Towards the identification of imaging biomarkers in schizophrenia, using multivariate pattern classification at a single-subject level

Individual prediction of long-term outcome in adolescents at ultra-high risk for psychosis: Applying machine learning techniques to brain imaging data

An important focus of studies of individuals at ultra-high risk (UHR) for psychosis has been to identify biomarkers to predict which individuals will transition to psychosis. However, the majority of individuals will prove to be resilient and go on to experience remission of their symptoms and function well. The aim of this study was to investigate the possibility of using structural MRI measures collected in UHR adolescents at baseline to quantitatively predict their long-term clinical outcome and level of functioning. We included 64 UHR individuals and 62 typically developing adolescents (12-18 years old at recruitment). At six-year follow-up, we determined resilience for 43 UHR individuals. Support Vector Regression analyses were performed to predict long-term functional and clinical outcome from baseline MRI measures on a continuous scale, instead of the more typical binary classification. This led to predictive correlations of baseline MR measures with level of functioning, and negative and disorganization symptoms. The highest correlation (r = 0.42) was found between baseline subcortical volumes and long-term level of functioning. In conclusion, our results show that structural MRI data can be used to quantitatively predict long-term functional and clinical outcome in UHR individuals with medium effect size, suggesting that there may be scope for predicting outcome at the individual level. Moreover, we recommend classifying individual outcome on a continuous scale, enabling the assessment of different functional and clinical scales separately without the need to set a threshold. Hum Brain Mapp 38:704-714, 2017. © 2016 Wiley Periodicals, Inc.

Classifying Schizophrenia Using Multimodal Multivariate Pattern Recognition Analysis: Evaluating the Impact of Individual Clinical Profiles on the Neurodiagnostic Performance

Previous studies have shown that structural brain changes are among the best-studied candidate markers for schizophrenia (SZ) along with functional connectivity (FC) alterations of resting-state (RS) patterns. This study aimed to investigate effects of clinical and sociodemographic variables on the classification by applying multivariate pattern analysis (MVPA) to both gray matter (GM) volume and FC measures in patients with SZ and healthy controls (HC). RS and structural magnetic resonance imaging data (sMRI) from 74 HC and 71 SZ patients were obtained from a Mind Research Network COBRE dataset available via COINS (http://coins.mrn.org/dx). We used a MVPA framework using support-vector machines embedded in a repeated, nested cross-validation to generate a multi-modal diagnostic system and evaluate its generalizability. The dependence of neurodiagnostic performance on clinical and sociodemographic variables was evaluated. The RS classifier showed a slightly higher accuracy (70.5%) compared to the structural classifier (69.7%). The combination of sMRI and RS outperformed single MRI modalities classification by reaching 75% accuracy. The RS based moderator analysis revealed that the neurodiagnostic performance was driven by older SZ patients with an earlier illness onset and more pronounced negative symptoms. In contrast, there was no linear relationship between the clinical variables and neuroanatomically derived group membership measures. This study achieved higher accuracy distinguishing HC from SZ patients by fusing 2 imaging modalities. In addition the results of RS based moderator analysis showed that age of patients, as well as their age at the illness onset were the most important clinical features.

Dynamic cerebral reorganization in the pathophysiology of schizophrenia: a MRI-derived cortical thickness study

BACKGROUND

A structural neuroanatomical change indicating a reduction in brain tissue is a notable feature of schizophrenia. Several pathophysiological processes such as aberrant cortical maturation, progressive tissue loss and compensatory tissue increase could contribute to the structural changes seen in schizophrenia.

METHOD

We studied cortical thickness using surface-based morphometry in 98 clinically stable patients with schizophrenia and 83 controls. Using a pattern classification approach, we studied whether the features that discriminate patients from controls vary across the different stages of the illness. Using a covariance analysis, we also investigated if concurrent increases accompany decreases in cortical thickness.

RESULTS

Very high levels of accuracy (96.3%), specificity (98.8%) and sensitivity (88%) were noted when classifying patients with <2 years of illness from controls. Within the patient group, reduced thickness was consistently accompanied by increased thickness in distributed brain regions. A pattern of cortical amelioration or normalization (i.e. reduced deviation from controls) was noted with increasing illness duration. While temporo-limbic and fronto-parietal regions showed reduced thickness, the occipital cortex showed increased thickness, especially in those with a long-standing illness.

CONCLUSION

A compensatory remodelling process might contribute to the cortical thickness variations in different stages of schizophrenia. Subtle cerebral reorganization reflecting the inherent plasticity of brain may occur concomitantly with processes contributing to tissue reduction in adult patients with schizophrenia.

The genetics of drug efficacy: opportunities and challenges

Lack of sufficient efficacy is the most common cause of attrition in late-phase drug development. It has long been envisioned that genetics could drive stratified drug development by identifying those patient subgroups that are most likely to respond. However, this vision has not been realized as only a small proportion of drugs have been found to have germline genetic predictors of efficacy with clinically meaningful effects, and so far all but one were found after drug approval. With the exception of oncology, systematic application of efficacy pharmacogenetics has not been integrated into drug discovery and development across the industry. Here, we argue for routine, early and cumulative screening for genetic predictors of efficacy, as an integrated component of clinical trial analysis. Such a strategy would identify clinically relevant predictors that may exist at the earliest possible opportunity, allow these predictors to be integrated into subsequent clinical development and provide mechanistic insights into drug disposition and patient-specific factors that influence response, therefore paving the way towards more personalized medicine.

Multivariate pattern analysis strategies in detection of remitted major depressive disorder using resting state functional connectivity

Understanding abnormal resting-state functional connectivity of distributed brain networks may aid in probing and targeting mechanisms involved in major depressive disorder (MDD). To date, few studies have used resting state functional magnetic resonance imaging (rs-fMRI) to attempt to discriminate individuals with MDD from individuals without MDD, and to our knowledge no investigations have examined a remitted (r) population. In this study, we examined the efficiency of support vector machine (SVM) classifier to successfully discriminate rMDD individuals from healthy controls (HCs) in a narrow early-adult age range. We empirically evaluated four feature selection methods including multivariate Least Absolute Shrinkage and Selection Operator (LASSO) and Elastic Net feature selection algorithms. Our results showed that SVM classification with Elastic Net feature selection achieved the highest classification accuracy of 76.1% (sensitivity of 81.5% and specificity of 68.9%) by leave-one-out cross-validation across subjects from a dataset consisting of 38 rMDD individuals and 29 healthy controls. The highest discriminating functional connections were between the left amygdala, left posterior cingulate cortex, bilateral dorso-lateral prefrontal cortex, and right ventral striatum. These appear to be key nodes in the etiopathophysiology of MDD, within and between default mode, salience and cognitive control networks. This technique demonstrates early promise for using rs-fMRI connectivity as a putative neurobiological marker capable of distinguishing between individuals with and without rMDD. These methods may be extended to periods of risk prior to illness onset, thereby allowing for earlier diagnosis, prevention, and intervention.

Detecting Neuroimaging Biomarkers for Psychiatric Disorders: Sample Size Matters

In a recent review, it was suggested that much larger cohorts are needed to prove the diagnostic value of neuroimaging biomarkers in psychiatry. While within a sample, an increase of diagnostic accuracy of schizophrenia (SZ) with number of subjects (N) has been shown, the relationship between N and accuracy is completely different between studies. Using data from a recent meta-analysis of machine learning (ML) in imaging SZ, we found that while low-N studies can reach 90% and higher accuracy, above N/2 = 50 the maximum accuracy achieved steadily drops to below 70% for N/2 > 150. We investigate the role N plays in the wide variability in accuracy results in SZ studies (63-97%). We hypothesize that the underlying cause of the decrease in accuracy with increasing N is sample heterogeneity. While smaller studies more easily include a homogeneous group of subjects (strict inclusion criteria are easily met; subjects live close to study site), larger studies inevitably need to relax the criteria/recruit from large geographic areas. A SZ prediction model based on a heterogeneous group of patients with presumably a heterogeneous pattern of structural or functional brain changes will not be able to capture the whole variety of changes, thus being limited to patterns shared by most patients. In addition to heterogeneity (sample size), we investigate other factors influencing accuracy and introduce a ML effect size. We derive a simple model of how the different factors, such as sample heterogeneity and study setup determine this ML effect size, and explain the variation in prediction accuracies found from the literature, both in cross-validation and independent sample testing. From this, we argue that smaller-N studies may reach high prediction accuracy at the cost of lower generalizability to other samples. Higher-N studies, on the other hand, will have more generalization power, but at the cost of lower accuracy. In conclusion, when comparing results from different ML studies, the sample sizes should be taken into account. To assess the generalizability of the models, validation (by direct application) of the prediction models should be tested in independent samples. The prediction of more complex measures such as outcome, which are expected to have an underlying pattern of more subtle brain abnormalities (lower effect size), will require large samples.

Retinal nerve fiber layer and macular thickness in patients with schizophrenia: Influence of recent illness episodes

Optical coherence tomography (OCT) has been recently used to investigate neuropsychiatric disorders. We aimed to study retinal OCT measures of patients with schizophrenia with respect to healthy controls, and to evaluate possible differences between recent illness episode (RIE) and non-recent illness episode (NRIE) patients. Thirty schizophrenia patients were classified as RIE (n=10) or NRIE (n=20), and compared with 30 matched controls. Statistical analyses included linear mixed-effects models to study the association between OCT measures and group membership. Multivariate models were used to control for potential confounders. In the adjusted linear mixed-effects regression model, patients had a significantly thinner retinal nerve fiber layer (RNFL) in overall measurements, and in the nasal, superior and inferior quadrants. Macular inner ring thickness and macular volume were also significantly smaller in patients than controls. Compared with controls, in the adjusted model only NRIE (but not RIE) patients had significantly reduced RNFL overall measures, superior RNFL, nasal RNFL, macular volume, and macular inner ring thickness. No significant correlation was found between illness duration and retinal measurements after controlling for age. In conclusion, retinal parameters observed using OCT in schizophrenia patients could be related to clinical status and merit attention as potential state biomarkers of the disorder.

The cortical thickness phenotype of individuals with DISC1 translocation resembles schizophrenia

BACKGROUND. The disrupted in schizophrenia 1 (DISC1) gene locus was originally identified in a Scottish pedigree with a high incidence of psychiatric disorders that is associated with a balanced t(1;11)(q42.1;q14.3) chromosomal translocation. Here, we investigated whether members of this family carrying the t(1;11)(q42.1;q14.3) translocation have a common brain-related phenotype and whether this phenotype is similar to that observed in schizophrenia (SCZ), using multivariate pattern recognition techniques.

METHODS. We measured cortical thickness, cortical surface area, subcortical volumes, and regional cerebral blood flow (rCBF) in healthy controls (HC) (n = 24), patients diagnosed with SCZ (n = 24), patients diagnosed with bipolar disorder (BP) (n = 19), and members of the original Scottish family (n = 30) who were either carriers (T+) or noncarriers (T–) of the DISC1 translocation. Binary classification models were developed to assess the differences and similarities across groups.

RESULTS. Based on cortical thickness, 72% of the T– group were assigned to the HC group, 83% of the T+ group were assigned to the SCZ group, and 45% of the BP group were classified as belonging to the SCZ group, suggesting high specificity of this measurement in predicting brain-related phenotypes. Shared brain-related phenotypes between SCZ and T+ individuals were found for cortical thickness only. Finally, a classification accuracy of 73% was achieved when directly comparing the pattern of cortical thickness of T+ and T– individuals.

CONCLUSION. Together, the results of this study suggest that the DISC1 translocation may increase the risk of psychiatric disorders in this pedigree by affecting neurostructural phenotypes such as cortical thickness.

FUNDING. This work was supported by the National Health Service Research Scotland, the Scottish Translational Medicine Research Collaboration, the Innovative Medicines Initiative (IMI), the Engineering and Physical Sciences Research Council (EPSRC), The Wellcome Trust, the National Institute of Health Research (NIHR), and Pfizer.

Overdiagnosis in the era of neuropsychiatric imaging

New guidelines proposed by the National Institute of Mental Health are intended to transform the management of patients with psychiatric disorders. It is anticipated that neuroimaging and other biomarkers will play a more prominent role in diagnosis and prognosis, especially in the prodromal phase of illness. Earlier treatment of psychiatric disorders has the potential to improve outcomes significantly. However, diagnosis in the absence of symptoms can lead to overdiagnosis. Overdiagnosis is a problem in many fields of medicine but could pose additional problems in psychiatry because of the stigmatization that often accompanies a diagnosis of mental illness. This review discusses the magnetic resonance imaging methods that hold the most promise for evaluating neuropsychiatric disorders, the likelihood that they could lead to overdiagnosis, and opportunities to minimize the impact of overdiagnosis in psychiatric disorders.

Development of a blood-based molecular biomarker test for identification of schizophrenia before disease onset

Recent research efforts have progressively shifted towards preventative psychiatry and prognostic identification of individuals before disease onset. We describe the development of a serum biomarker test for the identification of individuals at risk of developing schizophrenia based on multiplex immunoassay profiling analysis of 957 serum samples. First, we conducted a meta-analysis of five independent cohorts of 127 first-onset drug-naive schizophrenia patients and 204 controls. Using least absolute shrinkage and selection operator regression, we identified an optimal panel of 26 biomarkers that best discriminated patients and controls. Next, we successfully validated this biomarker panel using two independent validation cohorts of 93 patients and 88 controls, which yielded an area under the curve (AUC) of 0.97 (0.95-1.00) for schizophrenia detection. Finally, we tested its predictive performance for identifying patients before onset of psychosis using two cohorts of 445 pre-onset or at-risk individuals. The predictive performance achieved by the panel was excellent for identifying USA military personnel (AUC: 0.90 (0.86-0.95)) and help-seeking prodromal individuals (AUC: 0.82 (0.71-0.93)) who developed schizophrenia up to 2 years after baseline sampling. The performance increased further using the latter cohort following the incorporation of CAARMS (Comprehensive Assessment of At-Risk Mental State) positive subscale symptom scores into the model (AUC: 0.90 (0.82-0.98)). The current findings may represent the first successful step towards a test that could address the clinical need for early intervention in psychiatry. Further developments of a combined molecular/symptom-based test will aid clinicians in the identification of vulnerable patients early in the disease process, allowing more effective therapeutic intervention before overt disease onset.

Detecting neuroimaging biomarkers for schizophrenia: a meta-analysis of multivariate pattern recognition studies

Multivariate pattern recognition approaches have recently facilitated the search for reliable neuroimaging-based biomarkers in psychiatric disorders such as schizophrenia. By taking into account the multivariate nature of brain functional and structural changes as well as their distributed localization across the whole brain, they overcome drawbacks of traditional univariate approaches. To evaluate the overall reliability of neuroimaging-based biomarkers, we conducted a comprehensive literature search to identify all studies that used multivariate pattern recognition to identify patterns of brain alterations that differentiate patients with schizophrenia from healthy controls. A bivariate random-effects meta-analytic model was implemented to investigate the sensitivity and specificity across studies as well as to assess the robustness to potentially confounding variables. In the total sample of n=38 studies (1602 patients and 1637 healthy controls), patients were differentiated from controls with a sensitivity of 80.3% (95% CI: 76.7-83.5%) and a specificity of 80.3% (95% CI: 76.9-83.3%). Analysis of neuroimaging modality indicated higher sensitivity (84.46%, 95% CI: 79.9-88.2%) and similar specificity (76.9%, 95% CI: 71.3-81.6%) of rsfMRI studies as compared with structural MRI studies (sensitivity: 76.4%, 95% CI: 71.9-80.4%, specificity of 79.0%, 95% CI: 74.6-82.8%). Moderator analysis identified significant effects of age (p=0.029), imaging modality (p=0.019), and disease stage (p=0.025) on sensitivity as well as of positive-to-negative symptom ratio (p=0.022) and antipsychotic medication (p=0.016) on specificity. Our results underline the utility of multivariate pattern recognition approaches for the identification of reliable neuroimaging-based biomarkers. Despite the clinical heterogeneity of the schizophrenia phenotype, brain functional and structural alterations differentiate schizophrenic patients from healthy controls with 80% sensitivity and specificity.

Deep neural network with weight sparsity control and pre-training extracts hierarchical features and enhances classification performance: Evidence from whole-brain resting-state functional connectivity patterns of schizophrenia

Functional connectivity (FC) patterns obtained from resting-state functional magnetic resonance imaging data are commonly employed to study neuropsychiatric conditions by using pattern classifiers such as the support vector machine (SVM). Meanwhile, a deep neural network (DNN) with multiple hidden layers has shown its ability to systematically extract lower-to-higher level information of image and speech data from lower-to-higher hidden layers, markedly enhancing classification accuracy. The objective of this study was to adopt the DNN for whole-brain resting-state FC pattern classification of schizophrenia (SZ) patients vs. healthy controls (HCs) and identification of aberrant FC patterns associated with SZ. We hypothesized that the lower-to-higher level features learned via the DNN would significantly enhance the classification accuracy, and proposed an adaptive learning algorithm to explicitly control the weight sparsity in each hidden layer via L1-norm regularization. Furthermore, the weights were initialized via stacked autoencoder based pre-training to further improve the classification performance. Classification accuracy was systematically evaluated as a function of (1) the number of hidden layers/nodes, (2) the use of L1-norm regularization, (3) the use of the pre-training, (4) the use of framewise displacement (FD) removal, and (5) the use of anatomical/functional parcellation. Using FC patterns from anatomically parcellated regions without FD removal, an error rate of 14.2% was achieved by employing three hidden layers and 50 hidden nodes with both L1-norm regularization and pre-training, which was substantially lower than the error rate from the SVM (22.3%). Moreover, the trained DNN weights (i.e., the learned features) were found to represent the hierarchical organization of aberrant FC patterns in SZ compared with HC. Specifically, pairs of nodes extracted from the lower hidden layer represented sparse FC patterns implicated in SZ, which was quantified by using kurtosis/modularity measures and features from the higher hidden layer showed holistic/global FC patterns differentiating SZ from HC. Our proposed schemes and reported findings attained by using the DNN classifier and whole-brain FC data suggest that such approaches show improved ability to learn hidden patterns in brain imaging data, which may be useful for developing diagnostic tools for SZ and other neuropsychiatric disorders and identifying associated aberrant FC patterns.

Structural magnetic resonance imaging markers of susceptibility and transition to schizophrenia: a review of familial and clinical high risk population studies

There is a growing consensus that a symptomatology as complex and heterogeneous as schizophrenia is likely to be produced by widespread perturbations of brain structure, as opposed to isolated deficits in specific brain regions. Structural brain-imaging studies have shown that several features of the brain, such as grey matter, white matter integrity and the morphology of the cortex differ in individuals at high risk of the disorder compared to controls, but to a lesser extent than in patients, suggesting that structural abnormalities may form markers of vulnerability to the disorder. Research has had some success in delineating abnormalities specific to those individuals that transition to psychosis, compared to those at high risk that do not, suggesting that a general risk for the disorder can be distinguished from alterations specific to frank psychosis. In this paper, we review cross-sectional and longitudinal studies of individuals at familial or clinical high risk of the disorder. We conclude that the search for reliable markers of schizophrenia is likely to be enhanced by methods which amalgamate structural neuroimaging data into a coherent framework that takes into account the widespread distribution of brain alterations, and relates this to leading hypotheses of schizophrenia.

Fetal programming of schizophrenia: select mechanisms

Mounting evidence indicates that schizophrenia is associated with adverse intrauterine experiences. An adverse or suboptimal fetal environment can cause irreversible changes in brain that can subsequently exert long-lasting effects through resetting a diverse array of biological systems including endocrine, immune and nervous. It is evident from animal and imaging studies that subtle variations in the intrauterine environment can cause recognizable differences in brain structure and cognitive functions in the offspring. A wide variety of environmental factors may play a role in precipitating the emergent developmental dysregulation and the consequent evolution of psychiatric traits in early adulthood by inducing inflammatory, oxidative and nitrosative stress (IO&NS) pathways, mitochondrial dysfunction, apoptosis, and epigenetic dysregulation. However, the precise mechanisms behind such relationships and the specificity of the risk factors for schizophrenia remain exploratory. Considering the paucity of knowledge on fetal programming of schizophrenia, it is timely to consolidate the recent advances in the field and put forward an integrated overview of the mechanisms associated with fetal origin of schizophrenia.

Toward literature-based feature selection for diagnostic classification: a meta-analysis of resting-state fMRI in depression

Information derived from functional magnetic resonance imaging (fMRI) during wakeful rest has been introduced as a candidate diagnostic biomarker in unipolar major depressive disorder (MDD). Multiple reports of resting state fMRI in MDD describe group effects. Such prior knowledge can be adopted to pre-select potentially discriminating features for diagnostic classification models with the aim to improve diagnostic accuracy. Purpose of this analysis was to consolidate spatial information about alterations of spontaneous brain activity in MDD, primarily to serve as feature selection for multivariate pattern analysis techniques (MVPA). Thirty two studies were included in final analyses. Coordinates extracted from the original reports were assigned to two categories based on directionality of findings. Meta-analyses were calculated using the non-additive activation likelihood estimation approach with coordinates organized by subject group to account for non-independent samples. Converging evidence revealed a distributed pattern of brain regions with increased or decreased spontaneous activity in MDD. The most distinct finding was hyperactivity/hyperconnectivity presumably reflecting the interaction of cortical midline structures (posterior default mode network components including the precuneus and neighboring posterior cingulate cortices associated with self-referential processing and the subgenual anterior cingulate and neighboring medial frontal cortices) with lateral prefrontal areas related to externally-directed cognition. Other areas of hyperactivity/hyperconnectivity include the left lateral parietal cortex, right hippocampus and right cerebellum whereas hypoactivity/hypoconnectivity was observed mainly in the left temporal cortex, the insula, precuneus, superior frontal gyrus, lentiform nucleus and thalamus. Results are made available in two different data formats to be used as spatial hypotheses in future studies, particularly for diagnostic classification by MVPA.