dtiRIM: A generalisable deep learning method for diffusion tensor imaging.

Diffusion weighted MRI is an indispensable tool for routine patient screening and diagnostics of pathology. Recently, several deep learning methods have been proposed to quantify diffusion parameters, but poor generalisation to new data prevents broader use of these methods, as they require retraining of the neural network for each new scan protocol. In this work, we present the dtiRIM, a new deep learning method for Diffusion Tensor Imaging (DTI) based on the Recurrent Inference Machines. Thanks to its ability to learn how to solve inverse problems and to use the diffusion tensor model to promote data consistency, the dtiRIM can generalise to variations in the acquisition settings. This enables a single trained network to produce high quality tensor estimates for a variety of cases. We performed extensive validation of our method using simulation and in vivo data, and compared it to the Iterated Weighted Linear Least Squares (IWLLS), the approach of the state-of-the-art MRTrix3 software, and to an implementation of the Maximum Likelihood Estimator (MLE). Our results show that dtiRIM predictions present low dependency on tissue properties, anatomy and scanning parameters, with results comparable to or better than both IWLLS and MLE. Further, we demonstrate that a single dtiRIM model can be used for a diversity of data sets without significant loss in quality, representing, to our knowledge, the first generalisable deep learning based solver for DTI.

Anatomical integration and rich-club connectivity in euthymic bipolar disorder.

BACKGROUND: Although repeatedly associated with white matter microstructural alterations, bipolar disorder (BD) has been relatively unexplored using complex network analysis. This method combines structural and diffusion magnetic resonance imaging (MRI) to model the brain as a network and evaluate its topological properties. A group of highly interconnected high-density structures, termed the 'rich-club', represents an important network for integration of brain functioning. This study aimed to assess structural and rich-club connectivity properties in BD through graph theory analyses. METHOD: We obtained structural and diffusion MRI scans from 42 euthymic patients with BD type I and 43 age- and gender-matched healthy volunteers. Weighted fractional anisotropy connections mapped between cortical and subcortical structures defined the neuroanatomical networks. Next, we examined between-group differences in features of graph properties and sub-networks. RESULTS: Patients exhibited significantly reduced clustering coefficient and global efficiency, compared with controls globally and regionally in frontal and occipital regions. Additionally, patients displayed weaker sub-network connectivity in distributed regions. Rich-club analysis revealed subtly reduced density in patients, which did not withstand multiple comparison correction. However, hub identification in most participants indicated differentially affected rich-club membership in the BD group, with two hubs absent when compared with controls, namely the superior frontal gyrus and thalamus. CONCLUSIONS: This graph theory analysis presents a thorough investigation of topological features of connectivity in euthymic BD. Abnormalities of global and local measures and network components provide further neuroanatomically specific evidence for distributed dysconnectivity as a trait feature of BD.