Establishing Disorder-Specific and Transdiagnostic Neural Features of Psychiatric Disorders Through Large-Scale Functional Magnetic Resonance Imaging Meta-Analyses

Abstract

Introduction

Meta-analyses of functional magnetic resonance imaging (fMRI) studies have been used to elucidate the most reliable neural features associated with various psychiatric disorders. However, it has not been well-established whether each of these neural features is linked to a specific disorder or is transdiagnostic across multiple disorders and disorder categories, including mood, anxiety, and anxiety-related disorders.

Objectives

This project aims to advance our understanding of the disorder-specific and transdiagnostic neural features associated with mood, anxiety, and anxiety-related disorders as well as to refine the methodology used to compare multiple disorders.

Methods

We conducted an exhaustive PubMed literature search followed by double-screening, double-extraction, and cross-checking to identify all whole-brain, case-control fMRI activation studies of mood, anxiety, and anxiety-related disorders in order to construct a large-scale meta-analytic database of primary studies of these disorders. We then employed multilevel kernel density analysis (MKDA) with Monte-Carlo simulations to correct for multiple comparisons as well as ensemble thresholding to reduce cluster size bias to analyze primary fMRI studies of mood, anxiety, and anxiety-related disorders followed by application of triple subtraction techniques and a second-order analysis to elucidate the disorder-specificity of the previously identified neural features.

Results

We found that participants diagnosed with mood, anxiety, and anxiety-related disorders exhibited statistically significant (p < .05 – 0.0001; FWE-corrected) differences in neural activation relative to healthy controls throughout the cerebral cortex, limbic system, and basal ganglia. In addition, each of these psychiatric disorders exhibited a particular profile of neural features that ranged from disorder-specific, to category-specific, to transdiagnostic.

Conclusions

These findings indicate that psychiatric disorders exhibit a complex profile of neural features that vary in their disorder-specificity and can be detected with large-scale fMRI meta-analytic techniques. This approach has potential to fundamentally transform neuroimaging investigations of clinical disorders by providing a novel procedure for establishing disorder-specificity of observed results, which can be then used to advance our understanding of individual disorders as well as broader nosological issues related to diagnosis and classification of psychiatric disorders.

Disclosure of Interest

None Declared