The relationship between susceptibility weighted phase and white matter fiber orientation

Introduction: The origin of contrast in susceptibility weighted phase imaging (1) is not yet fully understood. Both grey and white matter show considerable heterogeneity on phase images (2) and white matter T2* was also found to be heterogeneous (3). Subjects and Method: Data from 2 healthy volunteers (1 female, 1 male, ages 29 and 31) were acquired on a Philips Achieva 3.0T system. SWI data were acquired with a flow-compensated 3D gradient echo method [4] (TR/TE/alpha=40/20/19, acquisition matrix = 480 x 231 x 32, reconstruction matrix 560 x 560 x 64). Phase images [7] were unwrapped [8,9] and high pass filtered. DTI data were acquired axial orientation with the following parameters: SE EPI, Voxel size = 2.2 x 2.2 2.2, FOV = 212 x 212 x 132, SENSE-factor = 2.4, TR = 7.4 s, TE = 75 ms, δ = 13.2ms, Δ = 27.4 ms, b = 0 & 1000s/mm, 16 directions, 2 averages. Magnitude and phase images were reconstructed for the SWI. Phase wraps were removed from the phase images by homodyne filtering in 2D with a Hanning window of 140 x 140 pixels. This filter also removes phase variations with low spatial frequencies The DTI data were corrected for eddy currents using FSL (6). The phase images were coregistered to the DTI data using FSL's FLIRT tool by coregistering the magnitude and applying the parameters to the phase images. DTI analysis was performed with FSL’s FDT diffusion The phase was compared to the orientation between the fibers and the main magnetic field obtained from the DTI data. To exclude noise, CSF and grey matter, only pixels with fractional anisotropy above 0.25 were evaluated. The phase of all pixels for a given fiber orientation was averaged and plotted against the fiber orientation for the range between 0 to 90 degrees. Results: A map of fiber orientation is shown in Fig 1. Blue corresponds to fibers parallel to the main magnetic field and red to fibers perpendicular to the magnetic field. The corresponding phase is shown in Fig 2. The realtionship between fiber orientation and phase was similar for both subjects, except for a constant offset. Phase was negative for fibers parallel and perpendicular to Bo and positive for fibres at intermediate angles