||<100% bgcolor="#FFFF99">'''Using a field-map to undistort EPIs'''<
>[[FmBackground|Background]]<
>[[FmPower|Improvement in power!]]<
>[[FmFieldmaps|Acquiring fieldmaps]]<
>[[FmDownload|Download]]<
>[[FmRequirements|What you'll need to undistort an EPI data series]]<
>[[FmReference|Guide and reference]]<
>[[FmExample|Step-by-step example]]<
>[[FmBatch|Batch mode]]|| == Siemens step-by-step example == First, get the [[http://imaging.mrc-cbu.cam.ac.uk/downloads/fmu/fmu_siemens_exampledataset.tar|Siemens example dataset (2MB)]] (thanks to Michael Benharrosh and Marlene Richter at Princeton University for this). Unzip it and you should have three images in analyze format: an EPI, and the phase and magnitude from a gre_fieldmapping sequence. When you've installed the tool, at the matlab prompt type >> fieldmap_undistort_v3 Then, in the SPM window, choose the following options ||Accept all defaults||No|| ||Start at which stage||Reco ''(default)''|| ||Which option?||Siemens gre_fieldmapping|| ||Realign fieldmaps to com||Yes ''(default)''|| ||Shift EPI forward by||0 ''(default)''|| ||Evolution time||2.46 ''(default)''|| ||Slice angle of EPIs relative to fieldmaps||0 ''(default)''|| ||Bandwidth||2604|| ||Axis of phase encoding (undistortion)||y ''(default)''|| ||Constrain coregistration||Translations only ''(default)''|| ||Type of interpolation||sinc ''(default)''|| ||Maximum shift allowed||9999 ''(default)''|| ||Adjust mean displacement?||Force mean to zero ''(default)''|| ||Directory for fieldmap files||Choose a blank directory (e.g., fieldmaps)|| ||Choose GRE phase map||PHANTOM_FIELDMAP_PHASE.img|| ||Choose GRE magnitude map||PHANTOM_FIELDMAP_MAG.img|| ||Choose EPI to coregister to||PHANTOM_EPI.img|| ||Choose EPIs to undistort||PHANTOM_EPI.img|| The EPI before and after undistortion should look like this: || {{http://imaging.mrc-cbu.cam.ac.uk/images/fm_phantom_before_undist.jpg}} || {{http://imaging.mrc-cbu.cam.ac.uk/images/fm_phantom_after_undist.jpg}} || There's nothing that can be done about the dropout, but even the large distortion produced by this mis-shimming is mostly corrected. If you are going to collect your own data on a Siemens machine and your system doesn't allow you to reconstruct both mag and phase simultaneously, you'll need to change one line in your pulse sequence. Contact Stefan Thesen at Siemens [his surname] @siemens.com who'll probably be able to help.) == Bruker step-by-step example == First, get the [[http://imaging.mrc-cbu.cam.ac.uk/downloads/fmu/fmu_exampledataset.tar|Bruker example dataset (34MB)]] Run by typing fieldmap_undistort at the SPM prompt, and choose the following options. Where requested, choose the directories and files from your copy of the data. ||Accept all defaults: ||<^ 62%>Yes || ||<^ 38%>Choose Directory of Phase Map: ||<^ 62%>example_dataset/phasemaps || ||<^ 38%>Choose Directory of Phase 1a: ||<^ 62%>example_dataset/rawprereco/10 || ||<^ 38%>Choose Directory of Phase 1b: ||<^ 62%>example_dataset/rawprereco/11 || ||<^ 38%>Choose Directory of Phase 1c: ||<^ 62%>example_dataset/rawprereco/12 || ||<^ 38%>Choose EPI header file (IMND or brkhdr): ||<^ 62%>example_dataset/epis/000484 m03_04_EPI.brkhdr || ||<^ 38%>Choose EPI to register to: ||<^ 62%>example_dataset/epis/meany000484-m03_04_EPI_0002.img || ||<^ 38%>Choose EPI to undistort: ||<^ 62%>example_dataset/epis/meany000484-m03_04_EPI_0002.img || The undistorted EPIs will be in the same directory as the raw EPIs, with the added suffix ā€˜uā€™.