A review of diffusion MRI in mood disorders: mechanisms and predictors of treatment response

Abstract

By measuring the molecular diffusion of water molecules in brain tissue, diffusion MRI (dMRI) provides unique insight into the microstructure and structural connections of the brain in living subjects. Since its inception, the application of dMRI in clinical research has expanded our understanding of the possible biological bases of psychiatric disorders and successful responses to different therapeutic interventions. Here, we review the past decade of diffusion imaging-based investigations with a specific focus on studies examining the mechanisms and predictors of therapeutic response in people with mood disorders. We present a brief overview of the general application of dMRI and key methodological developments in the field that afford increasingly detailed information concerning the macro- and micro-structural properties and connectivity patterns of white matter (WM) pathways and their perturbation over time in patients followed prospectively while undergoing treatment. This is followed by a more in-depth summary of particular studies using dMRI approaches to examine mechanisms and predictors of clinical outcomes in patients with unipolar or bipolar depression receiving pharmacological, neurostimulation, or behavioral treatments. Limitations associated with dMRI research in general and with treatment studies in mood disorders specifically are discussed, as are directions for future research. Despite limitations and the associated discrepancies in findings across individual studies, evolving research strongly indicates that the field is on the precipice of identifying and validating dMRI biomarkers that could lead to more successful personalized treatment approaches and could serve as targets for evaluating the neural effects of novel treatments.

Fig. 1. Diffusion-based models and methods.

a Diagrams illustrating some of the most common models used to extract dMRI metrics at the voxel level (from left to right): single diffusion tensor imaging (DTI) model depicting anisotropic (green) and isotropic (blue) water movement; multi-tensor fiber orientation distribution function model; Neurite orientation dispersion and density imaging (NODDI, adopted from [207]) [24] and free water tissue fraction models as compared to DTI and diffusion kurtosis imaging (DKI). b Post-processing methods of extracting diffusion metrics (from left to right): DTI skeleton maps using tract-based spatial statistics (TBSS [25]); JHU WM atlases for defining track-specific dMRI metrics [27, 66, 208]; whole-brain tractography using streamline representation; network modeling using graph theory; region-to-region structural connectome mapping. Tensor models and JHU atlas adapted from [209, 210].

Fig. 2. Summary of study inclusion for review.

The PRISMA (preferred reporting items for systematic reviews and meta-analyses) method was used to identify and evaluate original research articles for inclusion in this review.

Fig. 3. Major white matter tract pathways represented as tractography streamlines.

Though different WM atlases are available [211], here we provide a reference to the WM tracts most commonly used in dMRI studies of treatment in MDD and BD. Major association fibers include AF arcuate fasciculus, CG cingulum bundle, CgH parahippocampal cingulate, IFOF inferior fronto-occipital fasciculus, ILF inferior longitudinal fasciculus, SLF superior longitudinal fasciculus, and UNC uncinate fasciculus. Other association, projection, and commissural fibers include AR acoustic radiations, CST corticospinal tract, Fm forceps minor (genu of the corpus callosum), FM forceps major (splenium of the corpus callosum), fornix, MCP middle cerebral peduncle, OR option radiations, SCP superior cerebral peduncle. Thalamic radiations include ATR anterior thalamic radiations, STR superior thalamic radiations, and PTR posterior thalamic radiations. Corona radiata are split into anterior (ACR) and posterior (PCR) corona radiata. Tract atlas created from [212].