Phase Sensitive Reconstruction in Diffusion Spectrum Imaging Enabling Velocity Encoding and Unbiased Noise Distribution

Jonathan I. Sperl¹, Ek T. Tan², Tim Sprenger¹, ³, Vladimir Golkov¹, ³, Kevin F. King⁴, Christopher J. Hardy², Luca Marinelli², Marion I. Menzel¹

Standard diffusion MRI data processing is based on the magnitude of the data, while the phase is neglected. However, valuable information about coherent motion like flow or pulsatility is encoded in the phase. Moreover, by separating the phase information from the data, subsequent processing like tensor fitting or fiber tracking can be done based on the real part of the data avoiding the bias introduced by the Rician noise distribution of the magnitude data. This work presents a robust workflow for a phase sensitive reconstruction of DSI data allowing for the extraction of velocity components and bias-free diffusion information.