Lack of Gender-Related Differences in Childhood-Onset Schizophrenia

Abstract

Objective

Gender differences, including younger age of onset and greater premorbid deficits in men, have been reported in adult-onset schizophrenia (AOS). This study comprehensively evaluated gender differences in childhood-onset schizophrenia (COS), a rare variant of the disorder.

Method

We examined demographic, premorbid, clinical, familial, and cognitive characteristics, as well as presence of chromosomal abnormalities and brain magnetic resonance imaging (MRI) cortical volumes, in 133 patients with COS. Cortical analyses included age- and gender-matched healthy volunteers (n=124).

Results

Males with COS (n=72) had a slight but significant earlier age of onset than females with COS (mean age 9.51± 2.28 vs. 10.29 ± 1.63, t_[131] = 2.21, p = .03), higher verbal IQ scores (83.00 ± 15.97 vs. 75.58 ± 15.10, t_[89] = 2.24, p = .03), and higher rates of comorbid pervasive developmental disorder (PDD; 28.17% vs. 6.90%, χ²_[1] = 9.54, p < .01) and attention-deficit/hyperactivity disorder (ADHD; 43.86% vs. 21.43%, χ²_[1] = 5.40, p = .02). There were no significant gender differences across other demographic, IQ, or clinical measures, frequency of chromosomal abnormalities, family clinical measures, premorbid functioning, or in gender by disorder interactions for MRI brain measures.

Conclusion

After a comprehensive examination, we found few remarkable gender differences in COS. Although less striking than that seen in AOS, males with COS had an earlier age of onset. ADHD and PDD rates were high in COS overall, suggesting greater neurodevelopmental vulnerability in COS. However, the gender ratios of these comorbidities in COS mirror those of the general populations, indicating that these gender differences may be unrelated to COS.

INTRODUCTION

Gender differences in schizophrenia have been reported for several aspects of the disorder^1,2. These differences might contribute to the marked heterogeneity of disease presentation, course, and response to treatment^3,4. Childhood-onset schizophrenia (COS), a rare, severe, and possibly a more homogenous variant of the disorder^5,6, has not been comprehensively examined with respect to gender differences.

Variations in the incidence and prevalence of schizophrenia between men and women have been debated, with some reports of increased incidence in men⁷, while epidemiologic studies report no difference in prevalence^8,9. The significantly earlier age of onset in men (roughly 5-6 years earlier^10,11), as well as a second peak of onset in women over age 40, may contribute to these results^3,10,11. Findings on sex differences in symptom presentation and prognosis are less consistent. Several studies have found women with schizophrenia to have less severe negative symptoms and better prognosis than men^2,12-14, while others report minimal to no gender differences^15,16. Since an earlier onset of schizophrenia tends to relate to poorer prognosis¹⁷, the later onset in women might reflect some overall protective effect.

Deficits in premorbid social functioning and cognitive functioning in adult schizophrenia patients¹⁸, academic and occupational adjustment^19,20, marriage adjustment¹⁸, and general premorbid functioning^13,20 tend to further indicate a mitigating effect for females. However, findings in social functioning^21-23 and cognitive functioning are less conclusive^2,3. Moreover, while psychiatric comorbidities in schizophrenia are common^24,25 and may affect many of these domains, salient differential effects of gender on frequency of comorbidities in adult-onset schizophrenia (AOS) appear lacking. In COS, an earlier study of our group identified greater frequency of pervasive developmental disorder (PDD) diagnoses among male patients with COS²⁶, and others have reported greater comorbid ADHD in males with COS²⁵.

Gender differences have also been reported in brain structure of schizophrenia patients. Compared to controls, patients with schizophrenia tend to have greater total ventricular volume, lower absolute whole brain volume, and lower cerebral gray matter volume²⁷. Some reports indicate greater severity of these structural abnormalities in males compared to females, especially with regard to ventricular enlargement²⁸. However, findings are highly inconsistent,^1,2 with some reporting that females show greater reductions in prefrontal cortex, orbital regions, basal forebrain, anterior cingulate, and posterior supramarginal gyrus, whereas males show greater relative reduction in frontomedial and middle frontal cortices, paracingulate gyrus, and insula^29,30. Other studies have found no gender differences with respect to abnormalities in volumes of cortical gray matter, white matter and sulci, lateral and third ventricles, and key subcortical structures^31,32. This last point was highlighted in a meta-analysis of 58 magnetic resonance imaging (MRI) studies of patients with schizophrenia, which found that gender was not significantly related to structural differences²⁷.

Few studies have examined gender differences in brain development in early-onset or childhood-onset schizophrenia (COS). A small longitudinal study (n=21) found males with early-onset schizophrenia (EOS) to have more frontal gray matter loss compared to controls than females with EOS, although that study was underpowered to detect change in women due to small sample size (23.81% female n=5)³³. Earlier studies in our group have found no gender differences in cortical thickness and subcortical volume development^34,35 or development of total cortical gray matter³⁶ in patients with COS. Nevertheless, while changes in cortical brain volume are among the most consistently reported abnormalities in both AOS and COS, a thorough and systematic evaluation of potential gender differences remains lacking^37,38.

In the current study, we explore gender differences more comprehensively in a larger sample of patients with COS. Specifically, we examine demographics, clinical presentation, premorbid functioning, cognitive ability, presence of chromosomal abnormalities, brain MRI cortical volumes, and family clinical data. Because COS is in several respects continuous with later-onset schizophrenia^39,40, we expected that the more consistently reported gender differences, such as poorer premorbid functioning and younger age of onset in males, might be more prominent for patients with COS.

METHOD

Participants

Patients were recruited nationally as part of an ongoing longitudinal study of COS at the National Institute of Mental Health (NIMH)⁴¹. Criteria for selection and inclusion have been described previously^41,42. Briefly, children and adolescents ages 6-18 who met DSM-III-R/DSM-IV criteria for schizophrenia with onset prior to age 13 and with a prepsychotic IQ above 70 were nationally recruited. 133 patients with COS (72 male, 61 female) participated in the study. Additionally, 124 age- and gender-matched healthy volunteers formed the control group for analyses of brain structure.

Clinical and Neuropsychological Measures

A research team that included two psychiatrists collected the clinical data of patients with COS including a structured interview with the patient and their parents at screening or admission⁴³. Interviews included the Scale for the Assessment of Positive Symptoms (SAPS)⁴⁴, the Scale for the Assessment of Negative Symptoms (SANS)⁴⁵, the Brief Psychiatric Rating Scale (BPRS)⁴⁶, the Simpson-Angus Extrapyramidal Side Effect Scale (SIM)⁴⁷, and the Abnormal Involuntary Movement Scale (AIMS)⁴⁸. Rates of comorbid attention-deficit/hyperactivity disorder (ADHD) were determined using the Kiddie Schedule for Affective Disorders and Schizophrenia (K-SADS)⁴³. Because K-SADS does not include the pervasive developmental disorder (PDD) diagnosis, all children were screened with the Autism Screening Questionnaire (ASQ)⁴⁹ and further assessed by clinical interview. Comorbid PDD was given when history and thorough review of clinical records showed clear PDD/autistic spectrum symptoms before the onset of psychosis with symptoms still observable at the NIMH evaluation (see Sporn et al.²⁶ for more details). During an extensive clinical evaluation, parents of the participants with COS reported age of onset and premorbid developmental, academic, social, language, and motor problems or delays and provided corresponding medical and school records. These were combined into what we termed a “Hollis score,” following the premorbid developmental difficulties described previously⁵⁰.

Additionally, trained study staff administered the current, age-appropriate versions of Wechsler Intelligence Scales to participants with COS either at baseline or follow-up. Given that participants had varying degrees of psychotic symptoms at different time points, which affected their ability to engage in IQ testing, for participants who had IQ testing both at baseline and follow-up, we selected the highest score. The Wechsler Adult Intelligence Scale–Revised (WAIS-R)⁵¹, Wechsler Abbreviated Scale of Intelligence First Edition (WASI)⁵², and Second Edition (WASI-II)⁵³ were versions used along with the Wechsler Intelligence Scale for Children–Revised (WISC-R)⁵⁴ and Third Edition (WISC-III)⁵⁵.

Biological parents and full siblings of the participants with COS were evaluated with clinical and structural diagnostic interviews (details reported previously²⁶) to assess the presence or history of a DSM-IV diagnosis and presence or symptom counts of paranoid personality disorder, schizoid personality disorder, schizotypal personality disorder, and avoidant personality disorder.

Genetic Measures

As part of a broader study of the genetics of EOS, we obtained DNA from all participants with COS. Chromosome abnormalities in the participants with COS were determined from array-based single-nucleotide polymorphism genotyping from blood samples and is described elsewhere⁵⁶. In the present study, we compared for males and females the carrier status (yes/no) of chromosome abnormalities, including copy number variants (CNVs) and large chromosomal errors.

Image Acquisition and Processing

T1-weighted images were acquired with the same 1.5-T MRI scanner (Signa; GE Medical Systems, Milwaukee, Wisconsin). A 3-dimensional spoiled gradient-recalled echo in the steady-state sequence designed to optimize discrimination of gray matter (GM), white matter, and cerebrospinal fluid was used to acquire 124 contiguous 1.5-mm-thick sections in the axial plane. Imaging parameters were as follows: repetition time, 24 milliseconds; echo time, 5 milliseconds; flip angle, 45°; acquisition matrix, 256 × 192; and field of view, 24 cm. Each participant was scanned at approximately two-year intervals.

Freesurfer 5.3 (http://surfer.nmr.mgh.harvard.edu/fswiki) was used for surface-based cortical processing^57,58 and volumetric subcortical segmentation⁵⁹ of T1-weighted anatomical images. All individual images were transformed into Talairach space⁶⁰. On the basis of prior probabilistic information, each individual cortical hemispheric sheet was parcellated into 34 regions^61,62. A trained imaging specialist reviewed individual scans, and those with significant artifact or motion disturbance were excluded from analysis. A total 287 scans of patients with COS (60 male, 45 female), and 360 scans of healthy matched control participants (72 male, 52 female) were included in the final analysis. Volumes of each cortical and subcortical region were extracted. Total cortical volume was calculated as the sum of all 34 cortical regions in both hemispheres without including cerebrum and subcortical gray matter. Estimated total intracranial volume was used as intracranial volume (ICV)⁶³.

Statistical Analysis

Demographic and clinical data

Independent t-tests were used to determine significant differences between male and female patients with COS for continuous demographic, clinical, and neuropsychological variables. A Chi-square test or Fisher’s exact test was used to assess the significance of the differences in categorical variables. Missing data were excluded pairwise from the analyses. A p value < .05 (two tailed) was considered statistically significant. Bonferroni correction for multiple comparisons was used for imaging data but not for the clinical and demographic variables, given the numerous inconsistencies in gender differences on these dimensions in adult schizophrenia and the exploratory nature of these analyses in our COS sample. As Streiner and Norman⁶⁴ suggest, correcting for multiplicity in these situations results in an excess of type 2 errors, closing off potentially worthwhile research areas. All the demographic, clinical, and neurobiological analysis was done using STATA/SE version 14.0 statistical software (Stata Corp, College Station, TX, USA), and all brain imaging analysis was carried out in the Statistics Toolbox of Matlab (version R2014A, The Math Works Inc., Natick, Massachusetts).

Brain differences between groups in each gender

In the cross-sectional analysis, analysis of variance (ANOVA) models were used to examine group differences in volume measures for males and females separately when the first MRI scan was collected. Each volume measure was inputted as the dependent variable, and group as the independent variable. We examined total cortical, total ventricular volumes, and 34 cortical region volumes in each hemisphere. ICV was used as a covariate variable for correction of the head size when examining individual cortical regions. For each volume measure, Cohen’s d was calculated as the effect size of group difference for both male and female.

Interaction effect between gender and group on brain volumes

ANOVA models were used to further examine interaction effect between gender (male vs. female) and group (COS vs. control). Each volume measure was inputted as the dependent variable, gender and group as independent variables that interact with each other. Again, ICV was used as a covariate variable. The significant threshold of p was set at .00071 for Bonferroni’s correction of a total of 70 comparisons.

In parallel to the cross-sectional analysis, using all scans, we performed longitudinal analysis with linear mixed effect models to examine group differences in brain volume development for all regions. Brain volume was the dependent variable. Fixed effects included the group, gender, age (centered at the average age of 17.00), an interaction term of gender x group, and ICV. We included a random intercept per person to account for within-subject variance.

RESULTS

We examined gender differences across several domains in 133 patients with COS, of which 46% were female. For the IQ analysis, 42 participants (19 male, 23 female) were unable or unwilling to engage in IQ testing at any time point; there were no gender differences between the tested and untested group (χ²= 2.0, p = .16). There were no significant gender differences among race, ethnicity, parental socioeconomic status, or full scale IQ (Table 1). There were also no significant sex differences in the version of the IQ test administered (χ² = 10.50, p = .11) or the age at IQ test (t = 0.95, p = .11).

Table 1.

Demographic Data for Male and Female Childhood-Onset Schizophrenia Patients

Variables	Male (54.14%)		Female (45.86%)		Test Statistic (df)	p
	Mean ± SD/ n (%)	n	Mean ± SD/ n (%)	n
Age of Onset, years	9.51 ± 2.28	72	10.29 ± 1.63	61	t(131)=2.21	.03
Race
Caucasian	41 (56.94)	72	33 (55.00)	60	χ2(3)=1.17	.76
African American	19 (26.39)		20 (33.33)
Asian	4 (5.56)		2 (3.33)
Other	8 (11.11)		5 (8.33)
Ethnicity
Hispanic	7 (9.86)	71	9 (15.00)	60	χ2(1)=0.80	.37
Non-Hispanic	64 (90.14)		51 (85.00)
Parental SES
1	13 (18.31)	71	9 (15.00)	60	Fisher’s exact(4)	.13
2	24 (33.00)		17 (28.33)
3	18 (25.35)		23 (38.33)
4	16 (22.54)		8 (13.33)
5	0 (0.00)		3 (5.00)
IQ
Full IQ	79.02 ± 16.20	53	74.66 ± 16.33	38	t(89)=1.26	.21
Verbal IQ	83.00 ± 15.97		75.58 ± 15.10		t(89)=2.24	.03
Performance IQ	77.57 ± 16.74		77.05 ± 18.30		t(89)=0.14	.89

Note: Boldface type indicates statistical significance. Socioeconomic status (SES) was based on two-factor index of social position by Hollingshead (1958)⁸⁰; 1= highest, 5 = lowest SES.

However, female patients with COS had significantly lower verbal IQ than male patients with COS (t_[89] = 2.24, p = .03). Additionally, males had a slightly but significantly younger age of onset than females in the sample (t_[131] = 2.21, p = .03). There was greater density of earlier ages of onset in males compared to females (Figure 1). Clinical review of younger male charts did not reveal any striking clinical differences from the later-onset males. There were no significant differences between males and females in chromosomal abnormalities or in most clinical measures (Table 2). Most chromosomal abnormalities were distinct cases of CNVs, with three cases of two patients with COS having the same CNV (2q25.3, 15q13.3 deletion, 16p11.2 duplication) and 5 patients with COS with 22q11.21 deletion, as reported previously⁵⁶. However, males had significantly higher rates of comorbid PDD (χ ²_[1] = 9.54, p < .01) and ADHD ( χ²_[1] = 5.40, p = .02) than females. Males and females did not significantly differ in rates of premorbid abnormalities across academic, language, and motor domains. In the social domain, 55 of 71 males (76.4%) and 35 of 58 (58.3%) females reported premorbid social abnormalities (χ²_[1] = 4.92, p = .03). Nonetheless, further analysis showed that the presence of PDD, which was higher in males, accounted for this difference. In patients with COS without a PDD diagnosis, 35 of 51 males (68.6%) and 31 of 54 females (57.4%) reported premorbid social abnormalities (χ²_[1] = 1.41, p = .23).

Figure 1.

Age of onset of childhood-onset schizophrenia

Table 2.

Clinical Ratings for Male and Female Childhood-Onset Schizophrenia Patients

Variables	Male (54.14%)		Female (45.86%)		Test statistic (df)	p
	Mean ± SD/ n (%)	n	Mean ± SD/ n (%)	n
ASQ score	9.49 ± 8.35	69	8.26 ± 8.39	54	t(121)=0.81	.42
Hollis score	4.17 ± 2.90	72	3.35 ± 2.76	60	t(130)=1.65	.10
SAPS score	37.24 ± 18.64	62	41.40 ± 17.51	53	t(113)=1.22	.22
SANS score	47.81 ± 27.52	62	56.36 ± 23.84	55	t(115)=1.79	.08
BPRS score	46.99 ± 13.11	68	47.66 ± 12.64	58	t(124)=0.29	.77
SIM score	11.96 ± 2.74	48	15.31 ± 14.82	49	t(95)=1.54	.13
AIMS score	10.98 ± 2.09	44	10.69 ± 2.46	45	t(87)=0.60	.55
PDD	20 (28.17)	71	4 (6.90)	58	χ2(1)=9.54	<.01
ADHD	25 (43.86)	57	9 (21.43)	42	χ2(1)=5.40	.02
Abnormal chromosome	16 (22.22)	72	20 (32.79)	61	χ2(1)=1.87	.17

Note: Boldface type indicates statistical significance. ADHD = attention-deficit/hyperactivity disorder; AIMS = Abnormal Involuntary Movement Scale; ASQ = Autism Screening Questionnaire; BPRS = Brief Psychiatric Rating Scale for Children 18 Symptoms Version; PDD = pervasive developmental disorder; SANS = Scale for the Assessment of Negative Symptoms; SAPS = Scale for the Assessment of Positive Symptoms; SIM = Simpson-Angus Extrapyramidal Side Effect Scale.

There were no gender differences in the presence of schizophrenia, schizoaffective disorder, bipolar disorder, psychosis, mood disorders, anxiety disorders, alcohol disorders, and behavioral disorders in mothers, fathers, siblings, and family members independently (data not shown). There were similarly no differences in the presence or number of symptoms of paranoid personality disorder, schizoid personality disorders, schizotypal personality disorder, and avoidant personality disorder.

Finally, in a cross-sectional analysis of baseline images at the average age of 14.24, we examined gender differences in brain structural abnormalities. Total ventricular volume was significantly larger and total cortical volume significantly lower in participants with COS than control participants for both males and females (Table 3), but no significant interactions were found between gender and group. Effect sizes were similar between males and females, indicating that abnormalities in the overall brain structure were not different between genders.

Table 3.

Cross-Sectional Regional Brain Volumes for Male and Female Childhood-Onset Schizophrenia (COS) Patients and Healthy Matched Healthy Control (HC) Participants

Volume (mL)	Male (COS-HC)				Female (COS-HC)				Gender*Group Interactionb
	COS	Control	COS vs. Control	Effect size	COS	Control	COS vs. Control	Effect size	Male(COS-HC) vs. Female (COS- HC)
	Mean ± SD	Mean ± SD	p	Cohen D	Mean ± SD	Mean ± SD	p	Cohen D	p
Global Measure
Total ventricles	19.54±10.40	13.54±7.59	.00018	0.67	16.0±7.2	11.7±6.1	.00025	0.66	.57
Total cortex	514.59±72.26	563.89±47.46	.000000083	-0.82	450.1±46.2	497.2±36.5	.000000069	-1.14	.56
Individual Cortical Regiona
Effect Size (male<female)
Left primary
Auditory cortex	1.28±0.27	1.34±2.62	.15	-0.23	1.02±0.21	1.22±0.20	.0000073	-0.98	.04
Right primary
Auditory cortex	0.97±0.18	1.04±0.20	.039	-0.35	0.81±0.17	0.97±0.17	.000024	-0.92	.08
Effect Size (male>female)
Left caudal anterior
Cingulate cortex	1.83±0.47	2.17±0.53	.000081	-0.67	1.75±0.44	1.89±0.47	.17	-0.31	.08
Left fusiform	9.83±1.91	11.14±1.65	.0000016	-0.74	8.81±1.44	9.52±1.24	.014	-0.53	.06
Left lateral
Orbitofrontal cortex	7.86±1.38	8.72±1.11	.000016	-0.70	7.17±1.10	7.66±0.78	.017	-0.52	.10

Note: Boldface type indicates statistical significance.

^aOnly individual cortical regions with p of interaction effect (whether using intracranial volume [ICV] as a covariate or not) < 0.1 were included here.

^bNo significant interaction effect was found for both global measures and individual cortical regions (p<.00071 for Bonferroni’s correction).

In addition to these two global measures, we examined 34 cortical regions in each hemisphere and found several areas with a trend (p<.1) toward an interaction effect. Specifically, male patients with COS had more volume loss than female patients in the left lateral orbitofrontal, caudal anterior cingulate, and fusiform cortices (Table 3). However, we found no significant group by gender interaction effects in either global or regional cortical volumes.

We further examined gender differences in brain volume development in a longitudinal analysis (not shown). A total of 647 scans (287 of 105 COS at the average age 17.00, 360 of 124 controls at 16.42 the average age) were included. There were no significant differences in mean age at scan, or number of scans per participant between COS and healthy volunteers (data not shown). No significant interactions between gender and group were found either for both global and individual regional volumes (data available on request).

DISCUSSION

After an extensive examination of gender differences in patients with COS, including global and regional cortical volumes, presence of chromosomal abnormalities, numerous clinical ratings, family clinical ratings, abnormalities in premorbid functioning, and demographic variables, the only gender differences found in this study were a slightly but significantly earlier age of onset and higher comorbid ADHD and PDD diagnoses in males, as well as lower verbal IQ in females. Consistent with the approximate 4:1 ratio of PDD in the general population^65,66, males in our sample were four times more likely to have PDD than females. Similarly, the 2:1 male to female ratio of ADHD in our sample matches that found in the general population⁶⁷. Since these ratios mirror that of the non-COS population, these gender differences are likely unrelated to the presentation of COS and instead reflect general gender differences. It is noteworthy, however, that both comorbidities are much more prevalent in our COS sample than seen in the population at large. Higher prevalence of lifetime ADHD in COS was also observed in the study of Ross et al. (2006), who reported lifetime rates of comorbid ADHD in 16 of 23 females (70%) and 54 of 60 males (90%) with COS (gender effect p=.03)²⁵. These findings support a more general model of greater neurodevelopmental vulnerability underlying COS^68-70.

Additionally, in keeping with studies of adult schizophrenia¹⁰, males had a significantly earlier age of onset than females. While the mean age of onset was less than a year different between genders, as shown in Figure 1, 11 of the 12 youngest onset cases were males. After individual review of clinical records in this subgroup, however, we could find no clinical correlates for age of onset in this group. Presently, then, this finding remains unexplained and will benefit from future work with careful attention to possible causes or correlates of this apparent early difference.

We found no gender by group interactions for brain MRI measures at baseline or longitudinally. This is consistent with previous findings in our group, which indicated no gender differences in cortical thickness, subcortical volume development,^34,35 or cortical gray matter development³⁶. These findings also corroborate findings of AOS studies²⁷. While the number of studies assessing gender differences in EOS and COS are few overall, contrary to our finding, a small longitudinal study found males with EOS to have more frontal gray matter loss compared to controls than women. However, the study was limited by a small sample of women (n=5 females; total n=21) and lacked analysis of the gender by group interaction³³.

From a clinical perspective, the common finding that women have better premorbid functioning than men in adult schizophrenia does not hold in our COS sample, nor does the less-consistent finding that women have less severe negative symptoms^12-14. We found no significant gender difference in any clinical measures or in social, academic, motor, and language premorbid functioning after accounting for PDD.

Interestingly, females in our sample also scored lower on verbal IQ than males, but there was no significant difference in overall or performance IQ. Cognitive functioning is one of the most inconsistently reported sex differences in the adult schizophrenia literature, with many studies finding no differences^71-74. For example, Bonzikas et al. found women with schizophrenia to have better verbal learning and memory than men with schizophrenia⁷⁵. However, Lewine et al. found women with schizophrenia to score lower in the language domain, consisting of verbal IQ and oral fluency, than males, in line with our finding⁷⁶. Overall, our findings suggest that in COS, patients’ IQ tends to be similar for males and females.

Limitations of this study included the inevitable unknown referral and selection biases when recruiting for a very rare disorder. While we recruited nationally, this is not a population-based sample. Additionally, because males had significantly higher rates of ADHD and PDD, it is possible that the earlier age of onset finding is associated with earlier age of contact with mental health professionals. Another limitation is that all of our participants were taking antipsychotic medications at the time of their MRI scan, which could affect our findings; however, as both females and males were taking medications, it seems unlikely that these would affect the findings differentially. Furthermore, as some studies have suggested that exposure to typical⁷⁷ or atypical antipsychotics^78,79 is associated with increased basal ganglia and other subcortical gray matter volumes, we did not examine the subcortical gray matter volumes to avoid this confounding effect. Finally, while our numbers are small, the study represents the largest study to date of patients with COS.

In conclusion, after thorough review of premorbid, clinical, familial, cognitive, and structural imaging, this study found few remarkable gender differences in COS. Furthermore, our strongest finding of greater comorbid PDD and ADHD in males replicates the gender ratio observed in the general population and is unlikely to be specific to COS-related gender differences. While this study did not assess genetic data in depth, we found no significant or clinically meaningful gender differences in COS.