- Top of page
- Material and methods
- Declaration of interest
In the past years, several researchers have investigated the possible associations between brain structural abnormalities and schizophrenia, even at the early phases of the disease [1-5].
Structural abnormalities of gray matter (GM) found in first-episode schizophrenia (FES) are supposed to be unbiased by secondary processes, such as duration of illness, long-term treatment, and different outcome variants, and therefore ideally reflecting only the primary pathological changes. If the whole pattern of GM abnormalities reflects an aberrant neuronal network, a concomitant alteration of white matter (WM) tissue might also be observed [6-8]. Previous diffusion tensor imaging (DTI) studies in chronic schizophrenia have shown decreased fractional anisotropy (FA) in frontal  and parietal lobes , while some authors found even no differences between patients and controls . Interestingly, recent studies suggested that WM abnormalities were present since the onset of the disease .
Both GM and WM abnormalities are thought to have clinical and neuropsychological correlates, which could characterize neurobiological models of the disorder.
A few pioneering studies have addressed the relationship between psychopathological dimensions and neuroanatomical changes in FES: Some authors provided evidence that morphology of the temporal cortex is associated with positive as well as with negative symptoms [13, 14], while others failed to detect any relationships between brain structural deviations and symptoms .
Even less investigated was the association between brain structure and cognitive performance. Disruptions of WM networks have been proposed as potential mechanisms for cognitive dysfunction in schizophrenia and other neurological disorders . Few studies suggested that WM abnormalities of the temporal regions may account for memory functioning impairment in the early course of schizophrenia , while deficits of executive and motor functioning might depend on WM disarray of the major tracts connecting fronto-temporal cortices . Unfortunately, GM volume and WM integrity have not been explored together in correlation with neuropsychological performances, so far.
It is becoming more evident that psychopathological and neuropsychological profiles are interrelated aspects of a complex behavioural expression. For instance, in non-affective psychosis, significant correlations between negative and disorganization symptom clusters and the majority of neuropsychological measures have been reported, while positive and depressive symptoms have shown no relations with cognition . Accordingly, different symptoms might show different patterns of relations with brain structure.
Over the years, functional neuroimaging techniques have provided a great body of evidence about cognitive functions, but unfortunately the tasks used for brain activation often do not explore the same functions and are generally less standardized than neuropsychological tests. Consequently, a pure volumetric approach such as voxel-based morphometry (VBM) might present less methodological caveats than functional neuroimaging .
Aims of the study
We used VBM and tract-based spatial statistics to investigate GM volume and WM integrity in FES, correlating gray and WM structural changes with symptoms and neurocognitive scores.
- Top of page
- Material and methods
- Declaration of interest
In recent years, schizophrenia has been proposed as a neurodevelopmentally derived ‘misconnection syndrome’ involving connections between cortical regions and the cerebellum mediated through the thalamus (the cortico-cerebellar-thalamic-cortical circuit, CCTCC) . Abnormal CCTCC leads to misconnection in many aspects of mental activity, or to ‘cognitive dysmetria’, which is substantially a pattern of disorganization. Following this theory, significant correlations between symptoms, cognitive performances, and cerebral anatomy might be detected in schizophrenia. Inherently with the concept of ‘neurodevelopment’, it appears likely to find such correlations since the onset of the disease.
In the present study, whole-brain, rater-independent VBM was carried out to investigate GM and WM volume abnormalities at the onset of schizophrenia (FES), addressing possible neuroanatomical underpinnings of symptoms and cognitive impairment.
First, our results suggest the presence of a reduction in left temporal GM volume in patients as compared to controls, which has been advocated as a trait feature of chronic schizophrenia. Our findings are consistent with those of some authors , while others showed GM reduction in different regions [1, 2]. Factors that may underlie the differing patterns of morphological changes in FES are numerous, rendering difficult any comparison with previous studies. Because sociodemographic and clinical characteristics of our series were comparable with the previous investigations applying VBM in FES, differences between other reports may be due to different definition of first episode, duration of illness before assessment, and type of medication. We hypothesized that the influence of antipsychotic treatment in GM volume of our population was unlikely, as we had only patients treated for less than 14 days. Variations in VBM methodology, such as normalization parameters, may affect the sensitivity of the study to detect changes in some brain regions, particularly in medio-temporal regions; differences in image smoothing may affect the number of areas of volume reduction detected . The SPM8 model and the DARTEL registration method used in our study allow for the identification of structural changes with a more accurate intersubject alignment, obtaining a correct realignment of small inner structure .
The second finding of our study was a disarray of all major WM tracts in patients compared to controls. Although obtained in a small sample, this finding suggests the presence of WM disarray since the onset of the disease, sustaining that neurological aberrations lead to intra- and interhemispheric deregulated connectivity, which may explain the global nature and heterogeneity of cognitive deficits in schizophrenia . As far as imaging correlates are concerned, to the best of our knowledge this is the first study examining the relationship between GM volume, WM integrity, and both psychopathological and neuropsychological measures in FES. The psychopathological factorial model itself – which is the best model enabling to study the relationship between specific symptoms and underlining neural substrates – has been rarely applied to correlative analysis in schizophrenia [34-37].
First, the disorganized/cognitive PANSS factor was associated with decrease in GM volume and FA disarray in the left cerebellum. Cerebellum is involved in basic neurocognitive functions such as timing and associative learning and plays a significant role in cognition . It is activated in a variety of mental activities including facial recognition, emotion attribution, directed attention, and memory. To the best of our knowledge, this is the first study identifying, in the early course of schizophrenia, the disorganized psychopathological pattern related to volume deficits and microstructural disarray in that region, thus replicating findings from chronic population .
Second, negative symptoms displayed a significant inverse correlation with FA of the major WM cortico-cortical association tracts. This is consistent with previous studies and may be explained by the ‘functional disconnection’ hypothesis, according to which a disconnection between fronto-temporal WM areas might be related to negative symptoms of schizophrenia [17, 39]. Unfortunately, we failed to detect a correlation between negative symptoms and cortical GM volume. One possible interpretation is that disturbed neural circuits, rather than structural alterations per se, may play a role in the development of symptoms in FES . Interestingly, a disarray of WM seems to be specifically associated with the expression of negative symptoms in chronic patients . Unfortunately, the small sample size of the present study did not allow more definitive conclusions.
Correlative neuropsychological analysis showed additional interesting results. Social cognition impairment was related to a GM decrease in right temporo-occipital cortex and to a concurrent WM disarray of superior and middle temporal gyrus bilaterally, anterior thalamic radiation, and superior longitudinal fasciculus. The correlation with GM volume is in line with a previous study in first-episode psychosis . On the other hand, WM abnormalities could be partially included in the revised face-perception circuitry, composed of a core system (temporo-occipital regions) mediating the visual analysis of faces, and of an extended pathway (frontal-limbic system) deriving meaning from face perception . As supported by recent literature, impaired social cognition is a core feature of schizophrenia, representing a specific deficit rather than a result of a generalized cognitive dysfunction . Interestingly, it appears to be present even before the onset of psychosis  and to be significantly related to brain structural abnormalities .
The results at uncorrected cluster levels are also discussed below; low statistical power may represent a possible caveat, but is perhaps allowed by the preliminary nature of our study.
Both speed of processing and visual memory impairment were related to a fronto-temporal WM disarray. This common pattern of underlining anatomical abnormalities could possibly reflect the clinical association between them . However, we failed to detect a significant concurrent GM volume decrease, which is not surprising as some authors suggested that WM pathology may play a primary role in cognitive deficits .
In summary, this is an original exploratory study of anatomical underpinnings of neuropsychopathology in FES. While the small number of patients is of course a weakness of the study, one strength is represented by the selection of patients. All patients included had an untreated psychosis for less than 12 months before admission, and all had diagnostic confirmation of schizophrenia six months after study entry. This supports the accuracy of our diagnoses of FEP. Moreover, a possible confounding factor in the evaluation of cognitive function, such as the history of prior drug abuse, has been excluded in all our patients. Of note, controls were matched for age, gender, and education, which could be confounding factors when inferring cognitive function from brain structure.
Consistently with a neurodevelopmentally derived ‘misconnection syndrome’ , our results suggest how the structural development of key brain regions may relate to neuropsychopathological dysfunction even at the very early stages of schizophrenia. Overall, confirmatory longitudinal studies are needed.