Peer victimization and the association with hippocampal development and working memory in children with and without ADHD

Children with ADHD are at risk of experiencing peer victimization, which is associated with delayed brain development and cognitive diculties. We aimed to evaluate the relationship between problem behaviours, peer victimization, hippocampal morphology, and working memory in children with and without ADHD. Methods 218 typically developing participants (50.5% male) and 232 participants diagnosed with ADHD (77.6% male) were recruited. The ADHD group was subdivided into inattentive (ADHD-I) or combined (ADHD-C) type. The Child Behaviour Checklist measured problem behaviours and peer victimization. Hippocampal subeld volumes were obtained using Magnetic Resonance Imaging and the Wechsler Intelligence Scale for Children-fth edition measured working memory (WM).


Introduction
Attention De cit Hyperactivity Disorder (ADHD) affects approximately 5% of the child and youth population worldwide and is characterized by symptoms of hyperactivity-impulsivity, and inattention 1,2 .
ADHD is further characterized into three diagnostic subtypes, including ADHD hyperactive-impulsive type (ADHD-H), ADHD inattentive type (ADHD-I), and ADHD combined type (ADHD-C) 1,3 . ADHD-H is characterized by clinically elevated symptoms of hyperactivity-impulsivity 1 . ADHD-I is characterized by clinically elevated symptoms of inattention. Children and youth with ADHD-C display both clinically elevated hyperactivity-impulsivity and inattention symptoms 1 . compared to TD children 12 , some studies report decreased hippocampal volumes in ADHD compared to TD children 13,14 , and others report no differences 15 .
The hippocampus is a structure of particular interest in the adversity literature because it is involved in regulating hormonal stress responses and is composed of stress-sensitive sub elds including the Cornu Ammonis 1 (CA1), Cornu Ammonis 3 (CA3), and Cornu Ammonis 4 (CA4) 16,17 . It is also of particular interest of study in the ADHD population, considering children with ADHD are at an increased risk of experiencing chronic stress, including that related to peer victimization 12,13,14,15 .
Chronic stress has been hypothesized to impact the hippocampus in different ways. For one, during periods of increased stress, the adrenal gland releases glucocorticoids (GCs) 17 . In rodent and primate models, GCs have been observed to kill neurons in stress-sensitive sub elds of the hippocampus, for example in the CA3, thereby leading to decreased volumes 17 . Other research suggests that chronically elevated levels of stress might lead to a reduced hormonal stress responsivity that can mitigate the neurotoxic effects of GCs, resulting in unaltered or even enlarged hippocampal volumes 17,18,19 . In the literature, both decreased and enlarged hippocampal volumes have been observed. Some studies have demonstrated that early life stress is associated with decreased hippocampal volumes, although in some pediatric populations, such as in children with maltreatment-related PTSD, larger hippocampal volumes have been observed in relation to increased adversity 16,20,21 .
The research on peer victimization as a form of early life stress and its association with hippocampal volume is scarce. In some studies, smaller hippocampal volumes were found in TD children who reported more peer victimization, but further research is required to obtain a better understanding of this relationship 9,22 . Further, the association between peer victimization to hippocampal volume in the ADHD population and how this differs between ADHD subtypes has not been studied.
In addition to its involvement in regulating the stress response, the hippocampus is also a key structure in executive functioning 23 . In adult and aging populations, larger hippocampal volumes are associated with better performance on tasks of executive functioning, including working memory (WM) 23,24,25 . However, this same relationship is not consistently observed in children, with some evidence suggesting that children with larger hippocampal volumes perform worse on tasks of memory 24 . Children with ADHD exhibit de cits in WM, however the relationship between hippocampal volume and WM ability in the ADHD population remains largely unexplored 26 .
In the current work, we examined the association amongst social di culties, brain development and executive function in a large heterogenous sample of children with ADHD and TD children. We addressed three main research questions: 1) Are problem behaviours and peer victimization predicted by ADHD diagnostic category? 2) Do hippocampal sub eld volumes predict problem behaviours and peer victimization levels? 3) Do hippocampal sub eld volumes predict WM ability? Our central hypothesis is that ADHD participants will report higher levels of problem behaviours and peer victimization compared to TD children, and that hippocampal subregion volumes will signi cantly predict levels of peer victimization as well as WM ability.

Participants
Participants in this study ranged in age from 6.00 to 17.70 years. Data were collected by researchers at the Child Mind Institute and participants were recruited as part of the Healthy Brain Network initiative 27 .
Both typically developing (TD) participants and those with a clinical diagnosis of ADHD were tested at three sites: Rutgers University, New York, Citigroup Biomedical Imaging Center, Weill Cornell Medicine, New York, and Staten Island Diagnostic Research Center, New York. The research ethics boards at all respective institutions approved the study. Written informed consent was obtained from the participants' parents and written assent was obtained from the participants. A breakdown of participant demographics can be found in Table 1. Table 1 Characteristics of TD, ADHD-C and ADHD-I participants.  In combination with clinical judgment, the computerized version of the Kiddie Schedule for Affective Disorders and Schizophrenia (K-SADS-COMP) was administered under the supervision of a licensed clinician to make diagnoses of ADHD, and to con rm the absence of ADHD diagnoses in TD participants 28 . Participants diagnosed with ADHD were further divided into subgroups, those diagnosed with ADHD-I and those diagnosed with ADHD-C. Participants with an ADHD-H diagnosed were excluded from this study due to a small sample size of only 15 participants having received this subtype diagnosis. All participants in the study were uent in English, had a parent that was able to complete informant questionnaires, had an Intelligence Quotient (IQ) score of over 66, and were free of brain injury or disease. Participants were excluded if they presented with schizophrenia or bipolar disorder, psychosis, substance dependence, acute intoxication, a neurodegenerative disorder, or any other neurodevelopmental disorder. The study was conducted in accordance with the Declaration of Helsinki.

Procedures
The data for this study was collected in 4 study visits. In the rst visit, the WISC-V was administered by a trained psychometrist. The second visit consisted of the MRI protocol, including the T1 sequence conducted by a trained MRI technician. During the scan participants watched two cartoon movies, the rst movie was a short-lm titled 'The Present' and the second was a 10-minute clip from the full-length lm 'Despicable Me.' Parent questionnaires were completed during the third visit and K-SADS-COMP was completed during the fourth and nal visit.

Demographic Measures
The Barratt Simpli ed Measure of Social Status (BSMSS) was used as a measure of socioeconomic status (SES) and was completed by a parent (Barratt, 2006). The BSMSS assesses level of education and occupation for both of a child's parents, or from one parent if the child is from a single-parent household 29 . The scores were converted into a total score between 8-66 for each child, with a higher score indicative of a higher SES 29 .

Psychological Measures
To assess symptoms of hyperactivity-impulsivity and inattention, the Strengths and Weaknesses Assessment of ADHD and Normal Behaviour (SWAN) was completed by a parent (Swanson et al., 2001).
To measure levels of problem behaviours in the participants, the Child Behaviour Checklist (CBCL) was administered to parents 30 . The subscales used from the CBCL in this study included the aggressive behaviour, rule breaking behaviour, social problems and withdrawn scales. In addition, peer victimization was measured using the CBCL subscale developed by McCloskey and Stuewig 31 , comprised of four questions: "doesn't get along with other kids", "gets in many ghts", "gets teased a lot", and "not liked by other children," The Wechsler Intelligence Scale for Children-fth edition (WISC-V) was used to obtain a Working Memory Index (WMI) in all participants, including a digit span task and a picture span task 32 .

Magnetic Resonance Imaging
A Siemens 3T Trio scanner was used at the Rutgers University site 27 . At the Staten Island Diagnostic Research Center, participants were scanned on a 1.5T Siemens Avanto scanner and at Citigroup Biomedical Imaging Center participants were scanned on a Siemens 3T Prisma 27 . To adjust for the differences in scanners, statistical models included Magnetic Resonance Imaging (MRI) site as a covariate.

Statistical Analysis
Statistical analyses were completed the IBM SPSS Statistics software package (version 26, Statistical Package for the Social Sciences, IBM, Armonk, NY).
Our rst aim was to examine problem behaviour and peer victimization in children and adolescents with TD, ADHD-I, ADHD-C. In ve general linear models, we examined parent-reported problem behaviours and peer victimization (social problems, aggressive behaviour, rule breaking behaviour, withdrawal and peer victimization subscale raw scores; dependent variables) in related to diagnostic group (TD, ADHD-I, ADHD-C; independent variables). Age, sex, socioeconomic status (using the BSMSS), and study site/MRI site were adjusted for in each model.
Our second aim was to examine problem behaviour and peer victimization and their relationships to hippocampal sub eld volume. In four general linear models we examined peer victimization (subscale raw score; dependent variable), aggressive behaviour (subscale raw score; dependent variable), rule breaking behaviour (subscale raw score; dependent variable), and social problems (subscale raw score; dependent variable) in relation to hippocampal sub eld volumes (right and left hippocampal tail, subiculum, CA1, hippocampal ssure, presubiculum, parasubiculum, molecular layer, GCMLDG, CA3, CA4, mbria and HATA; independent variables), while adjusting for age, sex, socioeconomic status (using the BSMSS), diagnostic group, total cerebral volume (TCV), and MRI site.
Our third aim was to examine the association of hippocampal sub eld volumes and working memory ability. A general linear model was used to examine the relationship between the left and right hippocampal sub eld volumes (hippocampal tail, subiculum, CA1, hippocampal ssure, presubiculum, parasubiculum, molecular layer, GCMLDG, CA3, CA4, mbria and HATA; independent variables), and working memory ability (WISC-V: working memory subscale standardized score; dependent variable). We adjusted for age, sex, socioeconomic status (using the BSMSS), diagnostic group, TCV, and MRI site.

Problem Behaviours, Peer Victimization and Diagnostic Group
In our rst aim, we examined the relationship between problem behaviours from the CBCL, including social problems, rule breaking behaviour and aggressive behaviour, and diagnostic group (TD, ADHD-I, and ADHD-C). We also examined how the experience of peer victimization differs by diagnostic group (TD, ADHD-I, and ADHD-C). Age, sex, SES, and study site were adjusted for in all the models. In the problem behaviours models 433 participants were included in analysis, 204 TD children, 123 ADHD-I children, and 106 ADHD-C children.
Seventeen participants were excluded from the problem behaviour analyses due to missing data. In the peer victimization model a total of 431 total participants were included in the analysis, 203 TD children, 123 ADHD-I children, and 105 ADHD-C children.
Compared to the TD group, children and adolescents from both the ADHD-C (B = 5.42, 95%CI = 4.26-6.58, p < 0.001) and ADHD-I (B = 2.41, 95%CI = 1.32-3.51, p < 0.001) groups had signi cantly higher scores on the aggressive behaviour subscale. ADHD-C participants also had signi cantly higher scores than the TD participants on the rule breaking behaviour (

Post-Hoc Cluster Analysis
To examine the associations amongst diagnosis, hippocampal volumes, WM ability as well as ADHD symptomatology, a K-means cluster analysis was performed with 245 participants: 111 TD, 58 ADHD-C, and 76 ADHD-I. The analysis included variables of brain morphology, behaviour, and cognition (Z-scored).
The model included, left CA3 volume (p = 0.018), peer victimization score (p < 0.001), WMI (p < 0.001), and the hyperactivity-impulsivity and inattention subscales from the SWAN (both p < 0.001). Refer to Table 2 for a summary of the clusters. Cluster 1 included a total of 41 participants and is characterized as children with small left CA3 volumes, high WMI scores, low peer victimization scores, and low hyperactivity-impulsivity and inattention symptomatology scores. This group is made up of 40 TD participants, and 1 participant from the ADHD-I group. All but 3 participants in this cluster scored 0.5 SD below the mean on the peer victimization scale. In this cluster 56% of the participants had left CA3 volumes that were below the average and 75% of the participants had higher average WMI. In regard to symptoms of hyperactivity-impulsivity and inattention, 39% of the participants in cluster 1 scored 1 SD below the mean and 34% scored below 2 SD below the mean.
Cluster 2 included a total of 54 participants. Participants in this group have average left CA3 volumes, average WMI scores, high peer victimization scores, and high hyperactivity-impulsivity and inattention symptomatology. This group is made up of 19 TD participants, 24 ADHD-C participants and 11 ADHD-I participants. All participants in this cluster were above average in peer victimization scores, with approximately 41% of the cluster at least 2 SD above the mean, and 56% above 1 SD. In this cluster 54% of participants had left CA3 volumes that were below average and 20% of participants had WMI at least 1 SD below the average. In regard to symptomatology, 20% of the participants had hyperactivity-impulsivity scores at least 1 SD above the mean, and 28% had inattention scores at least 1 SD above the mean.
Cluster 3 included a total of 150 participants and can be described as participants with large left CA3 volumes, low WMI scores, low peer victimization scores, and average hyperactivity-impulsivity and inattention symptomatology. This cluster is made up of 52 TD participants, 34 ADHD-C participants, and 64 participants from the ADHD-I group. In cluster 3, approximately 79% of the participants had peer victimization scores that were 0.5 SD below the mean, A total of 60% of the participants had left CA3 volumes that were above the mean, with 21% being least 1 SD above the average. Approximately 30% of this cluster had WMI above the mean, and most participants in this cluster were above average in hyperactivity-impulsivity and inattention symptomatology, 57% and 60% respectively. When examining cluster membership by diagnostic groups, we found 36% of the total TD sample (n = 111) was grouped into cluster 1, 17% were grouped into cluster 2, and 47% were grouped into cluster 3. Of the 58 total ADHD-C participants, 41% were grouped into cluster 2, and 59% were grouped into cluster 3. Of the 76 total ADHD-I participants, 1% belonged to cluster 1, 14% were grouped into cluster 2, and 84% were grouped into cluster 3. Refer to Fig. 2 for a summary of the cluster analysis. Signi cant differences were found between Cluster 1 and Cluster 2 on WMI, peer victimization, hyperactivity-impulsivity, and inattention (all p < 0.001). Signi cant differences were found between Cluster 1 and 3 on Left CA3 volume (p = 0.015), WMI, hyperactivity-impulsivity, and inattention (all p < 0.001). Signi cant differences were found between Cluster 2 and 3 on peer victimization (p < 0.001) and hyperactivity-impulsivity (p = 0.001).

Discussion
The presence of problem behaviours, levels of peer victimization, hippocampal sub eld morphology, and working memory were assessed in a large heterogenous sample of children and adolescents further divided by diagnostic group: TD, ADHD-C and ADHD-I. We found that problem behaviours and levels of peer victimization differed between TD, ADHD-C, and ADHD-I groups. We also report that hippocampal sub eld volumes predict the presence of peer victimization levels in children with ADHD and TD children.
Lastly, hippocampal sub eld volumes were associated with working memory ability in children with ADHD and TD children. Peer victimization was highest in children that displayed high levels of hyperactivity-impulsivity. Findings suggest that children with ADHD-C who display elevated levels of hyperactivity-impulsivity may be at greater risk for peer victimization.

Problem Behaviours, Peer Victimization and Diagnostic Group
As hypothesized, children in the ADHD diagnostic groups had elevated levels problem behaviours and peer victimization compared to TD children. Parents of children with ADHD-C reported higher levels of aggressive behaviour, rule breaking behaviour, social problems, and higher levels of experiencing peer victimization in their children compared to parental reports for TD and ADHD-I children. The ADHD-I children only differed from the TD children when comparing levels of aggressive behaviour; with signi cantly higher reported levels in the ADHD-I group. No signi cant differences were found in levels of withdrawal between any of the diagnostic groups.
Children and adolescents with ADHD are typically not well-liked by their peers and often experience social rejection and victimization 5 . Children with ADHD who display more severe levels of externalizing behaviours have been observed to be the most at-risk pediatric population to experience peer victimization 5,6 . Investigation into how peer victimization differs amongst the diagnosis subtypes of ADHD is sparse. Recent ndings into the symptomatology of ADHD in relation to peer victimization suggest that more severe symptoms of hyperactivity-impulsivity are associated with higher rates of peer victimization among children 7 . Not only are hyperactive-impulsive symptoms related to experiencing more victimization, it has also been demonstrated that individuals with these symptoms are also more likely to also be perpetrators of bullying themselves 7 . This notion is consistent with the ndings from the present study. Although the peer victimization measure in the present study only accounted for experiencing victimization, the signi cantly elevated aggressive behaviour, rule breaking behaviour, and social problems scores in the ADHD-C group suggest that this group might also be perpetrators of victimization as well and might t the pro le of a bully victim 7 . As the ADHD-I group has levels of hyperactivity-impulsivity that are comparable to the TD group, it follows that they are victimized by their peers at rates similar to the TD group. Although the ADHD-I children also had signi cantly elevated aggression scores compared to TD children, the lack of hyperactivity-impulsivity symptoms combined with low rates of peer victimization suggest that this group might t into a different category than the ADHD-C group; that is, one of aggressive children and adolescents that are accepted by their peers 34 .
Experiencing peer victimization may be traumatizing, with a tremendous impact on later psychological development 35 . It is crucial that future research focuses on examining the impact of different types of victimization on child and adolescent psychological health. Determining how different types of victimization effect children can inform intervention strategies and can be useful for informing schoolbased interventions.

Hippocampal Sub eld Volume, Problem Behaviours and Peer Victimization
In the current study, contrary to our hypothesis, larger left CA3 volumes were associated with higher parent-reported rates of peer victimization in children with ADHD-C, ADHD-I and TD children. The CA3 region is a particularly stress-sensitive area and has previously been associated with early life adversity 8,36 . Typically, smaller hippocampal volumes have been associated with a greater incidence of trauma and adversity in TD children 16,17,36 . However, this association has been found to be different in children with mental health and neurodevelopmental disorders 18,21,37 . Some models suggest that chronically elevated levels of stress during childhood may cause reduced stress responsivity which in turn may mitigate the neurotoxic effects GCs typically have in atrophying the neurons of the hippocampus 17,18,19 . This cascade of events results in larger hippocampal volumes, as found in our study.
Research conducted with pediatric post-traumatic stress disorder (PTSD) populations has provided evidence for larger hippocampal volumes when compared to matched non-traumatized groups, suggesting that anxiety and stress may be associated with increased growth of the hippocampus 21 . A recent study found larger subiculum, presubiculum, and CA1 volumes to be associated with childhood trauma in bipolar children 18 . Children with ASD have also been found to have enlarged hippocampal volumes compared to TD children 37  Future research in this area should focus on examining if and how the relationship between hippocampal volume and peer victimization differs according to gender and age. To maintain the large sample size in the current study we were not able to split our population in these ways; however, these factors were used as covariates. Age was not a signi cant covariate in the model relating hippocampal volumes and peer victimization. In addition, future research is warranted in elucidating the relationship between ADHD subtype, symptomatology and peer victimization in relation to hippocampal volume, speci cally in including the ADHD-H subtype.

Hippocampal Sub eld Volume and Working Memory
The ADHD-I children had signi cantly lower scores on the WMI when compared to TD children. The ADHD-C group did not signi cantly differ from the ADHD-I group nor the TD group on WMI scores.
Working memory de cits are often seen in children with an ADHD diagnosis 26,38 . However, there is a lack of consensus on how the different ADHD subtypes differ in respect to working memory ability, and cognitive ability in general. Some studies report no differences between high and low inattention symptoms and working memory scores 39 . Other studies have found inattention to be a signi cant predictor of working memory ability in adolescents with ADHD 40 .As children with ADHD-C also have clinically elevated symptoms of inattention, it is unclear why the ADHD-C children did not display WM de cits when compared to the TD children in our study.
Larger left CA3 volumes and smaller left CA4 volumes were signi cantly predictive of higher WMI scores in all of the diagnostic groups. CA3 and CA4 are both key hippocampal sub elds that are involved in WM 41 . Larger volumes predict better cognitive function in adult samples 24,25,31 . Yet, ndings in children and adolescents provide inconsistent results to support this view. In a meta-analysis by Van Petten 24 multiple studies examining hippocampal volume and memory performance in child and adolescent populations found negative correlations between volume and memory abilities. This suggests that the positive relationship between cognitive ability and hippocampal brain volume may increase with age 24 . A recent study found that as children age, smaller hippocampal volumes are associated with superior memory abilities 42,43 . Previous studies examined whole hippocampal volumes. Future studies should investigate the relationship between hippocampal sub eld volumes and working memory ability in ADHD populations.
ADHD Subtypes, Hippocampal volumes, Working memory and Peer Victimization A three-cluster K-means model was used to characterize the present study population. Cluster 1 included 40 TD children and 1 child with ADHD-I who have low levels of inattentive and hyperactive-impulsive symptomatology, good working memory ability, and low peer victimization. Cluster 2 included 19 TD, 24 ADHD-C and 11 ADHD-I children who have signi cantly elevated levels of peer victimization and hyperactivity-impulsivity when compared to the two other clusters. Cluster 3 included 52 TD, 34 ADHD-C and 64 ADHD-I children who have signi cantly enlarged left CA3 volumes and signi cantly higher hyperactivity-impulsivity and WMIs compared to children in Cluster 1.
The results of the cluster analysis suggest that ADHD symptomatology, especially hyperactivityimpulsivity, is related to increased levels of peer victimization. Our model also suggests that larger left CA3 volumes are associated with ADHD symptomatology. In both Clusters 2 and 3, symptoms hyperactivity-impulsivity and inattention are signi cantly elevated and, in both groups, left CA3 volumes are enlarged compared to Cluster 1, although only signi cantly enlarged in Cluster 3. Similarities can be seen between our model and the inverse correlation Plessen 12 et al. observed between CA3 volume and symptoms of inattention in children with ADHD. In our model children in Cluster 3 have the most prominent enlargement of left CA3 volumes, but milder symptoms of inattention than children in Cluster 2. This suggests the possibility that enlarged left CA3 volumes are representative of a compensatory mechanism by which the hippocampus hypertrophies in response to the presence of ADHD symptoms and results in less severe symptomatology 12 . The relationship between symptoms of inattention in children with ADHD and hippocampal volume, speci cally in the CA3 sub eld should be further explored.
Our results suggest that deep phenotyping of brain morphology, cognition, and behaviour can identify subtle differences in ADHD subtypes.
A limitation of the current study is our use of a parent-report questionnaire, the CBCL, as a measure of problem behaviour and peer victimization. Children have many experiences without their parents present, such as in school, recreational activities, etc., thus child reports may be able to provide additional information that cannot be captured by parent-report alone.
Youth Self Report data was available however not enough children and adolescents completed this questionnaire, which drastically reduced the sample size and power of analyses. Another limitation of the current study was the absence of the ADHD hyperactive/impulsive subtype (ADHD-H). We only had access to data for 15 participants diagnosed with ADHD-H (16.5% of the total ADHD sample), which was not a large enough sample for valid analyses. In children and adolescents ranging from 3 to 18 years of age, ADHD-I is the most common ADHD subtype with a prevalence of between 2.2-5.7%, the prevalence of ADHD-C ranges from 1.1-2.4%, and the prevalence of ADHD-H ranges from 1.1-4.9% 49 . As children age, an ADHD-H diagnosis becomes less common and could be the reason we had access to the data of very few children with this subtype. Another limitation is that different scanners, with different strengths were used at the various MRI sites. However, a majority of the participants were scanned at CBIC and RU, which both had 3T scanners. The medication status of the children and adolescents in the study sample are unknown and this is a potential confound for behavioural reports, hippocampal volumes and WM performance.

Declarations Author Contributions
A.P., E.G.D., and D.S. were involved in study design and conception. A.P. conducted data analysis, interpreted the results, and wrote the manuscript text. S.C. pre-processed the Magnetic Resonance Imaging data. A.P, E.G.D., and D.S reviewed the manuscript.

Corresponding Author
Correspondence to Emma G. Duerden.

Figure 1
Problem Behaviours and Peer Victimization in Diagnostic Groups. CBCL subscale scores for children and adolescents from the ADHD-C, ADHD-I, and TD groups. Children and adolescents from the ADHD-C group had signi cantly higher Aggressive Behaviour, Rule Breaking Behaviour, Social Problems, and Peer Victimization scores compared to the TD group. The ADHD-C group also had signi cantly higher Aggressive Behaviour, Rule Breaking Behaviour, Social Problems and Peer Victimization scores compared to the ADHD-I group. Children and adolescents in the ADHD-I group had signi cantly higher Aggressive Behaviour scores than the TD group. None of the groups signi cantly differed on the Withdrawn scores.
Scores represent the estimated marginal means, adjusted for age, sex, study site, and SES. P values are Bonferroni corrected (pairwise) for multiple comparisons. Error bars re ect standard error. *p<0.001 K-means Clustering based on Peer Victimization, Left CA3, WMI, Hyperactivity-Impulsivity and Inattention. The three-cluster model is depicted above. Cluster 1 is characterized by small left CA3 volume, high WMI, low peer victimization, and low hyperactivity-impulsivity and inattention. Cluster 1 is almost entirely made up of TD participants. Cluster 2 is characterized as average left CA3 volumes, average WMI, high peer victimization, and high hyperactivity-impulsivity and inattention. Of the participants in cluster 2, 44% are from the ADHD-C group. Cluster 3 is characterized as large left CA3 volume, low WMI, low peer victimization, and average hyperactivity-impulsivity and inattention. Cluster 3 is approximately. 43% ADHD-I participants, 35% TD participants, and 22% ADHD-C participants.