Inference in the age of big data: Future perspectives on neuroscience

Atlases of cognition with large-scale human brain mapping

To map the neural substrate of mental function, cognitive neuroimaging relies on controlled psychological manipulations that engage brain systems associated with specific cognitive processes. In order to build comprehensive atlases of cognitive function in the brain, it must assemble maps for many different cognitive processes, which often evoke overlapping patterns of activation. Such data aggregation faces contrasting goals: on the one hand finding correspondences across vastly different cognitive experiments, while on the other hand precisely describing the function of any given brain region. Here we introduce a new analysis framework that tackles these difficulties and thereby enables the generation of brain atlases for cognitive function. The approach leverages ontologies of cognitive concepts and multi-label brain decoding to map the neural substrate of these concepts. We demonstrate the approach by building an atlas of functional brain organization based on 30 diverse functional neuroimaging studies, totaling 196 different experimental conditions. Unlike conventional brain mapping, this functional atlas supports robust reverse inference: predicting the mental processes from brain activity in the regions delineated by the atlas. To establish that this reverse inference is indeed governed by the corresponding concepts, and not idiosyncrasies of experimental designs, we show that it can accurately decode the cognitive concepts recruited in new tasks. These results demonstrate that aggregating independent task-fMRI studies can provide a more precise global atlas of selective associations between brain and cognition.

Cognitive neuroscience uses neuroimaging to identify brain systems engaged in specific cognitive tasks. However, linking unequivocally brain systems with cognitive functions is difficult: each task probes only a small number of facets of cognition, while brain systems are often engaged in many tasks. We develop a new approach to generate a functional atlas of cognition, demonstrating brain systems selectively associated with specific cognitive functions. This approach relies upon an ontology that defines specific cognitive functions and the relations between them, along with an analysis scheme tailored to this ontology. Using a database of thirty neuroimaging studies, we show that this approach provides a highly-specific atlas of mental functions, and that it can decode the mental processes engaged in new tasks.

Data-driven, voxel-based analysis of brain PET images: Application of PCA and LASSO methods to visualize and quantify patterns of neurodegeneration

Spatial patterns of radiotracer binding in positron emission tomography (PET) images may convey information related to the disease topology. However, this information is not captured by the standard PET image analysis that quantifies the mean radiotracer uptake within a region of interest (ROI). On the other hand, spatial analyses that use more advanced radiomic features may be difficult to interpret. Here we propose an alternative data-driven, voxel-based approach to spatial pattern analysis in brain PET, which can be easily interpreted. We apply principal component analysis (PCA) to identify voxel covariance patterns, and optimally combine several patterns using the least absolute shrinkage and selection operator (LASSO). The resulting models predict clinical disease metrics from raw voxel values, allowing for inclusion of clinical covariates. The analysis is performed on high-resolution PET images from healthy controls and subjects affected by Parkinson’s disease (PD), acquired with a pre-synaptic and a post-synaptic dopaminergic PET tracer. We demonstrate that PCA identifies robust and tracer-specific binding patterns in sub-cortical brain structures; the patterns evolve as a function of disease progression. Principal component LASSO (PC-LASSO) models of clinical disease metrics achieve higher predictive accuracy compared to the mean tracer binding ratio (BR) alone: the cross-validated test mean squared error of adjusted disease duration (motor impairment score) was 16.3 ± 0.17 years² (9.7 ± 0.15) with mean BR, versus 14.4 ± 0.18 years² (8.9 ± 0.16) with PC-LASSO. We interpret the best-performing PC-LASSO models in the spatial sense and discuss them with reference to the PD pathology and somatotopic organization of the striatum. PC-LASSO is thus shown to be a useful method to analyze clinically-relevant tracer binding patterns, and to construct interpretable, imaging-based predictive models of clinical metrics.

Pipeline validation for connectivity-based cortex parcellation

Structural connectivity plays a dominant role in brain function and arguably lies at the core of understanding the structure-function relationship in the cerebral cortex. Connectivity-based cortex parcellation (CCP), a framework to process structural connectivity information gained from diffusion MRI and diffusion tractography, identifies cortical subunits that furnish functional inference. The underlying pipeline of algorithms interprets similarity in structural connectivity as a segregation criterion. Validation of the CCP-pipeline is critical to gain scientific reliability of the algorithmic processing steps from dMRI data to voxel grouping. In this paper we provide a proof of concept based upon a novel model validation principle that characterizes the trade-off between informativeness and robustness to assess the validity of the CCP pipeline, including diffusion tractography and clustering. We ultimately identify a pipeline of algorithms and parameter settings that tolerate more noise and extract more information from the data than their alternatives.

Patterns of schizophrenia symptoms: hidden structure in the PANSS questionnaire

The clinical presentation of patients with schizophrenia has long been described to be very heterogeneous. Coherent symptom profiles can probably be directly derived from behavioral manifestations quantified in medical questionnaires. The combination of machine learning algorithms and an international multi-site dataset (n = 218 patients) identified distinctive patterns underlying schizophrenia from the widespread PANSS questionnaire. Our clustering approach revealed a negative symptom patient group as well as a moderate and a severe group, giving further support for the existence of schizophrenia subtypes. Additionally, emerging regression analyses uncovered the most clinically predictive questionnaire items. Small subsets of PANSS items showed convincing forecasting performance in single patients. These item subsets encompassed the entire symptom spectrum confirming that the different facets of schizophrenia can be shown to enable improved clinical diagnosis and medical action in patients. Finally, we did not find evidence for complicated relationships among the PANSS items in our sample. Our collective results suggest that identifying best treatment for a given individual may be grounded in subtle item combinations that transcend the long-trusted positive, negative, and cognitive categories.

Deconstructing multivariate decoding for the study of brain function

Multivariate decoding methods were developed originally as tools to enable accurate predictions in real-world applications. The realization that these methods can also be employed to study brain function has led to their widespread adoption in the neurosciences. However, prior to the rise of multivariate decoding, the study of brain function was firmly embedded in a statistical philosophy grounded on univariate methods of data analysis. In this way, multivariate decoding for brain interpretation grew out of two established frameworks: multivariate decoding for predictions in real-world applications, and classical univariate analysis based on the study and interpretation of brain activation. We argue that this led to two confusions, one reflecting a mixture of multivariate decoding for prediction or interpretation, and the other a mixture of the conceptual and statistical philosophies underlying multivariate decoding and classical univariate analysis. Here we attempt to systematically disambiguate multivariate decoding for the study of brain function from the frameworks it grew out of. After elaborating these confusions and their consequences, we describe six, often unappreciated, differences between classical univariate analysis and multivariate decoding. We then focus on how the common interpretation of what is signal and noise changes in multivariate decoding. Finally, we use four examples to illustrate where these confusions may impact the interpretation of neuroimaging data. We conclude with a discussion of potential strategies to help resolve these confusions in interpreting multivariate decoding results, including the potential departure from multivariate decoding methods for the study of brain function.

Linked dimensions of psychopathology and connectivity in functional brain networks

Neurobiological abnormalities associated with psychiatric disorders do not map well to existing diagnostic categories. High co-morbidity suggests dimensional circuit-level abnormalities that cross diagnoses. Here we seek to identify brain-based dimensions of psychopathology using sparse canonical correlation analysis in a sample of 663 youths. This analysis reveals correlated patterns of functional connectivity and psychiatric symptoms. We find that four dimensions of psychopathology - mood, psychosis, fear, and externalizing behavior - are associated (r = 0.68-0.71) with distinct patterns of connectivity. Loss of network segregation between the default mode network and executive networks emerges as a common feature across all dimensions. Connectivity linked to mood and psychosis becomes more prominent with development, and sex differences are present for connectivity related to mood and fear. Critically, findings largely replicate in an independent dataset (n = 336). These results delineate connectivity-guided dimensions of psychopathology that cross clinical diagnostic categories, which could serve as a foundation for developing network-based biomarkers in psychiatry.

Lost in translation

Translation in cognitive neuroscience remains beyond the horizon, brought no closer by supposed major advances in our understanding of the brain. Unless our explanatory models descend to the individual level-a cardinal requirement for any intervention-their real-world applications will always be limited. Drawing on an analysis of the informational properties of the brain, here we argue that adequate individualisation needs models of far greater dimensionality than has been usual in the field. This necessity arises from the widely distributed causality of neural systems, a consequence of the fundamentally adaptive nature of their developmental and physiological mechanisms. We discuss how recent advances in high-performance computing, combined with collections of large-scale data, enable the high-dimensional modelling we argue is critical to successful translation, and urge its adoption if the ultimate goal of impact on the lives of patients is to be achieved.

Lost in translation

Translation in cognitive neuroscience remains beyond the horizon, brought no closer by supposed major advances in our understanding of the brain. Unless our explanatory models descend to the individual level-a cardinal requirement for any intervention-their real-world applications will always be limited. Drawing on an analysis of the informational properties of the brain, here we argue that adequate individualisation needs models of far greater dimensionality than has been usual in the field. This necessity arises from the widely distributed causality of neural systems, a consequence of the fundamentally adaptive nature of their developmental and physiological mechanisms. We discuss how recent advances in high-performance computing, combined with collections of large-scale data, enable the high-dimensional modelling we argue is critical to successful translation, and urge its adoption if the ultimate goal of impact on the lives of patients is to be achieved.

Neuroimaging meta-analysis of cannabis use studies reveals convergent functional alterations in brain regions supporting cognitive control and reward processing

Lagging behind rapid changes to state laws, societal views, and medical practice is the scientific investigation of cannabis's impact on the human brain. While several brain imaging studies have contributed important insight into neurobiological alterations linked with cannabis use, our understanding remains limited. Here, we sought to delineate those brain regions that consistently demonstrate functional alterations among cannabis users versus non-users across neuroimaging studies using the activation likelihood estimation meta-analysis framework. In ancillary analyses, we characterized task-related brain networks that co-activate with cannabis-affected regions using data archived in a large neuroimaging repository, and then determined which psychological processes may be disrupted via functional decoding techniques. When considering convergent alterations among users, decreased activation was observed in the anterior cingulate cortex, which co-activated with frontal, parietal, and limbic areas and was linked with cognitive control processes. Similarly, decreased activation was observed in the dorsolateral prefrontal cortex, which co-activated with frontal and occipital areas and linked with attention-related processes. Conversely, increased activation among users was observed in the striatum, which co-activated with frontal, parietal, and other limbic areas and linked with reward processing. These meta-analytic outcomes indicate that cannabis use is linked with differential, region-specific effects across the brain.

Prediction complements explanation in understanding the developing brain

A central aim of human neuroscience is understanding the neurobiology of cognition and behavior. Although we have made significant progress towards this goal, reliance on group-level studies of the developed adult brain has limited our ability to explain population variability and developmental changes in neural circuitry and behavior. In this review, we suggest that predictive modeling, a method for predicting individual differences in behavior from brain features, can complement descriptive approaches and provide new ways to account for this variability. Highlighting the outsized scientific and clinical benefits of prediction in developmental populations including adolescence, we show that predictive brain-based models are already providing new insights on adolescent-specific risk-related behaviors. Together with large-scale developmental neuroimaging datasets and complementary analytic approaches, predictive modeling affords us the opportunity and obligation to identify novel treatment targets and individually tailor the course of interventions for developmental psychopathologies that impact so many young people today.

Machine Learning Approaches for Clinical Psychology and Psychiatry

Machine learning approaches for clinical psychology and psychiatry explicitly focus on learning statistical functions from multidimensional data sets to make generalizable predictions about individuals. The goal of this review is to provide an accessible understanding of why this approach is important for future practice given its potential to augment decisions associated with the diagnosis, prognosis, and treatment of people suffering from mental illness using clinical and biological data. To this end, the limitations of current statistical paradigms in mental health research are critiqued, and an introduction is provided to critical machine learning methods used in clinical studies. A selective literature review is then presented aiming to reinforce the usefulness of machine learning methods and provide evidence of their potential. In the context of promising initial results, the current limitations of machine learning approaches are addressed, and considerations for future clinical translation are outlined.

The hippocampal network model: A transdiagnostic metaconnectomic approach

Purpose

The hippocampus plays a central role in cognitive and affective processes and is commonly implicated in neurodegenerative diseases. Our study aimed to identify and describe a hippocampal network model (HNM) using trans-diagnostic MRI data from the BrainMap® database. We used meta-analysis to test the network degeneration hypothesis (NDH) (Seeley et al., 2009) by identifying structural and functional covariance in this hippocampal network.

Methods

To generate our network model, we used BrainMap's VBM database to perform a region-to-whole-brain (RtWB) meta-analysis of 269 VBM experiments from 165 published studies across a range of 38 psychiatric and neurological diseases reporting hippocampal gray matter density alterations. This step identified 11 significant gray matter foci, or nodes. We subsequently used meta-analytic connectivity modeling (MACM) to define edges of structural covariance between nodes from VBM data as well as functional covariance using the functional task-activation database, also from BrainMap. Finally, we applied a correlation analysis using Pearson's r to assess the similarities and differences between the structural and functional covariance models.

Key findings

Our hippocampal RtWB meta-analysis reported consistent and significant structural covariance in 11 key regions. The subsequent structural and functional MACMs showed a strong correlation between HNM nodes with a significant structural-functional covariance correlation of r = .377 (p = .000049).

Significance

This novel method of studying network covariance using VBM and functional meta-analytic techniques allows for the identification of generalizable patterns of functional and structural abnormalities pertaining to the hippocampus. In accordance with the NDH, this framework could have major implications in studying and predicting spatial disease patterns using network-based assays.

•

We derived regions that structurally co-vary with the hippocampus in a network model using a transdiagnostic meta-analytic approach.

•

We used meta-analytic connectivity mapping to assess inter-regional connectivity from BrainMap's structural and functional databases.

•

We tested the network degeneration hypothesis by identifying network correlations between structural and functional networks.

High-dimensional therapeutic inference in the focally damaged human brain

See Thiebaut de Schotten and Foulon (doi:10.1093/brain/awx332) for a scientific commentary on this article.Though consistency across the population renders the extraordinarily complex functional anatomy of the human brain surveyable, the inverse inference-from common functional maps to individual behaviour-is constrained by marked individual deviation from the population mean. Such inference is fundamental to the evaluation of therapeutic interventions in focal brain injury, where the impact of an induced structural change in the brain is quantified by its behavioural consequences, inevitably refracted through the lens of lesion-outcome relations. Current therapeutic evaluations do not incorporate inferences to the individual outcome derived from a detailed specification of the lesion anatomy, relying only on reductive parameters such as lesion volume and crudely discretised location. Examining 1172 patients with anatomically registered focal brain lesions, here we show that such low-dimensional models are highly insensitive to therapeutic effects. In contrast, high-dimensional models supported by machine learning dramatically improve sensitivity by leveraging complex individuating patterns in the functional architecture of the brain. The failure to replicate in humans positive interventional effects in experimental animals is thus revealed to have a remediable inferential cause, forcing a radical re-evaluation of therapeutic inference in the human brain.

Machine learning: supervised methods

Supervised learning algorithms extract general principles from observed examples guided by a specific prediction objective.

Predictive assessment of models for dynamic functional connectivity

In neuroimaging, it has become evident that models of dynamic functional connectivity (dFC), which characterize how intrinsic brain organization changes over time, can provide a more detailed representation of brain function than traditional static analyses. Many dFC models in the literature represent functional brain networks as a meta-stable process with a discrete number of states; however, there is a lack of consensus on how to perform model selection and learn the number of states, as well as a lack of understanding of how different modeling assumptions influence the estimated state dynamics. To address these issues, we consider a predictive likelihood approach to model assessment, where models are evaluated based on their predictive performance on held-out test data. Examining several prominent models of dFC (in their probabilistic formulations) we demonstrate our framework on synthetic data, and apply it on two real-world examples: a face recognition EEG experiment and resting-state fMRI. Our results evidence that both EEG and fMRI are better characterized using dynamic modeling approaches than by their static counterparts, but we also demonstrate that one must be cautious when interpreting dFC because parameter settings and modeling assumptions, such as window lengths and emission models, can have a large impact on the estimated states and consequently on the interpretation of the brain dynamics.

Machine Learning for Precision Psychiatry: Opportunities and Challenges

The nature of mental illness remains a conundrum. Traditional disease categories are increasingly suspected to misrepresent the causes underlying mental disturbance. Yet psychiatrists and investigators now have an unprecedented opportunity to benefit from complex patterns in brain, behavior, and genes using methods from machine learning (e.g., support vector machines, modern neural-network algorithms, cross-validation procedures). Combining these analysis techniques with a wealth of data from consortia and repositories has the potential to advance a biologically grounded redefinition of major psychiatric disorders. Increasing evidence suggests that data-derived subgroups of psychiatric patients can better predict treatment outcomes than DSM/ICD diagnoses can. In a new era of evidence-based psychiatry tailored to single patients, objectively measurable endophenotypes could allow for early disease detection, individualized treatment selection, and dosage adjustment to reduce the burden of disease. This primer aims to introduce clinicians and researchers to the opportunities and challenges in bringing machine intelligence into psychiatric practice.

Classical Statistics and Statistical Learning in Imaging Neuroscience

Brain-imaging research has predominantly generated insight by means of classical statistics, including regression-type analyses and null-hypothesis testing using t-test and ANOVA. Throughout recent years, statistical learning methods enjoy increasing popularity especially for applications in rich and complex data, including cross-validated out-of-sample prediction using pattern classification and sparsity-inducing regression. This concept paper discusses the implications of inferential justifications and algorithmic methodologies in common data analysis scenarios in neuroimaging. It is retraced how classical statistics and statistical learning originated from different historical contexts, build on different theoretical foundations, make different assumptions, and evaluate different outcome metrics to permit differently nuanced conclusions. The present considerations should help reduce current confusion between model-driven classical hypothesis testing and data-driven learning algorithms for investigating the brain with imaging techniques.

The independent influences of age and education on functional brain networks and cognition in healthy older adults

Healthy aging is accompanied by a constellation of changes in cognitive processes and alterations in functional brain networks. The relationships between brain networks and cognition during aging in later life are moderated by demographic and environmental factors, such as prior education, in a poorly understood manner. Using multivariate analyses, we identified three latent patterns (or modes) linking resting-state functional connectivity to demographic and cognitive measures in 101 cognitively normal elders. The first mode (P = 0.00043) captures an opposing association between age and core cognitive processes such as attention and processing speed on functional connectivity patterns. The functional subnetwork expressed by this mode links bilateral sensorimotor and visual regions through key areas such as the parietal operculum. A strong, independent association between years of education and functional connectivity loads onto a second mode (P = 0.012), characterized by the involvement of key hub regions. A third mode (P = 0.041) captures weak, residual brain-behavior relations. Our findings suggest that circuits supporting lower level cognitive processes are most sensitive to the influence of age in healthy older adults. Education, and to a lesser extent, executive functions, load independently onto functional networks-suggesting that the moderating effect of education acts upon networks distinct from those vulnerable with aging. This has important implications in understanding the contribution of education to cognitive reserve during healthy aging. Hum Brain Mapp 38:5094-5114, 2017. © 2017 Wiley Periodicals, Inc.