Grey matter morphometric biomarkers for classifying early schizophrenia and PD psychosis: a multicentre study

Participants

In this study, we used a cross-sectional dataset to investigate early schizophrenia and Parkinson’s disease, combining imaging data from six original projects: the Early Psychosis Human Connectome Project (EP-HCP, https://www.humanconnectome.org/study/human-connectome-project-for-early-psychosis), an early schizophrenia dataset collected in Cambridge, UK ^9,11, an at-risk for psychosis dataset collected in Singapore ⁴⁰, and three PD psychosis datasets, from Cambridge, UK ^8,10, Sydney, Australia ⁴¹ and Bangalore, India ²⁵. The final dataset included 722 participants, consisting of: individuals with an at-risk mental state for developing psychosis (ARMS), showing sub-threshold positive and negative symptoms of schizophrenia; individuals with a first episode of psychosis (FEP), consisting of first episode of schizophrenia and first episode of schizoaffective disorder; healthy controls matching FEP and ARMS (Con-Psy); PD without psychosis (PDN); PD with psychosis (PDP); healthy controls matching PDN and PDP (Con-PD). Various clinical scores were recorded. Symptoms related to psychosis and schizophrenia were measured using the Comprehensive Assessment of At-Risk Mental States (CAARMS) in ARMS ⁴² and the Positive and Negative Syndrome Scale (PANSS) in FEP ⁴³. In PD, the Hoehn and Yahr scale⁴⁴ was used to assess the disease stage, and the Unified Parkinson’s Disease Rating Scale (UPDRS, ⁴⁵) item 2 to assess psychotic symptoms and hallucinations. In PD, both the Mini-Mental State Examination (MMSE ⁴⁶) and the Montreal Cognitive Assessment (MoCA, ⁴⁷) were used to assess cognitive decline. MoCA scores were converted to MMSE using a validated conversion table ⁴⁸. Demographic and clinical details, as well as corresponding statistics are described in Table 1.

Ethical approval was obtained from local ethical committees for each original studies: The studies were approved by the Cambridgeshire 3 National Health Service research ethics committee ^8,10; by the ethics review board of the Singaporean National Healthcare Group ⁴⁰; by the ethical committee of the University of Sydney ⁴¹; and by the Institute Ethics Committee of NIMHANS, Bangalore ²⁵. Furthermore, freely available data was used from the Human Connectome Projects (https://www.humanconnectome.org/study/human-connectome-project-for-early-psychosis), for which ethical approval was waived by the Ethical Commission Board of the Technical University Munich. All subjects gave written informed consent in accordance with the Declaration of Helsinki.

MRI acquisition, image preprocessing and Independent component analysis

T1-weighted structural images were acquired for all individuals, at a field strength of 3T. The different MRI sequences are detailed in Supplementary Table 1. T1-weighted structural images were bias field corrected and segmented into grey matter, white matter, and CSF using Statistical Parametric Mapping (SPM12, http://www.fil.ion.ucl.ac.uk/spm/software/spm12/), running on MATLAB version 2018b. Diffeomorphic Anatomical Registration through Exponentiated Lie Algebra toolbox (DARTEL)¹⁰¹ was applied to grey matter images. This procedure created a sample-specific template representative of all 722 subjects by iterative alignment of all images. Subsequently, the template underwent non-linear registration with modulation for linear and non-linear deformations to the MNI-ICBM152 template. Each participant’s grey matter map was then registered to the group template and smoothed with an 8 mm³ isotropic Gaussian kernel.

Independent component analysis

The independent component analysis (ICA) was conducted according to ^49–51. As a first step, all individually modulated and smoothed grey matter maps were concatenated to create a 4D file, which served as the basis for the independent component analysis (ICA). To ensure that only grey matter voxels were retained for the ICA, an absolute grey matter threshold of 0.1 was applied to all images. ICA was performed using the Multivariate Exploratory Linear Optimized Decomposition into Independent Components (MELODIC) method (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/MELODIC) as implemented in the FSL analysis package ¹⁰² version 6.0. To derive data-driven population-based networks of grey matter covariance, the ICA was performed on all subjects (n=722) thus identifying common spatial components based on the covariation of grey matter patterns across all participants. In line with previous work which employed similar methods, we chose 30 components ^50,51,103 which allows for the investigation of a relatively detailed organization and represents one of the most frequent choices in resting state ICA analyses. To avoid spurious results, each of the 30 components or 30 morphometric networks was thresholded at z = 3.5 and binarized ^49–51. Finally, each participant’s grey matter volume was extracted from each of the 30 morphometric networks.

Statistical analyses

Grey matter volume differences between groups

Results of the repeated-measures ANCOVA, with GM volume of the 30 networks as the within-subject factor, group as the between-subject factor and age, TIV, gender and scan site as covariates, showed a significant main effect of group (F(1, 712)=11.55, p<0.001), significant main effect of network-related GM volume (F(8, 5826)=33.87, p<0.001), and significant interactions of network-related GM volume with age (F(8, 5826)=12.79, p<0.001), TIV (F(8, 5826)=36.98, p<0.001), gender (F(8, 5826)=5.29, p<0.001), scan site (F(8, 5826)=8.78, p<0.001) and group (F(41, 5826)=2.23, p<0.001). All within-subject effects were Greenhouse-Geisser corrected due to a significant result in the Mauchly sphericity test. The repeated-measures ANCOVA comparing FEP vs. Con-Psy, Con-PD vs. PDN, Con-PD vs. PDP, FEP vs. PDN, ARM vs. PDN, as well as young vs. elderly controls (Con-Psy vs. Con-PD) showed a significant main effect of group. Details of these results are presented in the Supplementary Materials.

The AUCs from the ROC analyses, representing the overall classification performance of each population-derived morphometric network to differentiate the different groups, are presented in Figure 2 and Table 2. Classification performances differed depending on group comparison. The Con-Psy were differentiated from FEP with an overall good performance in the training and in the validation (AUCs average: 0.80). The Con-PD were differentiated from PDN with a fair performance (AUCs average: 0.72) in the training set and a poor performance (AUCs average: 0.65) in the validation. Similarly, Con-PD were differentiated from PDP with a fair performance (AUCs average: 0.71) in the training, but also a fair performance (AUCs average: 0.71) in the validation. Classification of elderly (Con-PD) from young controls (Con-Psy), however, produced a mainly good to excellent performance in the training set (AUCs average: 0.94) and validation (AUCs average: 0.89; see ROC curves in Figure 2). These results indicate that morphometric networks are suitable parameters to differentiate Con-Psy from FEP, as well as younger (Con-Psy) from elderly controls (Con-PD), and to a lesser degree also for the differentiation of Con-PD from PDP. The best classifying networks for the comparison FEP vs. Con-Psy and PDP vs. Con-PD are presented in Figure 3, showing an overlap in the NW14, the precuneus.

• Download figure
• Open in new tab

Figure 2. Classification performance of group differentiation.

The 30 anatomically derived morphometric areas from the ICA networks thresholded at z=3.5 and overlaid on the ROC curves for each group differentiation. Model training results are presented in solid lines, model evaluation in dotted lines. Black ROC: Con-Psy vs. Con-PD, red ROC: Con-Psy vs. FEP, blue ROC: Con-PD vs PDP; purple ROC: Con-PD vs. PDN.

• Download figure
• Open in new tab

Figure 3. Best classifying networks for FEP and PDP versus controls, with overlap in the precuneus.

A. NW14 and NW16 produced the best classification performance (AUC=0.82) to discriminate FEP from Con-Psy; these NWs consist of the precuneus, temporal pole, parahippocampal gyrus, the orbitofrontal cortex, and the lingual gyrus. B. NW 14 and 18, consisting of the precuneus and the thalamus, produced the best classification performance (AUC>0.73) to discriminate PDP from Con-PD.

We computed a control analysis, in which we trained our model on one comparison (e.g., FEP vs. Con-Psy) and tested on another comparison (e.g., PDP vs. Con-PD). We conducted three main models, using the complete sample: Model 1 - FEP vs. Con-Psy, Model 2 PDN vs. Con-PD, Model 3 PDP vs. Con-PD. We then used Model 1-3 to predict the other two comparisons. Interestingly, while Model 1 (FEP vs. Con-Psy) itself classified well, it failed to classify PDP vs. Con-PD and PDN vs. Con-PD; Model 2 (PDN vs. Con-PD) also classified with fair performance, but the model performed poorly or failed in the classification of the other group comparisons; Model 3 (PDP vs. Con-PD) produced not only fair to good classification in itself, but also when classifying FEP vs. Con-Psy (see Supplementary Figure 1 and Supplementary Table 3).

Whole-brain grey matter pattern differences between groups

Assessment of GM pattern similarity (i.e., homogeneity) indicating how similar one’s whole-brain organization is with every other individual of the respective group revealed a lower homogeneity in all patient groups (Con-Psy vs. FEP (χ2(59)=689.59, p<0.001); Con-Psy vs. ARMS (χ2(59)=440.69 p<0.001); Con-PD vs. PDN (χ2(59)=532.95, p<0.001); Con-PD vs. PDP (χ2(59)=417.01, p<0.001)).

Differences in intra-network variability between groups

The MSLR test assessing differences in the coefficients of variation of GM volume between groups showed a highly significant group effect between psychosis controls and FEP (χ2(1)=18.57, p<0.0001), Con-PD and PDN (χ2(1)=15.63, p<0.0001), as well as Con-PD and PDP (χ2(1)=15.61, p<0.0001), indicating a higher variability in all patient groups across all networks (Figure 4). We furthermore assessed differences for each network using a Bonferroni corrected threshold for multiple comparisons (p<0.002), see Supplementary Table 2 for details. In summary, for the comparison between Con-Psy and FEP, we found significant differences in NW13, NW15 and NW23; between Con-PD and PDN in NW5, NW19, NW26, and NW28; and between Con-PD and PDP in NW19, NW21, and NW 28. All differences were based on an increased coefficient of variation (i.e., variability) in patients relative to healthy controls (see Figure 4).

• Download figure
• Open in new tab

Figure 4. A. Homogeneity of GM volume per network and individual, across all groups.

The GM volume of each network for each individual is correlated with the GM volume of each NW of any other individual. Lighter colours indicate lower correlations. Black squares indicate groups. B-D. Network-specific variability as assessed by the coefficient of variation for different group comparisons: B) Con-Psy versus FEP; C) Con-PD vs. PDN; D) Con-PD vs. PDP. Group differences were investigated using the modified signed-likelihood ratio (MSLR) test; * significant at p<0.002 corrected for multiple comparisons (i.e., 30 networks).

Association with clinical scores

In FEP, we found a significant correlation between the GM volume of NW23, and PANSS (r=-0.21, p=0.017, corrected for multiple comparison), indicating reduced GM volume with higher clinical scores. Correlations between NW13, and NW15 and clinical scores did not reveal any significant effects (p=0.1-0.17). Similarly, in PDP, we found a significant interaction between the GM volume of both NW21 and NW28 and psychosis severity (Hallucination and Psychosis, MDS-UPDRS, item 2; X²=11.26, p=0.0104, X²=11.31, p=0.0102, respectively and corrected for multiple comparison), which showed reduced GM volume with increasing psychosis severity. Furthermore, in PDN, but not in PDP, we found a significant correlation between the GM volume of NW5, 19, 26, 28, and MMSE in PDN, corrected for multiple comparison (r=0.24, p=0.0085; r=0.23, p=0.012; r=0.26, p=0.0053, r=0.3, p=0.00092, respectively). The correlation implies lower GM volume with lower cognitive scores. In Con-PD, the correlation between GM volume of NW19 and 28 and MMSE produced a trend towards significance for the GM (r=0.26, p=0.021, r=0.23, p=0.045, respectively), indicating the same relationship as in PDN – greater GM volume with higher cognitive scores. Importantly, Con-PDs show a smaller range of cognitive scores, pointing towards less cognitive decline. All clinical associations are presented in Figure 5.

• Download figure
• Open in new tab

Figure 5. Correlation of clinical and cognitive scores with specific GM NWs which showed significantly different variability between controls and patients.

A. shows a significant negative correlation between GM NW 23 and PANSS total, indicating lower GM volume with higher symptoms in FEP. B. reveals a significant interaction between both GM NWs, 21 and 28, and the Hallucination and Psychosis, MDS-UPDRS score, also showing reduced GM volume with higher psychotic symptoms. C/D/E show correlations of GM NWs and MMSE in Con-PD (C), PDN (D) and PDP (E). While there are significant positive correlations in PDN and Con-Psy, indicating higher GM volume with less cognitive decline, there is no such correlation in PDP (E). D. the violin plot shows the distribution of the MMSE scores across PDP, PDN and Con-PD, the box plots show individual scores, the median as a line and the mean as a dot. All analyses are controlled for multiple comparisons.

Global group differences in grey matter pattern

The ICA analysis identified 30 morphometric networks which clearly circumscribe cortical and subcortical areas using individual GM maps of all subjects. The structural covariance analysis revealed significant differences between patients and controls across both disorders - FEP vs. Con-Psy, PDN vs. Con-PD, PDP vs. Con-PD. Interestingly, the comparison between psychosis-risk (ARMS) and Con-Psy, as well as comparisons between the patient groups (FEP vs. PDP, ARMS vs. FEP, ARMS vs. PDP, and PDP vs. PDN) remained non-significant, potentially indicating similarities in the GM alterations across disease stages and disorders. Comparisons between ARMS vs. PDN, and FEP vs. PDN were conducted for completeness. As expected, those comparisons revealed significant results, most likely due to age related alterations. GM alterations across the whole brain found in FEP compared to Con-Psy are in line with the literature reporting GM reductions across large areas of the brain ^1,54, including areas such as the anterior cingulate cortex (ACC), thalamus, insula and inferior frontal gyrus (IFG), superior temporal gyrus (STG), middle temporal gyrus (MTG), precuneus, and dorsolateral prefrontal cortex (DLPFC). They are in line with previous studies using SBM in patients with psychosis ^35–37,55. These studies reported decreased grey matter volume in mainly frontal, temporal and parietal regions, although it should be noted that methodological details of the SBM approaches differed between the studies and only two of those (Kasparek et al., 2010; Li et al., 2019) investigated patients with a first episode psychosis. Similarly, we found global, not NW specific, reductions of GM volume across all NW in PDP and PDN compared to Con-PD. Psychosis, especially hallucinations in PD are associated with GM alterations in temporal and visual areas compared to non-psychotic PD patients ²³ and in the dorsal visual stream, the midbrain, cerebellar and limbic and paralimbic structures compared to healthy controls ^26,29,31. In this study, the structural covariance analysis did not reveal differences between PDN and PDP, as PD-associated changes might be prevailing and analysis was conducted at the network level such that the sensitivity to highly localised effects may have been reduced. Partly localised GM alterations, in PD in general, have been reported in fronto-temporo-parietal and occipital areas, as well as subcortical areas like the caudate, the putamen and limbic areas ^38,56–58. Interestingly, there are no overall differences between FEP and PDP or PDN in the age-corrected GM NWs, potentially indicating similarities in structural changes ^59,60.

In our study, we did not find GM differences between ARMS and Con-Psy, despite several studies indicating such differences, especially in the insula, prefrontal and temporal brain regions ^61–64. The following considerations may explain the lack of findings in our sample. First of all, GM changes especially in temporal and frontal areas have been linked to symptom severity particularly attenuated psychotic symptoms ⁶⁵, our sample of ARMS individuals is relatively mildly affected. Secondly, our sample combines European and Asian individuals (ratio 1:2); while all studies that report grey matter differences assess European, North-American or Australian participants ^61–64, a recent study reported no regional grey matter differences in an Asian sample ⁶⁶, discussing lower prevalence of illegal drug use as a potential reason ⁶⁷. While this might provide a potential explanation, the general heterogeneity of this group might be equally likely. In the ARMS group, we did not differentiate between those who transition, or have an increased genetic risk, and those who remit. A meta-analysis ²², however, showed that grey matter differences are more pronounced not only in high-risk individuals who transition into frank psychosis but also in those with a genetic risk compared to those who remitted, for whom it may also normalise. Thirdly, a recent meta-analysis in ARMS reported both increased and decreased GM volumes in different regions compared to healthy controls ⁶⁸. Given these inconsistent findings, it is possible that we did not find a significant overall (i.e., across all NW) group difference in GM volume in the current study for these reasons. As a proof of concept, we observed a strong decrease of GM volume, between young individuals and elderly individuals across all networks; as well as good to excellent classification performances ⁵⁰.

ROC-Classification using grey matter networks for FEP and PDP

Is GM volume in NWs a suitable characteristic to identify individuals with early psychosis (i.e., FEP) or Parkinson’s psychosis (i.e., PDP)? Using logistic regression analysis, we found that morphometric NW patterns are suitable for classification of FEP and Con-Psy with an overall good performance (AUC>0.8). Networks that discriminated best (NW3, 8, 14, 16, 21, 30) included the thalamus, putamen, insula, hippocampus, amygdala, n. accumbens, precuneus, temporal pole, parahippocampal gyrus, orbitofrontal cortex, posterior cingulate gyrus, and lingual gyrus. Those regions are highly relevant for the psychopathology in psychosis, and structural alterations are well described in these areas ^{14,18,19, 69–71}. Also, functional alterations have been detected in those areas, with regard to functional connectivity in general and in the default and salience networks specifically ^72–76, as well as cognitive, reward and salience processing ^{9,11, 77–79}.

Importantly, GM NWs also allowed fair classification performance when discriminating PDP from Con-PD (AUC>0.73). Brain regions of the best classifying networks (NW 8, 14, 18, 19, 23, 24, 28) include the middle temporal gyrus, postcentral gyrus, parahippocampal gyrus, hippocampus, precuneus, thalamus, n. accumbens, putamen, insula, temporal fusiform cortex, lateral occipital cortex, cerebellum crus I, II, cerebellum VIIb, VIIIa, frontal pole, and the Heschl’s gyrus. Again, these regions have been discussed reliably in the literature as core structures for functional and structural alterations in PD with psychotic symptoms^{23,25,26,29,31,80,81}.

Interestingly, the only study ²⁹ applying the structural covariance method to cortical thickness and surface area in PD patients with vs. without visual hallucinations found, amongst others, significant differences in interregional surface area covariance in frontal and inferior-superior parietal regions, temporal fusiform areas, the lateral occipital gyrus, and insula as well as differences in betweenness centrality in PD patients with visual hallucinations compared to those without in the left and right lingual gyrus, in the left lateral occipital gyrus and the right superior parietal lobe. Particularly the overlap in lateral occipital regions, middle temporal areas, fusiform cortex, and the insula between these findings and our results are of notice and point to alterations in areas critically involved in visual perception in PD patients with hallucinations that seem to manifest both in comparison to healthy controls as well as when compared to PD patients without hallucinations or psychotic symptoms.

There is a strong overlap in fairly well classifying regions between FEP and PDP, especially in the putamen, insula, hippocampus, parahippocampal gyrus precuneus, and thalamus. The presence of psychotic symptoms in this group of PD patients might introduce additional differentiating structural characteristics allowing for a better classification. Still, the specificity and sensitivity are reduced compared to the classification of early psychosis, which may result from a close association between age and illness duration in this particular group ^56,58. In a recent meta-analysis ⁸² investigating progressive grey matter atrophy in individuals with PD, significant grey matter reductions were detected in mainly in the caudate, putamen, n. accumbens, and amygdala. Our work shows that these regions overlap with areas affected and used for the classification not only in PD with psychosis but also in early psychosis.

The classification of PD alone, without psychotic symptoms, was poor (max. AUC 0.68) in our sample. This is in contrast to a recent study by Lee and colleagues ³⁸, who were able to classify between PD patients and healthy controls with an accuracy of 0.75 in the validation sample. This study, however, did not differentiate between PD patients with and without psychotic symptoms. Therefore, improved performance in Lee and colleagues ³⁸ compared to our work, could result from the inclusion of individuals with psychotic symptoms. Taken together, our results suggest that the presence of psychotic symptoms allows for a more precise differentiation between patients and healthy control subjects in general, independent of their primary diagnosis. Despite the overlap in brain regions involved that seem to link to the presence of psychotic symptoms, it is not possible in this dataset to differentiate the contribution of specific psychotic symptoms, e.g. visual vs. auditory hallucinations. Importantly, however, functional alterations in the precuneus has been associated with visual hallucinations in PD (see reviews ^83,84) as well as with auditory hallucinations in schizophrenia ^85,86, suggesting potential unifying mechanisms underlying hallucinations in both disorders. Larger studies with distinguishable subgroups of symptom expression are needed to fully understand this potential target area.

Decreased homogeneity and increased variability in patients links to symptoms

As expected, when investigating correlations of individual’s GM NW volumes to every other individual’s GM NW volumes, we found smaller homogeneity – or, in other terms, decreased inter-individual correlation in whole brain grey matter patterns – in all patient groups compared to healthy controls. This decreased homogeneity may be linked to clinical symptomatology. These results are in line with findings in schizophrenia ^87–89 or Alzheimer’s Disease using a similar approach ⁵⁰. Both, Parkinson’s disease and Psychosis are neurobiologically heterogeneous disorders ^88,90, having multiple clinical subtypes, occurring with co-morbidities, and diverse representations across behavior, genetics and brain morphometry. Relating to this, we, therefore, explored interindividual GM volume variability; the variability was increased globally in FEP, PDP and PDN compared to their control groups. Additionally, we found specific NWs that showed increased variability. Within the FEP patient group GM volume was significantly more variable in NWs 13, 15 and 23 comprising the temporal lobe, amygdala, n. accumbens, large areas of the cerebellum, occipital lobe and the frontal pole. In a meta-analysis Brugger and Howes ⁸⁸ investigated GM variability in specific regions and found increased variability in the putamen, thalamus, temporal lobe, and third ventricle, providing some overlap, but also decreased variability in the anterior cingulate cortex. Although the increased variability may be partially caused by secondary factors like medication, illness duration or comorbidities, inherent to all case-control, the most likely cause for the increase variability is, however, the heterogeneity of the neurobiological processes underlying the illness. This heterogeneity furthermore indicates that individuals develop different psychopathological profiles. In support of the latter explanation, we found an association between GM volume in NW 23, comprising temporal lobe, cerebellar areas, fontal pole, postcentral gyrus and occipital lobe, and symptom strength as measured by PANSS, indicating that the increased variability in this region may be explained by symptom expression ^65,91.

Findings in the PD group are consistent with this account. Here we found overall increased variability in PDP as well as PDN compared to Con-PD. In PDP compared to Con-PD variability was significantly greater in NW 19, 21, and 28, comprising areas such as the n. accumbens, putamen, insula, posterior cingulate gyrus, and temporal lobe, showing strong overlap with more heterogeneous areas in the FEP sample. Interestingly, GM volume in NW 21 and 28 showed an association with psychotic symptom strength measured using the MDS-UPDRS, but no correlation with cognitive decline. In contrast, PDN had increased variability in NW 5, 19, 26, and 28, including areas such as the cerebellum, n. accumbens, putamen, insula, thalamus and temporal lobe, which was, in turn, correlated with cognitive performance (i.e., MMSE score), indicating that reduced GM volume in PDN in these areas might be closely related to cognitive decline. Interestingly, in Con-PD, a trend for the same association was detected. Considering that the cognitive decline is lower in Con-PD and therefore the range decreased, the slightly lower correlation seems plausible. These findings are intriguing as they show that, while in FEP and PDP GM volume reduction in NWs with increased variability is linked to increased psychotic symptoms but not cognitive decline, in individuals not affected by psychotic symptoms, such as PDN and Con-PD, GM volume reduction in NWs with increased variability is linked to cognitive decline. Interestingly, one of the overlapping areas is the cerebellum, which has been reported in multiple studies discussing GM alterations in psychiatric and especially psychotic disorders ^89,92–94, but which has also been linked to symptom expression and development ⁹⁵. Temporal lobe alterations are common findings in psychosis, especially in the lateral ^62,96 and medial parts ^97,98, which have been linked to the neurobiological basis of psychosis ^99,100, providing some commonality with PD psychosis, as alterations in these areas may be linked to developing psychotic symptoms.