Correction of spatially dependent phase shifts for partial Fourier imaging.

Published

Journal Article

Partial Fourier MR images (PFI) are constructed from data that have fewer phase encoding views than are conventionally acquired using direct Fourier transform spin echo acquisition. The PFI data acquisition is structured to obtain the same spatial resolution as conventional acquisition, trading off signal-to-noise reduction for acquisition time improvement. The "missing" views can be zero filled or, if the data are Hermitian, supplied by symmetry (basic algorithm). The effect of spatially dependent phase shifts (SDPS) on PFI constructed with zero-fill or the basic algorithm is illustrated. The causes and typical magnitudes of such SDPS are discussed. In spin echo data only the low order, slowly varying SDPS, is shown to be significant. Through use of simulated and actual data sets these typical SDPS are shown to produce significant artifacts in PFI, when the amount of missing data is close to one-half. The artifacts are reduced when less data are missing. Good images can be generated with the zero-fill algorithm if less than 25% of the data is missing. Several methods of correcting phase shifts in PFI are developed: the basic Hermitian algorithm with frequency (x) direction correction (BAX), basic Fourier correction algorithm (BFC) and an improved iterative Fourier correction algorithm (IFC). The BFC and IFC can produce good images when as much as 46% of the data is missing. Data with rapidly varying SDPS, for example, gradient refocused data, make the phase correction task more difficult. With less than 25% of the data missing, however, acceptable gradient refocused PFI images can be created.

Full Text

Cited Authors

  • MacFall, JR; Pelc, NJ; Vavrek, RM

Published Date

  • March 1988

Published In

Volume / Issue

  • 6 / 2

Start / End Page

  • 143 - 155

PubMed ID

  • 3374286

Pubmed Central ID

  • 3374286

International Standard Serial Number (ISSN)

  • 0730-725X

Language

  • eng

Conference Location

  • Netherlands