Resting-state functional connectivity patterns associated with childhood maltreatment in a large bicentric cohort of adults with and without major depression.

BACKGROUND: Childhood maltreatment (CM) represents a potent risk factor for major depressive disorder (MDD), including poorer treatment response. Altered resting-state connectivity in the fronto-limbic system has been reported in maltreated individuals. However, previous results in smaller samples differ largely regarding localization and direction of effects. METHODS: We included healthy and depressed samples [n = 624 participants with MDD; n = 701 healthy control (HC) participants] that underwent resting-state functional MRI measurements and provided retrospective self-reports of maltreatment using the Childhood Trauma Questionnaire. A-priori defined regions of interest [ROI; amygdala, hippocampus, anterior cingulate cortex (ACC)] were used to calculate seed-to-voxel connectivities. RESULTS: No significant associations between maltreatment and resting-state connectivity of any ROI were found across MDD and HC participants and no interaction effect with diagnosis became significant. Investigating MDD patients only yielded maltreatment-associated increased connectivity between the amygdala and dorsolateral frontal areas [pFDR < 0.001; η2partial = 0.050; 95%-CI (0.023-0.085)]. This effect was robust across various sensitivity analyses and was associated with concurrent and previous symptom severity. Particularly strong amygdala-frontal associations with maltreatment were observed in acutely depressed individuals [n = 264; pFDR < 0.001; η2partial = 0.091; 95%-CI (0.038-0.166)). Weaker evidence - not surviving correction for multiple ROI analyses - was found for altered supracallosal ACC connectivity in HC individuals associated with maltreatment. CONCLUSIONS: The majority of previous resting-state connectivity correlates of CM could not be replicated in this large-scale study. The strongest evidence was found for clinically relevant maltreatment associations with altered adult amygdala-dorsolateral frontal connectivity in depression. Future studies should explore the relevance of this pathway for a maltreated subgroup of MDD patients.

Cognitive performance and brain structural connectome alterations in major depressive disorder.

BACKGROUND: Cognitive dysfunction and brain structural connectivity alterations have been observed in major depressive disorder (MDD). However, little is known about their interrelation. The present study follows a network approach to evaluate alterations in cognition-related brain structural networks. METHODS: Cognitive performance of n = 805 healthy and n = 679 acutely depressed or remitted individuals was assessed using 14 cognitive tests aggregated into cognitive factors. The structural connectome was reconstructed from structural and diffusion-weighted magnetic resonance imaging. Associations between global connectivity strength and cognitive factors were established using linear regressions. Network-based statistics were applied to identify subnetworks of connections underlying these global-level associations. In exploratory analyses, effects of depression were assessed by evaluating remission status-related group differences in subnetwork-specific connectivity. Partial correlations were employed to directly test the complete triad of cognitive factors, depressive symptom severity, and subnetwork-specific connectivity strength. RESULTS: All cognitive factors were associated with global connectivity strength. For each cognitive factor, network-based statistics identified a subnetwork of connections, revealing, for example, a subnetwork positively associated with processing speed. Within that subnetwork, acutely depressed patients showed significantly reduced connectivity strength compared to healthy controls. Moreover, connectivity strength in that subnetwork was associated to current depressive symptom severity independent of the previous disease course. CONCLUSIONS: Our study is the first to identify cognition-related structural brain networks in MDD patients, thereby revealing associations between cognitive deficits, depressive symptoms, and reduced structural connectivity. This supports the hypothesis that structural connectome alterations may mediate the association of cognitive deficits and depression severity.

[Big Data, AI and Machine Learning for Precision Psychiatry: How are they changing the clinical practice?] / Big Data' KI und Maschinenlernen auf dem Weg zur Precision-Psychiatry ­ wie verändern sie den therapeutischen Alltag?

Currently, we are witnessing an increasing interest in predictive models and personalized diagnosis and treatment choice in psychiatric research. Against this background, the emerging field of Precision Psychiatry is trying to establish precise diagnostics and personalized therapy through Big Data. Electronic Health Records (EHR), smartphone-based data collection and advances in genotyping and imaging allow for a detailed clinical and neurobiological characterization of numerous patients. In order to revolutionize the treatment of psychiatric disorders, a personalization of psychiatry through machine learning (ML) and artificial intelligence (AI) is needed. We must therefore establish an AI ecosystem to develop and strictly validate custom-tailored AI and ML solutions. Furthermore, personalized predictions and detailed patient information must be integrated in AI-based Clinical Decision Support systems. Only in this way can Big Data, ML and AI support the clinician most effectively and help personalize treatment in psychiatry.

Shared and Specific Patterns of Structural Brain Connectivity Across Affective and Psychotic Disorders.

BACKGROUND: Altered brain structural connectivity has been implicated in the pathophysiology of psychiatric disorders including schizophrenia (SZ), bipolar disorder (BD), and major depressive disorder (MDD). However, it is unknown which part of these connectivity abnormalities are disorder specific and which are shared across the spectrum of psychotic and affective disorders. We investigated common and distinct brain connectivity alterations in a large sample (N = 1743) of patients with SZ, BD, or MDD and healthy control (HC) subjects. METHODS: This study examined diffusion-weighted imaging-based structural connectome topology in 720 patients with MDD, 112 patients with BD, 69 patients with SZ, and 842 HC subjects (mean age of all subjects: 35.7 years). Graph theory-based network analysis was used to investigate connectome organization. Machine learning algorithms were trained to classify groups based on their structural connectivity matrices. RESULTS: Groups differed significantly in the network metrics global efficiency, clustering, present edges, and global connectivity strength with a converging pattern of alterations between diagnoses (e.g., efficiency: HC > MDD > BD > SZ, false discovery rate-corrected p = .028). Subnetwork analysis revealed a common core of edges that were affected across all 3 disorders, but also revealed differences between disorders. Machine learning algorithms could not discriminate between disorders but could discriminate each diagnosis from HC. Furthermore, dysconnectivity patterns were found most pronounced in patients with an early disease onset irrespective of diagnosis. CONCLUSIONS: We found shared and specific signatures of structural white matter dysconnectivity in SZ, BD, and MDD, leading to commonly reduced network efficiency. These results showed a compromised brain communication across a spectrum of major psychiatric disorders.

PHOTONAI-A Python API for rapid machine learning model development.

PHOTONAI is a high-level Python API designed to simplify and accelerate machine learning model development. It functions as a unifying framework allowing the user to easily access and combine algorithms from different toolboxes into custom algorithm sequences. It is especially designed to support the iterative model development process and automates the repetitive training, hyperparameter optimization and evaluation tasks. Importantly, the workflow ensures unbiased performance estimates while still allowing the user to fully customize the machine learning analysis. PHOTONAI extends existing solutions with a novel pipeline implementation supporting more complex data streams, feature combinations, and algorithm selection. Metrics and results can be conveniently visualized using the PHOTONAI Explorer and predictive models are shareable in a standardized format for further external validation or application. A growing add-on ecosystem allows researchers to offer data modality specific algorithms to the community and enhance machine learning in the areas of the life sciences. Its practical utility is demonstrated on an exemplary medical machine learning problem, achieving a state-of-the-art solution in few lines of code. Source code is publicly available on Github, while examples and documentation can be found at www.photon-ai.com.