Cognitive subgroups of affective and non-affective psychosis show differences in medication and cortico-subcortical brain networks

Katharina M Bracher, Afra Wohlschläger, Kathrin Koch, Franziska Knolle
doi: https://doi.org/10.1101/2022.08.03.22278370

Abstract

Cognitive deficits are prevalent in individuals with psychosis and are associated with neurobiological changes, potentially serving as an endophenotype for psychosis. Using the HCP Early Psychosis dataset (n=226), we aimed to replicate cognitive subtypes (deficit, intermediate, spared) through data-driven clustering on affective and non-affective psychosis patients and controls. We explored differences between the clusters in symptom manifestation, cognition, medication, and grey matter volume, comparing patients to controls. Fuzzy K-Means clustering on PCA-selected features revealed three cognitive subgroups significantly varying in clinical symptoms and cognitive impair-ment, and importantly also in medication and grey matter volume in fronto-parietal and subcortical networks. The spared cluster (86% controls, 37% affective psychosis, 17% non-affective psychosis) exhibited unimpaired cognition, lowest symptoms/medication, and grey matter comparable to controls. The deficit cluster (4% controls, 10% affec-tive psychosis, 47% non-affective psychosis) had impairments across domains, highest symptoms/medication, and pronounced grey matter alterations. The intermediate clus-ter (11% controls, 54% affective psychosis, 36% non-affective psychosis) showed fewer deficits than the second cluster, but similar symptoms/medication/grey matter to the first. Controlling for medication, cognitive scores correlated with grey matter changes and negative symptoms across all patients. Our findings generally emphasize the inter-play between cognition, brain structure, symptoms, and medication, and specifically suggest a possible mediating role of cognition linking alterations in brain structure and symptoms, which highlights the potential of screening cognitive changes to aid in tailoring treatments and interventions.

1 Introduction

Cognitive alterations are core symptoms of psychosis [1–4], which have been described in areas of working memory [5, 6], attention [7, 8], reasoning [9, 10], decision making [11–13], salience processing [14, 15], learning (e.g., lower learning rate or increased forgetting, [16–18]) and problem solving [19, 20] and across all stages of the disease [21–25]. Cognitive deficits furthermore precede the clinical onset of psychosis [26], and predict functional out-come in later stages of the disease [27, 28], impacting employment status, independent living and social functioning [28, 29]. Although cognitive impairments are present in about 80% of patients suffering from psychotic disorders [30–32], therapeutic interventions are limited. A meta-analysis including 93 studies using different agents targeting mainly glutamatergic and cholinergic neurotransmitter systems, but also serotonin, dopamine, GABA and nora-drenaline agents [33] reported a significant, although very small (g=0.10) improvement of cognition in general. However, this meta-analysis failed to find significant improvement for cognitive subdomains [33]. Cognitive training, also called cognitive remediation therapy, has produced more promising effects [34, 35]. A recent meta-analysis found that cognitive reme-diation showed significant small-to-moderate cognitive improvements in all domains studied (g=.19–.33) and a small improvement in function (g=.21). Furthermore, research has shown that fewer cognitive deficits and higher cognitive reserve during prodromal and first episode psychosis are generally, diagnosis-independently, associated with better functioning and re-covery [36, 37]. Indicating that maintaining and improving cognitive functioning is crucial in the interventional and therapeutic processes.

Cognitive deficits in psychosis in general have been linked to alterations in the cortico-cerebellar-thalamic-cortical circuits [38]. Here, dysfunctional GABA (gamma-amino-butyric acid) inter-neurons, the main inhibitory neurons of the central nervous system, may disrupt the balance between excitatory and inhibitory processes in the cortex [39]. A recent review [25] summarizes the association between functional brain alterations and cognitive deficits in individuals at-risk for psychosis, early onset psychosis, and chronic schizophrenia. They report a clear association between altered cortical (e.g., prefrontal cortex, anterior cingulate cortex, insula) and subcortical (e.g., thalamus, striatum, hippocampus, cerebellum) brain signalling and aberrant cognition across the different stages. Further support is provided by imaging studies showing reduced gray matter volume and altered network organization which correlates with cognitive deficits, at illness onset [40, 41], early psychosis [42] and chronic schizophrenia [43].

Interestingly, psychosis patients with different diagnoses, e.g., affective vs non-affective psychosis, show varying cognitive deficits [44–46]. In a review, Barch and Sheffield [46] summarized that while the severity of cognitive impairment is stronger in non-affective com-pared to affective psychosis, the relative impairments across different cognitive domains are very similar. Other studies however do not differentiate between affective vs non-affective psychosis when investigating cognitive deficits [47]. Despite these differences in cognitive alterations with regard to specific diagnoses, cognitive deficits in psychosis have been de-scribed and investigated as intermediate phenotypes [48]. In a recent study, Shafee et al.[49] pointed out that cognitive phenotypes may vary grossly depending on specific types of psy-chosis (e.g. affective vs non-affective), suggesting that certain domains of cognition (e.g., working memory vs face processing) may be more etiologically linked to psychosis than oth-ers. Using a K-means clustering approach in a cross-diagnostic sample, Lewandowski et al. [50] identified four cognitive subgroups combining different psychosis groups. Importantly, they identified one cognitively intact cluster including healthy controls and patients with different diagnoses, while the other three clusters were dominated by different cognitive im-pairment profiles [50]. A recent systematic review of data-driven identification of cognitive subtypes [51] highlighted that despite the heterogeneity of clustering methods used and cog-nitive domains studied, there is some commonality in the identification of a severe cognitive deficit phenotype showing deficits across multiple domains and a spared cognitive deficit phenotype with similar performances to controls. It is, however, unclear how these cogni-tive subtypes vary or are linked to differences in symptom expression, medication status and brain structure (i.e., grey matter volume). These open questions are however crucial in order to understand whether cognitive subtypes are clinically relevant, and could increase our mechanistic understanding of the disorder.

In the current study, we, therefore, aimed at exploring these important open ques-tions. First, we wished to replicate three cognitive subtypes using the HCP Early Psy-chosis dataset (https://www.humanconnectome.org/study/human-connectome-project-for-early-psychosis, [52]) using data-driven clustering on standardized cognitive, perceptual and emotional task and score data, but no clinical data. Second, we explored whether the pa-tients in the three clusters differed in symptom expression, cognition, medication and grey matter volume. And, third, depending on results for the first two questions, we wished to understand if symptoms, alterations in cognition and brain morphometry were associated when controlling for medication within and across the clusters.

2 Methods

2.1 Participants

We analyzed data collected by the “Human Connectome Project for Early Psychosis” (HCP-EP, [53], [54]). The HCP-EP 1.1 release (August 2021 HCP-EP Release 1.1 on NDA) contains 251 subjects consisting of 68 healthy control individuals, 57 patients with affective and 126 patients with non-affective psychosis, both patient groups were within the first three years of the onset of psychotic symptoms. The Structured Clinical Interview for DSM-5: Research Version (SCID-5-RV) ([55]) was used to confirm diagnoses of non-affective (i.e., schizophrenia, schizophreniform, schizoaffective, psychosis not otherwise specified, delu-sional disorder, brief psychotic disorder) or affective psychosis (i.e., major depression with psychosis or bipolar disorder with psychosis). Clinical symptoms were assessed using the Positive and Negative Syndrome Scale (PANSS, [56]). Disease onset for all patients was within the last five years prior to study enrollment. For a comprehensive cognitive, percep-tual and emotional assessment, the NIH toolbox ([57], [58]; i.e., cognition (Picture Sequence test, Dimensional Change test, Flanker test, Picture Vocabulary test, Pattern Completion test, List Scoring test, and Oral Reading test), emotion (Self-report emotion), perception (Words in Noise, Odor Identification, and Dynamic Visual Acuity), sensory-motor functions (9-Hole Pegboard, and Grip Strength), the HCP Lifespan Measures ([59]; i.e., Delay Dis-counting and Penn Emotion Recognition), the WASI-II ([60]) and the Seidman Auditory Continuous Performance Test ([61]) was used. Structural brain imaging data was available for 183 of the 251 subjects. After ensuring that there were sufficient data for both, features and subjects (see description of analysis below), the analysis was performed on 226 subjects (i.e., 56 healthy controls, 52 affective psychosis group, 118 non-affective psychosis group). Demographics and clinical scores for the three groups are presented in Tab. 1. Detailed analysis of PANSS items are presented in Suppl. Fig. 1.

View this table:
Table 1: Group demographics and clinical scores of final sample (N=226)

2.2 Variable Selection and Preprocessing

Initially, all cognitive, perceptual and social/emotional functioning scores available in the HCP-EP data-set were selected, resulting in 70 variables as potentially relevant to our analysis (Suppl. Tab. 1). As covariates, we chose age, gender, socio-economic status and mother’s level of education. Variables containing information of primary diagnosis for affective and non-affective psychosis, such as the Positive and Negative Symptom Score or the Clinical Assessment Interview for Negative Symptoms, as well as variables describing medication dosage, usage or equivalent doses, were not included in the clustering analysis, but only used for subsequent analyses.

Many of the selected variables contained missing data. We therefore excluded variables when more than 10% of the entries were missing, and we excluded subjects with more than 20% missing variables (Suppl. Tab. 1). As a result, 33 of initially 70 variables remained in the data set and 226 subjects. The distribution of the remaining subjects reflected the distribution of the original data with regard to diagnosis type. For an overview of the selected variables, and group comparison, see Suppl. Tab. 2. In the final dataset used for analysis, subjects were missing a maximum of seven variables (Suppl. Tab. 4). Missing data were imputed using the mean for continuous and mode for categorical variables (e.g. demographic control variables). Since the ratio of patients and controls was not balanced, mean or mode for data imputation were calculated separately for patients and controls. Both patient groups were combined in order to minimize the bias of classical group membership (Suppl. Tab. 3).

Prior to our analysis, all continuous features and covariates were normalized using z-score normalization. Ordinal covariates such as socio-economic status and mother’s education were scaled between 0 and 1, and treated gender binary.

The HCP-EP brain imaging data contain structural magnetic resonance imaging (MRI) data. We used T1-weighted structural images recorded at a 3T SIEMENS MAGNETOM Prisma scanner using a MPRAGE sequence (TR=2400ms, TE=2.22ms, FoV read=256mm, FoV phase=93.8%, flip angle=8 deg, slices per slap=208, slice thickness=0.8mm). In order to receive individual gray matter volume for structural covariance networks generated across all subjects (Suppl. Tab. 2), the structural brain data was preprocessed as described in the supplement section supplementary material 1.4 and elsewhere [62]. The structural covariance networks were used for analyses following the clustering.

2.3 Correlation and Homogeneity

To investigate correlation between all features, we performed a Pearson’s correlation be-tween all subjects’ non-standardized variables. Differences between correlations between the groups were evaluated with a T-test for means of two independent samples. We then tested homogeneity between groups, using a Levene’s test for equal variances between groups. Bonferroni corrections were applied.

2.4 Feature Selection

To reduce the dimensionality of our data set and at the same time keeping its maximal variance, we used a principal component analysis (PCA), which is suitable for our data type [63], and was applied to all variables and covariates. Significant principal components were identified using a permutation test (5000 random permutations of each feature across subjects). Components which survived permutation testing were considered significant.

2.5 Fuzzy Clustering

Instead of a hard clustering approach that assumes well separated clusters and assigns each data point to only one cluster (e.g. K-Means), we used a soft clustering approach. This approach accounts for fuzzy boundaries between subgroups, and is thus better suited for overlapping subgroups [64], as would be expected for clinical groups. We, therefore, used Fuzzy K-Means clustering as described in [65]. As input to our clustering analysis, we used the dimensionality-reduced data which included demographic control variables for all three groups (i.e., controls, affective and non-affective patients). We specified the number of clusters prior to our analysis. We used a priori knowledge about the number of cognitive clusters (i.e., three different cognitive cluster were reported in the literature [51]). As a control analysis, we also used three clusters on patient data only. This analysis was used to determine influence of control subjects on the clustering, and is presented in the supplements.

Performance of the clustering analysis was determined by the ratio of the subjects of one group (healthy controls, affective psychosis group and non-affective psychosis group) in each cluster. The ratio of subjects in cluster j for group i in {controls, affective and non-affective} Embedded Image

2.6 Cluster exploration

After the identification of the three clusters, we explored differences in the patients dis-tributed over those three clusters and controls in cognitive scores, clinical scores, medication and grey matter volume. We want to point out that the control subjects were removed from each of the clusters and combined in one healthy group. Thus, a group comparison across four groups was computed. For comparing groups, we used SciPy’s Kruskal-Wallis test with Dunn’s test for post-hoc analyses, a Chi-square test of independence or ranked analysis of variance with Bonferroni corrected post-hoc tests. As a control analysis, we performed par-tial Pearson’s correlations between cognitive scores, clinical scores, and grey matter volume controlled for medication across all patients, with multiple comparison corrections.

2.7 Statistical implementation

Preprocessing and data analysis was performed in Python 3.9.7. We used scikit-learn 1.0.2, SciPy 1.7.2 for all analyses and the fuzzy clustering implementation of [66]. Partial correlations were performed using the ppcor 1.1 [67]. For clustering analyses, brain data was corrected for total intacranial volume (TIV), age and sex, using the R stats package, version 4.0.5 R [68].

3 Results

3.1 Homogeneity of data across groups

Correlations of all cognitive, perceptual and emotional functioning data revealed, that correlations were highest within controls compared to affective (T-test: t5838=16.34, p<0.001) and non-affective (T-test: t17058=26.03, p<0.001) patients. For non-affective patients, corre-lations within subjects were lowest (non-affective vs. affective patients: T-test: t16626=8.47, p<0.001), indicating greatest variability. Confirming these results, highest heterogeneity was found in non-affective patients compared to controls (Levene’s test for equal variance: F1,174=187.3, p<0.001) and affective (Levene’s: F1,168=70.39, p<0.001). Controls were the most homogeneous (Levene’s: F1,106=22.34, p<0.001) (Fig. 1a).

Figure 1:
Figure 1: Variable correlations and dimensionality reduction of cognitive data

(a) All variables (i.e., cognitive, emotional, perceptual data) correlated across subjects using Pearson correlation. Correlations within groups are displayed in squares on the diagonal, and correlations between groups are displayed in off-diagonal squares. (b) Variance ex-plained by each of the PCs in % of a PCA performed on all variables and covariates across 226 subjects. The first five PCs (blue) survived permutation testing (p<0.05, 5000 permu-tations). Significant components captured 55.8 % of all variance. (c) Individual data points represent relevant variables for each subject, displayed on the first two principal components and colored according to subject group affiliation.

3.2 Feature selection using principal component analysis

During the data dimensionality reduction, we identified five significant PCs, using all data from all groups, that captured 55.8 % of total variance (Fig. 1b). Explained variance of PCAs was consistent with other symptom-reduction studies [69]. Fig. 1c illustrates the data reduction using the first two principal components. The two patient groups and the control group are predominantly distributed across the dimension of P1. However, no clear boundaries between groups could be detected. The features that explained most variance were Fluid Intelligence and Crystallized Intelligence, Total IQ, the Picture Vocabulary Test Picture vocabulary test, Oral reading recognition, Auditory attention, Working memory, WASI -Verbal comprehension, WASI -Matrix reasoning as well as DCCS -Executive functioning (Suppl. Fig. 3). To check the influence of combining patient and control data, the same analysis was performed on patients only data, which also resulted in five significant components that captured 53.7 % of all variance (p<0.05, 5000 permutations) (Suppl. Fig. 4a, b). The top ten features contributing to the explained variance were the same across both analysis -with and without controls. Those feature were taken for subsequent statistical analyses.

3.3 Group representation in the three clusters

We performed clustering on the selected features, including controls and both patient groups, with three clusters. Fig. 2a and b present the results. One cluster (cluster 0) contained 86% (48/56) of all controls and 37% (19/52) of all affective subjects. Non-affective subjects were represented only in small proportion of 17% (20/118). A mixed, second cluster (cluster 1) consisted of mostly patients, with the majority of affective individuals: 54% (28/52) of non-affective and 36% (42/118) of affective patients and only 11% (6/56) control. Most non-affective subjects were contained in the third cluster (cluster 2) with 47% (56/118) of the non-affective, and only 10% of affective patients (5/52) and 4% (2/56) of controls. (Fig. 2b).

Figure 2:
Figure 2: Cluster analysis: group representations and cognitive differences

(a) Result of clustering. Cluster affiliation of each subject is displayed on the first two PCs. Colors correspond to cluster 0, 1 or 2. (b) Percentage of subjects of each group in each cluster. E.g. 47 % (56/118) of all non-affective subjects are in cluster 2.

3.4 Exploring cognition, symptoms, medication and gray matter volume in the clusters

To identify possible cognitive subgroups represented in the clusters, we explored cognitive scores and symptom expression for patients assigned to the clusters (see comprehensive display of statistical differences between all clusters and cognitive items in Suppl. Tab. 5). We compared cognitive scores across patients in respective clusters. Scores of control subjects were used for comparison. Patients in cluster 2 (mainly non-affective patients) showed a significant decrease in all cognitive features compared to patients assigned to other clusters, as well as controls. Patients in cluster 1, which contains 28 affective and 42 non-affective patients, showed significant decrease of Fluid Cognition, Auditory attention (%correct) and DCCS -Executive functioning compared to controls. Patients (19 affective and 20 non-affective) in cluster 0 showed most similar scores compared to control subjects, with Auditory attention (%correct) being the only score that is significantly lower compared to controls (Fig. 3c) and group comparisons are presented in Tab. 5.

Figure 3:
Figure 3: Symptom expression and medication status

(a) Clinical scores of affective and non-affective patient within each cluster are displayed in boxplots. Individual boxplots show data minimum, first quartile, median, third quartile, and data maximum. Individual subjects are overlaid as dots. Outliers are indicated outside the minimum or maximum. (b) Medication dosage and status of affective and non-affective patients within each cluster. (c) Cognitive score of affective and non-affective patient within each cluster are displayed and compared to all control subjects (c, light blue). The plots display the ten features contributing most to explained variance of the PCs. (d) Differences in three grey matter volume comparing affective and non-affective patient within each cluster are displayed and compared to all control subjects (c, light blue). NW 18 comprises the putamen and the amygdala; NW 29 comprises the paracingulate gyrus, the juxtapositional lobule, the supe-rior parietal lobule, and the precentral gyrus;1a1nd the NW 30 comprises the superior fornatl gyrus, the frontal pole, the putamen, the postcentral gyrus and the cerebellum crus. The plots display the ten features contributing most to explained variance of the PCs. Signifi-cance is indicated as p ≥ 0.05 (*), p ≥ 0.01 (**) and p ≥ 0.001 (***).

Patients in cluster 2, who showed the strongest cognitive deficits and who were mainly diagnosed with non-affective psychosis, were significantly increased in PANSS total com-pared to patients in both, cluster 0 and 1. They were also significantly increased in PANSS positive, PANSS negative and general psychopathology compared to patients in cluster 0. Even tough patients in cluster 0 and cluster 1 did not show as a strong difference in cognitive scores, patients in cluster 1 were significantly increased in PANSS total, PANSS positive and general psychopathology compared to patients in cluster 0 (Fig. 3a) and group comparisons are presented in Tab. 5.

Further, we explored the medication status of patients within the clusters. We found that patients assigned to cluster 2 had a significantly higher Chlorpromazine equivalence dose in general and at the scanning date compared to patients in both cluster 0 and 1, and a significant increase in lifetime antipsychotic drug exposure compared to patients in cluster 1 (Fig. 3b) and group comparisons are presented in Tab. 5.

Furthermore, we explored differences in grey matter networks across the patients in the clusters and the controls. The ranked analysis of variance revealed a highly significant inter-action effect (F(87)=2.564, p=1.79e-13) between the patients in the three clusters and the controls (four groups) and the grey matter volume in each network (30 networks). Bonferroni corrected post-hoc analyses revealed significant differences in network 18 (F(3,48.64)=9.32, p=0.00027), network 29 (F(3,45.13)=7.8, p=6e-05) and network 30 (F(3,45)=10.95, p=2e-05) (Fig. 3d) and group comparisons are presented in Tab. 5.

Finally, we investigated partial Pearson correlations between the ten cognitive features, the three significant grey matter volume networks, and the clinical scores, controlling for Chlorpromazine equivalent dose within each cluster and across all patients. Not with in the clusters, but across all patients we found several significant, multiple-comparison-corrected associations, see Fig. 4 and Suppl. Tab. 6.

Figure 4:
Figure 4: Partial correlations between brain, cognition and clinical scores

Correcting for the equivalent dose of medication, we used Partial Pearson correlations to investigate the interaction between cognitive scores (i.e., Fluid Intelligence and Crystallized Intelligence, Total IQ, the Picture Vocabulary Test Picture vocabulary test, Oral reading recognition, Auditory attention, Working memory, WASI -Verbal comprehension, WASI -Matrix reasoning as well as DCCS -Executive functioning), brain networks (i.e., NW18, NW29, NW30), and clinical scores (i.e., PANSS total, PANSS positive, PANSS negative, General psychopathology score) across all patients independent of cluster.

View this table:
Table 2:

Comparison between patients in all clusters and control subjects for cognitive features that contributed most to variance, as well as medication and PANSS comparison between patients in different clusters. For the group comparison, a Kruskal-Wallis was used with a Dunn’s test for pairwise comparisons (statistics and p-value displayed or ns -non significant).

4 Discussion

The aim of this study was (1) to investigate cognitive subtypes using the HCP Early Psychosis dataset and data-driven clustering on standardized cognitive, perceptual and emo-tional task and score data, but no clinical data; to explore (2) differences in and (3) associ-ations between cognition, symptom expression, medication and grey matter volume. Using a data driven parameter selection and clustering approach, we were able to identify three clusters that significantly differed in across cognitive deficits forming a cognitively intact cluster, an intermediately affected cluster and an cognitively affected cluster, and poten-tially characterise subgroups previously described in the literature [36, 50, 70–75]. Impor-tantly, our results extend those findings, showing that patients within those clusters also differ in medication dosage, in specific grey matter brain networks and in clinical symptoms. Interestingly, across all patients but not within clusters we found that decreased grey mat-ter volume in frontal, parietal and subcortical regions was linked to higher cognitive scores including crystallized cognition, verbal comprehension or matrix reasoning, when control-ling for medication, and that decreased cognitive scores were linked to increased negative symptoms, when controlling for medication.

Using a three cluster solution on all participants (i.e., controls, affective and non-affective psychosis), allowed the identification of cognitive subtypes, which significantly varied in clinical and cognitive impairment. Patients of cluster 2, consisting of nearly 50% of the non-affective psychosis individuals, expressed the highest symptom scores across PANSS total, PANSS negative and positive and general psychopathology and impaired cognition in all domains compared to cluster 0 containing 17% of the non-affective and 37% of the affective individuals, and partially also compared to cluster 1 which consists of the majority of affective individuals and 36% of non-affective individuals. Patients in cluster 0 had the lowest symptom scores and globally spared cognitive abilities, which were similar to those of controls. Cluster 1 was intermediate, with cognitive impairments in several but not all domains and slightly increased symptoms compared to the cognitively spared cluster. This finding confirms results from Lewandowski and colleagues [50] who found a four-cluster solu-tion to provide the best fit to their data containing three diagnostic patient groups, with one globally impaired cluster, for which cognitive deficits were associated with symptom severity and poorer functioning, one cognitively spared cluster and two intermediate clusters [50]. The overall structure of cognitive clusters identified in the present study supports findings discussed in a recent meta-analysis of data-driven identification of cognitive phenotypes in schizophrenia [51]. Green and colleagues [51] describe that what is characteristic to all cluster solutions is the presence of a cognitively spared, one or multiple intermediate and a deficit subgroup [51]. Importantly, the current study replicates three cognitive clusters us-ing standard cognitive, perceptual and emotional assessments, selected for general cognitive screening purposes but not necessarily to detected the largest or most consistent cognitive deficits in early stages of psychosis, suggesting generalisability among these clusters.

In addition to the differences across many cognitive domains between the patients of the clusters, we also found differences with regard to amount of medication using Chlorpro-mazin equivalent doses and grey matter volume in frontal, parietal and subcortical brain areas. This emphasizes the complexity of the inter-relationships of cognitive deficits, brain alteration, medication usage and symptom expression, especially when considering that the clustering is based on task and questionnaire data only, and still, differences across all do-mains -cognition, brain scores, medication and symptoms -have been identified. Several studies indicate an association between higher doses of medication and stronger cognitive deficits [76–78]. In a birth cohort study, for example, Husa and colleagues [76] showed that a higher lifetime dose of anti-psychotics was associated with lower cognitive performance in schizophrenia patients at the age of 43. Interestingly, a longitudinal study [77] investigating the effect of anti-psychotic treatment discontinuation showed that those individuals who did not remain on their medication after a 3.5 year follow up had improved significantly more than those who stayed on their medication even when controlling for symptom severity and cognitive scores at baseline. General non-adherence of medication use, however, does not have the same positive effect on cognition [79]. Especially, anticholinergic medication has been associated with a high cognitive burden [78], which is being supported by our results. Changes in gray matter volume have been associated with an increased risk for psychosis and disease development [80–82], and provided the basis for good classification in a recent multicohort-study [83], as well as in earlier studies [84, 85], although classification results are inconsistent [86]. Our results show that the cluster with the strongest cognitive deficits has increased grey matter volume in three brain networks spanning fronto-parietal and sub-cortical areas. Across all participants and when correcting for medication we also found a negative correlation between grey matter volume and cognitive performance in several cognitive tests, including general cognition, verbal cognition and reasoning. Interestingly, grey matter volume alterations, especially reductions have been reported in association with schizophrenia [41, 87–89]. Results however depend on the specific region [90], illness stage and medication [91, 92]. Interactions between grey matter alterations and various cognitive scores have not been studied extensively. Most studies investigated cognitive alterations in specific domains (e.g., working memory) and often reported positive correlations [93–95]. Very few studies report negative correlations -increased grey matter volume being linked to reduced cognitive scores. Zhang and colleagues [96] for example showed, comparable with our results, that the performance in the Stroop Color-Word Test’s Card C was negatively correlated with grey matter volumes of frontal and middle frontal brain areas. In a cohort of at-risk mental state for psychosis individuals, Koutsouleris and colleagues [97] found positive and negative correlations between grey matter volume and performance in the trail-making test, with negative correlation reported for cerebellar regions, which is comparable to what we report in the current study. We suspect that the structural covariance networks anal-ysis, which specifically aims at finding similarity networks between participants, and then analyses grey matter volume differences within these similar networks might contribute to the difference in directionality of the correlations. Our results however clearly demonstrate differences in the cluster containing the most strongly affected patients -clinically and cog-nitively. These differences reported in the literature may suggest that grey matter changes are not generally linked to cognitive changes, but rather play a mediating role. This argu-ment would be in the same line of thought as put forward by Palaniyappan [98], stating that grey matter changes in multimodal brain regions which have a supervisory function on sensory, emotional and language processing, may link to symptom expression when occur-ring with functional impairments. Future studies should therefore aim at the combination of additional structural imaging data, such as structural or white matter connectivity, to com-plement their analysis and potentially identify underlying neuropathological mechanisms.

Our data furthermore reveals a strong link between negative symptoms and cognitive impairments. Investigating PANSS total, we found a step-wise increase in symptom severity from cluster 0 to cluster 2, with the cognitive deficit cluster 2 to show highest symptom scores. We found, also, that cluster 0, the cognitively spared cluster, showed significantly lower positive symptoms compared to cluster 1 and 2. Furthermore, the globally impaired cluster (cluster 2) revealed increased negative symptoms compared to both other clusters, and increased global psychopathology compared to the cognitively spared cluster (cluster 0), indicating that the cognitive deficits occur in those subjects with strong negative symptoms and a higher severity of general psychopathology [99]. This is in congruence with Oomen and colleagues [100] who reported three clusters based on only cognitive data with one severely cognitively impaired cluster, which showed general functioning being significantly lower compared to the patients in the other clusters. Interestingly, patients of the severely impaired cluster also showed lower general functioning scores at a trend at 6- and 12-month follow-up. Similar results were reported by Haining and colleagues [74]. Tan and colleagues [75] on the other hand did not find associations between the three cognitive clusters and symptom expression, but found that the cognitively impaired subgroup already showed worse academic performance at the level of childhood, early and late adolescence. These findings confirm the critical relevance of cognitive deficits for early detection and functional prediction [26–28, 74]. This often replicated distribution of a severely impaired cluster, indicating that early interventions based on such cluster analysis would be suitable too.

Finally, as our results confirm that patients with non-affective psychosis show stronger cognitive deficits compared to patients with affective psychosis, it is not surprising that non-affective patients are more likely to be in the intermediate and impaired subgroups. Still both patient groups are present in all clusters, with the cognitively intact subgroup consisting of 19/52 and 20/118 patients with affective and non-affective psychosis, respectively, the intermediate group with 28/52 affective and 42/118 non-affective psychosis patients, and the cognitively impaired group with 5/52 affective and 56/118 non-affective psychosis patients. Our results indicate that clustering extends classical patient classification and diagnosis solely based on International Statistical Classification of Diseases (ICD)/Diagnostic and Statistical Manual of Mental Disorders (DSM) [101, 102] and may provide an additional characterisation of patients which may emphasise additional targets of interventions and treatment, such as cognitive remediation, especially for those individuals in the intermediate and in the deficit cluster.

This study has several limitations: Generally, K-Means [103] is a commonly used cluster-ing algorithm that performed well on our behavioral, such as cognitive, data. Nevertheless, there are some drawbacks of this method: First, the algorithm requires a predefined number of clusters. Based on the diagnosis, data contained at least three groups. We used this prior information as basis for the specification of the number of clusters in our analysis. Based on the literature, where a spared, an impaired and an intermediate cluster are regularly de-tected [51] we predefined three clusters for the identification of subgroups. Identification of four or more clusters require large sample sizes in order to produce reliable and interpretable results. Our sample size does not allow this, because of the restricted number of subjects. Second, K-Means clustering does not work well with non-spherical cluster or clusters with different sizes [104]. We, therefore, also performed fuzzy K-means clustering approach which accounts for fuzzy boundaries between subgroups, and is suited for potentially overlapping subgroups [64]. The age difference between the groups might constitute another limitation of the study. However, since age was added as a covariate to all analyses, we believe that this aspect did not affect the results to a significant degree. Third, the basis of our clus-tering provided cognitive, perceptual and emotional test scores, which were taken from the standardised NIH toolbox ([57], [58]). Further improvement of clustering and identification of cognitive subgroups may be achieved through the selection of specific behavioral tasks and cognitive domains. Moreover, advanced analysis strategies, e.g., computational mod-elling may improve clustering [16, 17, 105], as it provides the opportunity to identify and mathematically differentiate behavioral parameters, which were found to be reliable and unique across individuals [106].

In conclusion, our results provide evidence for the presence of three cognitive subgroups -one cognitively intact, one intermediate and one deficit group -replicating previous find-ings. This study however extends those results showing that patients within those three clusters also differ with respect to current medication dosage and grey matter volume in fronto-parietal and subcortical regions. Our results therefore emphasize the complex inter-relations between cognition, symptoms, brain structure and medication, drawing attention to the pivotal role alterations in cognition as a factor for the selection of treatments and interventions.

Data Availability

All data produced in the present study are available upon reasonable request to the authors

Acknowledgement

Research using Human Connectome Project for Early Psychosis (HCP-EP) data reported in this publication was supported by the National Institute of Mental Health of the National Institutes of Health under Award Number U01MH109977. The HCP-EP 1.1 Release data used in this report came from DOI: 10.15154/1522899.

Footnotes

  • Conflict of Interest: None of the authors declares a conflict of interest.

  • Author contributions: FK, KB, AW conceptualisation; KK preprocessing of structural brain data; FK and KB formal analysis of cognitive data and clustering analysis; FK and KB writing of first draft; FK, KB, AW and KK editing and revising of manuscript.

  • We defined our clusters on data only containing cognitive, emotional and perceptual task and questionnaire data, while before we included structural brain data. We the analysed the clusters with regard to alterations in cognition, medication, symptoms and structural brain data. This we had not done before.

References

  1. [1].↵
    P. Fusar-Poli, G. Deste, R. Smieskova, S. Barlati, A. R. Yung, O. Howes, R.-D. Stieglitz, A. Vita, P. McGuire, and S. Borgwardt. “Cognitive Functioning in Prodro-mal Psychosis: A Meta-analysis”. Archives of General Psychiatry 69.6 (June 2012), pp. 562–571. issn: 0003-990X. doi: 10.1001/archgenpsychiatry.2011.1592.
    OpenUrlCrossRefPubMedWeb of Science
  2. [2].
    D. M. Barch and A. Ceaser. “Cognition in schizophrenia: core psychological and neural mechanisms”. Trends in cognitive sciences 16.1 (2012), pp. 27–34.
    OpenUrlCrossRefPubMedWeb of Science
  3. [3].
    S. J. Dienel and D. A. Lewis. “Alterations in cortical interneurons and cognitive function in schizophrenia”. Neurobiology of Disease 131 (2019), p. 104208. issn: 0969-9961. doi: 10.1016/j.nbd.2018.06.020.
    OpenUrlCrossRefPubMed
  4. [4].↵
    M. F. Green and P. D. Harvey. “Cognition in schizophrenia: Past, present, and fu-ture”. Schizophrenia Research: Cognition 1.1 (2014), e1–e9. issn: 2215-0013.
    OpenUrlCrossRef
  5. [5].↵
    A. Zanello, L. Curtis, M. B. Bâ, and M. C. Merlo. “Working memory impairments in first-episode psychosis and chronic schizophrenia”. Psychiatry research 165.1–2 (2009), pp. 10–18.
    OpenUrl
  6. [6].↵
    S. J. Wood, C. Pantelis, T. Proffitt, L. Phillips, G. Stuart, J.-A. Buchanan, K. Ma-hony, W. Brewer, D. Smith, and P. McGorry. “Spatial working memory ability is a marker of risk-for-psychosis”. Psychological medicine 33.7 (2003), pp. 1239–1247.
    OpenUrlCrossRefPubMedWeb of Science
  7. [7].↵
    B. A. Cornblatt and A. K. Malhotra. “Impaired attention as an endophenotype for molecular genetic studies of schizophrenia”. American journal of medical genetics 105.1 (2001), pp. 11–15.
    OpenUrlCrossRefPubMedWeb of Science
  8. [8].↵
    I. Aase, K. Kompus, J. Gisselgård, I. Joa, J. O. Johannessen, and K. Brønnick. “Language lateralization and auditory attention impairment in young adults at ultra-high risk for psychosis: A dichotic listening study”. Frontiers in psychology 9 (2018), p. 608.
    OpenUrl
  9. [9].↵
    P. Garety, E. Joyce, S. Jolley, R. Emsley, H. Waller, E. Kuipers, P. Bebbington, D. Fowler, G. Dunn, and D. Freeman. “Neuropsychological functioning and jumping to conclusions in delusions”. Schizophrenia Research 150.2-3 (2013), pp. 570–574. issn: 09209964. doi: 10.1016/j.schres.2013.08.035.
    OpenUrlCrossRef
  10. [10.].↵
    A. O. Ermakova, N. Gileadi, F. Knolle, A. Justicia, R. Anderson, P. C. Fletcher, M. Moutoussis, and G. K. Murray. “Cost evaluation during decision-making in patients at early stages of psychosis”. Computational Psychiatry 3 (2019), pp. 18–39. issn: 2379-6227.
    OpenUrl
  11. [11].↵
    P. Sterzer, M. Voss, F. Schlagenhauf, and A. Heinz. “Decision-making in schizophre-nia: A predictive-coding perspective”. NeuroImage 190 (2019), pp. 133–143. issn: 1053-8119.
    OpenUrl
  12. [12].
    A. O. Ermakova, F. Knolle, A. Justicia, E. T. Bullmore, P. B. Jones, T. W. Robbins, P. C. Fletcher, and G. K. Murray. “Abnormal reward prediction-error signalling in antipsychotic naive individuals with first-episode psychosis or clinical risk for psy-chosis”. Neuropsychopharmacology 43.8 (2018), pp. 1691–1699.
    OpenUrlCrossRef
  13. [13].↵
    F. Brandl, F. Knolle, M. Avram, C. Leucht, I. Yakushev, J. Priller, S. Leucht, S. Ziegler, K. Wunderlich, and C. Sorg. “Negative symptoms, striatal dopamine and model-free reward decision-making in schizophrenia”. Brain 146.2 (2023), pp. 767– 777.
    OpenUrl
  14. [14].↵
    F. Knolle, A. O. Ermakova, A. Justicia, P. C. Fletcher, N. Bunzeck, E. Düzel, and G. K. Murray. “Brain responses to different types of salience in antipsychotic naıve first episode psychosis: An fMRI study”. Translational psychiatry 8.1 (2018), pp. 1– 13.
    OpenUrl
  15. [15].↵
    J. P. Kesby, G. K. Murray, and F. Knolle. “Neural circuitry of salience and reward processing in psychosis”. Biological Psychiatry Global Open Science (2021). issn: 2667-1743.
  16. [16].↵
    F. Knolle, E. Sterner, M. Moutoussis, R. A. Adams, J. D. Griffin, J. Haarsma, H. Taverne, I. M. Goodyer, P. C. Fletcher, and G. K. Murray. “Action selection in early stages of psychosis: an active inference approach”. Journal of Psychiatry and Neuroscience 48.1 (2023), E78–E89.
    OpenUrlAbstract/FREE Full Text
  17. [17].↵
    M. Montagnese, F. Knolle, J. Haarsma, J. D. Griffin, A. Richards, P. E. Vertes, B. Kiddle, P. C. Fletcher, P. B. Jones, M. J. Owen, et al. “Reinforcement learning as an intermediate phenotype in psychosis? Deficits sensitive to illness stage but not asso-ciated with polygenic risk of schizophrenia in the general population”. Schizophrenia research 222 (2020), pp. 389–396.
    OpenUrlCrossRef
  18. [18.].↵
    F. Knolle, P. Sen, A. J. Culbreth, K. Koch, B. Schmitz-Koep, D. A. Guersel, K. Wunderlich, M. Avram, C. Sorg, and F. Brandl. “Modelling model-based and model-free decision-making in a transdiagnostic sample to investigate disorder-specific and transdiagnostic alterations”. medRxiv (2023), pp. 2023–12.
  19. [19].↵
    R. Morris, T. Rushe, P. Woodruffe, and R. Murray. “Problem solving in schizophre-nia: A specific deficit in planning ability”. Schizophrenia Research 14.3 (1995), pp. 235–246. issn: 0920-9964. doi: 10.1016/0920-9964(94)00044-9.
    OpenUrlCrossRef
  20. [20].↵
    N. Revheim, I. Schechter, D. Kim, G. Silipo, B. Allingham, P. Butler, and D. C. Javitt. “Neurocognitive and symptom correlates of daily problem-solving skills in schizophrenia”. Schizophrenia research 83.2–3 (2006), pp. 237–245.
    OpenUrl
  21. [21].↵
    Y. Liu, G. Wang, H. Jin, H. Lyu, Y. Liu, W. Guo, C. Shi, J. Meyers, J. Wang, J. Zhao, et al. “Cognitive deficits in subjects at risk for psychosis, first-episode and chronic schizophrenia patients”. Psychiatry Research 274 (2019), pp. 235–242.
    OpenUrlPubMed
  22. [22].
    J. M. Sheffield, N. R. Karcher, and D. M. Barch. “Cognitive deficits in psychotic disorders: a lifespan perspective”. Neuropsychology review 28.4 (2018), pp. 509–533.
    OpenUrlCrossRef
  23. [23].
    J. M. Gold and P. D. Harvey. “Cognitive deficits in schizophrenia”. Psychiatric Clin-ics of North America 16.2 (1993), pp. 295–312.
    OpenUrlPubMedWeb of Science
  24. [24].
    M. Fioravanti, O. Carlone, B. Vitale, M. E. Cinti, and L. Clare. “A meta-analysis of cognitive deficits in adults with a diagnosis of schizophrenia”. Neuropsychology review 15.2 (2005), pp. 73–95.
    OpenUrlCrossRefPubMedWeb of Science
  25. [25].↵
    S. Kelly, S. Guimond, A. Lyall, W. S. Stone, M. E. Shenton, M. Keshavan, and L. J. Seidman. “Neural correlates of cognitive deficits across developmental phases of schizophrenia”. Neurobiology of disease 131 (2019), p. 104353.
    OpenUrl
  26. [26].↵
    M. F. Green. “Impact of cognitive and social cognitive impairment on functional outcomes in patients with schizophrenia”. The Journal of clinical psychiatry 77.suppl 2 (2016), p. 8569. issn: 0160-6689.
    OpenUrl
  27. [27].↵
    G. Fervaha, H. Takeuchi, G. Foussias, M. K. Hahn, O. Agid, and G. Remington. “Achievement motivation in early schizophrenia: Relationship with symptoms, cog-nition and functional outcome”. eng. Early Interv Psychiatry 12.6 (2018), pp. 1038– 1044. doi: 10.1111/eip.12405.
    OpenUrlCrossRef
  28. [28].↵
    K. Allott, P. Liu, T.-M. Proffitt, and E. Killackey. “Cognition at illness onset as a predictor of later functional outcome in early psychosis: systematic review and methodological critique”. Schizophrenia research 125.2-3 (2011), pp. 221–235. issn: 0920-9964.
    OpenUrl
  29. [29].↵
    R. Van Winkel, I. Myin-Germeys, M. De Hert, P. Delespaul, J. Peuskens, and J. Van Os. “The association between cognition and functional outcome in first-episode patients with schizophrenia: mystery resolved?” Acta Psychiatrica Scandinavica 116.2 (2007), pp. 119–124.
    OpenUrlCrossRefPubMedWeb of Science
  30. [30].↵
    A. Reichenberg, P. D. Harvey, C. R. Bowie, R. Mojtabai, J. Rabinowitz, R. K. Heaton, and E. Bromet. “Neuropsychological function and dysfunction in schizophrenia and psychotic affective disorders”. Schizophrenia bulletin 35.5 (2009), pp. 1022–1029.
    OpenUrlCrossRefPubMedWeb of Science
  31. [31].
    A. McCleery and K. H. Nuechterlein. “Cognitive impairment in psychotic illness: prevalence, profile of impairment, developmental course, and treatment considera-tions”. Dialogues in clinical neuroscience (2022).
  32. [32].↵
    R. S. Keefe and W. S. Fenton. “How should DSM-V criteria for schizophrenia include cognitive impairment?” Schizophrenia bulletin 33.4 (2007), pp. 912–920.
    OpenUrlCrossRefPubMedWeb of Science
  33. [33].↵
    I. Sinkeviciute, M. Begemann, M. Prikken, B. Oranje, E. Johnsen, W. U. Lei, K. Hug-dahl, R. A. Kroken, C. Rau, J. D. Jacobs, et al. “Efficacy of different types of cognitive enhancers for patients with schizophrenia: a meta-analysis”. NPJ schizophrenia 4.1 (2018), pp. 1–14.
    OpenUrl
  34. [34].↵
    A. Tripathi, S. K. Kar, and R. Shukla. “Cognitive deficits in schizophrenia: under-standing the biological correlates and remediation strategies”. Clinical Psychophar-macology and Neuroscience 16.1 (2018), p. 7.
    OpenUrl
  35. [35].↵
    A. Vita, S. Barlati, A. Ceraso, G. Nibbio, C. Ariu, G. Deste, and T. Wykes. “Ef-fectiveness, core elements, and moderators of response of cognitive remediation for schizophrenia: a systematic review and meta-analysis of randomized clinical trials”. JAMA psychiatry 78.8 (2021), pp. 848–858.
    OpenUrl
  36. [36].↵
    S. Amoretti, A. R. Rosa, G. Mezquida, B. Cabrera, M. Ribeiro, M. Molina, M. Bioque, A. Lobo, A. Gonźalez-Pinto, D. Fraguas, and, et al. “The impact of cogni-tive reserve, cognition and clinical symptoms on psychosocial functioning in first-episode psychoses”. Psychological Medicine 52.3 (2022), pp. 526–537. doi:
    OpenUrlCrossRef
  37. [37].↵
    O. Santesteban-Echarri, M. Paino, S. Rice, C. Gonźalez-Blanch, P. McGorry, J. Glee-son, and M. Alvarez-Jimenez. “Predictors of functional recovery in first-episode psy-chosis: a systematic review and meta-analysis of longitudinal studies”. Clinical Psy-chology Review 58 (2017), pp. 59–75.
    OpenUrl
  38. [38].↵
    D. A. Lewis. “Cortical circuit dysfunction and cognitive deficits in schizophrenia– implications for preemptive interventions”. European Journal of Neuroscience 35.12 (2012), pp. 1871–1878.
    OpenUrlCrossRefPubMed
  39. [39].↵
    M.-y. Xu and A. H. Wong. “GABAergic inhibitory neurons as therapeutic targets for cognitive impairment in schizophrenia”. Acta Pharmacologica Sinica 39.5 (2018), pp. 733–753.
    OpenUrl
  40. [40].↵
    J. Kambeitz, L. Kambeitz-Ilankovic, C. Cabral, D. B. Dwyer, V. D. Calhoun, M. P. Van Den Heuvel, P. Falkai, N. Koutsouleris, and B. Malchow. “Aberrant functional whole-brain network architecture in patients with schizophrenia: a meta-analysis”. Schizophrenia Bulletin 42.suppl 1 (2016), S13–S21.
    OpenUrlCrossRefPubMed
  41. [41].↵
    P. Fusar-Poli, J. Radua, P. McGuire, and S. Borgwardt. “Neuroanatomical maps of psychosis onset: voxel-wise meta-analysis of antipsychotic-naive VBM studies”. Schizophrenia bulletin 38.6 (2012), pp. 1297–1307.
    OpenUrl
  42. [42].↵
    T. M. Minatogawa-Chang, M. S. Schaufelberger, A. M. Ayres, F. L. Duran, E. K. Gutt, R. M. Murray, T. M. Rushe, P. K. McGuire, P. R. Menezes, M. Scazufca, et al. “Cognitive performance is related to cortical grey matter volumes in early stages of schizophrenia: a population-based study of first-episode psychosis”. Schizophrenia research 113.2–3 (2009), pp. 200–209.
    OpenUrl
  43. [43].↵
    N. Banaj, F. Piras, F. Piras, V. Ciullo, M. Iorio, C. Battaglia, D. Pantoli, G. Ducci, and G. Spalletta. “Cognitive and psychopathology correlates of brain white/grey mat-ter structure in severely psychotic schizophrenic inpatients”. Schizophrenia research: cognition 12 (2018), pp. 29–36.
    OpenUrl
  44. [44].↵
    B. Fagerlund, A. K. Pagsberg, and R. P. Hemmingsen. “Cognitive deficits and levels of IQ in adolescent onset schizophrenia and other psychotic disorders”. Schizophrenia research 85.1–3 (2006), pp. 30–39.
    OpenUrl
  45. [45].
    M. Aas, P. Dazzan, H. L. Fisher, C. Morgan, K. Morgan, A. Reichenberg, J. Zanelli, P. Fearon, P. B. Jones, R. M. Murray, et al. “Childhood trauma and cognitive function in first-episode affective and non-affective psychosis”. Schizophrenia research 129.1 (2011), pp. 12–19.
    OpenUrlCrossRefPubMed
  46. [46].↵
    D. M. Barch and J. M. Sheffield. “Cognitive impairments in psychotic disorders: common mechanisms and measurement”. World Psychiatry 13.3 (2014), pp. 224– 232.
    OpenUrlCrossRefPubMed
  47. [47].↵
    N. Craddock, M. C. ODonovan, and M. J. Owen. “Psychosis genetics: modeling the relationship between schizophrenia, bipolar disorder, and mixed (or “schizoaffective”) psychoses”. Schizophrenia bulletin 35.3 (2009), pp. 482–490.
    OpenUrlCrossRefPubMedWeb of Science
  48. [48].↵
    W. Mark and T. Toulopoulou. “Cognitive intermediate phenotype and genetic risk for psychosis”. Current Opinion in Neurobiology 36 (2016). Neurobiology of disease, pp. 23–30. issn: 0959-4388. doi: 10.1016/j.conb.2015.08.008.
    OpenUrlCrossRef
  49. [49].↵
    R. Shafee, P. Nanda, J. L. Padmanabhan, N. Tandon, N. Alliey-Rodriguez, S. Kala-purakkel, D. J. Weiner, R. E. Gur, R. S. Keefe, S. K. Hill, et al. “Polygenic risk for schizophrenia and measured domains of cognition in individuals with psychosis and controls”. Translational psychiatry 8.1 (2018), pp. 1–9.
    OpenUrl
  50. [50].↵
    K. E. Lewandowski, J. T. Baker, J. M. McCarthy, L. A. Norris, and D. OÖngür. “Reproducibility of Cognitive Profiles in Psychosis Using Cluster Analysis”. Journal of the International Neuropsychological Society 24.4 (2018), pp. 382–390. doi:
    OpenUrlCrossRef
  51. [51].↵
    M. J. Green, L. Girshkin, K. Kremerskothen, O. Watkeys, and Y. Quidé. “A system-atic review of studies reporting data-driven cognitive subtypes across the psychosis spectrum”. Neuropsychology Review 30.4 (2020), pp. 446–460.
    OpenUrl
  52. [52].↵
    K. E. Lewandowski, S. Bouix, D. Ongur, and M. E. Shenton. “Neuroprogression across the early course of psychosis”. Journal of psychiatry and brain science 5 (2020).
  53. [53].↵
    A. Breier, M. Shenton, D. Holt, M. Keshavan, D. Ö ngür, and L. Seidman. HCP Early Psychosis 1.0 Data Release: Reference Manual 9. Tech. rep. August. Indianapolis: Boston, MA: Brigham and Womeńs Hospital, Beth Israel Deaconess-Massachusetts Mental Health Center, McLean Hospital, Massachusetts General Hospital; 2021.
  54. [54].↵
    K. E. Lewandowski, S. Bouix, and D. Ongur. “Neuroprogression across the Early Course of Psychosis”. Journal of Psychiatry and Brain Science (2020), pp. 1–25. doi: 10.20900/jpbs.20200002.
    OpenUrlCrossRef
  55. [55.].↵
    M. First, J. Williams, R. Karg, and R. Spitzer. “ructured Clinical Interview for DSM-5-Research Version (SCID-5 for DSM-5, Research Version; SCID-5-RV). (2015) Arlington”. VA: American Psychiatric Association ().
  56. [56].↵
    S. R. Kay, A. Fiszbein, and L. A. Opler. “The positive and negative syndrome scale (PANSS) for schizophrenia”. Schizophrenia bulletin 13.2 (1987), pp. 261–276.
    OpenUrlCrossRefPubMedWeb of Science
  57. [57].↵
    R. C. Gershon, M. V. Wagster, H. C. Hendrie, N. A. Fox, K. F. Cook, and C. J. Nowinski. “NIH toolbox for assessment of neurological and behavioral function”. Neurology 80.11 Supplement 3 (2013), S2–S6.
    OpenUrl
  58. [58].↵
    [Anonymous]. “NATIONAL INSTITUTES OF HEALTH TOOLBOX COGNITION BATTERY (NIH TOOLBOX CB): VALIDATION FOR CHILDREN BETWEEN 3 AND 15 YEARS”. MONOGRAPHS OF THE SOCIETY FOR RESEARCH IN CHILD DEVELOPMENT 78.4 (Aug. 2013), pp. 147–149. issn: 0037-976X. doi:
    OpenUrl
  59. [59].↵
    R. C. Gur, J. Richard, P. Hughett, M. E. Calkins, L. Macy, W. B. Bilker, C. Brensinger, and R. E. Gur. “A cognitive neuroscience-based computerized battery for efficient measurement of individual differences: standardization and initial construct validation”. Journal of neuroscience methods 187.2 (2010), pp. 254–262.
    OpenUrlCrossRefPubMedWeb of Science
  60. [60.].↵
    D. Wechsler. “Wechsler abbreviated scale of intelligence” (1999).
  61. [61].↵
    L. J. Seidman, E. C. Meyer, A. J. Giuliano, H. C. Breiter, J. M. Goldstein, W. S. Kremen, H. W. Thermenos, R. Toomey, W. S. Stone, M. T. Tsuang, et al. “Auditory working memory impairments in individuals at familial high risk for schizophrenia.” Neuropsychology 26.3 (2012), p. 288.
    OpenUrl
  62. [62].↵
    F. Knolle, S. S. Arumugham, R. A. Barker, M. W. Chee, A. Justicia, N. Kamble, J. Lee, S. Liu, A. Lenka, S. J. Lewis, et al. “A multicentre study on grey matter morphometric biomarkers for classifying early schizophrenia and parkinson’s disease psychosis”. npj Parkinson’s Disease 9.1 (2023), p. 87.
    OpenUrl
  63. [63.].↵
    C. Sperber. “The use of principal component and factor analysis to measure fun-damental cognitive processes in neuropsychological data”. bioRxiv (2022). doi:
  64. [64].↵
    K. Majumdar. “Human scalp EEG processing: Various soft computing approaches”. Applied Soft Computing 11.8 (2011), pp. 4433–4447. doi:.
    OpenUrl
  65. [65].↵
    J. C. Bezdek, R. Ehrlich, and W. Full. “FCM: The fuzzy c-means clustering algo-rithm”. Computers & Geosciences 10.2–3 (1984), pp. 191–203. doi:.
    OpenUrl
  66. [66].↵
    M. L. D. Dias. fuzzy-c-means: An implementation of Fuzzy C-means clustering algo-rithm. May 2019. doi: 10.5281/zenodo.3066222.
    OpenUrlCrossRef
  67. [67].↵
    S. Kim. “ppcor: an R package for a fast calculation to semi-partial correlation coeffi-cients”. Communications for statistical applications and methods 22.6 (2015), p. 665.
    OpenUrl
  68. [68].↵
    R Core Team. R: A Language and Environment for Statistical Computing. ISBN 3-900051-07-0. R Foundation for Statistical Computing. Vienna, Austria, 2013.
  69. [69].↵
    J. L. Ji, M. Helmer, C. Fonteneau, J. B. Burt, Z. Tamayo, J. Dem̌sar, B. D. Adkinson, A. Savíc, K. H. Preller, F. Moujaes, F. X. Vollenweider, W. J. Martin, G. Repov̌s, J. D. Murray, and A. Anticevic. “Mapping brain-behavior space relationships along the psychosis spectrum”. eLife 10 (July 2021). issn: 2050084X. doi: 10.7554/eLife.66968.
    OpenUrlCrossRef
  70. [70].↵
    K. Lewandowski, S. Sperry, B. Cohen, and D. OÖngür. “Cognitive variability in psy-chotic disorders: a cross-diagnostic cluster analysis”. Psychological medicine 44.15 (2014), pp. 3239–3248.
    OpenUrl
  71. [71].
    M. P. Reser, K. A. Allott, E. Killackey, J. Farhall, and S. M. Cotton. “Exploring cognitive heterogeneity in first-episode psychosis: What cluster analysis can reveal”. Psychiatry Research 229.3 (2015), pp. 819–827.
    OpenUrlCrossRef
  72. [72].
    J. Uren, S. M. Cotton, E. Killackey, M. M. Saling, and K. Allott. “Cognitive clusters in first-episode psychosis: Overlap with healthy controls and relationship to con-current and prospective symptoms and functioning.” Neuropsychology 31.7 (2017), p. 787.
    OpenUrl
  73. [73].
    J. J. Crouse, A. A. Moustafa, S. E. Bogaty, I. B. Hickie, and D. F. Hermens. “Par-cellating cognitive heterogeneity in early psychosis-spectrum illnesses: a cluster anal-ysis”. Schizophrenia Research 202 (2018), pp. 91–98.
    OpenUrlPubMed
  74. [74].↵
    K. Haining, R. Gajwani, J. Gross, A. I. Gumley, R. A. Ince, S. M. Lawrie, F. Schultze-Lutter, M. Schwannauer, and P. J. Uhlhaas. “Characterising cognitive heterogeneity in individuals at clinical high-risk for psychosis: a cluster analysis with clinical and functional outcome prediction”. European Archives of Psychiatry and Clinical Neu-roscience 272.3 (2022), pp. 437–448.
    OpenUrl
  75. [75].↵
    E. J. Tan, S. L. Rossell, K. L. Subotnik, J. Ventura, and K. H. Nuechterlein. “Cog-nitive heterogeneity in first-episode psychosis and its relationship with premorbid developmental adjustment”. Psychological Medicine 52.16 (2022), pp. 3885–3894.
    OpenUrl
  76. [76].↵
    A. P. Husa, J. Moilanen, G. K. Murray, R. Marttila, M. Haapea, I. Rannikko, J. H. Barnett, P. B. Jones, M. Isohanni, A. M. Remes, et al. “Lifetime antipsychotic med-ication and cognitive performance in schizophrenia at age 43 years in a general pop-ulation birth cohort”. Psychiatry research 247 (2017), pp. 130–138.
    OpenUrl
  77. [77].↵
    N. Albert, L. Randers, K. Allott, H. D. Jensen, M. Melau, C. Hjorthøj, and M. Nordentoft. “Cognitive functioning following discontinuation of antipsychotic med-ication. A naturalistic sub-group analysis from the OPUS II trial”. Psychological medicine 49.7 (2019), pp. 1138–1147.
    OpenUrl
  78. [78].↵
    Y. B. Joshi, M. L. Thomas, D. L. Braff, M. F. Green, R. C. Gur, R. E. Gur, K. H. Nuechterlein, W. S. Stone, T. A. Greenwood, L. C. Lazzeroni, et al. “Anticholiner-gic medication burden–associated cognitive impairment in schizophrenia”. American Journal of Psychiatry 178.9 (2021), pp. 838–847.
    OpenUrlPubMed
  79. [79].↵
    F. Senner, L. Hiendl, S. Bengesser, K. Adorjan, I.-G. Anghelescu, B. T. Baune, M. Budde, U. Dannlowski, D. E. Dietrich, P. Falkai, et al. “Medication adherence and cognitive performance in schizophrenia-spectrum and bipolar disorder: results from the PsyCourse Study”. Translational Psychiatry 13.1 (2023), p. 99.
    OpenUrl
  80. [80].↵
    R. Gallardo-Ruiz, B. Crespo-Facorro, E. Setíen-Suero, and D. Tordesillas-Gutierrez. “Long-term grey matter changes in first episode psychosis: a systematic review”. Psychiatry Investigation 16.5 (2019), p. 336.
    OpenUrl
  81. [81].
    A. Vita, L. De Peri, G. Deste, and E. Sacchetti. “Progressive loss of cortical gray matter in schizophrenia: a meta-analysis and meta-regression of longitudinal MRI studies”. Translational psychiatry 2.11 (2012), e190–e190.
    OpenUrl
  82. [82].↵
    P. Fusar-Poli, R. Smieskova, G. Serafini, P. Politi, and S. Borgwardt. “Neuroanatom-ical markers of genetic liability to psychosis and first episode psychosis: a voxel-wise meta-analytical comparison”. The World Journal of Biological Psychiatry 15.3 (2014), pp. 219–228.
    OpenUrlCrossRefPubMed
  83. [83].↵
    F. Knolle, S. S. Arumugham, R. A. Barker, M. W. Chee, A. Justicia, N. Kamble, J. Lee, S. Liu, A. Lenka, S. J. Lewis, et al. “Grey matter morphometric biomarkers for classifying early schizophrenia and PD psychosis: a multicentre study”. medRxiv (2022).
  84. [84].↵
    S. Iwabuchi, P. F. Liddle, and L. Palaniyappan. “Clinical utility of machine-learning approaches in schizophrenia: improving diagnostic confidence for translational neu-roimaging”. Frontiers in psychiatry 4 (2013), p. 95.
    OpenUrl
  85. [85].↵
    H. G. Schnack, M. Nieuwenhuis, N. E. van Haren, L. Abramovic, T. W. Scheewe, R. M. Brouwer, H. E. H. Pol, and R. S. Kahn. “Can structural MRI aid in clini-cal classification? A machine learning study in two independent samples of patients with schizophrenia, bipolar disorder and healthy subjects”. Neuroimage 84 (2014), pp. 299–306.
    OpenUrlCrossRefPubMedWeb of Science
  86. [86].↵
    S. Vieira, Q.-y. Gong, W. H. Pinaya, C. Scarpazza, S. Tognin, B. Crespo-Facorro, D. Tordesillas-Gutierrez, V. Ortiz-Garcıa, E. Setien-Suero, F. E. Scheepers, et al. “Us-ing machine learning and structural neuroimaging to detect first episode psychosis: reconsidering the evidence”. Schizophrenia bulletin 46.1 (2020), pp. 17–26.
    OpenUrl
  87. [87].↵
    K. Dempster, R. Norman, J. Théberge, M. Densmore, B. Schaefer, and P. Williamson. “Cognitive performance is associated with gray matter decline in first-episode psy-chosis”. Psychiatry Research: Neuroimaging 264 (2017), pp. 46–51.
    OpenUrl
  88. [88].
    E. I. Ivleva, A. S. Bidesi, M. S. Keshavan, G. D. Pearlson, S. A. Meda, D. Dodig, A. F. Moates, H. Lu, A. N. Francis, N. Tandon, et al. “Gray matter volume as an in-termediate phenotype for psychosis: Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP)”. American Journal of Psychiatry 170.11 (2013), pp. 1285– 1296.
    OpenUrlCrossRefPubMedWeb of Science
  89. [89].↵
    N. Liu, Y. Xiao, W. Zhang, B. Tang, J. Zeng, N. Hu, S. Chandan, Q. Gong, and S. Lui. “Characteristics of gray matter alterations in never-treated and treated chronic schizophrenia patients”. Translational psychiatry 10.1 (2020), p. 136.
    OpenUrl
  90. [90].↵
    D. C. Glahn, A. R. Laird, I. Ellison-Wright, S. M. Thelen, J. L. Robinson, J. L. Lancaster, E. Bullmore, and P. T. Fox. “Meta-analysis of gray matter anomalies in schizophrenia: application of anatomic likelihood estimation and network analysis”. Biological psychiatry 64.9 (2008), pp. 774–781.
    OpenUrlCrossRefPubMedWeb of Science
  91. [91].↵
    X.-X. Shan, Y.-P. Ou, P. Pan, Y.-D. Ding, J. Zhao, F. Liu, J.-D. Chen, W.-B. Guo, and J.-P. Zhao. “Increased frontal gray matter volume in individuals with prodromal psychosis”. CNS Neuroscience & Therapeutics 25.9 (2019), pp. 987–994.
    OpenUrl
  92. [92].↵
    S. V. Haijma, N. Van Haren, W. Cahn, P. C. M. Koolschijn, H. E. Hulshoff Pol, and R. S. Kahn. “Brain volumes in schizophrenia: a meta-analysis in over 18 000 subjects”. Schizophrenia bulletin 39.5 (2013), pp. 1129–1138.
    OpenUrlCrossRefPubMedWeb of Science
  93. [93].↵
    M. Rapado-Castro, M. Villar-Arenzana, J. Janssen, D. Fraguas, I. Bombin, J. Castro-Fornieles, M. Mayoral, A. Gonźalez-Pinto, E. de la Serna, M. Parellada, et al. “Fronto-parietal gray matter volume loss is associated with decreased working memory perfor-mance in adolescents with a first episode of psychosis”. Journal of Clinical Medicine 10.17 (2021), p. 3929.
    OpenUrl
  94. [94].
    M. Pićo-Pérez, R. Vieira, M. Fernández-Rodrıguez, M. A. P. De Barros, J. Radua, and P. Morgado. “Multimodal meta-analysis of structural gray matter, neurocognitive and social cognitive fMRI findings in schizophrenia patients”. Psychological Medicine 52.4 (2022), pp. 614–624.
    OpenUrl
  95. [95].↵
    R. C. Wolf, A. Höse, K. Frasch, H. Walter, and N. Vasic. “Volumetric abnormalities associated with cognitive deficits in patients with schizophrenia”. European Psychi-atry 23.8 (2008), pp. 541–548.
    OpenUrl
  96. [96].↵
    Y. Zhu, F. Y. Womer, H. Leng, M. Chang, Z. Yin, Y. Wei, Q. Zhou, S. Fu, X. Deng, J. Lv, et al. “The relationship between cognitive dysfunction and symptom dimensions across schizophrenia, bipolar disorder, and major depressive disorder”. Frontiers in psychiatry 10 (2019), p. 253.
    OpenUrl
  97. [97].↵
    N. Koutsouleris, K. Patschurek-Kliche, J. Scheuerecker, P. Decker, R. Bottlender, G. Schmitt, D. Rujescu, I. Giegling, C. Gaser, M. Reiser, et al. “Neuroanatomical corre-lates of executive dysfunction in the at-risk mental state for psychosis”. Schizophrenia research 123.2–3 (2010), pp. 160–174.
    OpenUrl
  98. [98].↵
    L. Palaniyappan, V. Balain, and P. F. Liddle. “The neuroanatomy of psychotic diathesis: A meta-analytic review”. Journal of Psychiatric Research 46.10 (2012), pp. 1249–1256. issn: 0022-3956. doi: 10.1016/j.jpsychires.
    OpenUrlCrossRefPubMed
  99. [99].↵
    M. Carbon and C. U. Correll. “Thinking and acting beyond the positive: the role of the cognitive and negative symptoms in schizophrenia”. CNS Spectrums 19.S1 (2014), pp. 35–53. doi: 10.1017/S1092852914000601.
    OpenUrlCrossRef
  100. [100].↵
    P. P. Oomen, M. J. Begemann, B. A. Brand, L. de Haan, W. Veling, S. Koops, J. van Os, F. Smit, P. R. Bakker, N. van Beveren, et al. “Longitudinal clinical and functional outcome in distinct cognitive subgroups of first-episode psychosis: a cluster analysis”. Psychological Medicine (2021), pp. 1–11.
  101. [101].↵
    M. J. Kas, B. Penninx, B. Sommer, A. Serretti, C. Arango, and H. Marston. “A quantitative approach to neuropsychiatry: The why and the how.” eng. Neuroscience and biobehavioral reviews 97 (Feb. 2019), pp. 3–9. issn: 1873-7528 (Electronic). doi: 10.1016/j.neubiorev.2017.12.008.
    OpenUrlCrossRef
  102. [102].↵
    P. F. Hitchcock, E. I. Fried, and M. J. Frank. “Computational psychiatry needs time and context”. Annual review of psychology 73 (2022), p. 243.
    OpenUrlPubMed
  103. [103].↵
    J. A. Hartigan and M. A. Wong. “A K-Means Clustering Algorithm”. Journal of the Royal Statistical Society. Series C (Applied Statistics) 28.1 (Mar. 1979), pp. 100–108. issn: 00359254. doi: 10.2307/2346830.
    OpenUrlCrossRef
  104. [104].↵
    L. Ertöz, M. Steinbach, and V. Kumar. “Finding Clusters of Different Sizes, Shapes, and Densities in Noisy, High Dimensional Data”. Proceedings. Society for Indus-trial & Applied Mathematics (SIAM), May 2003, pp. 47–58. doi: 9781611972733.5.
  105. [105].↵
    Q. J. Huys, T. V. Maia, and M. J. Frank. “Computational psychiatry as a bridge from neuroscience to clinical applications”. Nature neuroscience 19.3 (2016), pp. 404–413.
    OpenUrlCrossRefPubMed
  106. [106].↵
    A. Mkrtchian, V. Valton, and J. P. Roiser. “Reliability of Decision-Making and Re-inforcement Learning Computational Parameters”. bioRxiv (2021). doi: 2021.06.30.450026.eprint: 11/06/2021.06.30.450026.full.pdf.
  107. [107].
    J. Ashburner. “A fast diffeomorphic image registration algorithm”. Neuroimage 38.1 (2007), pp. 95–113.
    OpenUrlCrossRefPubMedWeb of Science
  108. [108].
    K. Koch, D. Rodriguez-Manrique, O. G. Rus-Oswald, D. A. Gürsel, G. Berberich, M. Kunz, and C. Zimmer. “Homogeneous grey matter patterns in patients with obsessive-compulsive disorder”. NeuroImage: Clinical 31 (2021), p. 102727.
    OpenUrl
  109. [109].
    A. Pichet Binette, J. Gonneaud, J. W. Vogel, R. La Joie, P. Rosa-Neto, D. L. Collins, J. Poirier, J. C. Breitner, S. Villeneuve, E. Vachon-Presseau, et al. “Morphomet-ric network differences in ageing versus Alzheimer’s disease dementia”. Brain 143.2 (2020), pp. 635–649.
    OpenUrlCrossRefPubMed
  110. [110].
    Y. Zeighami, M. Ulla, Y. Iturria-Medina, M. Dadar, Y. Zhang, K. M.-H. Larcher, V. Fonov, A. C. Evans, D. L. Collins, and A. Dagher. “Network structure of brain atrophy in de novo Parkinsońs disease”. Elife 4 (2015), e08440.
    OpenUrlCrossRefPubMed
  111. [111].
    M. Beckmann, H. Johansen-Berg, and M. F. Rushworth. “Connectivity-based par-cellation of human cingulate cortex and its relation to functional specialization”. Journal of Neuroscience 29.4 (2009), pp. 1175–1190.
    OpenUrlAbstract/FREE Full Text
Back to top
Posted February 12, 2024.
Download PDF

Supplementary Material

Data/Code
Email
Cognitive subgroups of affective and non-affective psychosis show differences in medication and cortico-subcortical brain networks
Katharina M Bracher, Afra Wohlschläger, Kathrin Koch, Franziska Knolle
medRxiv 2022.08.03.22278370; doi: https://doi.org/10.1101/2022.08.03.22278370
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area