Abstract
This thesis presents some innovative techniques useful for the diffusion spectrum imaging (DSI) in the sense that they can be applied together to significantly reduce the data acquisition time up to 60% theoretically.In order to shorten the data acquisition time of diffusion spectrum imaging (DSI), homogeneous q-space sampling using body-center-cubic (BCC) lattice is proposed. For spherical band-limited 3D signal, the sampling efficiency of BCC is 30% higher than that of conventional Cartesian scheme. Because probability density function (PDF) of water molecular motion of DSI is spherical band-limited, BCC sampling scheme is applicable and may reduce sampling data up to 30% theoretically, which means 30% reduction of data acquisition time.Since probability density function of water molecular motion is positive, half-Fourier method can be introduced, i.e. only q-space data within half sphere is sufficient to reconstruct PDF, for additional 40~50% of acquisition time reduction. Because there might be center-shift problem for DSI data, half-and-interlaced-spherical acquisition scheme (HIS) is proposed for correcting this artifacts when half-Fourier method is used.Discrete spectrum Fourier transform (DSFT) is proposed to reconstruct PDF directly on arbitrarily selected positions without interpolation and therefore solves the encountered sub-voxel-reconstruction problem.Experiments show that reasonable results of tractography can be achieved with 80% reduction of data acquisition time using BCC and HIS sampling scheme and DSFT reconstruction.