmoniemi RJ, Virtanen J, Naatanen R. Separate time behaviors of the temporal and frontal mismatch negativity sources. Neuroimage 2000;12:14-19. Rinne T, Degerman A, Alho K. Superior temporal and inferior frontal cortices are activated by infrequent sound duration decrements: an fMRI study. Neuroimage 2005;26:66-72. Rissling AJ, Braff DL, Swerdlow NR, Hellemann G, Rassovsky Y, Sprock J, et al. Disentangling early sensory information processing deficits in schizophrenia. Clin Neurophysiol 2012;123:19421949. Rosburg T, Trautner P, Dietl T, Korzyukov OA, Boutros NN, Schaller C, et al. Subdural recordings of the mismatch negativity (MMN) in patients with focal epilepsy. Brain 2005;128:819-828. Semlitsch HV, Anderer P, Schuster P, Presslich O. A solution for reliable and valid reduction of ocular artifacts, applied to the P300 ERP. Psychophysiology 1986;23:695-703. Shin KS, Kim JS, Kang DH, Koh Y, Choi JS, O'Donnell BF, et al. Pre-attentive auditory processing in ultra-high-risk for schizophrenia with magnetoencephalography. Biol Psychiatry 2009;65:1071-1078. Takahashi H, Rissling AJ, Pascual-Marqui R, Kirihara K, Pela M, Sprock J, et al. Neural substrates of normal and impaired preattentive sensory discrimination in large cohorts of nonpsychiatric subjects and schizophrenia patients as indexed by MMN and P3a change detection responses. Neuroimage 2013;66:594-603. Treen D, Batlle S, Molla L, Forcadell E, Chamorro J, Bulbena A, et al. Are there glutamate abnormalities in subjects at high risk mental state for psychosis? A review of the evidence. Schizophr Res 2016;171:166-175. 20

Umbricht D, Javitt D, Novak G, Bates J, Pollack S, Lieberman J, et al. Effects of clozapine on auditory event-related potentials in schizophrenia. Biol Psychiatry 1998;44:716-725. Umbricht D, Javitt D, Novak G, Bates J, Pollack S, Lieberman J, et al. Effects of risperidone on auditory event-related potentials in schizophrenia. Int J Neuropsychopharmacol 1999;2:299304. Umbricht D, Krljes S. Mismatch negativity in schizophrenia: a meta-analysis. Schizophr Res 2005;76:123. von Hohenberg CC, Pasternak O, Kubicki M, Ballinger T, Vu MA, Swisher T, et al. White matter microstructure in individuals at clinical high risk of psychosis: a whole-brain diffusion tensor imaging study. Schizophr Bull 2014;40:895-903. Wheeler AL, Voineskos AN. A review of structural neuroimaging in schizophrenia: from connectivity to connectomics. Front Hum Neurosci 2014;8:653. Wu H, Wang X, Gao Y, Lin F, Song T, Zou Y, et al. NMDA receptor antagonism by repetitive MK801 administration induces schizophrenia-like structural changes in the rat brain as revealed by voxel-based morphometry and diffusion tensor imaging. Neuroscience 2016;322:221-233. Wynn JK, Sugar C, Horan WP, Kern R, Green MF. Mismatch negativity, social cognition, and functioning in schizophrenia patients. Biol Psychiatry 2010;67:940-947. Ziermans TB, Schothorst PF, Schnack HG, Koolschijn PC, Kahn RS, van Engeland H, et al. Progressive structural brain changes during development of psychosis. Schizophr Bull 2012;38:519-530.

Figure Legends

Fig. 1. (a) Grand-averaged mismatch negativity (MMN) waveforms across the three groups at the Fz and FCz electrode sites. (b) Grand-averaged MMN current density reconstruction 21

(CDR) in the brain using the standard MNI space across the three groups.

Fig. 2. Comparison of mismatch negativity (MMN) current source density (CSD) using statistical parametric mapping (SPM), thresholded at p<0.01 (uncorrected); The left hemisphere is depicted on the left in the axial slices. (a) presents regions in which healthy controls (HCs) showed greater CSD than schizophrenia patients, (b) presents regions in which HCs showed greater CSD than subjects at clinical high risk (CHR) for psychosis. There were no clusters in which schizophrenia patients or CHR subjects showed greater CSD than HCs.

Fig. 3. (a) Time course of mismatch negativity (MMN) current source density (CSD) strength in the frontal and temporal cortices in both hemispheres across the three groups. (b) Latencies in CSD activation peaks in the temporal and frontal cortices in both hemispheres for each subject. Individuals with an aberrant sequence of MMN generator activity, reflecting that MMN response activation in the temporal cortex peaked after that in the frontal cortex, are marked in orange color.

22

Table 1. Demographic and clinical characteristics of the samples. HC

CHR

SZ

Statistical analysisa

(N=47)

(N=40)

(N=29)

χ2, F or T (df)

P

Sex (Male/Female)

29/18

28/12

13/16

4.512 (2)

0.105

Handedness (Right/Left)

46/1

37/3

25/4

3.836 (2)

0.147

Age (years)

24.6 (5.3)

19.9 (3.0)

24.4 (5.1)

13.738 (2, 113)

<0.001**

IQ

115.3 (13.4)

108.5 (12.1)

98.0 (13.2)

16.011 (2, 113)

<0.001**

Education (years)

14.5 (1.8)

12.3 (1.6)

14.0 (2.1)

15.791 (2, 113)

<0.001**

Positive symptoms

NA

NA

15.8 (5.5)

NA

NA

Negative symptoms

NA

NA

17.6 (5.7)

NA

NA

General symptoms

NA

NA

33.5 (6.8)

NA

NA

Positive symptoms

NA

9.2 (2.8)

NA

NA

NA

Negative symptoms

NA

14.2 (5.9)

NA

NA

NA

Disorganization

NA

4.2 (2.5)

NA

NA

NA

General symptoms

NA

6.4 (4.0)

NA

NA

NA

GAF

NA

50.5 (7.6)

48.1 (10.9)

-1.040 (1, 67)

0.302

HAM-A

NA

9.2 (6.3)

6.0 (3.8)

2.407 (1, 65)

0.019*

HAM-D

NA

11.3 (6.5)

8.8 (4.4)

1.731 (1, 65)

0.088

PANSS

SOPS

Abbreviations: HC, Healthy Control; CHR, Clinical High Risk; SZ, Schizophrenia; IQ, Intelligent Quotient; PANSS, Positive and Negative Syndrome Scale; SOPS, Scale of Prodromal Symptoms; GAF, Global Assessment of Functioning; HAM-A, Hamilton Rating Scale for Anxiety; HAM-D, Hamilton Rating Scale for Anxiety; NA, not applicable. Data are given as the mean (standard deviation). ** P<.005. * P<.05 a Analysis of variance, Independent t test or Welch's t test if the variances were not equal, χ 2 analysis or Fisher's exact test for categorical data. b Data from two schizophrenia patients were missing.

23

Table 2. Comparison of mismatch negativity (MMN) amplitudes, latencies, and current source density (CSD) across groups. HC

CHR

SZ

Statistical analysisa

(N=47)

(N=40)

(N=29)

F or χ2 (df)

P

Fz electrode site

-2.696 (1.241)

-2.538 (1.139)

-1.861 (0.932)

5.972 (3, 72)

0.003**

FCz electrode site

-2.713 (1.214)

-2.411 (0.900)

-1.911 (1.071)

6.393 (3, 72)

0.002**

Cz electrode site

-2.018 (1.055)

-1.762 (0.662)

-1.615 (0.856)

2.863 (3, 72)

0.061

Fz electrode site

182.4 (25.1)

175.3 (23.0)

183.7 (29.7)

0.755 (3, 72)

0.463

FCz electrode site

183.3 (22.5)

179.8 (21.5)

179.8 (29.5)

0.315 (3, 72)

0.730

Cz electrode site

182.7 (22.2)

178.2 (22.5)

178.6 (26.7)

0.415 (3, 72)

0.661

Left temporal cortex

0.014 (0.017)

0.009 (0.010)

0.007 (0.006)

2.672 (3, 72)

0.074

Left frontal cortex

0.018 (0.031)

0.009 (0.012)

0.008 (0.010)

2.817 (3, 72)

0.064

Right temporal cortex

0.014 (0.015)

0.010 (0.010)

0.009 (0.009)

2.740 (3, 72)

0.069

Right frontal cortex

0.017 (0.022)

0.009 (0.009)

0.008 (0.006)

5.332 (3, 72)

0.006*

Left temporal cortex

183.3 (33.2)

184.6 (36.4)

189.1 (37.5)

0.678 (3, 72)

0.510

Left frontal cortex

204.0 (33.8)

194.0 (38.4)

204.0 (33.8)

1.060 (3, 72)

0.350

Right temporal cortex

186.1 (29.5)

179.1 (28.0)

196.0 (39.4)

2.223 (3, 72)

0.113

Right frontal cortex

200.2 (31.5)

192.4 (29.6)

203.6 (33.4)

2.001 (3, 72)

0.140

Delay of left frontal cortex activation relative to left temporal cortex (ms)b

20.7 (35.5)**

9.4 (44.1)

13.7 (52.9)

0.393 (3, 72)

0.676

Delay of right frontal cortex activation relative to right temporal cortex (ms)b

14.1 (30.9)**

13.4 (35.3)*

7.7 (47.9)

0.284 (3, 72)

0.753

Left hemisphere (Aberrantc/Normald)

2/45

6/34

7/22

6.529 (2)

0.038*

5/42

4/36

11/18

11.606 (2)

0.003**

MMN peak amplitude (μV)

MMN peak latency (ms)

MMN CSD peak strength

MMN CSD peak latency (ms)

Right hemisphere

(Aberrantc/Normald)

Abbreviations: HC, Healthy Control; CHR, Clinical High Risk; SZ, Schizophrenia. Data are given as the mean (standard deviation). * P<.05. ** P<.005. a Analysis of variance with age and sex as covariates, χ 2 analysis or Fisher's exact test for categorical data. b Paired t-test for within group comparison of current source density (CSD) peak latencies between regions of interest. c Number of subjects who showed delay of frontal lobe CSD peak activation relative to temporal lobe CSD peak activation below the lower level of one standard deviation of the activation of the healthy subjects. d Number of subjects who showed delay of frontal lobe CSD peak activation relative to temporal lobe CSD peak activation above the lower level of one standard deviation of the activation of the healthy subjects.

24

25

26

27