'Time-Series' Analysis of the Diffusion Weighted Signal as a Model-Free Approach to Segmenting Tissue
Jones D, Deoni S
Institute of Psychiatry, King's College
In this work we investigate the use of a pseudo time-series analysis approach to characterize and segment diffusion-weighted (DW) MRI data. By treating the diffusion data acquired over a range of encoding gradients as a time-series wave-form we hypothesize that similar tissue (shape and orientation) will share common characteristics in their time-series and therefore be maximally correlated. As proof of principle, we apply this approach to in vivo thalamic data to segment the thalamic nuclei and show that this simple approach, free of model assumptions, provides reliable delineation of the DW-MRI data.