Development and validation of a multimodal neuroimaging biomarker for electroconvulsive therapy outcome in depression: a multicenter machine learning analysis.

BACKGROUND: Electroconvulsive therapy (ECT) is the most effective intervention for patients with treatment resistant depression. A clinical decision support tool could guide patient selection to improve the overall response rate and avoid ineffective treatments with adverse effects. Initial small-scale, monocenter studies indicate that both structural magnetic resonance imaging (sMRI) and functional MRI (fMRI) biomarkers may predict ECT outcome, but it is not known whether those results can generalize to data from other centers. The objective of this study was to develop and validate neuroimaging biomarkers for ECT outcome in a multicenter setting. METHODS: Multimodal data (i.e. clinical, sMRI and resting-state fMRI) were collected from seven centers of the Global ECT-MRI Research Collaboration (GEMRIC). We used data from 189 depressed patients to evaluate which data modalities or combinations thereof could provide the best predictions for treatment remission (HAM-D score ⩽7) using a support vector machine classifier. RESULTS: Remission classification using a combination of gray matter volume and functional connectivity led to good performing models with average 0.82-0.83 area under the curve (AUC) when trained and tested on samples coming from the three largest centers (N = 109), and remained acceptable when validated using leave-one-site-out cross-validation (0.70-0.73 AUC). CONCLUSIONS: These results show that multimodal neuroimaging data can be used to predict remission with ECT for individual patients across different treatment centers, despite significant variability in clinical characteristics across centers. Future development of a clinical decision support tool applying these biomarkers may be feasible.

The functional connectome in obsessive-compulsive disorder: resting-state mega-analysis and machine learning classification for the ENIGMA-OCD consortium.

Current knowledge about functional connectivity in obsessive-compulsive disorder (OCD) is based on small-scale studies, limiting the generalizability of results. Moreover, the majority of studies have focused only on predefined regions or functional networks rather than connectivity throughout the entire brain. Here, we investigated differences in resting-state functional connectivity between OCD patients and healthy controls (HC) using mega-analysis of data from 1024 OCD patients and 1028 HC from 28 independent samples of the ENIGMA-OCD consortium. We assessed group differences in whole-brain functional connectivity at both the regional and network level, and investigated whether functional connectivity could serve as biomarker to identify patient status at the individual level using machine learning analysis. The mega-analyses revealed widespread abnormalities in functional connectivity in OCD, with global hypo-connectivity (Cohen's d: -0.27 to -0.13) and few hyper-connections, mainly with the thalamus (Cohen's d: 0.19 to 0.22). Most hypo-connections were located within the sensorimotor network and no fronto-striatal abnormalities were found. Overall, classification performances were poor, with area-under-the-receiver-operating-characteristic curve (AUC) scores ranging between 0.567 and 0.673, with better classification for medicated (AUC = 0.702) than unmedicated (AUC = 0.608) patients versus healthy controls. These findings provide partial support for existing pathophysiological models of OCD and highlight the important role of the sensorimotor network in OCD. However, resting-state connectivity does not so far provide an accurate biomarker for identifying patients at the individual level.

An overview of the first 5 years of the ENIGMA obsessive-compulsive disorder working group: The power of worldwide collaboration.

Neuroimaging has played an important part in advancing our understanding of the neurobiology of obsessive-compulsive disorder (OCD). At the same time, neuroimaging studies of OCD have had notable limitations, including reliance on relatively small samples. International collaborative efforts to increase statistical power by combining samples from across sites have been bolstered by the ENIGMA consortium; this provides specific technical expertise for conducting multi-site analyses, as well as access to a collaborative community of neuroimaging scientists. In this article, we outline the background to, development of, and initial findings from ENIGMA's OCD working group, which currently consists of 47 samples from 34 institutes in 15 countries on 5 continents, with a total sample of 2,323 OCD patients and 2,325 healthy controls. Initial work has focused on studies of cortical thickness and subcortical volumes, structural connectivity, and brain lateralization in children, adolescents and adults with OCD, also including the study on the commonalities and distinctions across different neurodevelopment disorders. Additional work is ongoing, employing machine learning techniques. Findings to date have contributed to the development of neurobiological models of OCD, have provided an important model of global scientific collaboration, and have had a number of clinical implications. Importantly, our work has shed new light on questions about whether structural and functional alterations found in OCD reflect neurodevelopmental changes, effects of the disease process, or medication impacts. We conclude with a summary of ongoing work by ENIGMA-OCD, and a consideration of future directions for neuroimaging research on OCD within and beyond ENIGMA.

Longitudinal resting-state network connectivity changes in electroconvulsive therapy patients compared to healthy controls.

OBJECTIVE: Electroconvulsive therapy (ECT) is effective for major depressive episodes. Understanding of underlying mechanisms has been increased by examining changes of brain connectivity but studies often do not correct for test-retest variability in healthy controls (HC). In this study, we investigated changes in resting-state networks after ECT in a multicenter study. METHODS: Functional resting-state magnetic resonance imaging data, acquired before start and within one week after ECT, from 90 depressed patients were analyzed, as well as longitudinal data of 24 HC. Group-information guided independent component analysis (GIG-ICA) was used to spatially restrict decomposition to twelve canonical resting-state networks. Selected networks of interest were the default mode network (DMN), salience network (SN), and left and right frontoparietal network (LFPN, and RFPN). Whole-brain voxel-wise analyses were used to assess group differences at baseline, group by time interactions, and correlations with treatment effectiveness. In addition, between-network connectivity and within-network strengths were computed. RESULTS: Within-network strength of the DMN was lower at baseline in ECT patients which increased after ECT compared to HC, after which no differences were detected. At baseline, ECT patients showed lower whole-brain voxel-wise DMN connectivity in the precuneus. Increase of within-network strength of the LFPN was correlated with treatment effectiveness. We did not find whole-brain voxel-wise or between-network changes. CONCLUSION: DMN within-network connectivity normalized after ECT. Within-network increase of the LFPN in ECT patients was correlated with higher treatment effectiveness. In contrast to earlier studies, we found no whole-brain voxel-wise changes, which highlights the necessity to account for test-retest effects.

Effective resting-state connectivity in severe unipolar depression before and after electroconvulsive therapy.

BACKGROUND: Electroconvulsive therapy (ECT) is one of the most effective treatments for severe depressive disorders. A recent multi-center study found no consistent changes in correlation-based (undirected) resting-state connectivity after ECT. Effective (directed) connectivity may provide more insight into the working mechanism of ECT. OBJECTIVE: We investigated whether there are consistent changes in effective resting-state connectivity. METHODS: This multi-center study included data from 189 patients suffering from severe unipolar depression and 59 healthy control participants. Longitudinal data were available for 81 patients and 24 healthy controls. We used dynamic causal modeling for resting-state functional magnetic resonance imaging to determine effective connectivity in the default mode, salience and central executive networks before and after a course of ECT. Bayesian general linear models were used to examine differences in baseline and longitudinal effective connectivity effects associated with ECT and its effectiveness. RESULTS: Compared to controls, depressed patients showed many differences in effective connectivity at baseline, which varied according to the presence of psychotic features and later treatment outcome. Additionally, effective connectivity changed after ECT, which was related to ECT effectiveness. Notably, treatment effectiveness was associated with decreasing and increasing effective connectivity from the posterior default mode network to the left and right insula, respectively. No effects were found using correlation-based (undirected) connectivity. CONCLUSIONS: A beneficial response to ECT may depend on how brain regions influence each other in networks important for emotion and cognition. These findings further elucidate the working mechanisms of ECT and may provide directions for future non-invasive brain stimulation research.

Neural Substrates of Psychotic Depression: Findings From the Global ECT-MRI Research Collaboration.

Psychotic major depression (PMD) is hypothesized to be a distinct clinical entity from nonpsychotic major depression (NPMD). However, neurobiological evidence supporting this notion is scarce. The aim of this study is to identify gray matter volume (GMV) differences between PMD and NPMD and their longitudinal change following electroconvulsive therapy (ECT). Structural magnetic resonance imaging (MRI) data from 8 independent sites in the Global ECT-MRI Research Collaboration (GEMRIC) database (n = 108; 56 PMD and 52 NPMD; mean age 71.7 in PMD and 70.2 in NPMD) were analyzed. All participants underwent MRI before and after ECT. First, cross-sectional whole-brain voxel-wise GMV comparisons between PMD and NPMD were conducted at both time points. Second, in a flexible factorial model, a main effect of time and a group-by-time interaction were examined to identify longitudinal effects of ECT on GMV and longitudinal differential effects of ECT between PMD and NPMD, respectively. Compared with NPMD, PMD showed lower GMV in the prefrontal, temporal and parietal cortex before ECT; PMD showed lower GMV in the medial prefrontal cortex (MPFC) after ECT. Although there was a significant main effect of time on GMV in several brain regions in both PMD and NPMD, there was no significant group-by-time interaction. Lower GMV in the MPFC was consistently identified in PMD, suggesting this may be a trait-like neural substrate of PMD. Longitudinal effect of ECT on GMV may not explain superior ECT response in PMD, and further investigation is needed.

Structural neuroimaging biomarkers for obsessive-compulsive disorder in the ENIGMA-OCD consortium: medication matters.

No diagnostic biomarkers are available for obsessive-compulsive disorder (OCD). Here, we aimed to identify magnetic resonance imaging (MRI) biomarkers for OCD, using 46 data sets with 2304 OCD patients and 2068 healthy controls from the ENIGMA consortium. We performed machine learning analysis of regional measures of cortical thickness, surface area and subcortical volume and tested classification performance using cross-validation. Classification performance for OCD vs. controls using the complete sample with different classifiers and cross-validation strategies was poor. When models were validated on data from other sites, model performance did not exceed chance-level. In contrast, fair classification performance was achieved when patients were grouped according to their medication status. These results indicate that medication use is associated with substantial differences in brain anatomy that are widely distributed, and indicate that clinical heterogeneity contributes to the poor performance of structural MRI as a disease marker.

ENIGMA and global neuroscience: A decade of large-scale studies of the brain in health and disease across more than 40 countries.

This review summarizes the last decade of work by the ENIGMA (Enhancing NeuroImaging Genetics through Meta Analysis) Consortium, a global alliance of over 1400 scientists across 43 countries, studying the human brain in health and disease. Building on large-scale genetic studies that discovered the first robustly replicated genetic loci associated with brain metrics, ENIGMA has diversified into over 50 working groups (WGs), pooling worldwide data and expertise to answer fundamental questions in neuroscience, psychiatry, neurology, and genetics. Most ENIGMA WGs focus on specific psychiatric and neurological conditions, other WGs study normal variation due to sex and gender differences, or development and aging; still other WGs develop methodological pipelines and tools to facilitate harmonized analyses of "big data" (i.e., genetic and epigenetic data, multimodal MRI, and electroencephalography data). These international efforts have yielded the largest neuroimaging studies to date in schizophrenia, bipolar disorder, major depressive disorder, post-traumatic stress disorder, substance use disorders, obsessive-compulsive disorder, attention-deficit/hyperactivity disorder, autism spectrum disorders, epilepsy, and 22q11.2 deletion syndrome. More recent ENIGMA WGs have formed to study anxiety disorders, suicidal thoughts and behavior, sleep and insomnia, eating disorders, irritability, brain injury, antisocial personality and conduct disorder, and dissociative identity disorder. Here, we summarize the first decade of ENIGMA's activities and ongoing projects, and describe the successes and challenges encountered along the way. We highlight the advantages of collaborative large-scale coordinated data analyses for testing reproducibility and robustness of findings, offering the opportunity to identify brain systems involved in clinical syndromes across diverse samples and associated genetic, environmental, demographic, cognitive, and psychosocial factors.

Diagnostic neuroimaging markers of obsessive-compulsive disorder: Initial evidence from structural and functional MRI studies.

As of yet, no diagnostic biomarkers are available for obsessive-compulsive disorder (OCD), and its diagnosis relies entirely upon the recognition of behavioural features assessed through clinical interview. Neuroimaging studies have shown that various brain structures are abnormal in OCD patients compared to healthy controls. However, the majority of these results are based on average differences between groups, which limits diagnostic usage in clinical practice. In recent years, a growing number of studies have applied multivariate pattern analysis (MVPA) techniques on neuroimaging data to extract patterns of altered brain structure, function and connectivity typical for OCD. MVPA techniques can be used to develop predictive models that extract regularities in data to classify individual subjects based on their diagnosis. In the present paper, we reviewed the literature of MVPA studies using data from different imaging modalities to distinguish OCD patients from controls. A systematic search retrieved twelve articles that fulfilled the inclusion and exclusion criteria. Reviewed studies have been able to classify OCD diagnosis with accuracies ranging from 66% up to 100%. Features important for classification were different across imaging modalities and widespread throughout the brain. Although studies have shown promising results, sample sizes used are typically small which can lead to high variance of the estimated model accuracy, cohort-specific solutions and lack of generalizability of findings. Some of the challenges are discussed that need to be overcome in order to move forward toward clinical applications.

Dose-dependent effects of the selective serotonin reuptake inhibitor citalopram: A combined SPECT and phMRI study.

BACKGROUND: Serotonin transporter blockers, like citalopram, dose-dependently bind to the serotonin transporter. Pharmacological magnetic resonance imaging (phMRI) can be used to non-invasively monitor effects of serotonergic medication. Although previous studies showed that phMRI can measure the effect of a single dose of serotoninergic medication, it is currently unclear whether it can also detect dose-dependent effects. AIMS: To investigate the dose-dependent phMRI response to citalopram and compared this with serotonin transporter occupancy, measured with single photon emission computed tomography (SPECT). METHODS: Forty-five healthy females were randomized to pre-treatment with placebo, a low (4 mg) or clinically standard (16 mg) oral citalopram dose. Prior to citalopram, and 3 h after, subjects underwent SPECT scanning. Subsequently, a phMRI scan with a citalopram challenge (7.5 mg intravenously) was conducted. Change in cerebral blood flow in response to the citalopram challenge was assessed in the thalamus and occipital cortex (control region). RESULTS: Citalopram dose-dependently affected serotonin transporter occupancy, as measured with SPECT. In addition, citalopram dose-dependently affected the phMRI response to intravenous citalopram in the thalamus (but not occipital cortex), but phMRI was less sensitive in distinguishing between groups than SPECT. Serotonin transporter occupancy showed a trend-significant correlation to thalamic cerebral blood flow change. CONCLUSION: These results suggest that phMRI likely suffers from higher variation than SPECT, but that these techniques probably also assess different functional aspects of the serotonergic synapse; therefore phMRI could complement positron emission tomography/SPECT for measuring effects of serotonergic medication.

Individual Prediction of Behavioral Variant Frontotemporal Dementia Development Using Multivariate Pattern Analysis of Magnetic Resonance Imaging Data.

BACKGROUND: Patients with behavioral variant of frontotemporal dementia (bvFTD) initially may only show behavioral and/or cognitive symptoms that overlap with other neurological and psychiatric disorders. The diagnostic accuracy is dependent on progressive symptoms worsening and frontotemporal abnormalities on neuroimaging findings. Predictive biomarkers could facilitate the early detection of bvFTD. OBJECTIVE: To determine the prognostic accuracy of clinical and structural MRI data using a support vector machine (SVM) classification to predict the 2-year clinical follow-up diagnosis in a group of patients presenting late-onset behavioral changes. METHODS: Data from 73 patients were included and divided into probable/definite bvFTD (n = 18), neurological (n = 28), and psychiatric (n = 27) groups based on 2-year follow-up diagnosis. Grey-matter volumes were extracted from baseline structural MRI scans. SVM classifiers were used to perform three binary classifications: bvFTD versus neurological and psychiatric, bvFTD versus neurological, and bvFTD versus psychiatric group(s), and one multi-class classification. Classification performance was determined for clinical and neuroimaging data separately and their combination using 5-fold cross-validation. RESULTS: Accuracy of the binary classification tasks ranged from 72-82% (p < 0.001) with adequate sensitivity (67-79%), specificity (77-88%), and area-under-the-receiver-operator-curve (0.80-0.9). Multi-class accuracy ranged between 55-59% (p < 0.001). The combination of clinical and voxel-wise whole brain data showed the best performance overall. CONCLUSION: These results show the potential for automated early confirmation of diagnosis for bvFTD using machine learning analysis of clinical and neuroimaging data in a diverse and clinically relevant sample of patients.