Learning stable and predictive network-based patterns of schizophrenia and its clinical symptoms

What Remains to Be Discovered in Schizophrenia Therapeutics: Contributions by Advancing the Molecular Mechanisms of Drugs for Psychosis and Schizophrenia

Schizophrenia is a frequently debilitating and complex mental disorder affecting approximately 1% of the global population, characterized by symptoms such as hallucinations, delusions, disorganized thoughts and behaviors, cognitive dysfunction, and negative symptoms. Traditional treatment has centered on postsynaptic dopamine antagonists, commonly known as antipsychotic drugs, which aim to alleviate symptoms and improve functioning and the quality of life. Despite the availability of these medications, significant challenges remain in schizophrenia therapeutics, including incomplete symptom relief, treatment resistance, and medication side effects. This opinion article explores advancements in schizophrenia treatment, emphasizing molecular mechanisms, novel drug targets, and innovative delivery methods. One promising approach is novel strategies that target neural networks and circuits rather than single neurotransmitters, acknowledging the complexity of brain region interconnections involved in schizophrenia. Another promising approach is the development of biased agonists, which selectively activate specific signaling pathways downstream of receptors, offering potential for more precise pharmacological interventions with fewer side effects. The concept of molecular polypharmacy, where a single drug targets multiple molecular pathways, is exemplified by KarXT, a novel drug combining xanomeline and trospium to address both psychosis and cognitive dysfunction. This approach represents a comprehensive strategy for schizophrenia treatment, potentially improving outcomes for patients. In conclusion, advancing the molecular understanding of schizophrenia and exploring innovative therapeutic strategies hold promise for addressing the unmet needs in schizophrenia treatment, aiming for more effective and tailored interventions. Future research should focus on these novel approaches to achieve better clinical outcomes and improve the functional level and quality of life for individuals with schizophrenia.

Sampling inequalities affect generalization of neuroimaging-based diagnostic classifiers in psychiatry

BACKGROUND

The development of machine learning models for aiding in the diagnosis of mental disorder is recognized as a significant breakthrough in the field of psychiatry. However, clinical practice of such models remains a challenge, with poor generalizability being a major limitation.

METHODS

Here, we conducted a pre-registered meta-research assessment on neuroimaging-based models in the psychiatric literature, quantitatively examining global and regional sampling issues over recent decades, from a view that has been relatively underexplored. A total of 476 studies (n = 118,137) were included in the current assessment. Based on these findings, we built a comprehensive 5-star rating system to quantitatively evaluate the quality of existing machine learning models for psychiatric diagnoses.

RESULTS

A global sampling inequality in these models was revealed quantitatively (sampling Gini coefficient (G) = 0.81, p < .01), varying across different countries (regions) (e.g., China, G = 0.47; the USA, G = 0.58; Germany, G = 0.78; the UK, G = 0.87). Furthermore, the severity of this sampling inequality was significantly predicted by national economic levels (β =  - 2.75, p < .001, R²_adj = 0.40; r =  - .84, 95% CI: - .41 to - .97), and was plausibly predictable for model performance, with higher sampling inequality for reporting higher classification accuracy. Further analyses showed that lack of independent testing (84.24% of models, 95% CI: 81.0-87.5%), improper cross-validation (51.68% of models, 95% CI: 47.2-56.2%), and poor technical transparency (87.8% of models, 95% CI: 84.9-90.8%)/availability (80.88% of models, 95% CI: 77.3-84.4%) are prevailing in current diagnostic classifiers despite improvements over time. Relating to these observations, model performances were found decreased in studies with independent cross-country sampling validations (all p < .001, BF₁₀ > 15). In light of this, we proposed a purpose-built quantitative assessment checklist, which demonstrated that the overall ratings of these models increased by publication year but were negatively associated with model performance.

CONCLUSIONS

Together, improving sampling economic equality and hence the quality of machine learning models may be a crucial facet to plausibly translating neuroimaging-based diagnostic classifiers into clinical practice.

Evaluation of Risk of Bias in Neuroimaging-Based Artificial Intelligence Models for Psychiatric Diagnosis: A Systematic Review

IMPORTANCE

Neuroimaging-based artificial intelligence (AI) diagnostic models have proliferated in psychiatry. However, their clinical applicability and reporting quality (ie, feasibility) for clinical practice have not been systematically evaluated.

OBJECTIVE

To systematically assess the risk of bias (ROB) and reporting quality of neuroimaging-based AI models for psychiatric diagnosis.

EVIDENCE REVIEW

PubMed was searched for peer-reviewed, full-length articles published between January 1, 1990, and March 16, 2022. Studies aimed at developing or validating neuroimaging-based AI models for clinical diagnosis of psychiatric disorders were included. Reference lists were further searched for suitable original studies. Data extraction followed the CHARMS (Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modeling Studies) and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines. A closed-loop cross-sequential design was used for quality control. The PROBAST (Prediction Model Risk of Bias Assessment Tool) and modified CLEAR (Checklist for Evaluation of Image-Based Artificial Intelligence Reports) benchmarks were used to systematically evaluate ROB and reporting quality.

FINDINGS

A total of 517 studies presenting 555 AI models were included and evaluated. Of these models, 461 (83.1%; 95% CI, 80.0%-86.2%) were rated as having a high overall ROB based on the PROBAST. The ROB was particular high in the analysis domain, including inadequate sample size (398 of 555 models [71.7%; 95% CI, 68.0%-75.6%]), poor model performance examination (with 100% of models lacking calibration examination), and lack of handling data complexity (550 of 555 models [99.1%; 95% CI, 98.3%-99.9%]). None of the AI models was perceived to be applicable to clinical practices. Overall reporting completeness (ie, number of reported items/number of total items) for the AI models was 61.2% (95% CI, 60.6%-61.8%), and the completeness was poorest for the technical assessment domain with 39.9% (95% CI, 38.8%-41.1%).

CONCLUSIONS AND RELEVANCE

This systematic review found that the clinical applicability and feasibility of neuroimaging-based AI models for psychiatric diagnosis were challenged by a high ROB and poor reporting quality. Particularly in the analysis domain, ROB in AI diagnostic models should be addressed before clinical application.

Translating Big Data to Clinical Outcomes in Anxiety: Potential for Multimodal Integration

PURPOSE OF THE REVIEW

This review describes approaches to research on anxiety that attempt to link neural correlates to treatment response and novel therapies. The review emphasizes pediatric anxiety disorders since most anxiety disorders begin before adulthood.

RECENT FINDINGS

Recent literature illustrates how current treatments for anxiety manifest diverse relations with a range of neural markers. While some studies demonstrate post-treatment normalization of markers in anxious individuals, others find persistence of group differences. For other markers, which show no pretreatment association with anxiety, the markers nevertheless distinguish treatment-responders from non-responders. Heightened error related negativity represents the risk marker discussed in the most depth; however, limitations in measures related to error responding necessitate multimodal and big-data approaches. Single risk markers show limits as correlates of treatment response. Large-scale, multimodal data analyzed with predictive models may illuminate additional risk markers related to anxiety disorder treatment outcomes. Such work may identify novel targets and eventually guide improvements in treatment response/outcomes.

Unraveling schizophrenia replicable functional connectivity disruption patterns across sites

Functional connectivity (FC) disruption is a remarkable characteristic of schizophrenia. However, heterogeneous patterns reported across sites severely hindered its clinical generalization. Based on qualified nodal-based FC of 340 schizophrenia patients (SZ) and 348 normal controls (NC) acquired from seven different scanners, this study compared four commonly used site-effect correction methods in removing the site-related heterogeneities, and then tried to cluster the abnormal FCs into several replicable and independent disrupted subnets across sites, related them to clinical symptoms, and evaluated their potentials in schizophrenia classification. Among the four site-related heterogeneity correction methods, ComBat harmonization (F1 score: 0.806 ± 0.145) achieved the overall best balance between sensitivity and false discovery rate in unraveling the aberrant FCs of schizophrenia in the local and public data sets. Hierarchical clustering analysis identified three replicable FC disruption subnets across the local and public data sets: hypo-connectivity within sensory areas (Net1), hypo-connectivity within thalamus, striatum, and ventral attention network (Net2), and hyper-connectivity between thalamus and sensory processing system (Net3). Notably, the derived composite FC within Net1 was negatively correlated with hostility and disorientation in the public validation set (p < .05). Finally, the three subnet-specific composite FCs (Best area under the receiver operating characteristic curve [AUC] = 0.728) can robustly and meaningfully discriminate the SZ from NC with comparable performance with the full identified FCs features (best AUC = 0.765) in the out-of-sample public data set (Z = -1.583, p = .114). In conclusion, ComBat harmonization was most robust in detecting aberrant connectivity for schizophrenia. Besides, the three subnet-specific composite FC measures might be replicable neuroimaging markers for schizophrenia.

Individual-specific networks for prediction modelling – A scoping review of methods

Background

Recent advances in biotechnology enable the acquisition of high-dimensional data on individuals, posing challenges for prediction models which traditionally use covariates such as clinical patient characteristics. Alternative forms of covariate representations for the features derived from these modern data modalities should be considered that can utilize their intrinsic interconnection. The connectivity information between these features can be represented as an individual-specific network defined by a set of nodes and edges, the strength of which can vary from individual to individual. Global or local graph-theoretical features describing the network may constitute potential prognostic biomarkers instead of or in addition to traditional covariates and may replace the often unsuccessful search for individual biomarkers in a high-dimensional predictor space.

Methods

We conducted a scoping review to identify, collate and critically appraise the state-of-art in the use of individual-specific networks for prediction modelling in medicine and applied health research, published during 2000–2020 in the electronic databases PubMed, Scopus and Embase.

Results

Our scoping review revealed the main application areas namely neurology and pathopsychology, followed by cancer research, cardiology and pathology (N = 148). Network construction was mainly based on Pearson correlation coefficients of repeated measurements, but also alternative approaches (e.g. partial correlation, visibility graphs) were found. For covariates measured only once per individual, network construction was mostly based on quantifying an individual’s contribution to the overall group-level structure. Despite the multitude of identified methodological approaches for individual-specific network inference, the number of studies that were intended to enable the prediction of clinical outcomes for future individuals was quite limited, and most of the models served as proof of concept that network characteristics can in principle be useful for prediction.

Conclusion

The current body of research clearly demonstrates the value of individual-specific network analysis for prediction modelling, but it has not yet been considered as a general tool outside the current areas of application. More methodological research is still needed on well-founded strategies for network inference, especially on adequate network sparsification and outcome-guided graph-theoretical feature extraction and selection, and on how networks can be exploited efficiently for prediction modelling.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-022-01544-6.

Machine learning classification of schizophrenia patients and healthy controls using diverse neuroanatomical markers and Ensemble methods

Schizophrenia is a major psychiatric disorder that imposes enormous clinical burden on patients and their caregivers. Determining classification biomarkers can complement clinical measures and improve understanding of the neural basis underlying schizophrenia. Using neuroanatomical features, several machine learning based investigations have attempted to classify schizophrenia from healthy controls but the range of neuroanatomical measures employed have been limited in range to date. In this study, we sought to classify schizophrenia and healthy control cohorts using a diverse set of neuroanatomical measures (cortical and subcortical volumes, cortical areas and thickness, cortical mean curvature) and adopted Ensemble methods for better performance. Additionally, we correlated such neuroanatomical features with Quality of Life (QoL) assessment scores within the schizophrenia cohort. With Ensemble methods and diverse neuroanatomical measures, we achieved classification accuracies ranging from 83 to 87%, sensitivities and specificities varying between 90-98% and 65-70% respectively. In addition to lower QoL scores within schizophrenia cohort, significant correlations were found between specific neuroanatomical measures and psychological health, social relationship subscale domains of QoL. Our results suggest the utility of inclusion of subcortical and cortical measures and Ensemble methods to achieve better classification performance and their potential impact of parsing out neurobiological correlates of quality of life in schizophrenia.

Diagnosis of schizophrenia with functional connectome data: a graph-based convolutional neural network approach

Previous deep learning methods have not captured graph or network representations of brain structural or functional connectome data. To address this, we developed the BrainNet-Global Covariance Pooling-Attention Convolutional Neural Network (BrainNet-GA CNN) by incorporating BrainNetCNN and global covariance pooling into the self-attention mechanism. Resting-state functional magnetic resonance imaging data were obtained from 171 patients with schizophrenia spectrum disorders (SSDs) and 161 healthy controls (HCs). We conducted an ablation analysis of the proposed BrainNet-GA CNN and quantitative performance comparisons with competing methods using the nested tenfold cross validation strategy. The performance of our model was compared with competing methods. Discriminative connections were visualized using the gradient-based explanation method and compared with the results obtained using functional connectivity analysis. The BrainNet-GA CNN showed an accuracy of 83.13%, outperforming other competing methods. Among the top 10 discriminative connections, some were associated with the default mode network and auditory network. Interestingly, these regions were also significant in the functional connectivity analysis. Our findings suggest that the proposed BrainNet-GA CNN can classify patients with SSDs and HCs with higher accuracy than other models. Visualization of salient regions provides important clinical information. These results highlight the potential use of the BrainNet-GA CNN in the diagnosis of schizophrenia.

Improving the predictive potential of diffusion MRI in schizophrenia using normative models-Towards subject-level classification

Diffusion MRI studies consistently report group differences in white matter between individuals diagnosed with schizophrenia and healthy controls. Nevertheless, the abnormalities found at the group-level are often not observed at the individual level. Among the different approaches aiming to study white matter abnormalities at the subject level, normative modeling analysis takes a step towards subject-level predictions by identifying affected brain locations in individual subjects based on extreme deviations from a normative range. Here, we leveraged a large harmonized diffusion MRI dataset from 512 healthy controls and 601 individuals diagnosed with schizophrenia, to study whether normative modeling can improve subject-level predictions from a binary classifier. To this aim, individual deviations from a normative model of standard (fractional anisotropy) and advanced (free-water) dMRI measures, were calculated by means of age and sex-adjusted z-scores relative to control data, in 18 white matter regions. Even though larger effect sizes are found when testing for group differences in z-scores than are found with raw values (p < .001), predictions based on summary z-score measures achieved low predictive power (AUC < 0.63). Instead, we find that combining information from the different white matter tracts, while using multiple imaging measures simultaneously, improves prediction performance (the best predictor achieved AUC = 0.726). Our findings suggest that extreme deviations from a normative model are not optimal features for prediction. However, including the complete distribution of deviations across multiple imaging measures improves prediction, and could aid in subject-level classification.

DeepNeuro: an open-source deep learning toolbox for neuroimaging

Translating deep learning research from theory into clinical practice has unique challenges, specifically in the field of neuroimaging. In this paper, we present DeepNeuro, a Python-based deep learning framework that puts deep neural networks for neuroimaging into practical usage with a minimum of friction during implementation. We show how this framework can be used to design deep learning pipelines that can load and preprocess data, design and train various neural network architectures, and evaluate and visualize the results of trained networks on evaluation data. We present a way of reproducibly packaging data pre- and postprocessing functions common in the neuroimaging community, which facilitates consistent performance of networks across variable users, institutions, and scanners. We show how deep learning pipelines created with DeepNeuro can be concisely packaged into shareable Docker and Singularity containers with user-friendly command-line interfaces.

Artificial intelligence (AI) in medicine as a strategic valuable tool

Humans' creativity led to machines that outperform human capabilities in terms of workload, effectiveness, precision, endurance, strength, and repetitiveness. It has always been a vision and a way to transcend the existence and to give more sense to life, which is precious. The common denominator of all these creations was that they were meant to replace, enhance or go beyond the mechanical capabilities of the human body. The story takes another bifurcation when Alan Turing introduced the concept of a machine that could think, in 1950. Artificial intelligence, presented as a term in 1956, describes the use of computers to imitate intelligence and critical thinking comparable to humans. However, the revolution began in 1943, when artificial neural networks was an attempt to exploit the architecture of the human brain to perform tasks that conventional algorithms had little success with. Artificial intelligence is becoming a research focus and a tool of strategic value. The same observations apply in the field of healthcare, too. In this manuscript, we try to address key questions regarding artificial intelligence in medicine, such as what artificial intelligence is and how it works, what is its value in terms of application in medicine, and what are the prospects?

Multimodal MRI assessment for first episode psychosis: A major change in the thalamus and an efficient stratification of a subgroup

Multi‐institutional brain imaging studies have emerged to resolve conflicting results among individual studies. However, adjusting multiple variables at the technical and cohort levels is challenging. Therefore, it is important to explore approaches that provide meaningful results from relatively small samples at institutional levels. We studied 87 first episode psychosis (FEP) patients and 62 healthy subjects by combining supervised integrated factor analysis (SIFA) with a novel pipeline for automated structure‐based analysis, an efficient and comprehensive method for dimensional data reduction that our group recently established. We integrated multiple MRI features (volume, DTI indices, resting state fMRI—rsfMRI) in the whole brain of each participant in an unbiased manner. The automated structure‐based analysis showed widespread DTI abnormalities in FEP and rs‐fMRI differences between FEP and healthy subjects mostly centered in thalamus. The combination of multiple modalities with SIFA was more efficient than the use of single modalities to stratify a subgroup of FEP (individuals with schizophrenia or schizoaffective disorder) that had more robust deficits from the overall FEP group. The information from multiple MRI modalities and analytical methods highlighted the thalamus as significantly abnormal in FEP. This study serves as a proof‐of‐concept for the potential of this methodology to reveal disease underpins and to stratify populations into more homogeneous sub‐groups.

Towards a brain-based predictome of mental illness

Neuroimaging-based approaches have been extensively applied to study mental illness in recent years and have deepened our understanding of both cognitively healthy and disordered brain structure and function. Recent advancements in machine learning techniques have shown promising outcomes for individualized prediction and characterization of patients with psychiatric disorders. Studies have utilized features from a variety of neuroimaging modalities, including structural, functional, and diffusion magnetic resonance imaging data, as well as jointly estimated features from multiple modalities, to assess patients with heterogeneous mental disorders, such as schizophrenia and autism. We use the term "predictome" to describe the use of multivariate brain network features from one or more neuroimaging modalities to predict mental illness. In the predictome, multiple brain network-based features (either from the same modality or multiple modalities) are incorporated into a predictive model to jointly estimate features that are unique to a disorder and predict subjects accordingly. To date, more than 650 studies have been published on subject-level prediction focusing on psychiatric disorders. We have surveyed about 250 studies including schizophrenia, major depression, bipolar disorder, autism spectrum disorder, attention-deficit hyperactivity disorder, obsessive-compulsive disorder, social anxiety disorder, posttraumatic stress disorder, and substance dependence. In this review, we present a comprehensive review of recent neuroimaging-based predictomic approaches, current trends, and common shortcomings and share our vision for future directions.

Treatment response prediction and individualized identification of first-episode drug-naïve schizophrenia using brain functional connectivity

Identifying biomarkers in schizophrenia during the first episode without the confounding effects of treatment has been challenging. Leveraging these biomarkers to establish diagnosis and make individualized predictions of future treatment responses to antipsychotics would be of great value, but there has been limited progress. In this study, by using machine learning algorithms and the functional connections of the superior temporal cortex, we successfully identified the first-episode drug-naive (FEDN) schizophrenia patients (accuracy 78.6%) and predict their responses to antipsychotic treatment (accuracy 82.5%) at an individual level. The functional connections (FC) were derived using the mutual information and the correlations, between the blood-oxygen-level dependent signals of the superior temporal cortex and other cortical regions acquired with the resting-state functional magnetic resonance imaging. We also found that the mutual information and correlation FC was informative in identifying individual FEDN schizophrenia and prediction of treatment response, respectively. The methods and findings in this paper could provide a critical step toward individualized identification and treatment response prediction in first-episode drug-naive schizophrenia, which could complement other biomarkers in the development of precision medicine approaches for this severe mental disorder.

The Eminence of Co-Expressed Ties in Schizophrenia Network Communities

Exploring gene networks is crucial for identifying significant biological interactions occurring in a disease condition. These interactions can be acknowledged by modeling the tie structure of networks. Such tie orientations are often detected within embedded community structures. However, most of the prevailing community detection modules are intended to capture information from nodes and its attributes, usually ignoring the ties. In this study, a modularity maximization algorithm is proposed based on nonlinear representation of local tangent space alignment (LTSA). Initially, the tangent coordinates are computed locally to identify k-nearest neighbors across the genes. These local neighbors are further optimized by generating a nonlinear network embedding function for detecting gene communities based on eigenvector decomposition. Experimental results suggest that this algorithm detects gene modules with a better modularity index of 0.9256, compared to other traditional community detection algorithms. Furthermore, co-expressed genes across these communities are identified by discovering the characteristic tie structures. These detected ties are known to have substantial biological influence in the progression of schizophrenia, thereby signifying the influence of tie patterns in biological networks. This technique can be extended logically on other diseases networks for detecting substantial gene “hotspots”.

Modeling Research Topics for Artificial Intelligence Applications in Medicine: Latent Dirichlet Allocation Application Study

BACKGROUND

Artificial intelligence (AI)-based technologies develop rapidly and have myriad applications in medicine and health care. However, there is a lack of comprehensive reporting on the productivity, workflow, topics, and research landscape of AI in this field.

OBJECTIVE

This study aimed to evaluate the global development of scientific publications and constructed interdisciplinary research topics on the theory and practice of AI in medicine from 1977 to 2018.

METHODS

We obtained bibliographic data and abstract contents of publications published between 1977 and 2018 from the Web of Science database. A total of 27,451 eligible articles were analyzed. Research topics were classified by latent Dirichlet allocation, and principal component analysis was used to identify the construct of the research landscape.

RESULTS

The applications of AI have mainly impacted clinical settings (enhanced prognosis and diagnosis, robot-assisted surgery, and rehabilitation), data science and precision medicine (collecting individual data for precision medicine), and policy making (raising ethical and legal issues, especially regarding privacy and confidentiality of data). However, AI applications have not been commonly used in resource-poor settings due to the limit in infrastructure and human resources.

CONCLUSIONS

The application of AI in medicine has grown rapidly and focuses on three leading platforms: clinical practices, clinical material, and policies. AI might be one of the methods to narrow down the inequality in health care and medicine between developing and developed countries. Technology transfer and support from developed countries are essential measures for the advancement of AI application in health care in developing countries.

Machine learning in schizophrenia genomics, a case-control study using 5,090 exomes

Our hypothesis is that machine learning (ML) analysis of whole exome sequencing (WES) data can be used to identify individuals at high risk for schizophrenia (SCZ). This study applies ML to WES data from 2,545 individuals with SCZ and 2,545 unaffected individuals, accessed via the database of genotypes and phenotypes (dbGaP). Single nucleotide variants and small insertions and deletions were annotated by ANNOVAR using the reference genome hg19/GRCh37. Rare (predicted functional) variants with a minor allele frequency ≤1% and genotype quality ≥90 including missense, frameshift, stop gain, stop loss, intronic, and exonic splicing variants were selected. A file containing all cases and controls, the names of genes with variants meeting our criteria, and the number of variants per gene for each individual, was used for ML analysis. The supervised machine-learning algorithm used the patterns of variants observed in the different genes to determine which subset of genes can best predict that an individual is affected. Seventy percent of the data was used to train the algorithm and the remaining 30% of data (n = 1,526) was used to evaluate its efficiency. The supervised ML algorithm, gradient boosted trees with regularization (eXtreme Gradient Boosting implementation) was the best performing algorithm yielding promising results (accuracy: 85.7%, specificity: 86.6%, sensitivity: 84.9%, area under the receiver-operator characteristic curve: 0.95). The top 50 features (genes) of the algorithm were analyzed using bioinformatics resources for new insights about the pathophysiology of SCZ. This manuscript presents a novel predictor which could potentially enable studies exploring disease-modifying intervention in the early stages of the disease.

Classification and Prediction of Brain Disorders Using Functional Connectivity: Promising but Challenging

Brain functional imaging data, especially functional magnetic resonance imaging (fMRI) data, have been employed to reflect functional integration of the brain. Alteration in brain functional connectivity (FC) is expected to provide potential biomarkers for classifying or predicting brain disorders. In this paper, we present a comprehensive review in order to provide guidance about the available brain FC measures and typical classification strategies. We survey the state-of-the-art FC analysis methods including widely used static functional connectivity (SFC) and more recently proposed dynamic functional connectivity (DFC). Temporal correlations among regions of interest (ROIs), data-driven spatial network and functional network connectivity (FNC) are often computed to reflect SFC from different angles. SFC can be extended to DFC using a sliding-window framework, and intrinsic connectivity states along the time-varying connectivity patterns are typically extracted using clustering or decomposition approaches. We also briefly summarize window-less DFC approaches. Subsequently, we highlight various strategies for feature selection including the filter, wrapper and embedded methods. In terms of model building, we include traditional classifiers as well as more recently applied deep learning methods. Moreover, we review representative applications with remarkable classification accuracy for psychosis and mood disorders, neurodevelopmental disorder, and neurological disorders using fMRI data. Schizophrenia, bipolar disorder, autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), Alzheimer's disease and mild cognitive impairment (MCI) are discussed. Finally, challenges in the field are pointed out with respect to the inaccurate diagnosis labeling, the abundant number of possible features and the difficulty in validation. Some suggestions for future work are also provided.

Medicine and the rise of the robots: a qualitative review of recent advances of artificial intelligence in health

Artificial intelligence (AI) has the potential to significantly transform the role of the doctor and revolutionise the practice of medicine. This qualitative review paper summarises the past 12 months of health research in AI, across different medical specialties, and discusses the current strengths as well as challenges, relating to this emerging technology. Doctors, especially those in leadership roles, need to be aware of how quickly AI is advancing in health, so that they are ready to lead the change required for its adoption by the health system. Key points: ‘AI has now been shown to be as effective as humans in the diagnosis of various medical conditions, and in some cases, more effective.’ When it comes to predicting suicide attempts, recent research suggest AI is better than human beings. ‘AI’s current strength is in its ability to learn from a large dataset and recognise patterns that can be used to diagnose conditions, putting it in direct competition with medical specialties that are involved in diagnostic tests that involve pattern recognition, such as pathology and radiology’. The current challenges in AI include legal liability and attribution of negligence when errors occur, and the ethical issues relating to patient choices. ‘AI systems can also be developed with, or learn, biases, that will need to be identified and mitigated’. As doctors and health leaders, we need to start preparing the profession to be supported by, partnered with, and, in future, potentially be replaced by, AI and advanced robotics systems.