Multi-center MRI prediction models: Predicting sex and illness course in first episode psychosis patients

Psychotic relapse prediction via biomarker monitoring: a systematic review

Background

Associating temporal variation of biomarkers with the onset of psychotic relapse could help demystify the pathogenesis of psychosis as a pathological brain state, while allowing for timely intervention, thus ameliorating clinical outcome. In this systematic review, we evaluated the predictive accuracy of a broad spectrum of biomarkers for psychotic relapse. We also underline methodological concerns, focusing on the value of prospective studies for relapse onset estimation.

Methods

Following the PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) guidelines, a list of search strings related to biomarkers and relapse was assimilated and run against the PubMed and Scopus databases, yielding a total of 808 unique records. After exclusion of studies related to the distinction of patients from controls or treatment effects, the 42 remaining studies were divided into 5 groups, based on the type of biomarker used as a predictor: the genetic biomarker subgroup (n = 4, or 9%), the blood-based biomarker subgroup (n = 15, or 36%), the neuroimaging biomarker subgroup (n = 10, or 24%), the cognitive-behavioral biomarker subgroup (n = 5, or 12%) and the wearables biomarker subgroup (n = 8, or 19%).

Results

In the first 4 groups, several factors were found to correlate with the state of relapse, such as the genetic risk profile, Interleukin-6, Vitamin D or panels consisting of multiple markers (blood-based), ventricular volume, grey matter volume in the right hippocampus, various functional connectivity metrics (neuroimaging), working memory and executive function (cognition). In the wearables group, machine learning models were trained based on features such as heart rate, acceleration, and geolocation, which were measured continuously. While the achieved predictive accuracy differed compared to chance, its power was moderate (max reported AUC = 0.77).

Discussion

The first 4 groups revealed risk factors, but cross-sectional designs or sparse sampling in prospective studies did not allow for relapse onset estimations. Studies involving wearables provide more concrete predictions of relapse but utilized markers such as geolocation do not advance pathophysiological understanding. A combination of the two approaches is warranted to fully understand and predict relapse.

Deep learning of structural MRI predicts fluid, crystallized, and general intelligence

Can brain structure predict human intelligence? T1-weighted structural brain magnetic resonance images (sMRI) have been correlated with intelligence. However, the population-level association does not fully account for individual variability in intelligence. To address this, studies have emerged recently to predict individual subject's intelligence or neurocognitive scores. However, they are mostly on predicting fluid intelligence (the ability to solve new problems). Studies are lacking to predict crystallized intelligence (the ability to accumulate knowledge) or general intelligence (fluid and crystallized intelligence combined). This study tests whether deep learning of sMRI can predict an individual subject's verbal, comprehensive, and full-scale intelligence quotients (VIQ, PIQ, and FSIQ), which reflect fluid and crystallized intelligence. We performed a comprehensive set of 432 experiments, using different input image channels, six deep learning models, and two outcome settings, in 850 healthy and autistic subjects 6-64 years of age. Our findings indicate a statistically significant potential of T1-weighted sMRI in predicting intelligence, with a Pearson correlation exceeding 0.21 (p < 0.001). Interestingly, we observed that an increase in the complexity of deep learning models does not necessarily translate to higher accuracy in intelligence prediction. The interpretations of our 2D and 3D CNNs, based on GradCAM, align well with the Parieto-Frontal Integration Theory (P-FIT), reinforcing the theory's suggestion that human intelligence is a result of interactions among various brain regions, including the occipital, temporal, parietal, and frontal lobes. These promising results invite further studies and open new questions in the field.

Multivariable prediction of functional outcome after first-episode psychosis: a crossover validation approach in EUFEST and PSYSCAN

Several multivariate prognostic models have been published to predict outcomes in patients with first episode psychosis (FEP), but it remains unclear whether those predictions generalize to independent populations. Using a subset of demographic and clinical baseline predictors, we aimed to develop and externally validate different models predicting functional outcome after a FEP in the context of a schizophrenia-spectrum disorder (FES), based on a previously published cross-validation and machine learning pipeline. A crossover validation approach was adopted in two large, international cohorts (EUFEST, n = 338, and the PSYSCAN FES cohort, n = 226). Scores on the Global Assessment of Functioning scale (GAF) at 12 month follow-up were dichotomized to differentiate between poor (GAF current < 65) and good outcome (GAF current ≥ 65). Pooled non-linear support vector machine (SVM) classifiers trained on the separate cohorts identified patients with a poor outcome with cross-validated balanced accuracies (BAC) of 65-66%, but BAC dropped substantially when the models were applied to patients from a different FES cohort (BAC = 50-56%). A leave-site-out analysis on the merged sample yielded better performance (BAC = 72%), highlighting the effect of combining data from different study designs to overcome calibration issues and improve model transportability. In conclusion, our results indicate that validation of prediction models in an independent sample is essential in assessing the true value of the model. Future external validation studies, as well as attempts to harmonize data collection across studies, are recommended.

Harmonizing three-dimensional MRI using pseudo-warping field guided GAN

In pursuit of cultivating automated models for magnetic resonance imaging (MRI) to aid in diagnostics, an escalating demand for extensive, multisite, and heterogeneous brain imaging datasets has emerged. This potentially introduces biased outcomes when directly applied for subsequent analysis. Researchers have endeavored to address this issue by pursuing the harmonization of MRIs. However, most existing image-based harmonization methods for MRI are tailored for 2D slices, which may introduce inter-slice variations when they are combined into a 3D volume. In this study, we aim to resolve inconsistencies between slices by introducing a pseudo-warping field. This field is created randomly and utilized to transform a slice into an artificially warped subsequent slice. The objective of this pseudo-warping field is to ensure that generators can consistently harmonize adjacent slices to another domain, without being affected by the varying content present in different slices. Furthermore, we construct unsupervised spatial and recycle loss to enhance the spatial accuracy and slice-wise consistency across the 3D images. The results demonstrate that our model effectively mitigates inter-slice variations and successfully preserves the anatomical details of the images during the harmonization process. Compared to generative harmonization models that employ 3D operators, our model exhibits greater computational efficiency and flexibility.

Evaluating the relationship between magnetic resonance image quality metrics and deep learning-based segmentation accuracy of brain tumors

BACKGROUND

Magnetic resonance imaging (MRI) scans are known to suffer from a variety of acquisition artifacts as well as equipment-based variations that impact image appearance and segmentation performance. It is still unclear whether a direct relationship exists between magnetic resonance (MR) image quality metrics (IQMs) (e.g., signal-to-noise, contrast-to-noise) and segmentation accuracy.

PURPOSE

Deep learning (DL) approaches have shown significant promise for automated segmentation of brain tumors on MRI but depend on the quality of input training images. We sought to evaluate the relationship between IQMs of input training images and DL-based brain tumor segmentation accuracy toward developing more generalizable models for multi-institutional data.

METHODS

We trained a 3D DenseNet model on the BraTS 2020 cohorts for segmentation of tumor subregions enhancing tumor (ET), peritumoral edematous, and necrotic and non-ET on MRI; with performance quantified via a 5-fold cross-validated Dice coefficient. MRI scans were evaluated through the open-source quality control tool MRQy, to yield 13 IQMs per scan. The Pearson correlation coefficient was computed between whole tumor (WT) dice values and IQM measures in the training cohorts to identify quality measures most correlated with segmentation performance. Each selected IQM was used to group MRI scans as "better" quality (BQ) or "worse" quality (WQ), via relative thresholding. Segmentation performance was re-evaluated for the DenseNet model when (i) training on BQ MRI images with validation on WQ images, as well as (ii) training on WQ images, and validation on BQ images. Trends were further validated on independent test sets derived from the BraTS 2021 training cohorts.

RESULTS

For this study, multimodal MRI scans from the BraTS 2020 training cohorts were used to train the segmentation model and validated on independent test sets derived from the BraTS 2021 cohort. Among the selected IQMs, models trained on BQ images based on inhomogeneity measurements (coefficient of variance, coefficient of joint variation, coefficient of variation of the foreground patch) and the models trained on WQ images based on noise measurement peak signal-to-noise ratio (SNR) yielded significantly improved tumor segmentation accuracy compared to their inverse models.

CONCLUSIONS

Our results suggest that a significant correlation may exist between specific MR IQMs and DenseNet-based brain tumor segmentation performance. The selection of MRI scans for model training based on IQMs may yield more accurate and generalizable models in unseen validation.

Brain identification of IBS patients based on GBDT and multiple imaging techniques

The brain biomarker of irritable bowel syndrome (IBS) patients is still lacking. The study aims to explore a new technology studying the brain alterations of IBS patients based on multi-source brain data. In the study, a decision-level fusion method based on gradient boosting decision tree (GBDT) was proposed. Next, 100 healthy subjects were used to validate the effectiveness of the method. Finally, the identification of brain alterations and the pain evaluation in IBS patients were carried out by the fusion method based on the resting-state fMRI and DWI for 46 patients and 46 controls selected randomly from 100 healthy subjects. The results showed that the method can achieve good classification between IBS patients and controls (accuracy = 95%) and pain evaluation of IBS patients (mean absolute error = 0.1977). Moreover, both the gain-based and the permutation-based evaluation instead of statistical analysis showed that left cingulum bundle contributed most significantly to the classification, and right precuneus contributed most significantly to the evaluation of abdominal pain intensity in the IBS patients. The differences seem to suggest a probable but unexplored separation about the central regions between the identification and progression of IBS. This finding may provide one new thought and technology for brain alteration related to IBS.

Disentangling psychosis: The challenges of informing precision medicine for what is not a single disorder

The last few decades of psychosis research have focused on the first episode. Studying the illness at onset offers a better understanding of its social and biological risk factors, and outcome correlates, without the confounding effects of chronicity on brain or social functioning. Significant efforts have been devoted to the identification of predictors of both illness onset and subsequent clinical and functional outcomes using different approaches. Among these, neuroimaging has provided important findings on brain neuromorphological differences between individuals with psychosis who have different outcomes. This is the main focus of this commentary. It is important to note that the neuromorphological differences reported in the literature between subgroups of individuals with different outcomes have not been of clinical utility so far. Rather, these findings have highlighted the presence of high heterogeneity in the brain biology that underlies psychosis. Mindful of this challenge, researchers have been experimenting with new analytical approaches, such as those that bypass the need to compare groups defined by a priori clinical labels. Our biggest challenge in the future will be to identify measures which could be used, alone or in combination, for a more precise stratification in clinical trials of new compounds or more personalized interventions.

Methodological and Quality Flaws in the Use of Artificial Intelligence in Mental Health Research: Systematic Review

BACKGROUND

Artificial intelligence (AI) is giving rise to a revolution in medicine and health care. Mental health conditions are highly prevalent in many countries, and the COVID-19 pandemic has increased the risk of further erosion of the mental well-being in the population. Therefore, it is relevant to assess the current status of the application of AI toward mental health research to inform about trends, gaps, opportunities, and challenges.

OBJECTIVE

This study aims to perform a systematic overview of AI applications in mental health in terms of methodologies, data, outcomes, performance, and quality.

METHODS

A systematic search in PubMed, Scopus, IEEE Xplore, and Cochrane databases was conducted to collect records of use cases of AI for mental health disorder studies from January 2016 to November 2021. Records were screened for eligibility if they were a practical implementation of AI in clinical trials involving mental health conditions. Records of AI study cases were evaluated and categorized by the International Classification of Diseases 11th Revision (ICD-11). Data related to trial settings, collection methodology, features, outcomes, and model development and evaluation were extracted following the CHARMS (Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies) guideline. Further, evaluation of risk of bias is provided.

RESULTS

A total of 429 nonduplicated records were retrieved from the databases and 129 were included for a full assessment-18 of which were manually added. The distribution of AI applications in mental health was found unbalanced between ICD-11 mental health categories. Predominant categories were Depressive disorders (n=70) and Schizophrenia or other primary psychotic disorders (n=26). Most interventions were based on randomized controlled trials (n=62), followed by prospective cohorts (n=24) among observational studies. AI was typically applied to evaluate quality of treatments (n=44) or stratify patients into subgroups and clusters (n=31). Models usually applied a combination of questionnaires and scales to assess symptom severity using electronic health records (n=49) as well as medical images (n=33). Quality assessment revealed important flaws in the process of AI application and data preprocessing pipelines. One-third of the studies (n=56) did not report any preprocessing or data preparation. One-fifth of the models were developed by comparing several methods (n=35) without assessing their suitability in advance and a small proportion reported external validation (n=21). Only 1 paper reported a second assessment of a previous AI model. Risk of bias and transparent reporting yielded low scores due to a poor reporting of the strategy for adjusting hyperparameters, coefficients, and the explainability of the models. International collaboration was anecdotal (n=17) and data and developed models mostly remained private (n=126).

CONCLUSIONS

These significant shortcomings, alongside the lack of information to ensure reproducibility and transparency, are indicative of the challenges that AI in mental health needs to face before contributing to a solid base for knowledge generation and for being a support tool in mental health management.

Individual variation in brain network topology is linked to course of illness in major depressive disorder

Major depressive disorder (MDD) is a chronic and highly recurrent disorder. The functional connectivity in depression is affected by the cumulative effect of course of illness. However, previous neuroimaging studies on abnormal functional connection have not mainly focused on the disease duration, which is seen as a secondary factor. Here, we used a data-driven analysis (multivariate distance matrix regression) to examine the relationship between the course of illness and resting-state functional dysconnectivity in MDD. This method identified a region in the anterior cingulate cortex, which is most linked to course of illness. Specifically, follow-up seed analyses show this phenomenon resulted from the individual differences in the topological distribution of three networks. In individuals with short-duration MDD, the connection to the default mode network was strong. By contrast, individuals with long-duration MDD showed hyperconnectivity to the ventral attention network and the frontoparietal network. These results emphasized the centrality of the anterior cingulate cortex in the pathophysiology of the increased course of illness and implied critical links between network topography and pathological duration. Thus, dissociable patterns of connectivity of the anterior cingulate cortex is an important dimension feature of the disease process of depression.

Combining MRI and clinical data to detect high relapse risk after the first episode of psychosis

Detecting patients at high relapse risk after the first episode of psychosis (HRR-FEP) could help the clinician adjust the preventive treatment. To develop a tool to detect patients at HRR using their baseline clinical and structural MRI, we followed 227 patients with FEP for 18–24 months and applied MRIPredict. We previously optimized the MRI-based machine-learning parameters (combining unmodulated and modulated gray and white matter and using voxel-based ensemble) in two independent datasets. Patients estimated to be at HRR-FEP showed a substantially increased risk of relapse (hazard ratio = 4.58, P < 0.05). Accuracy was poorer when we only used clinical or MRI data. We thus show the potential of combining clinical and MRI data to detect which individuals are more likely to relapse, who may benefit from increased frequency of visits, and which are unlikely, who may be currently receiving unnecessary prophylactic treatments. We also provide an updated version of the MRIPredict software.

Modern views of machine learning for precision psychiatry

In light of the National Institute of Mental Health (NIMH)'s Research Domain Criteria (RDoC), the advent of functional neuroimaging, novel technologies and methods provide new opportunities to develop precise and personalized prognosis and diagnosis of mental disorders. Machine learning (ML) and artificial intelligence (AI) technologies are playing an increasingly critical role in the new era of precision psychiatry. Combining ML/AI with neuromodulation technologies can potentially provide explainable solutions in clinical practice and effective therapeutic treatment. Advanced wearable and mobile technologies also call for the new role of ML/AI for digital phenotyping in mobile mental health. In this review, we provide a comprehensive review of ML methodologies and applications by combining neuroimaging, neuromodulation, and advanced mobile technologies in psychiatry practice. We further review the role of ML in molecular phenotyping and cross-species biomarker identification in precision psychiatry. We also discuss explainable AI (XAI) and neuromodulation in a closed human-in-the-loop manner and highlight the ML potential in multi-media information extraction and multi-modal data fusion. Finally, we discuss conceptual and practical challenges in precision psychiatry and highlight ML opportunities in future research.

The impact of harmonization on radiomic features in Parkinson's disease and healthy controls: A multicenter study

Radiomics is a challenging development area in imaging field that is greatly capturing interest of radiologists and neuroscientists. However, radiomics features show a strong non-biological variability determined by different facilities and imaging protocols, limiting the reproducibility and generalizability of analysis frameworks. Our study aimed to investigate the usefulness of harmonization to reduce site-effects on radiomics features over specific brain regions. We selected T1-weighted magnetic resonance imaging (MRI) by using the MRI dataset Parkinson's Progression Markers Initiative (PPMI) from different sites with healthy controls (HC) and Parkinson's disease (PD) patients. First, the investigation of radiomics measure discrepancies were assessed on healthy brain regions-of-interest (ROIs) via a classification pipeline based on LASSO feature selection and support vector machine (SVM) model. Then, a ComBat-based harmonization approach was applied to correct site-effects. Finally, a validation step on PD subjects evaluated diagnostic accuracy before and after harmonization of radiomics data. Results on healthy subjects demonstrated a dependence from site-effects that could be corrected with ComBat harmonization. LASSO regressor after harmonization was unable to select any feature to distinguish controls by site. Moreover, harmonized radiomics features achieved an area under the receiving operating characteristic curve (AUC) of 0.77 (compared to AUC of 0.71 for raw radiomics measures) in distinguish Parkinson's patients from HC. We found a not-negligible site-effect studying radiomics of HC pre- and post-harmonization of features. Our validation study on PD patients demonstrated a significant influence of non-biological noise source in diagnostic performances. Finally, harmonization of multicenter radiomic data represent a necessary step to make analysis pipelines reliable and replicable for multisite neuroimaging studies.

Bipolar spectrum disorders are associated with increased gray matter volume in the medial orbitofrontal cortex and nucleus accumbens

Objective

Elevated sensitivity to rewards prospectively predicts Bipolar Spectrum Disorder (BSD) onset; however, it is unclear whether volumetric abnormalities also reflect BSD risk. BSDs emerge when critical neurodevelopment in frontal and striatal regions occurs in sex-specific ways. The current paper examined the volume of frontal and striatal brain regions in both individuals with and at risk for a BSD with exploratory analyses examining sex-specificity.

Methods

One hundred fourteen medication-free individuals ages 18-27 at low-risk for BSD (moderate-reward sensitivity; N = 37), at high-risk without a BSD (high-reward sensitivity; N = 47), or with a BSD (N = 30) completed a structural MRI scan of the brain. We examined group differences in gray matter volume in a priori medial orbitofrontal cortex (mOFC) and nucleus accumbens (NAcc) regions-of-interest.

Results

The BSD group had enlarged frontostriatal volumes (mOFC, NAcc) compared to low individuals (d = 1.01). The mOFC volume in BSD was larger than low-risk (d = 1.01) and the high-risk groups (d = 0.74). This effect was driven by males with a BSD, who showed an enlarged mOFC compared to low (d = 1.01) and high-risk males (d = 0.74). Males with a BSD also showed a greater NAcc volume compared to males at low-risk (d = 0.49), but not high-risk males.

Conclusions

An enlarged frontostriatal volume (averaged mOFC, NAcc) is associated with the presence of a BSD, while subvolumes (mOFC vs. NAcc) showed unique patterning in relation to risk. We report preliminary evidence that sex moderates frontostriatal volume in BSD, highlighting the need for larger longitudinal risk studies examining the role of sex-specific neurodevelopmental trajectories in emerging BSDs.

Development of machine learning models for mortality risk prediction after cardiac surgery

BACKGROUND

We developed machine learning models that combine preoperative and intraoperative risk factors to predict mortality after cardiac surgery.

METHODS

Machine learning involving random forest, neural network, support vector machine, and gradient boosting machine was developed and compared with the risk scores of EuroSCORE I and II, Society of Thoracic Surgeons (STS), as well as a logistic regression model. Clinical data were collected from patients undergoing adult cardiac surgery at the First Medical Centre of Chinese PLA General Hospital between December 2008 and December 2017. The primary outcome was post-operative mortality. Model performance was estimated using several metrics, including sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC). The visualization algorithm was implemented using Shapley's additive explanations.

RESULTS

A total of 5,443 patients were enrolled during the study period. The mean EuroSCORE II score was 3.7%, and the actual in-hospital mortality rate was 2.7%. For predicting operative mortality after cardiac surgery, the AUC scores were 0.87, 0.79, 0.81, and 0.82 for random forest, neural network, support vector machine, and gradient boosting machine, compared with 0.70, 0.73, 0.71, and 0.74 for EuroSCORE I and II, STS, and logistic regression model. Shapley's additive explanations analysis of random forest yielded the top-20 predictors and individual-level explanations for each prediction.

CONCLUSIONS

Machine learning models based on available clinical data may be superior to clinical scoring tools in predicting postoperative mortality in patients following cardiac surgery. Explanatory models show the potential to provide personalized risk profiles for individuals by accounting for the contribution of influencing factors. Additional prospective multicenter studies are warranted to confirm the clinical benefit of these machine learning-driven models.

Deep Learning in Large and Multi-Site Structural Brain MR Imaging Datasets

Large, multi-site, heterogeneous brain imaging datasets are increasingly required for the training, validation, and testing of advanced deep learning (DL)-based automated tools, including structural magnetic resonance (MR) image-based diagnostic and treatment monitoring approaches. When assembling a number of smaller datasets to form a larger dataset, understanding the underlying variability between different acquisition and processing protocols across the aggregated dataset (termed “batch effects”) is critical. The presence of variation in the training dataset is important as it more closely reflects the true underlying data distribution and, thus, may enhance the overall generalizability of the tool. However, the impact of batch effects must be carefully evaluated in order to avoid undesirable effects that, for example, may reduce performance measures. Batch effects can result from many sources, including differences in acquisition equipment, imaging technique and parameters, as well as applied processing methodologies. Their impact, both beneficial and adversarial, must be considered when developing tools to ensure that their outputs are related to the proposed clinical or research question (i.e., actual disease-related or pathological changes) and are not simply due to the peculiarities of underlying batch effects in the aggregated dataset. We reviewed applications of DL in structural brain MR imaging that aggregated images from neuroimaging datasets, typically acquired at multiple sites. We examined datasets containing both healthy control participants and patients that were acquired using varying acquisition protocols. First, we discussed issues around Data Access and enumerated the key characteristics of some commonly used publicly available brain datasets. Then we reviewed methods for correcting batch effects by exploring the two main classes of approaches: Data Harmonization that uses data standardization, quality control protocols or other similar algorithms and procedures to explicitly understand and minimize unwanted batch effects; and Domain Adaptation that develops DL tools that implicitly handle the batch effects by using approaches to achieve reliable and robust results. In this narrative review, we highlighted the advantages and disadvantages of both classes of DL approaches, and described key challenges to be addressed in future studies.

Reduced TMS-evoked fast oscillations in the motor cortex predict the severity of positive symptoms in first-episode psychosis

Accumulating evidence points to neurophysiological abnormalities of the motor cortex in Schizophrenia (SCZ). However, whether these abnormalities represent a core biological feature of psychosis rather than a superimposed neurodegenerative process is yet to be defined, as it is their putative relationship with clinical symptoms. in this study, we used Transcranial Magnetic Stimulation coupled with electroencephalography (TMS-EEG) to probe the intrinsic oscillatory properties of motor (Brodmann Area 4, BA4) and non-motor (posterior parietal, BA7) cortical areas in twenty-three first-episode psychosis (FEP) patients and thirteen age and gender-matched healthy comparison (HC) subjects. Patients underwent clinical evaluation at baseline and six-months after the TMS-EEG session. We found that FEP patients had reduced EEG activity evoked by TMS of the motor cortex in the beta-2 (25-34 Hz) frequency band in a cluster of electrodes overlying BA4, relative to HC participants. Beta-2 deficits in the TMS-evoked EEG response correlated with worse positive psychotic symptoms at baseline and also predicted positive symptoms severity at six-month follow-up assessments. Altogether, these findings indicate that reduced TMS-evoked fast oscillatory activity in the motor cortex is an early neural abnormality that: 1) is present at illness onset; 2) may represent a state marker of psychosis; and 3) could play a role in the development of new tools of outcome prediction in psychotic patients.

Geometric deep learning on brain shape predicts sex and age

The complex relationship between the shape and function of the human brain remains elusive despite extensive studies of cortical folding over many decades. The analysis of cortical gyrification presents an opportunity to advance our knowledge about this relationship, and better understand the etiology of a variety of pathologies involving diverse degrees of cortical folding abnormalities. Hypothesis-driven surface-based approaches have been shown to be particularly efficient in their ability to accurately describe unique features of the folded sheet topology of the cortical ribbon. However, the utility of these approaches has been blunted by their reliance on manually defined features aiming to capture the relevant geometric properties of cortical folding. In this paper, we propose an entirely novel, data-driven deep-learning based method to analyze the brain's shape that eliminates this reliance on manual feature definition. This method builds on the emerging field of geometric deep-learning and uses traditional convolutional neural network architecture uniquely adapted to the surface representation of the cortical ribbon. This method is a complete departure from prior brain MRI CNN investigations, all of which have relied on three dimensional MRI data and interpreted features of the MRI signal for prediction. MRI data from 6410 healthy subjects obtained from 11 publicly available data repositories were used for analysis. Ages ranged from 6 to 89 years. Both inner and outer cortical surfaces were extracted using Freesurfer and then registered into MNI space. For purposes of method development, both a classification and regression challenge were introduced for network learning including sex and age prediction, respectively. Two independent graph convolutional neural networks (gCNNs) were trained, the first of which to predict subject's self-identified sex, the second of which to predict subject's age. Class Activation Maps (CAM) and Regression Activation Maps (RAM) were constructed respectively to map the topographic distribution of the most influential brain regions involved in the decision process for each gCNN. Using this approach, the gCNN was able to predict a subject's sex with an average accuracy of 87.99 % and achieved a Person's coefficient of correlation of 0.93 with an average absolute error 4.58 years when predicting a subject's age. We believe this shape-based convolutional classifier offers a novel, data-driven approach to define biomedically relevant features from the brain at both the population and single subject levels and therefore lays a critical foundation for future precision medicine applications.

Deep learning identifies morphological determinants of sex differences in the pre-adolescent brain

The application of data-driven deep learning to identify sex differences in developing brain structures of pre-adolescents has heretofore not been accomplished. Here, the approach identifies sex differences by analyzing the minimally processed MRIs of the first 8144 participants (age 9 and 10 years) recruited by the Adolescent Brain Cognitive Development (ABCD) study. The identified pattern accounted for confounding factors (i.e., head size, age, puberty development, socioeconomic status) and comprised cerebellar (corpus medullare, lobules III, IV/V, and VI) and subcortical (pallidum, amygdala, hippocampus, parahippocampus, insula, putamen) structures. While these have been individually linked to expressing sex differences, a novel discovery was that their grouping accurately predicted the sex in individual pre-adolescents. Another novelty was relating differences specific to the cerebellum to pubertal development. Finally, we found that reducing the pattern to a single score not only accurately predicted sex but also correlated with cognitive behavior linked to working memory. The predictive power of this score and the constellation of identified brain structures provide evidence for sex differences in pre-adolescent neurodevelopment and may augment understanding of sex-specific vulnerability or resilience to psychiatric disorders and presage sex-linked learning disabilities.

Predicting individual improvement in schizophrenia symptom severity at 1-year follow-up: Comparison of connectomic, structural, and clinical predictors

In a machine learning setting, this study aims to compare the prognostic utility of connectomic, brain structural, and clinical/demographic predictors of individual change in symptom severity in individuals with schizophrenia. Symptom severity at baseline and 1-year follow-up was assessed in 30 individuals with a schizophrenia-spectrum disorder using the Brief Psychiatric Rating Scale. Structural and functional neuroimaging was acquired in all individuals at baseline. Machine learning classifiers were trained to predict whether individuals improved or worsened with respect to positive, negative, and overall symptom severity. Classifiers were trained using various combinations of predictors, including regional cortical thickness and gray matter volume, static and dynamic resting-state connectivity, and/or baseline clinical and demographic variables. Relative change in overall symptom severity between baseline and 1-year follow-up varied markedly among individuals (interquartile range: 55%). Dynamic resting-state connectivity measured within the default-mode network was the most accurate single predictor of change in positive (accuracy: 87%), negative (83%), and overall symptom severity (77%) at follow-up. Incorporating predictors based on regional cortical thickness, gray matter volume, and baseline clinical variables did not markedly improve prediction accuracy and the prognostic utility of these predictors in isolation was moderate (<70%). Worsening negative symptoms at 1-year follow-up were predicted by hyper-connectivity and hypo-dynamism within the default-mode network at baseline assessment, while hypo-connectivity and hyper-dynamism predicted worsening positive symptoms. Given the modest sample size investigated, we recommend giving precedence to the relative ranking of the predictors investigated in this study, rather than the prediction accuracy estimates.

Multivariate patterns of gray matter volume in thalamic nuclei are associated with positive schizotypy in healthy individuals

BACKGROUND

Previous models suggest biological and behavioral continua among healthy individuals (HC), at-risk condition, and full-blown schizophrenia (SCZ). Part of these continua may be captured by schizotypy, which shares subclinical traits and biological phenotypes with SCZ, including thalamic structural abnormalities. In this regard, previous findings have suggested that multivariate volumetric patterns of individual thalamic nuclei discriminate HC from SCZ. These results were obtained using machine learning, which allows case-control classification at the single-subject level. However, machine learning accuracy is usually unsatisfactory possibly due to phenotype heterogeneity. Indeed, a source of misclassification may be related to thalamic structural characteristics of those HC with high schizotypy, which may resemble structural abnormalities of SCZ. We hypothesized that thalamic structural heterogeneity is related to schizotypy, such that high schizotypal burden would implicate misclassification of those HC whose thalamic patterns resemble SCZ abnormalities.

METHODS

Following a previous report, we used Random Forests to predict diagnosis in a case-control sample (SCZ = 131, HC = 255) based on thalamic nuclei gray matter volumes estimates. Then, we investigated whether the likelihood to be classified as SCZ (π-SCZ) was associated with schizotypy in 174 HC, evaluated with the Schizotypal Personality Questionnaire.

RESULTS

Prediction accuracy was 72.5%. Misclassified HC had higher positive schizotypy scores, which were correlated with π-SCZ. Results were specific to thalamic rather than whole-brain structural features.

CONCLUSIONS

These findings strengthen the relevance of thalamic structural abnormalities to SCZ and suggest that multivariate thalamic patterns are correlates of the continuum between schizotypy in HC and the full-blown disease.

Towards Precision Medicine in Psychosis: Benefits and Challenges of Multimodal Multicenter Studies-PSYSCAN: Translating Neuroimaging Findings From Research into Clinical Practice

In the last 2 decades, several neuroimaging studies investigated brain abnormalities associated with the early stages of psychosis in the hope that these could aid the prediction of onset and clinical outcome. Despite advancements in the field, neuroimaging has yet to deliver. This is in part explained by the use of univariate analytical techniques, small samples and lack of statistical power, lack of external validation of potential biomarkers, and lack of integration of nonimaging measures (eg, genetic, clinical, cognitive data). PSYSCAN is an international, longitudinal, multicenter study on the early stages of psychosis which uses machine learning techniques to analyze imaging, clinical, cognitive, and biological data with the aim of facilitating the prediction of psychosis onset and outcome. In this article, we provide an overview of the PSYSCAN protocol and we discuss benefits and methodological challenges of large multicenter studies that employ neuroimaging measures.

Quantifying the Impact of Chronic Ischemic Injury on Clinical Outcomes in Acute Stroke With Machine Learning

Acute stroke is often superimposed on chronic damage from previous cerebrovascular events. This background will inevitably modulate the impact of acute injury on clinical outcomes to an extent that will depend on the precise anatomical pattern of damage. Previous attempts to quantify such modulation have employed only reductive models that ignore anatomical detail. The combination of automated image processing, large-scale data, and machine learning now enables us to quantify the impact of this with high-dimensional multivariate models sensitive to individual variations in the detailed anatomical pattern. We introduce and validate a new automated chronic lesion segmentation routine for use with non-contrast CT brain scans, combining non-parametric outlier-detection score, Zeta, with an unsupervised 3-dimensional maximum-flow, minimum-cut algorithm. The routine was then applied to a dataset of 1,704 stroke patient scans, obtained at their presentation to a hyper-acute stroke unit (St George's Hospital, London, UK), and used to train a support vector machine (SVM) model to predict between low (0-2) and high (3-6) pre-admission and discharge modified Rankin Scale (mRS) scores, quantifying performance by the area under the receiver operating curve (AUROC). In this single center retrospective observational study, our SVM models were able to differentiate between low (0-2) and high (3-6) pre-admission and discharge mRS scores with an AUROC of 0.77 (95% confidence interval of 0.74-0.79), and 0.76 (0.74-0.78), respectively. The chronic lesion segmentation routine achieved a mean (standard deviation) sensitivity, specificity and Dice similarity coefficient of 0.746 (0.069), 0.999 (0.001), and 0.717 (0.091), respectively. We have demonstrated that machine learning models capable of capturing the high-dimensional features of chronic injuries are able to stratify patients-at the time of presentation-by pre-admission and discharge mRS scores. Our fully automated chronic stroke lesion segmentation routine simplifies this process, and utilizes routinely collected CT head scans, thereby facilitating future large-scale studies to develop supportive clinical decision tools.

Improving individual predictions: Machine learning approaches for detecting and attacking heterogeneity in schizophrenia (and other psychiatric diseases)

Psychiatric diseases are very heterogeneous both in clinical manifestation and etiology. With the recent rise of using machine learning techniques to attempt to diagnose and prognose these disorders, the issue of heterogeneity becomes increasingly important. With the growing interest in personalized medicine, it becomes even more important to not only classify someone as a patient with a certain disorder, its treatment needs a more precise definition of the underlying neurobiology, since different biological origins of the same disease may require (very) different treatments. We review the possible contributions that machine learning techniques could make to explore the heterogeneous nature of psychiatric disorders with a focus on schizophrenia. First we will review how heterogeneity shows up and how machine learning, or multivariate pattern recognition methods in general, can be used to discover it. Secondly, we will discuss the possible uses of these techniques to attack heterogeneity, leading to improved predictions and understanding of the neurobiological background of the disorder.

Progress in psychoradiology, the clinical application of psychiatric neuroimaging

Psychoradiology is an emerging field that applies radiological imaging technologies to psychiatric conditions. In the past three decades, brain imaging techniques have rapidly advanced understanding of illness and treatment effects in psychiatry. Based on these advances, radiologists have become increasingly interested in applying these advances for differential diagnosis and individualized patient care selection for common psychiatric illnesses. This shift from research to clinical practice represents the beginning evolution of psychoradiology. In this review, we provide a summary of recent progress relevant to this field based on their clinical functions, namely the (1) classification and subtyping; (2) prediction and monitoring of treatment outcomes; and (3) treatment selection. In addition, we provide guidelines for the practice of psychoradiology in clinical settings and suggestions for future research to validate broader clinical applications. Given the high prevalence of psychiatric disorders and the importance of increased participation of radiologists in this field, a guide regarding advances in this field and a description of relevant clinical work flow patterns help radiologists contribute to this fast-evolving field.

Harmonization of resting-state functional MRI data across multiple imaging sites via the separation of site differences into sampling bias and measurement bias

When collecting large amounts of neuroimaging data associated with psychiatric disorders, images must be acquired from multiple sites because of the limited capacity of a single site. However, site differences represent a barrier when acquiring multisite neuroimaging data. We utilized a traveling-subject dataset in conjunction with a multisite, multidisorder dataset to demonstrate that site differences are composed of biological sampling bias and engineering measurement bias. The effects on resting-state functional MRI connectivity based on pairwise correlations because of both bias types were greater than or equal to psychiatric disorder differences. Furthermore, our findings indicated that each site can sample only from a subpopulation of participants. This result suggests that it is essential to collect large amounts of neuroimaging data from as many sites as possible to appropriately estimate the distribution of the grand population. Finally, we developed a novel harmonization method that removed only the measurement bias by using a traveling-subject dataset and achieved the reduction of the measurement bias by 29% and improvement of the signal-to-noise ratios by 40%. Our results provide fundamental knowledge regarding site effects, which is important for future research using multisite, multidisorder resting-state functional MRI data.

Predicting anxiety from wholebrain activity patterns to emotional faces in young adults: a machine learning approach

BACKGROUND

It is becoming increasingly clear that pathophysiological processes underlying psychiatric disorders categories are heterogeneous on many levels, including symptoms, disease course, comorbidity and biological underpinnings. This heterogeneity poses challenges for identifying biological markers associated with dimensions of symptoms and behaviour that could provide targets to guide treatment choice and novel treatment. In response, the research domain criteria (RDoC) (Insel et al., 2010) was developed to advocate a dimensional approach which omits any disease definitions, disorder thresholds, or cut-points for various levels of psychopathology to understanding the pathophysiological processes underlying psychiatry disorders. In the present study we aimed to apply pattern regression analysis to identify brain signatures during dynamic emotional face processing that are predictive of anxiety and depression symptoms in a continuum that ranges from normal to pathological levels, cutting across categorically-defined diagnoses.

METHODS

The sample was composed of one-hundred and fifty-four young adults (mean age=21.6 and s.d.=2.0, 103 females) consisting of eighty-two young adults seeking treatment for psychological distress that cut across categorically-defined diagnoses and 72 matched healthy young adults. Participants performed a dynamic face task involving fearful, angry and happy faces (and geometric shapes) while undergoing functional Magnetic Resonance Imaging (fMRI). Pattern regression analyses consisted of Gaussian Process Regression (GPR) implemented in the Pattern Recognition for Neuroimaging toolbox (PRoNTo). Predicted and actual clinical scores were compared using Pearson's correlation coefficient (r) and normalized mean squared error (MSE) to evaluate the models' performance. Permutation test was applied to estimate significance levels.

RESULTS

GPR identified patterns of neural activity to dynamic emotional face processing predictive of self-report anxiety in the whole sample, which covered a continuum that ranged from healthy to different levels of distress, including subthreshold to fully-syndromal psychiatric diagnoses. Results were significant using two different cross validation strategies (two-fold: r=0.28 (p-value=0.001), MSE=4.47 (p-value=0.001) and five fold r=0.28 (p-value=0.002), MSE=4.62 (p-value=0.003). The contributions of individual regions to the predictive model were very small, demonstrating that predictions were based on the overall pattern rather than on a small combination of regions.

CONCLUSIONS

These findings represent early evidence that neuroimaging techniques may inform clinical assessment of young adults irrespective of diagnoses by allowing accurate and objective quantitative estimation of psychopathology.

Brain connectivity alterations in early psychosis: from clinical to neuroimaging staging

Early in the course of psychosis, alterations in brain connectivity accompany the emergence of psychiatric symptoms and cognitive impairments, including processing speed. The clinical-staging model is a refined form of diagnosis that places the patient along a continuum of illness conditions, which allows stage-specific interventions with the potential of improving patient care and outcome. This cross-sectional study investigates brain connectivity features that characterize the clinical stages following a first psychotic episode. Structural brain networks were derived from diffusion-weighted MRI for 71 early-psychosis patients and 76 healthy controls. Patients were classified into stage II (first-episode), IIIa (incomplete remission), IIIb (one relapse), and IIIc (two or more relapses), according to the course of the illness until the time of scanning. Brain connectivity measures and diffusion parameters (fractional anisotropy, apparent diffusion coefficient) were investigated using general linear models and sparse linear discriminant analysis (sLDA), studying distinct subgroups of patients who were at specific stages of early psychosis. We found that brain connectivity impairments were more severe in clinical stages following the first-psychosis episode (stages IIIa, IIIb, IIIc) than in first-episode psychosis (stage II) patients. These alterations were spatially diffuse but converged on a set of vulnerable regions, whose inter-connectivity selectively correlated with processing speed in patients and controls. The sLDA suggested that relapsing-remitting (stages IIIb, IIIc) and non-remitting (stage IIIa) patients are characterized by distinct dysconnectivity profiles. Our results indicate that neuroimaging markers of brain dysconnectivity in early psychosis may reflect the heterogeneity of the illness and provide a connectomics signature of the clinical-staging model.

Enhancing the representation of functional connectivity networks by fusing multi-view information for autism spectrum disorder diagnosis

Functional connectivity network provides novel insights on how distributed brain regions are functionally integrated, and its deviations from healthy brain have recently been employed to identify biomarkers for neuropsychiatric disorders. However, most of brain network analysis methods utilized features extracted only from one functional connectivity network for brain disease detection and cannot provide a comprehensive representation on the subtle disruptions of brain functional organization induced by neuropsychiatric disorders. Inspired by the principles of multi-view learning which utilizes information from multiple views to enhance object representation, we propose a novel multiple network based framework to enhance the representation of functional connectivity networks by fusing the common and complementary information conveyed in multiple networks. Specifically, four functional connectivity networks corresponding to the four adjacent values of regularization parameter are generated via a sparse regression model with group constraint ( l_2,1 -norm), to enhance the common intrinsic topological structure and limit the error rate caused by different views. To obtain a set of more meaningful and discriminative features, we propose using a modified version of weighted clustering coefficients to quantify the subtle differences of each group-sparse network at local level. We then linearly fuse the selected features from each individual network via a multi-kernel support vector machine for autism spectrum disorder (ASD) diagnosis. The proposed framework achieves an accuracy of 79.35%, outperforming all the compared single network methods for at least 7% improvement. Moreover, compared with other multiple network methods, our method also achieves the best performance, that is, with at least 11% improvement in accuracy.

Baseline brain structural and functional predictors of clinical outcome in the early course of schizophrenia

Although schizophrenia is considered a brain disorder, the role of brain organization for symptomatic improvement remains inadequately defined. We investigated the relationship between baseline brain morphology, resting-state network connectivity and clinical response after 24-weeks of antipsychotic treatment in patients with schizophrenia (n = 95) using integrated multivariate analyses. There was no significant association between clinical response and measures of cortical thickness (r = 0.37, p = 0.98) and subcortical volume (r = 0.56, p = 0.15). By contrast, we identified a strong mode of covariation linking functional network connectivity to clinical response (r = 0.70; p = 0.04), and particularly to improvement in positive (weight = 0.62) and anxious/depressive symptoms (weight = 0.49). Higher internal cohesiveness of the default mode network was the single most important positive predictor. Key negative predictors involved the functional cohesiveness of central executive subnetworks anchored in the frontoparietal cortices and subcortical regions (including the thalamus and striatum) and the inter-network integration between the default mode and sensorimotor networks. The present findings establish links between clinical response and the functional organization of brain networks involved both in perception and in spontaneous and goal-directed cognition, thereby advancing our understanding of the pathophysiology of schizophrenia.

Neuromarkers for Mental Disorders: Harnessing Population Neuroscience

Despite abundant research into the neurobiology of mental disorders, to date neurobiological insights have had very little impact on psychiatric diagnosis or treatment. In this review, we contend that the search for neuroimaging biomarkers-neuromarkers-of mental disorders is a highly promising avenue toward improved psychiatric healthcare. However, many of the traditional tools used for psychiatric neuroimaging are inadequate for the identification of neuromarkers. Specifically, we highlight the need for larger samples and for multivariate analysis. Approaches such as machine learning are likely to be beneficial for interrogating high-dimensional neuroimaging data. We suggest that broad, population-based study designs will be important for developing neuromarkers of mental disorders, and will facilitate a move away from a phenomenological definition of mental disorder categories and toward psychiatric nosology based on biological evidence. We provide an outline of how the development of neuromarkers should occur, emphasizing the need for tests of external and construct validity, and for collaborative research efforts. Finally, we highlight some concerns regarding the development, and use of, neuromarkers in psychiatric healthcare.

Is It Possible to Predict the Future in First-Episode Psychosis?

The outcome of first-episode psychosis (FEP) is highly variable, ranging from early sustained recovery to antipsychotic treatment resistance from the onset of illness. For clinicians, a possibility to predict patient outcomes would be highly valuable for the selection of antipsychotic treatment and in tailoring psychosocial treatments and psychoeducation. This selective review summarizes current knowledge of prognostic markers in FEP. We sought potential outcome predictors from clinical and sociodemographic factors, cognition, brain imaging, genetics, and blood-based biomarkers, and we considered different outcomes, like remission, recovery, physical comorbidities, and suicide risk. Based on the review, it is currently possible to predict the future for FEP patients to some extent. Some clinical features-like the longer duration of untreated psychosis (DUP), poor premorbid adjustment, the insidious mode of onset, the greater severity of negative symptoms, comorbid substance use disorders (SUDs), a history of suicide attempts and suicidal ideation and having non-affective psychosis-are associated with a worse outcome. Of the social and demographic factors, male gender, social disadvantage, neighborhood deprivation, dysfunctional family environment, and ethnicity may be relevant. Treatment non-adherence is a substantial risk factor for relapse, but a small minority of patients with acute onset of FEP and early remission may benefit from antipsychotic discontinuation. Cognitive functioning is associated with functional outcomes. Brain imaging currently has limited utility as an outcome predictor, but this may change with methodological advancements. Polygenic risk scores (PRSs) might be useful as one component of a predictive tool, and pharmacogenetic testing is already available and valuable for patients who have problems in treatment response or with side effects. Most blood-based biomarkers need further validation. None of the currently available predictive markers has adequate sensitivity or specificity used alone. However, personalized treatment of FEP will need predictive tools. We discuss some methodologies, such as machine learning (ML), and tools that could lead to the improved prediction and clinical utility of different prognostic markers in FEP. Combination of different markers in ML models with a user friendly interface, or novel findings from e.g., molecular genetics or neuroimaging, may result in computer-assisted clinical applications in the near future.