Background
Schizophrenia is a mental disorder with typical onset in early adulthood [
1]. Although the disorder has also been identified in children and early adolescents, such occurrences are rare [
2,
3]. Several studies have gathered and scrutinised brain samples from early onset schizophrenia (EOS) patients in various domains. In particular, since brain morphological abnormalities are consistently found in the general samples of patients with schizophrenia (that is, adult onset populations), researchers have focused on the exploration of brain morphology in patients with EOS, defined herein as schizophrenia with onset under age 18 [
4‐
6]. Using magnetic resonance imaging (MRI), some research groups have reported enlargement of the lateral ventricles [
7,
8], and regional volume reduction in the superior temporal gyrus [
9,
10], thalamus [
8,
11,
12], and frontal lobe [
12‐
14] in EOS patients, mirroring findings reported in individuals with adult onset schizophrenia (AOS) [
15‐
17]. However, compared with AOS studies, the number of EOS studies, especially by separate research groups, is still small, and thus studies using an independent sample of EOS patients are in demand.
Previous studies of AOS patients have reported the relationships between structural brain volume alterations, and positive and negative symptoms in various brain regions, such as the superior temporal gyrus [
18‐
20], insula [
21], fusiform gyrus [
20], parahippocampal gyrus [
22], basal ganglia [
23], and prefrontal gyri [
24]. As for EOS patients, a few studies have investigated the relationships between regional brain changes and the symptom severity, and the regions found to be related to the symptoms in EOS patients, such as the hippocampal [
5], occipital, and parietal cortices [
25], are not entirely consistent with the reports on AOS patients.
Brain development in early life is thought to be dynamic, with the patterns of growth being diverse across different brain regions [
26]. In particular, brain regions that are inherently linked with the pathology of schizophrenia may undergo disproportional changes during the vulnerable period of brain development (that is, the period before adolescence) [
27,
28]. In the present study, we conducted voxel-based structural MRI analyses to explore any pattern of regional brain tissue volume abnormalities, and to elucidate the relationships between the regional brain volume and the severity of clinical symptoms in EOS patients.
Discussion
To our knowledge, this is the first voxel-based morphometry study indicating significant relationships between regional brain volume alterations and clinical symptoms in EOS patients (that is, with onset under 16 years). We found a significant GM volume reduction and a significant increase of CSF in EOS patients. The EOS patients had a significant reduction of regional GM in the left parahippocampal gyrus, the left inferior frontal gyrus, and the left superior temporal gyrus. In addition, they had reduced regional WM in the left posterior limb of the internal capsule, and in the left inferior longitudinal fasciculus. When the correlations between regional GM volume and positive symptoms were examined within the EOS patient group, positive symptoms were found to be significantly correlated with the reduced GM volume in the bilateral posterior cingulate gyrus, with the increased WM volume in the cerebellum (vermis region), and with the reduced WM volume in the brain stem and the bilateral cerebellum (hemisphere region). We also found a relationship between the severity of negative symptoms and the increased GM volume in the right thalamus.
The findings of a GM volume reduction and CSF increase in individuals with EOS in this study are compatible with previous volumetric studies of both EOS and adult onset schizophrenia (AOS) patients [
7,
13,
28,
40‐
42], suggesting that the pattern of volumetric alterations (that is, GM volume reduction and CSF increase) may be an inherent feature of schizophrenia irrespective of age at onset. The GM volume reduction in schizophrenia as a whole could reflect an exaggeration of a normal maturational process of synaptic/dendritic pruning during adolescence [
28].
The voxel-based analysis revealed that in the EOS group, the regional volume reduction of GM was evident in the left parahippocampal gyrus, the left inferior frontal gyrus, and the left superior temporal gyrus. These findings were consistent with those of the previous studies in patients with AOS [
15,
43‐
45]. However, because regional grey matter alterations have frequently been observed in various brain regions of AOS patients, rather than being limited to these three brain regions alone [
15], and because grey matter abnormalities are deemed to be a fundamental part of the pathophysiology of schizophrenia, the observation of grey matter changes in these three regions (that is, the parahippocampal, inferior frontal, and superior temporal gyrus) in individuals with EOS in the present study may indicate that the disease originates in these areas. Another possible explanation is that these three regions may be more subject to brain insults that are related to the predisposition to schizophrenia.
We also found white matter morphological changes in EOS patients. The reduction of WM volume was located in the left posterior limb of the internal capsule, and the left inferior longitudinal fasciculus. Although no prior study has addressed the regional WM volume alterations in EOS, the current findings are similar to those reported by Sigmundsson
et al. [
23], who investigated WM volume changes in AOS patients. In addition, diffusion tensor imaging (DTI) studies of AOS [
46] and EOS [
47] patients have also shown WM abnormalities in similar regions. Therefore, there is evidence emerging from a variety of sources to suggest that WM abnormalities are part of the patterns of brain parenchymal aberration associated with schizophrenia, although they may be a secondary process. In addition, we found a relationship between the brain stem (pons) volume reduction and positive symptoms. This may be intriguing in view of the evidence, from a positron emission tomography (PET) study, showing dysregulation of dopaminergic transmission in the midbrain in schizophrenia [
48].
In our sample of EOS patients, the brain region that was found to be associated with the severity of the positive symptoms was the posterior cingulate gyrus. Although similar findings have been reported in AOS patients [
49], most AOS studies have shown an association with other brain regions, such as the superior temporal gyrus [
18‐
20], insula [
21], fusiform gyrus [
20], and parahippocampal gyrus [
22]. It is of note, however, that the region of the posterior cingulate gyrus has also been reported to be associated with the positive symptoms in other neuroimaging studies of patients with AOS, such as pharmacological functional MRI [
50] and positron emission tomography (PET) studies [
51]. Furthermore, additional evidence from some AOS studies that the posterior cingulate gyrus volume reduction is associated with other clinical features, such as poor clinical outcome [
52], and the lack of insight and judgment [
53], suggests that neural circuits involving the posterior cingulate gyrus may have an important role in the pathophysiology of schizophrenia. Given that the posterior cingulate is related to processing of emotionally salient stimuli and spatial attention [
54], our finding that the posterior cingulate gyrus became prominent in relation to positive symptoms suggests that this region may be involved in the formation of delusions in individuals with a very early age of onset and at an early stage of the illness.
In the present study, the severity of the positive symptoms of schizophrenia was found to be associated with increased WM volume in the cerebellum (vermis region) and with decreased WM volume in the cerebellum (hemisphere region). The finding in the cerebellar vermis region is consistent with that of a previous AOS study [
55] showing a positive correlation between vermis WM volume and the severity of positive symptoms. As regards the cerebellar hemisphere region, neither previous studies employing EOS patients nor those with AOS patients have reported significant correlations with positive symptoms. This is the first study, to our knowledge, to demonstrate a relationship between positive symptoms and a decreased WM volume of the cerebellar hemisphere region.
The severity of the negative symptoms was found to be associated with the increased GM volume of the right thalamus. This finding is similar to the results of prior AOS studies showing a positive correlation between thalamic volume and the severity of negative symptoms [
56‐
58]. The thalamus is thought to play a role in sensory gating, a disruption of which has been reported to be involved in schizophrenia [
59,
60]. Previous AOS studies have also reported correlations between the negative symptoms and structural alterations in other brain regions: namely, reduced volume in the fusiform gyrus [
19], frontal lobe WM [
24,
61,
62], prefrontal GM [
63], cingulate WM and internal capsule [
64], and increased volume in the posterior superior temporal gyrus [
65]. However, in our sample of EOS patients, there were no associations between morphological measures in any of these regions and the severity score of negative symptoms. It could be that at an early stage of brain development (that is, in adolescence), the thalamus, which has reciprocal connectivity with the frontal regions, may be predominantly involved in generating negative symptoms (that is, cognitive deficits) via disturbed connectivity.
A question may arise as to why no relationships were evident between the clinical symptoms and the morphological measures in the three main brain regions (the frontal, temporal, and parahippocampal gyri) showing significant volume reductions in our EOS patients. What is puzzling is that other regions (that is, the posterior cingulate gyrus and the thalamus) were found to be associated with the symptoms in this study. One possible interpretation is that disturbed neural circuits rather than structural alterations
per se may play a role in the formation of symptoms in early onset schizophrenia patients (that is, patients with onset prior to brain maturation). Impaired circuits may involve interconnections between the posterior cingulate gyrus and the temporal lobe [
66] and between the thalamus and frontal lobe [
67,
68].
The results of this study should be interpreted in the context of the following limitations. First, the number of subjects was relatively small. In spite of this, the fact that we were able to detect the regional brain volume alterations in a unique sample of patients with EOS and the finding that most of the abnormalities found were identical to those for the general population of adult onset schizophrenia patients, may support the robustness of the current findings. Second, IQ score was not matched between the case and control groups; that is, the mean IQ was significantly lower in the patients than in the controls. Thus, we conducted an additional analysis in which IQ was adjusted for as a covariate, and found that the regional brain volume differences (for example, the three main GM regions) between the groups remained significant. As a result, the effects of IQ on the findings can be taken as minimal, especially with respect to the regional brain changes. Third, the effect of medication was not considered when regional brain changes were compared between the case and control groups. Antipsychotic medication can affect regional brain morphology in schizophrenia, particularly in the thalamus [
69] and basal ganglia [
70], resulting in increased volumes in these regions. If physicians tended to administer greater dosages of antipsychotic medication to combat the negative symptoms, then the relationship between increased volume in the thalamus and the negative symptoms found in this study would be accounted for by the medication effect. We performed an analysis in which medication dose was entered as a covariate, and found that the correlation between the severity of negative symptoms and the increase of volume in the thalamus remained significant.
Pathophysiological changes in schizophrenia – including brain morphological changes – may be drastic, especially in schizophrenic patients with onset before brain maturation. Therefore, studying schizophrenia patients whose age of onset is as early as childhood is valuable in clarifying the pathophysiological dynamics of the disorder. Large-scale longitudinal studies are also needed to elucidate brain morphological changes in young populations with early onset schizophrenia.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
YY, GS, and TT designed the study. YY and TT contributed to recruitment of study subjects and MRI data collection. HM and TT were involved in clinical evaluation of the participants and procedures of the MRI data acquisition. JS provided assistance of the MRI data collection and performed data analyses. KNi, HI, KJT, KT, KS, HS, KNa and NM participated in the stage of designing the study and recruitment of the study subjects. NT supervised the study and refined the analyses. YY, GS, and NT wrote the manuscript. All the authors read the paper and approved the final form of the manuscript.