Is there a symptomatic distinction between the affective psychoses and schizophrenia? A machine learning approach.

Dubiety exists over whether clinical symptoms of schizophrenia can be distinguished from affective psychosis, the assumption being that absence of a "point of rarity" indicates lack of nosological distinction, based on prior group-level analyses. Advanced machine learning techniques, using unsupervised (hierarchical clustering) and supervised (regularized logistic regression algorithm and nested-cross-validation) were applied to a dataset of 202 patients with functional psychosis (schizophrenia n = 120, affective psychosis, n = 82). Patients were initially assessed with the Present State Examination (PSE), and followed up 2.5 years later, when DSM III diagnoses were applied (independent of initial PSE). Based on PSE syndromes, unsupervised learning discriminated depressive (approximately unbiased probability, AUP = 0.92) and mania/psychosis (AUP = 0.94) clusters. The mania/psychosis cluster further split into two groups - a mania (AUP = 0.84) and a psychosis cluster (AUP = 0.88). Supervised machine learning classified schizophrenia or affective psychosis with 83.66% (95% CI = 77.83% to 88.48%) accuracy. Area under the ROC curve (AUROC) was 89.14%. True positive rate for schizophrenia was 88.24% (95%CI = 81.05-93.42%) and affective psychosis 77.11% (95%CI = 66.58-85.62). Classification accuracy and AUROC remained high when PSE syndromes corresponding to affective symptoms (those that corresponded to the depressive and mania clusters) were removed. PSE syndromes, based on clinical symptoms, therefore discriminated between schizophrenia and affective psychosis, suggesting validity to these diagnostic constructs.

Polygenic risk for schizophrenia, transition and cortical gyrification: a high-risk study.

BACKGROUND: Schizophrenia is a highly heritable disorder, linked to several structural abnormalities of the brain. More specifically, previous findings have suggested that increased gyrification in frontal and temporal regions are implicated in the pathogenesis of schizophrenia. METHODS: The current study included participants at high familial risk of schizophrenia who remained well (n = 31), who developed sub-diagnostic symptoms (n = 28) and who developed schizophrenia (n = 9) as well as healthy controls (HC) (n = 16). We first tested whether individuals at high familial risk of schizophrenia carried an increased burden of trait-associated alleles using polygenic risk score analysis. We then assessed the extent to which polygenic risk was associated with gyral folding in the frontal and temporal lobes. RESULTS: We found that individuals at high familial risk of schizophrenia who developed schizophrenia carried a significantly greater burden of risk-conferring variants for the disorder compared to those at high risk (HR) who developed sub-diagnostic symptoms or remained well and HC. Furthermore, within the HR cohort, there was a significant and positive association between schizophrenia polygenic risk score and bilateral frontal gyrification. CONCLUSIONS: These results suggest that polygenic risk for schizophrenia impacts upon early neurodevelopment to confer greater gyral folding in adulthood and an increased risk of developing the disorder.

Childhood adversity and hippocampal and amygdala volumes in a population at familial high risk of schizophrenia.

BACKGROUND: There is an established link between childhood adversity (CA) and schizophrenia. Hippocampus and amygdala abnormalities pre-date onset in those at high familial risk (fHR) of schizophrenia, but it is not clear whether these alterations are associated with CA in those at elevated risk of schizophrenia. METHODS: We examined hippocampal and amygdala volumes in those at fHR who had been referred to a social worker or the Children's Panel compared to those who had not. RESULTS: The right hippocampus and left amygdala were significantly smaller in those that had been referred to social work and Children's Panel. CONCLUSIONS: Our findings suggest that CA can influence structural changes in the brain in a cohort at fHR of schizophrenia. These findings provide further evidence that while genetic factors contribute to the structural changes found in schizophrenia, environmental factors such as CA can have a lasting impact on specific brain regions.

Childhood adversity and cortical thickness and surface area in a population at familial high risk of schizophrenia.

BACKGROUND: There is now a well-established link between childhood adversity (CA) and schizophrenia. Similar structural abnormalities to those found in schizophrenia including alterations in grey-matter volume have also been shown in those who experience CA. METHOD: We examined whether global estimates of cortical thickness or surface area were altered in those familial high-risk subjects who had been referred to a social worker or the Children's Panel compared to those who had not. RESULTS: We found that the cortical surface area of those who were referred to the Children's Panel was significantly smaller than those who had not been referred, but cortical thickness was not significantly altered. There was also an effect of social work referral on cortical surface area but not on thickness. CONCLUSIONS: Cortical surface area increases post-natally more than cortical thickness. Our findings suggest that CA can influence structural changes in the brain and it is likely to have a greater impact on cortical surface area than on cortical thickness.

Longitudinal changes in hippocampal volume in the Edinburgh High Risk Study of Schizophrenia.

Schizophrenia is associated with structural brain abnormalities that are likely to be present before disease onset. It remains unclear to what extent these represent general vulnerability indicators or are associated with the developing clinical state itself. It also remains unclear whether such state or trait alterations may be evident at any given time-point, or whether they progress over time. To investigate this, structural brain scans were acquired at two time-points (mean scan-interval 1.87years) in a cohort of young unaffected individuals at high familial risk of schizophrenia (baseline, n=142; follow-up, n=64) and healthy controls (baseline, n=36; follow-up, n=18). Sub-cortical reconstructions of the hippocampus and amygdala were generated using the longitudinal pipeline available with Freesurfer. The high risk cohort was subdivided into individuals that remained well during the study (HR[well], baseline, n=68; follow-up, n=30), transient and/or partial symptoms that were insufficient to support a formal diagnosis (HR[symp], baseline, n=57; follow-up, n=26) and individuals that subsequently developed schizophrenia according to ICD-10 criteria (HR[ill], baseline, n=17; follow-up, n=8). Longitudinal change in the hippocampus and amygdala was compared, focusing first on overall differences between high-risk individuals and controls and then on sub-group differences within the high-risk cohort. We found a significantly altered developmental trajectory for all high risk individuals compared to controls, with controls showing a significant increase in hippocampal volume over time compared to those at high risk. We did not find evidence of altered longitudinal trajectories based on clinical outcome within the high risk cohort. These results suggest that an altered developmental trajectory of hippocampal volume is associated with a general familial predisposition to develop schizophrenia, as this alteration was not related to subsequent clinical outcome.

Determinants of adult functional outcome in adolescents receiving special educational assistance.

BACKGROUND: This study investigates the role of IQ, autistic traits and challenging behaviours in affecting adult outcomes among adolescents who receive special educational assistance. METHODS: A total of 58 participants were recruited from an ongoing longitudinal study. All received assessments of IQ, behavioural patterns (using the Childhood Behaviour Checklist - CBCL) and autistic traits (using the Social Communication Questionnaire - SCQ) during adolescence and were followed up 6 years later (at a mean age of 22 years) using the World Health Organization Disability Assessment Schedule II (WHO-DAS II) to assess functional outcome. RESULTS: A significant positive relationship was found between CBCL score and WHO-DAS II score (ß = 0.511, P = 0.001). IQ score showed a negative relationship with total WHO-DAS II score (ß = -0.247, P = 0.04). SCQ score was not found to significantly influence total WHO-DAS II score (ß = -0.028, P = 0.84). CONCLUSIONS: Although the role of global intellectual ability is important, these results stress the highly predictive value of adolescent behaviours on functional outcomes in adult life among young adults receiving special educational assistance.

The proportion of tumor-stroma as a strong prognosticator for stage II and III colon cancer patients: validation in the VICTOR trial.

BACKGROUND: The intra-tumor stroma percentage in colon cancer (CC) patients has previously been reported by our group as a strong independent prognostic parameter. Patients with a high stroma percentage within the primary tumor have a poor prognosis. PATIENTS AND METHODS: Tissue samples from the most invasive part of the primary tumor of 710 patients (52% Stage II, 48% Stage III) participating in the VICTOR trial were analyzed for their tumor-stroma percentage. Stroma-high (>50%) and stroma-low (≤50%) groups were evaluated with respect to survival times. RESULTS: Overall and disease-free survival times (OS and DFS) were significantly lower in the stroma-high group (OS P<0.0001, hazard ratio (HR)=1.96; DFS P<0.0001, HR=2.15). The 5-year OS was 69.0% versus 83.4% and DFS 58.6% versus 77.3% for stroma-high versus stroma-low patients. CONCLUSION: This study confirms the intra-tumor stroma ratio as a prognostic factor. This parameter could be a valuable and low cost addition to the TNM status and next to current high-risk parameters such as microsatellite instability status used in routine pathology reporting. When adding the stroma-parameter to the ASCO criteria, the rate of 'undertreated' patients dropped from 5.9% to 4.3%, the 'overtreated' increased with 6.8% but the correctly classified increased with an additional 14%.

Tensor-based morphometry of cannabis use on brain structure in individuals at elevated genetic risk of schizophrenia.

BACKGROUND: Schizophrenia is associated with various brain structural abnormalities, including reduced volume of the hippocampi, prefrontal lobes and thalami. Cannabis use increases the risk of schizophrenia but reports of brain structural abnormalities in the cannabis-using population have not been consistent. We used automated image analysis to compare brain structural changes over time in people at elevated risk of schizophrenia for familial reasons who did and did not use cannabis. METHOD: Magnetic resonance imaging (MRI) scans were obtained from subjects at high familial risk of schizophrenia at entry to the Edinburgh High Risk Study (EHRS) and approximately 2 years later. Differential grey matter (GM) loss in those exposed (n=23) and not exposed to cannabis (n=32) in the intervening period was compared using tensor-based morphometry (TBM). RESULTS: Cannabis exposure was associated with significantly greater loss of right anterior hippocampal (pcorrected=0.029, t=3.88) and left superior frontal lobe GM (pcorrected=0.026, t=4.68). The former finding remained significant even after the exclusion of individuals who had used other drugs during the inter-scan interval. CONCLUSIONS: Using an automated analysis of longitudinal data, we demonstrate an association between cannabis use and GM loss in currently well people at familial risk of developing schizophrenia. This observation may be important in understanding the link between cannabis exposure and the subsequent development of schizophrenia.

Anterior cingulate morphology in people at genetic high-risk of schizophrenia.

BACKGROUND: Morphological abnormalities of the anterior cingulate (AC) occur in patients with schizophrenia and in symptomatic high-risk individuals, and may be predictive of subsequent psychosis. We investigated AC sulcal morphology in the Edinburgh High Risk Study cohort to see if such abnormalities are evident and predict psychosis in patients' relatives. We also investigated the association of the cingulate sulcus (CS) and paracingulate sulcus (PCS) variants with intelligence quotient (IQ). PATIENTS AND METHODS: We compared cingulate and paracingulate sulcal anatomy, using reliable standardised measurements, blind to group membership, in those at high genetic risk (n=146), first episode patients (n=34) and healthy controls (n=36); and compared high-risk subjects who did (n=17) or did not develop schizophrenia. RESULTS: Interruptions of the cingulate sulcus were more common in high-risk individuals and in those with schizophrenia, in both hemispheres, compared to controls. When separated by gender, these results were only present in males in the left hemisphere and only in females in the right hemisphere. A well-formed paracingulate sulcus was less common in high-risk participants and patients with schizophrenia, compared to controls; but this association was only present in males. These morphological variants of the paracingulate sulcus and the continuous cingulate sulcus were also associated with the higher IQ in male high-risk individuals. CONCLUSIONS: An interrupted cingulate sulcus pattern in both males and females and paracingulate morphology in males are associated with increased genetic risk of schizophrenia. Associations between cingulate and paracingulate morphology and premorbid IQ scores provide evidence that intellectual ability could be related to particular cytoarchitectural brain regions. Given that these sulci develop in early fetal life, such findings presumably reflect early neurodevelopmental abnormalities of genetic origin, although environmental effects and interactions cannot be ruled out.

Orbitofrontal morphology in people at high risk of developing schizophrenia.

BACKGROUND: Abnormalities of orbitofrontal cortex (OFC) sulcogyral patterns have been reported in schizophrenia, but it is not known if these predate psychosis. METHODS: Hundred and forty-six subjects at high genetic risk of schizophrenia, 34 first episode of schizophrenia patients (SZ) and 36 healthy controls were scanned and clinically assessed. Utilising the classification system proposed by Chiavaras, we categorised OFC patterns and compared their distribution between the groups, as well as between those high risk subjects who did, and did not develop schizophrenia. The relationship between OFC pattern and schizotypy was explored in high risk subjects. RESULTS: We refined Chiavaras' classification system, with the identification of a previously unreported variant of OFC surface structure. There were significant differences in distribution of OFC patterns between high risk subjects who did or did not develop schizophrenia as well as between the first episode of schizophrenia group and healthy controls. Within the high risk group, possession of OFC Type III was associated with higher ratings on the Structured Inventory for Schizotypy (SIS) psychotic factor. CONCLUSIONS: Our results suggest that OFC Type III is associated with psychotic features before the development of schizophrenia. Characterisation of OFC morphology may have a role in the identification of those at greatest risk of developing schizophrenia.

Deficits in facial, body movement and vocal emotional processing in autism spectrum disorders.

BACKGROUND: Previous behavioural and neuroimaging studies of emotion processing in autistic spectrum disorder (ASD) have focused on the use of facial stimuli. To date, however, no studies have examined emotion processing in autism across a broad range of social signals. METHOD: This study addressed this issue by investigating emotion processing in a group of 23 adults with ASD and 23 age- and gender-matched controls. Recognition of basic emotions ('happiness', 'sadness', 'anger', disgust' and 'fear') was assessed from facial, body movement and vocal stimuli. The ability to make social judgements (such as approachability) from facial stimuli was also investigated. RESULTS: Significant deficits in emotion recognition were found in the ASD group relative to the control group across all stimulus domains (faces, body movements and voices). These deficits were seen across a range of emotions. The ASD group were also impaired in making social judgements compared to the control group and this correlated with impairments in basic emotion recognition. CONCLUSIONS: This study demonstrates that there are significant and broad-ranging deficits in emotion processing in ASD present across a range of stimulus domains and in the auditory and visual modality; they cannot therefore be accounted for simply in terms of impairments in face processing or in the visual modality alone. These results identify a core deficit affecting the processing of a wide range of emotional information in ASD, which contributes to the impairments in social function seen in people with this condition.

Amygdala volume in a population with special educational needs at high risk of schizophrenia.

BACKGROUND: The mildly learning disabled population has a three-fold elevated risk for schizophrenia. It has been proposed that in some individuals this cognitive limitation is a pre-psychotic manifestation of early onset schizophrenia. We examined clinical and neuroanatomical measures of a putative extended phenotype of schizophrenia in an adolescent population receiving special educational assistance. We predicted that people with intellectual impairment and schizotypal features would exhibit amygdala volume reduction as one of the neuroanatomical abnormalities associated with schizophrenia. METHOD: Assessment by clinical interview, neuropsychological assessment and magnetic resonance imaging scanning was carried out in 28 intellectually impaired individuals identified as being at elevated risk of schizophrenia due to the presence of schizotypal traits, 39 intellectually impaired controls and 29 non-intellectually impaired controls. Amygdala volume was compared in these three groups and the relationship between symptomatology and amygdala volume investigated. RESULTS: Right amygdala volume was significantly increased in the elevated risk group compared with the intellectually impaired controls (p=0.05). A significant negative correlation was seen between left amygdala volume and severity of negative symptoms within this group (p<0.05), but not in either control group. CONCLUSIONS: Intellectually impaired subjects judged to be at elevated risk of schizophrenia on the basis of clinical assessment exhibit structural imaging findings which distinguish them from the generality of learning disabled subjects. Within this population reduced amygdala volume may be associated with negative-type symptoms and be part of an extended phenotype that reflects particularly elevated risk and/or early manifestations of the development of psychosis.

Hippocampal function in schizophrenia and bipolar disorder.

BACKGROUND: The hippocampus plays a central role in memory formation. There is considerable evidence of abnormalities in hippocampal structure and function in schizophrenia, which may differentiate it from bipolar disorder. However, no previous studies have compared hippocampal activation in schizophrenia and bipolar disorder directly. METHOD: Fifteen patients with schizophrenia, 14 patients with bipolar disorder and 14 healthy comparison subjects took part in the study. Subjects performed a face-name pair memory task during functional magnetic resonance imaging (fMRI). Differences in blood oxygen level-dependent (BOLD) activity were determined during encoding and retrieval of the face-name pairs. RESULTS: The patient groups showed significant differences in hippocampal and prefrontal cortex (PFC) activation during face-name pair learning. During encoding, patients with schizophrenia showed decreased anterior hippocampal activation relative to subjects with bipolar disorder, whereas patients with bipolar disorder showed decreased dorsal PFC activation relative to patients with schizophrenia. During retrieval, patients with schizophrenia showed greater activation of the dorsal PFC than patients with bipolar disorder. Patients with schizophrenia also differed from healthy control subjects in the activation of several brain regions, showing impaired superior temporal cortex activation during encoding and greater dorsal PFC activation during retrieval. These effects were evident despite matched task performance. CONCLUSIONS: Patients with schizophrenia showed deficits in hippocampal activation during a memory task relative to patients with bipolar disorder. The disorders were further distinguished by differences in PFC activation. The results demonstrate that these disorders can distinguished at a group level using non-invasive neuroimaging.

A common neural system mediating two different forms of social judgement.

BACKGROUND: A wide range of neuropsychiatric conditions, including schizophrenia and autistic spectrum disorder (ASD), are associated with impairments in social function. Previous studies have shown that individuals with schizophrenia and ASD have deficits in making a wide range of social judgements from faces, including decisions related to threat (such as judgements of approachability) and decisions not related to physical threat (such as judgements of intelligence). We have investigated healthy control participants to see whether there is a common neural system activated during such social decisions, on the basis that deficits in this system may contribute to the impairments seen in these disorders. METHOD: We investigated the neural basis of social decision making during judgements of approachability and intelligence from faces in 24 healthy participants using functional magnetic resonance imaging (fMRI). We used conjunction analysis to identify common brain regions activated during both tasks. RESULTS: Activation of the amygdala, medial prefrontal cortex, inferior prefrontal cortex and cerebellum was seen during performance of both social tasks, compared to simple gender judgements from the same stimuli. Task-specific activations were present in the dorsolateral prefrontal cortex in the intelligence task and in the inferior and middle temporal cortex in the approachability task. CONCLUSIONS: The present study identified a common network of brain regions activated during the performance of two different forms of social judgement from faces. Dysfunction of this network is likely to contribute to the broad-ranging deficits in social function seen in psychiatric disorders such as schizophrenia and ASD.

Specific cognitive deficits in a group at genetic high risk of schizophrenia.

BACKGROUND: Neuropsychological deficits in schizophrenia patients and their relatives have been thought to represent possible genetic vulnerability markers or endophenotypes of the disorder. The present study describes results from the Edinburgh High Risk Study of computerized testing using the Cambridge Neuropsychological Test Automated Battery (CANTAB) on a group at genetic high risk (HR) of schizophrenia and a control group. METHOD: A total of 97 HR and 25 control participants were assessed on three tests from the CANTAB - spatial span, spatial working memory, and Stockings of Cambridge. Analyses of covariance were used to compare the HR and control groups on the main outcome measures whilst controlling for intelligence quotient (IQ). Subsequent analysis examined the effects of the presence of symptoms on group differences. RESULTS: HR participants had significantly reduced spatial memory capacity [F(1, 118)=4.06, p=0.046] and significantly reduced planning processing speed [F(1, 116)=4.16, p=0.044] compared with controls even after controlling for general intelligence (IQ). Although HR individuals made more errors and showed poorer problem-solving and strategy performance compared with controls, these differences were not significant after controlling for IQ. Subsequent analysis indicated that the presence or absence of psychotic symptoms in the HR group did not influence these specific cognitive deficits. CONCLUSIONS: Spatial memory capacity and planning processing speed may represent cognitive endophenotypes characterising the genetic predisposition to schizophrenia in this HR group.

fMRI changes over time and reproducibility in unmedicated subjects at high genetic risk of schizophrenia.

BACKGROUND: Functional brain abnormalities have been repeatedly demonstrated in schizophrenia but there is little data concerning their progression. For such studies to have credibility it is first important to establish the reproducibility of functional imaging techniques. The current study aimed to examine these factors in healthy controls and in unmedicated subjects at high genetic risk of the disorder: (i) to examine the reproducibility of task-related activation patterns, (ii) to determine if there were any progressive functional changes in high-risk subjects versus controls reflecting inheritance of the schizophrenic trait, and (iii) to examine changes over time in relation to fluctuating positive psychotic symptoms (i.e. state effects). METHOD: Subjects were scanned performing the Hayling sentence completion test on two occasions 18 months apart. Changes in activation were examined in controls and high-risk subjects (n=16, n=63). Reproducibility was assessed for controls and high-risk subjects who remained asymptomatic at both time points (n=16, n=32). RESULTS: Intra-class correlation values indicated good agreement between scanning sessions. No significant differences over time were seen between the high-risk and control group; however, comparison of high-risk subjects who developed symptoms versus those who remained asymptomatic revealed activation increases in the left middle temporal gyrus (p=0.026). CONCLUSIONS: The current results suggest that functional changes over time occur in the lateral temporal cortex as high genetic risk subjects become symptomatic, further, they indicate the usefulness of functional imaging tools for investigating progressive changes associated with state and trait effects in schizophrenia.

Set shifting and reversal learning in patients with bipolar disorder or schizophrenia.

BACKGROUND: Bipolar disorder and schizophrenia have both been associated with deficits in extra-dimensional set shifting (EDS). Deficits in reversal learning (RL) have also been shown in schizophrenia but not in bipolar disorder. This study sought to assess the specificity of these findings in a direct comparison of clinically stable patients with each disorder. METHOD: The intra-dimensional/extra-dimensional (IDED) set-shifting task, part of the Cambridge Neuropsychological Test Automated Battery (CANTAB), was administered to 30 patients with schizophrenia, 47 with bipolar disorder and a group of 44 unaffected controls. EDS and RL errors were compared between the groups and related to measures of current and past psychiatric symptoms and medication. RESULTS: Both groups of patients with schizophrenia or bipolar disorder made more EDS and RL errors than controls. Neither measure separated the two disorders, even when the analysis was restricted to euthymic patients. No relationship was found with prescribed medication. CONCLUSIONS: Patients with bipolar disorder or schizophrenia show common deficits in EDS and RL. These deficits do not seem to be attributable to current symptoms and are consistent with disrupted networks involving the ventral prefrontal cortex.