Introduction
Klinefelter syndrome (47, XXY) is a genetic condition in which boys have an additional X chromosome, leading to the 47, XXY chromosomal pattern. Physical consequences include gonadal hormone deficiency, infertility, and long extremities [
11]. Many studies have focused on the physical, cognitive, and behavioral consequences of 47, XXY [
10,
25,
32]. In recent years, attention has also shifted to the neural basis underlying cognitive and behavioral characteristics of 47, XXY. Structural imaging studies thus far indicate that compared with nonclinical controls, males with 47, XXY have decreased total brain volume, enlarged ventricles, as well as smaller caudate, cerebellar, temporal, and frontal volumes [
26,
27,
30,
37,
46,
57]. Other studies have reported regions of decreased gray matter in the amygdala, insular cortex, hippocampus, cingulate, occipital lobe, parietal lobe, temporal pole, inferior frontal lobe, and superior temporal gyrus (STG) in males with 47, XXY [
15,
22,
42,
43,
46]. Additionally, functional MRI studies have reported deviant neural activation in the superior temporal gyrus, superior temporal sulcus, supramarginal gyrus, amygdala, insula, fusiform gyrus, and middle frontal gyrus [
12,
51,
53].
One would expect that these brain abnormalities give rise to cognitive difficulties such as language and reading disorders, executive dysfunction, and social-cognitive deficits. These problems have indeed been reported in 47, XXY [
10,
31,
49,
50,
52,
56]. A few of the main social-cognitive deficits associated with 47, XXY are difficulties in facial affect recognition, identification, and verbalization of emotions, theory of mind, and interpreting tone of voice [
49,
50,
56] Such neural and cognitive deficits may help explain an increased risk of developmental psychopathology that is reported in individuals with 47, XXY, including psychotic disorders, bipolar disorder, ADHD, and autism spectrum disorder (ASD) [
9,
13,
17,
21,
47,
48]. A formal ASD diagnosis requires clinically significant and persistent deficits in social communication and social interaction, as well as restricted, repetitive patterns of behavior, interests, or activities [
6]. However, even in the absence of a formal ASD diagnosis, prominent social and communicative difficulties may be present. Difficulty in social interactions, shyness, social withdrawal, problems with assertiveness, and increased levels of autistic traits (as measured with the Social Responsiveness Scale) [
19] have been reported [
9,
47,
49,
53,
55]. As the development of social competence during childhood is essential for successful participation in a relatively complex and fast-paced society as an adult, it is important to gain insight into the mechanisms underlying social dysfunction in 47, XXY.
Although there is overlap in social behavioral problems between individuals with 47, XXY and individuals with ASD, recent studies have highlighted variability in the underlying mechanisms driving these problems. For example, boys with 47, XXY show more social anxiety than boys with ASD [
55], and they display a specific cluster of autism symptoms that differs from idiopathic ASD in the ability to employ social verbalization/chat, offering to share objects, and interest in other children [
14]. Additionally, theory of mind impairments in 47, XXY appear to be related to executive dysfunction, while these impairments appear to be related to language and face recognition problems in boys with ASD [
56]. In line with these results, both boys with 47, XXY and boys with ASD show deviations in neural networks associated with facial affect labeling. However, while these deviations are located in the amygdala in boys with ASD, boys with 47, XXY show deviant frontal activation [
12]. It is important to also study the underlying mechanisms in terms of brain development. In this respect, it would be relevant to know to what degree neural architecture (in terms of gray matter volume) of structures related to social functioning is affected in boys with 47, XXY, and to assess to what degree morphological deviations are similar to, or different from, those of boys with ASD.
Many different brain structures are involved in decoding social stimuli. For instance, the basic processing of facial information (e.g., gaze shifts and mouth movements, important aspects of emotional expressions) and the recognition of biological motion involve the superior temporal cortex (i.e., both the superior temporal gyrus and sulcus). The amygdala is important for more complex social judgments based on facial information, assessing the significance of social information and recognizing threat or danger [
2]. The orbitofrontal cortex (OFC) is involved in many aspects of social competence and social-cognitive information processing including the evaluation of sensory stimuli, reward and punishment-related behavior, social decision making, theory of mind, self-reflection, and the representation of facial expressions and identity [
36,
39,
40]. The insular cortex mediates recognition of, and responses to, emotional stimuli, regulation of autonomic states related to emotional processes, as well as the representation of one’s internal state [
2]. Lastly, a structure involved in more complex interpretation of social-cognitive stimuli including face familiarity, theory of mind, and the distinction between self and other is the medial frontal cortex [
3].
The current study is based on the hypothesis that the anatomical maturation of this network may be adversely affected in boys with an extra X chromosome. Imaging techniques may help uncover different etiologies of risk of social dysfunction. Nevertheless, structural imaging studies focusing on males with 47, XXY are relatively sparse. In contrast, many studies have addressed structural brain abnormalities in regions related to social functioning in individuals with ASD. A recent review evaluated structural imaging studies in ASD. Relevant to the current study, lower gray matter volume in the superior temporal cortex and higher gray matter volume in the medial frontal cortex were systematically reported in boys with ASD compared with nonclinical, age-matched, male controls [
18]. However, to date, no studies have assessed regional morphologic brain differences between boys with 47, XXY and boys with ASD, which would be very relevant considering the differences in brain activation during social information processing between boys with 47, XXY and boys with ASD that were reported in a recent fMRI study [
12].
The main aim of this study was to assess the neural architecture of brain regions associated with social functioning in boys with 47, XXY and boys with ASD, using voxel-based morphometry. The focus was on the superior temporal cortex, amygdala, OFC, insular cortex, and medial frontal cortex. Significant differences would suggest social dysfunction in Klinefelter syndrome may be anchored in anatomical brain deviations. The secondary aim was to investigate gray matter volumetric differences between boys with 47, XXY and boys with ASD in these structures, to assess whether neural architecture of brain areas important for social information processing is differentially affected.
Discussion
In this imaging study, the neural architecture of brain regions associated with social functioning, i.e., the superior temporal cortex, amygdala, orbitofrontal cortex (OFC), insular cortex, and medial frontal cortex, was investigated in boys with 47, XXY, compared with boys with ASD and nonclinical, male controls. The results showed that boys with 47, XXY have significant gray matter volume reductions in the left and right insular cortices, as well as the left OFC, compared with healthy controls. These structures were also found to be significantly smaller in boys with 47, XXY than in boys with ASD. Additionally, boys with 47, XXY had significantly smaller gray matter volumes in the right superior temporal gyrus (STG) compared with controls.
In general, the OFC is associated with the evaluation of sensory stimuli, reward and punishment-related behavior, social decision making, and the representation of facial expressions and identity [
36,
39,
40]. Importantly, the OFC is also involved in processing mental state concepts and self-reflection [
7], related to both theory of mind and emotional literacy. Both theory of mind impairments and alexithymia have been reported in boys with 47, XXY [
49,
56]. Although speculative, the OFC volume reductions reported in the current study might be one of the anatomical underpinnings of these problems. Previous studies have not reported volume reductions in the OFC specifically, but reduced gray matter in frontal areas has been found repeatedly [
22,
26,
27,
43,
46]. Additionally, one study reported reductions in white matter regions near the OFC in their sample [
15].
The insular cortex is involved in recognition of and responses to emotional stimuli, regulation of autonomic states related to emotional processes, as well as the representation of one’s internal state [
2]. The finding of deviating gray matter volume in the insular cortex fits with the finding that individuals with 47, XXY show problems in identification and verbalization of emotional states, are more easily emotionally aroused than controls, and are significantly more influenced by their emotional states than controls when making decisions [
49,
54]. Another study suggests the insular cortex is functionally involved in decision-making processes [
41], and an fMRI study using a social judgment task reported less neural activation in the insula in individuals with 47, XXY [
53]. The bilateral reductions in insular cortex volume found in the current study are in line with results from earlier structural studies [
15,
42,
43] and might be an anatomical underpinning of these functional findings.
The STG is involved in the processing of facial information, e.g., small changes in facial expression and gaze direction [
2]. The finding of reduced right STG volume is in line with earlier studies reporting impairments in face processing in individuals with 47, XXY [
49,
53]. Additionally, Wernicke’s area and Heschl’s gyri are part of the Wernicke–Geschwind model of language processing and are located in the STG, making it a key structure in this model [
8,
24]. Although language functions are usually located in the left hemisphere, a functional MRI study reported decreased functional asymmetry in the STG during a language task in individuals with 47, XXY [
51], due to increased activation in the right hemisphere. This suggests individuals with 47, XXY may have decreased left-sided dominance for language, which could imply the Wernicke–Geschwind model of language processing also involves right-sided brain regions in this condition. The reported deviations in STG function [
51] support the suggestion of a relationship between anatomical deviations and altered functionality. Earlier structural MRI studies consistently reported abnormalities in the temporal lobe as a whole in individuals with 47, XXY [
26,
27,
30,
46]. In line with other studies [
15,
22,
37,
42], the current study suggests these abnormalities may be located specifically in the STG.
There has been much debate within the field of neuroscience about the relationship between brain anatomy and cognitive and behavioral function in general. This debate also extends to the field of 47, XXY, with the largest reported study by Skakkebaek et al. [
43] finding no association between specific differences in gray matter volume and the neuropsychological profile of 47, XXY. It is important to continue research in this area, to gain more insight into the impact of individual brain structures on behavioral function. In addition to assessing linear relationships between volume differences in individual brain structures and specific cognitive or behavioral problems, it may also prove essential to focus on neural networks. The deviations found in the current study (that focused on specific regions of interest) may very well be part of a larger network of structural and functional brain deviations that together lead to the cognitive and behavioral issues associated with 47, XXY. Imaging studies like this may have important theoretical implications, as they offer the opportunity for understanding the underlying mechanisms along a gene–brain–behavior pathway. The results are in line with the notion that a high number of X chromosome genes are expressed in the brain [
33]. It has been suggested that X chromosome genes are involved in sexually dimorphic brain development [
38], making it more likely that individuals with sex chromosome disorders, such as 47, XXY, show deviations in brain anatomy. Another possibility is that these deviations are the result of gene–gene interactions, or hormonal influences on brain development. Future studies focused on these mechanisms may aid in disentangling these complex issues. However, environmental factors may also have a marked influence on shaping the brain [
35], stressing the notion that a deterministic approach toward brain development in 47, XXY is not useful. Instead, early positive experiences and interventions (e.g., theory of mind training [
45] or social skills training) may have a positive influence on brain development and may even (partially) ameliorate or prevent severe social dysfunction later in development. Neuroimaging may not only help in identifying targets for treatment, but also offer sensitive tools to design and assess the effectiveness of such interventions in 47, XXY.
The secondary aim of this study was to investigate gray matter volumetric differences between boys with 47, XXY and boys with ASD in brain regions associated with social functioning. The SRS scores suggest boys with 47, XXY have increased levels of autism symptoms, and they could not be distinguished from boys with ASD in terms of overall level of these symptoms, but they showed significant reductions in left and right insular cortex volume, as well as the left OFC, compared with boys with ASD. These findings support the notion that the autism spectrum may be extraordinarily heterogeneous. Various neural etiologies may result in similar behavioral symptoms that are classified as autistic. Because boys with 47, XXY represent a genetically homogenous high-risk group for ASD, these findings might help in identifying biomarkers of social dysfunction related to X chromosomal genes, although much more research is needed in this area. The current findings may stimulate such lines of research.
A limitation of the current study was the relatively small sample size. Additionally, the 47, XXY sample appeared to consist of boys with more severe behavioral issues than samples from earlier studies, e.g., Tartaglia et al. [
47], as demonstrated by higher scores on the SRS. It is possible that parents of more severely affected boys with 47, XXY were more motivated to participate in neuroimaging research. However, the behavioral characteristics of the current sample were similar to those of the larger cohort from which they were drawn (for an extensive behavioral description of this cohort, see Van Rijn et al. [
55]). In addition, there may be cultural differences in the prevalence of behavioral issues associated with 47, XXY, due to differences in mental health care systems, including protocols for interventions or support over the course of development. Voxel-based morphometry has some inherent limitations, e.g., individual differences in cortical folding patterns influencing the results. The results of the current study are in line with those from earlier imaging studies on 47, XXY, with the exception of a lack of significant differences in amygdala volume. These have been reported in earlier neuroimaging studies [
43,
46], but were not present in our sample. It is possible this is due to age-related changes in brain morphology. However, the fact that the significant differences that were found in the current study are similar to those from earlier studies in adult individuals with 47, XXY, seems to suggest deviations in brain anatomy develop early and are relatively stable over time. It is therefore possible that differences in imaging analysis methodology or lack of statistical power have led to discrepancies between volume deviations in the amygdala in earlier studies, and a lack of such deviations in the current sample. Furthermore, the lack of significant volumetric differences between the ASD group and controls may imply they were relatively high functioning. IQ scores seem to corroborate this suggestion. A problem inherent to participating in imaging research is that typically, only a selective subgroup of ASD participants (e.g., those with high functioning autism) is included in a study. Additionally, findings regarding structural brain abnormalities in ASD are often conflicting and depend on many (study-specific) factors, such as inclusion and exclusion criteria, MR acquisition parameters, and details of the image processing pipeline [
18]. In the current study, small sample size may have played a role, although the 47, XXY group was equally small, but did show significant volume reductions. Lastly, because this study was explorative in nature and the number of contrasts was limited to six, no correction for multiple comparisons was applied. However, all but one significant result would have survived such a correction.