On Tue, 1 Apr 2008 21:53:45 +0100, Adrian Koziak <[log in to unmask]> wrote:
>Hello all,
>
>I have been attempting to apply B0 unwarping using a modified ucsd script
>(epidewarp) and B0 unwarping within fugue. We use an 8-channel head coil on
>a GE Signa HDX system. After finally obtaining properly scaled phase images
>from the scanner (many user CV's later), I re-ran both my modified ucsd
>script and fugue within feat, however, I have yet to obtain artifact free
>images. I have re-run tests on a single channel coil and the results seem to
>work quite well (I have used the fieldmap reconstruction steps as
>recommended in
>http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0803&L=FSL&P=R70049&D=0&I=-3&X=189EFD2A907E02D552&Y)
>Also, I have attached the prestat reports from FEAT for both the single
>channel and multi 8-channel acquisitions.
>
>I have a feeling that my method for acquiring the 8-channel scaled phase
>images is flawed. For those GE users out there, this is what I have done:
>
> -use two SPGR, TR 50, FA = 45, BW = 31.25, TE 5 and 8 ms, manual prescan
>between the two echo times
> -same slice locs and FOV as EPI acquisition, however higher resolution
>(cannot sample at 64x64 therefore I interpolate to match the EPI)
> -User CV's
> - rhrcctrl = 15 (output real, imag, phasex1000, magnitude images)
> - nofermi and nograd = 1(turn off filters applied to images)
> - rhimsize = 128 (recon at 128 x 128)
> The next CV's trick GE to use ASSET, as ASSET is not available in the
>package to use on a SPGR, however, with ASSET, properly scaled phase maps
>can be collected.
> - opphaseasset_factor = 1 (do ASSET in the phase direction)
> - rhasset_R =1 (turn ASSET on, but collect all the phase information)
>
>Phase maps obtained from this method seem correct, but due to the unwarping
>results, I fear they may be innaccurate. Has anyone else had this problem ??
>I could output the phase, real, imag, and mag images from each coil and
>recombine them, but I am trying to avoid this to decrease reconstruction
>time. Also, I am not 100% certain of how to combine either the phase or real
>and magnitude images from each coil. Regardless, this process does not seem
>trivial.
>
>I would appreciate any input.
>
>Thanks very much
>
>Adrian Koziak
>Research Assistant
>Brain-Body Insitute,
>Imaging Research Centre
>St. Joseph's Healthcare
>Hamilton, ON
>
|