The emerging field of 'predictive analytics in mental health' has recently generated tremendous interest with the bold promise to revolutionize clinical practice in psychiatry paralleling similar developments in personalized and precision medicine. Here, we provide an overview of the key questions and challenges in the field, aiming to (1) propose general guidelines for predictive analytics projects in psychiatry, (2) provide a conceptual introduction to core aspects of predictive modeling technology, and (3) foster a broad and informed discussion involving all stakeholders including researchers, clinicians, patients, funding bodies and policymakers.
Neuroimaging studies of patients with major depression have revealed abnormal intrinsic functional connectivity measured during the resting state in multiple distributed networks. However, it is unclear whether these findings reflect the state of major depression or reflect trait neurobiological underpinnings of risk for major depression.
We asked whether brain connectomics can predict response to treatment for a neuropsychiatric disorder better than conventional clinical measures. Pre-treatment resting-state brain functional connectivity and diffusion-weighted structural connectivity were measured in 38 patients with social anxiety disorder (SAD) to predict subsequent treatment response to cognitive behavioral therapy (CBT). We used a priori bilateral anatomical amygdala seed-driven resting connectivity and probabilistic tractography of the right inferior longitudinal fasciculus together with a data-driven multivoxel pattern analysis of whole-brain resting-state connectivity before treatment to predict improvement in social anxiety after CBT. Each connectomic measure improved the prediction of individuals' treatment outcomes significantly better than a clinical measure of initial severity, and combining the multimodal connectomics yielded a fivefold improvement in predicting treatment response. Generalization of the findings was supported by leave-one-out cross-validation. After dividing patients into better or worse responders, logistic regression of connectomic predictors and initial severity combined with leave-one-out cross-validation yielded a categorical prediction of clinical improvement with 81% accuracy, 84% sensitivity and 78% specificity. Connectomics of the human brain, measured by widely available imaging methods, may provide brain-based biomarkers (neuromarkers) supporting precision medicine that better guide patients with neuropsychiatric diseases to optimal available treatments, and thus translate basic neuroimaging into medical practice.
Despite growing evidence for atypical amygdala function and structure in major depression, it remains uncertain as to whether these brain differences reflect the clinical state of depression or neurobiological traits that predispose individuals to major depression. We examined function and structure of the amygdala and associated areas in a group of unaffected children of depressed parents (at-risk group) and a group of children of parents without a history of major depression (control group). Compared to the control group, the at-risk group showed increased activation to fearful relative to neutral facial expressions in the amygdala and multiple cortical regions, and decreased activation to happy relative to neutral facial expressions in the anterior cingulate cortex and supramarginal gyrus. At-risk children also exhibited reduced amygdala volume. The extensive hyperactivation to negative facial expressions and hypoactivation to positive facial expressions in at-risk children are consistent with behavioral evidence that risk for major depression involves a bias to attend to negative information. These functional and structural brain differences between at-risk children and controls suggest that there are trait neurobiological underpinnings of risk for major depression.
Neuroimaging has greatly enhanced the cognitive neuroscience understanding of the human brain and its variation across individuals (neurodiversity) in both health and disease. Such progress has not yet, however, propelled changes in educational or medical practices that improve people's lives. We review neuroimaging findings in which initial brain measures (neuromarkers) are correlated with or predict future education, learning, and performance in children and adults; criminality; health-related behaviors; and responses to pharmacological or behavioral treatments. Neuromarkers often provide better predictions (neuroprognosis), alone or in combination with other measures, than traditional behavioral measures. With further advances in study designs and analyses, neuromarkers may offer opportunities to personalize educational and clinical practices that lead to better outcomes for people.
We examined the normal development of intrinsic functional connectivity of the default network (brain regions typically deactivated for attention-demanding tasks) as measured by resting-state fMRI in children, adolescents, and young adults ages 8-24 years. We investigated both positive and negative correlations and employed analysis methods that allowed for valid interpretation of negative correlations and that also minimized the influence of motion artifacts that are often confounds in developmental neuroimaging. As age increased, there were robust developmental increases in negative correlations, including those between medial pFC (MPFC) and dorsolateral pFC (DLPFC) and between lateral parietal cortices and brain regions associated with the dorsal attention network. Between multiple regions, these correlations reversed from being positive in children to negative in adults. Age-related changes in positive correlations within the default network were below statistical threshold after controlling for motion. Given evidence in adults that greater negative correlation between MPFC and DLPFC is associated with superior cognitive performance, the development of an intrinsic anticorrelation between MPFC and DLPFC may be a marker of the large growth of working memory and executive functions that occurs from childhood to young adulthood.
Current behavioral measures poorly predict treatment outcome in social anxiety disorder (SAD). To our knowledge, this is the first study to examine neuroimaging-based treatment prediction in SAD.