Mismatch negativity of preschool children at risk of developing mental health problems

Abstract This study examined the relationship between mismatch negativity (MMN) during the passive oddball task and clinical assessment using a behavioral scale in nonclinical preschool children to identify neurobiological endophenotypes associated with the risk of developing mental health problems. We assessed the risk of developing mental health problems in preschool children using the Strengths and Difficulties Questionnaire, which is used worldwide as a behavior‐based screening tool for assessing mental health risks, and examined its relevance to amplitude and latency MMN. As a result, we found that children at a higher risk of mental health problems had smaller MMN amplitudes than those at lower risk. It was also found that MMN amplitude was negatively correlated with the assessed higher risk of mental health problems. Although it is not clear what neural mechanisms underlie the functional association between MMN and risk of mental health problems in preschool children, the findings of this study indicate that there is an involvement of individual differences in auditory processing in childhood mental health problems. The findings suggest that such neurological changes may be prodromal symptoms of the onset of psychiatric disorders and applicable as endophenotypic markers for the early detection of various psychiatric disorders.

had smaller MMN amplitudes than those at lower risk. It was also found that MMN amplitude was negatively correlated with the assessed higher risk of mental health problems. Although it is not clear what neural mechanisms underlie the functional association between MMN and risk of mental health problems in preschool children, the findings of this study indicate that there is an involvement of individual differences in auditory processing in childhood mental health problems. The findings suggest that such neurological changes may be prodromal symptoms of the onset of psychiatric disorders and applicable as endophenotypic markers for the early detection of various psychiatric disorders.

K E Y W O R D S
event-related potential, mismatch negativity, preschool children, strengths and difficulties questionnaire establishing effective screening methods for timely preventive interventions. It has been pointed out that alterations of mismatch negativity (MMN), an event-related potential (ERP) component, based on brain response to the detection of changes in auditory information (differences between input stimulus and the sensory memory of preceding stimulus), may be endophenotypes or biomarkers of the risk of developing psychiatric disorders. 6 Establishing effective endophenotypes or biomarkers for subclinical individuals to prevent the onset of later mental disorders should be prioritized to provide objective and cost-effective screening. 7,8 Moreover, these can be applied to the monitoring of treatment effects and the development of new treatment methods and provide new insights into the association between genetic vulnerability and the clinical phenotype of mental health problems. 7,8 The Strengths and Difficulties Questionnaire (SDQ) is one of the most frequently used questionnaires worldwide for screening mental health and difficulties in childhood and adolescence. 9 It is a brief questionnaire comprising 25 questions and has been translated into more than 80 languages. 10 A review of 48 studies of school-aged children, screened using the SDQ, found that internal consistency, test-retest reliability, and inter-rater agreement were satisfactory for the parent and teacher versions. 9 Regarding validity, it has been confirmed that it has a five-factor structure; correlations with other measures of child psychopathology were high; and evidence for the screening ability of the SDQ was convincing. 9 Furthermore, the relationship between the SDQ assessment and the prognosis has been investigated by longitudinal studies, and it has been suggested that childhood SDQ scores are significantly associated with subsequent mental health. [11][12][13][14][15][16][17] However, the neural basis for various subclinical symptoms associated with the SDQ, especially in preschool children, remains insufficiently investigated.
The purpose of this study was to examine neurobiological endophenotypes associated with the risk of developing mental health problems in preschool children for the early detection of mental health problems. We assessed the risk of developing mental health problems in preschool children, using the SDQ-used worldwide as a behavior-based screening tool for assessing mental health risks-and examined its relevance to MMN during passive oddball tasks. Our main hypothesis is that children at high risk of developing mental health problems have lower amplitude and longer latency of MMN than children at low risk, and the increased risk is negatively correlated with the amplitude and latency of MMN, as suggested by previous studies. [18][19][20] 2 | ME THODS

| Participants
Forty-three preschool children aged four to five years (23 boys, 20 girls; mean age: 60.0 ± 5.3 months; age range: 49-69 months), recruited from public nursery schools in Fukui prefecture via advertisements, participated in the study. All children were confirmed as being typically developed by a licensed clinical psychologist during the study period. None were confirmed to have had any developmental difficulties at the age of three. The race/ethnicity of all the children was Japanese. The exclusion criteria for the participants were as follows: physical problems; diagnosis of any otorhinolaryngology disorder within the last 6 months; subjects who had undergone treatment with antiallergic agents; past diagnosis of any psychiatric or neurodevelopmental disorder; head trauma with loss of consciousness; any history of epilepsy; considerable fetal exposure to alcohol or drugs; and perinatal complications, such as premature birth (gestational age <37 weeks, birth weight <2500 g). All children that participated in the experiment had no reported vision or hearing problems, previous brain injuries, known genetic disorders, or other neurological disorders. Handedness was assessed using the Edinburgh Handedness Inventory. 21

| Assessment of clinical symptoms
Behavioral characteristics and psychotic symptoms were assessed using the Strengths and Difficulties Questionnaire (SDQ) for preschool-aged children. Parents completed the SDQ, 11,22 a 25-item questionnaire, to assess children's internalizing and externalizing behavior problems as well as prosocial behavior tendencies. The SDQ is a brief behavioral screening questionnaire for 3-to 16-year-olds.
It exists in several versions to meet the needs of researchers, clinicians, and educationalists. All versions of the SDQ examine 25 attributes, some positive, and others negative. These 25 items are divided into five subscales: 1) emotional symptoms, 2) conduct problems, 3) hyperactivity/inattention, 4) peer relationship problems, and 5) prosocial behavior. It was also suggested that these subscales can be combined into groups for higher phenotypes by adding two categories: "internalizing problems" (emotional symptoms and peer relationship problems) and "externalizing problems" (conduct problems and hyperactivity/inattention). 11 To examine the association between risk to mental health and ERP components, participants were divided into two groups, high risk and low risk, according to the total difficulty score assessed by the SDQ. Based on the criteria of the Japanese version of the SDQ, 23 children with a cutoff value of 13 points or above were classified as "borderline" or "abnormal" and were included in the high-risk group, and those with 12 points or below were classified as "normal" and were included in the lowrisk group.

| MMN paradigm
The MMN components were measured from ERP induced using the auditory oddball task based on a paradigm from previous research. 24,25 Participants were presented with 1000 auditory stimuli consisting of 850 (85%) standard tones of 1000 Hz at 50 ms and 150 (15%) deviant tones of 2000 Hz at 50 ms, with a rise and fall time of 5 ms The interval between each stimulus was 600 ms The tone stimuli were presented at 75db through a speaker situated behind the participant, while participants were asked to ignore the sound and watch a silent movie animation presented by the forward visual display. The task was only to passively listen to the auditory stimuli, and no behavioral response to the stimuli was required.

| ERP recording
Electroencephalogram (EEG) data were measured using Ag/AgCl electrodes (Nihon Kohden) at five sites according to the 10-20 system of electrode placement. The electrodes were attached to an elastic electrode cap that was fastened with a chin strap. Horizontal electro-oculograms (EOGs) were recorded from electrodes placed at the outer canthus of each eye, and vertical EOGs were recorded from electrodes positioned above and below the left eye. Electrodes on bilateral ears were used as a common reference. The electrodes were attached after lightly cleaning the skin to maintain an impedance level below 10 kΩ. The EEG data were recorded by a biological amplifier (Polymate II AP216; TEAC Corp., Japan) with VitalTracer software (KISSEI COMTEC Co., Ltd., Japan). The data were preprocessed by a band-pass filter of 0.16-100 Hz and digitized at a sampling rate of 500Hz. Filters were set at a low cutoff of 0.53 Hz and a high cutoff of 30 Hz. The recordings were notch-filtered offline at 60 Hz.

| ERP data
EEG data were analyzed using EPLYZER II software (KISSEI COMTEC Co., Ltd., Japan). EEG data were preprocessed to remove artifacts due to eye movements using EOG signals and then segmented into stimulus-locked epochs of 600 ms duration (100 ms pre-stimulus to 500 ms post-stimulus). Across all conditions, the average number of epochs accepted after artifact exclusion was 831. were obtained for the three central and two temporal electrode sites (Fz, Cz, Pz, T3, and T4) and compared between groups (high risk, low risk).

| Statistical analyses
All variables were tested for normal distribution using the Kolmogorov-Smirnov test. Demographics variables were tested using chi-squared and t tests, including age, sex, and handedness.

| Demographic characteristics
The demographic and clinical characteristics of the study participants are shown in Table 1. No significant differences were found in

| MMN amplitude and latency
Means and standard deviations of MMN amplitude and latency at the different electrode sites by risk group are shown in Table 2 F(2,78) = 0.50; ps > .10; Figure 1B). These results suggest that a higher risk of mental health problems is associated with a lower amplitude rather than latency alterations in MMN.

| Relationship between MMN amplitude and behavioral adaptation levels
To investigate the relationship between atypical neural response and behavioral risk of mental health problems in children, we performed a correlation analysis with the SDQ total score and MMN amplitudes observed for their attenuation in the high-risk group.
The results indicated a significant negative correlation with the SDQ total score at Cz and Pz (Cz: r = −0.48, P < .001; Pz: r = −0.33, P = .03; Figure 2B,C), whereas no correlation was observed for Fz (r = −0.33, P = .03; Figure 2A). These results suggest that a higher risk of mental health problems is associated with lower MMN amplitudes at Cz and Pz. Next, when the relationship with the SDQ subscale was examined, a significant negative correlation was confirmed between the externalizing problems and both Cz and Pz (Cz: r = −0.50, P < .001; Pz: r = −0.38, P = .01), but the internalizing problems only correlated with Cz (r = −0.37, P = .01).

| D ISCUSS I ON
This study investigated the relationship between MMN during the passive oddball task and clinical assessment using the behavioral scale in preschool children to identify neurobiological endophenotypes associated with the risk of developing mental health problems.
As a result, we found that children at a higher risk of mental health problems had smaller MMN amplitudes than those at lower risk. It was also found that MMN amplitude was negatively correlated with To the best of our knowledge, this is the first study to investigate the relationship between MMN characteristics and the risk of mental health problems in a community sample of preschool children. As mentioned above, MMN is induced in the frontal cortex, in association with attentional shift based on the automatic detection of changes in auditory stimuli. 6,28 MMN is an ERP component of preattentive processing, and it has been suggested that MMN is involved in preattentive cognitive processing. 29 In the context of mental health problems, it has been reported that the amplitude of MMN is attenuated in children with attention-deficit/hyperactivity disorder (ADHD), compared with the control-group children. 19,30,31 It has also been indicated that MMN alterations are associated with several psychiatric disorders, such as schizophrenia, bipolar disorder, and major depression. [32][33][34][35] In addition, MMN has been suggested as The numbers in parentheses denote standard error.

TA B L E 2
Average MMN amplitude and latency in high-risk and low-risk children an indicator of cognitive decline observed in most psychiatric disorders. 36 The results of the present study are consistent with these findings. In this study, we found that the reduction in MMN amplitude at Cz and Pz was associated with behavioral symptoms in mental health problems, as measured by the SDQ scores, suggesting that MMN amplitude may be a precursor indicator for detecting abnormalities in cognitive function in preschool children at risk of developing mental health problems.
It is an interesting question as to what symptoms in psychiatric health problems are specifically associated with reduced MMN amplitude. In this study, externalizing problems were correlated with both Cz and Pz amplitudes on the SDQ subscale, while internalizing problems were only correlated with Cz amplitude. Although hyperactivity/inattention is a major component of externalizing problems, previous studies have reported that children with ADHD not only have reduced MMN amplitude, but their severity, including both hyperactivity and inattention, is also negatively correlated with Pz amplitude. 19,31 On the other hand, reduced MMN has also been suggested in association with internalizing problems in children. 20,37 Previous studies have reported that children with socio-emotional difficulties, such as autism spectrum disorders and social with- It should also be noted that we could not find any significant differences in the amplitude at the Fz or the latency of all the elec- In conclusion, in the general population of preschool children, we found that children at a higher risk of developing mental health problems had reduced MMN amplitudes in the frontal and parietal lobe regions, compared to those at a lower risk. Since such neurological changes may be prodromal symptoms of the onset of psychiatric disorders, they may be applicable as endophenotypic markers for the early detection of various psychiatric disorders. In the future, it is important to clarify how the association between the identified neurological changes and the risk of mental illness follows in subsequent development, especially in adolescence.

ACK N OWLED G M ENTS
This work was supported by JSPS KAKENHI Grant Numbers JP17K19898, JP19K21755, JP19H00617 (to AT), and AMED under Grant Number JP20gk0110052 (to AT).

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
TA and AT conceived the project and designed the experiments. TA

I N FO R M E D CO N S E NT
All parents of the participants provided written informed consent for participation in this study.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data cannot be made publicly available as data sharing was not included in the consent.