TARQUIN full changelog

Tentative support for DICOM, SPAR/SDAT, IMA and RDA formats for 3D MRSI.
Dynamics scans can now be filtered prior to averaging by specifying a csv file (--av_list). Useful for fMRS studies.
Bug fix for GUI where program crashes following fitting.
Bug fix for GE p-files Rdb reader rev : 14.3 and 16.0 where the offset to data was incorrect by one frame. This caused a crash at the reconstruction stage.

FIDs can be truncated prior to pre-processing to remove unnecessary noisy data points - saving memory and disk space for large MRSI datasets (--trunc_pts).
Improvement to reduce memory use when exporting large csv fit data files.
Bug fix where Cit was included in the default basis set for the GUI.
Bug fix for GE p-files where the offset to data was incorrect for rdb_header_rev 24,20. - Thanks to Felix Raschke for pointing this out.

FIDs can be truncated prior to pre-processing to remove unnecessary noisy data points - saving memory and disk space for large MRSI datasets (--trunc_pts).
Improvement to reduce memory use when exporting large csv fit data files.
Bug fix where Cit was included in the default basis set for the GUI.
Bug fix for GE p-files where the offset to data was incorrect for rdb_header_rev 24,20. - Thanks to Felix Raschke for pointing this out.
Modest SNR improvement for multi-channel GE SVS coil combination (now weighted by water amplitude where available). Thanks to Felix Raschke and Jamie Near for letting me know that this is a better approach than un-weighted combination.
Improvements to default MRI window level and window width parameters.
Improvements to SPAR/SDAT SVS and 2D MRSI display on MRI.
The number of components used in the HSVD water removal can now be specified (--hsvd_comps).

Bug fix should improve MRSI water removal quality.

Bug fix should improve MRSI water removal quality.

2D MRSI gui visualisation is much improved:
· New interactive grid display.
· SVS and 2D MRSI VOI's can now be overlayed with MRI. This should work for Philips/Siemens DICOM and Siemens IMA.
· Simple metabolite ratio maps can now be displayed.
· Water removal has been improved for spectra where the water peak is not at 0Hz.
· Eddy current correction is now applied before water removal, this improves the water removal quality in some cases.
· Fitting is now approx. 4x faster.

· Added the option to export and read LCModel basis files (details given in the user guide
· Bug fix for DICOM MRS with explicit sequence lengths.
· New packaging method for Linux to fix compatibility issues with newer distributions.
· Corrected peak shift output to be the correct sign.
· Added some QA type parameters to csv output.
· Updated eigen to version 3.1.1.
· Tuned soft constraints to give an improved baseline for some short TE spectra.
· Bug fix for reading Bruker data where the data file is not called quot;fidquot;
· Bug fix for old Philips SPAR/SDAT files with quot;SUN_dim1_extquot; style parameter names.
· Added svs_only parameter to c/l. When set to true, TARQUIN will exit rather than try and fit data where more than one fid is in the file. This can be handy when MRSI datasets mistakenly end up in SVS batch jobs.
· Added write_raw_w option to c/l. This allows automated generation of ascii format SVS data.
· Added final Phi0 and Phi1 values to pdf and csv output.

· A TARQUIN application bundle has been developed for Mac OS X, see the documentation for how to use the command-line version in OS X.
· Basis set shifts and dampings are now exported in the CSV results file.
· Dynamic frequency shifts are now exported in the CSV results file when correction is performed.
· Regression fix where an odd number of dynamic FIDs could not be loaded.
· Regression fix where jmrui txt files could not be loaded.
· Crash fix for k-space zero-filling a non-square acquisition.
· Better error handling for bad CSV basis files.
· DICOM MRS Columns and Rows tags now take precedence over SpectroscopyAcuisitionColumns and SpectroscopyAcuisitionRows.
· Once data has been successfully loaded in the gui, a new data set can not be loaded quot;over the topquot;. This should prevent some crashes.
· gnuplot_cex command line parameter added. For some reason font sizes are differed between gnuplot 4.4 and 4.6 so this option has been added.

· Added editing for PRESS TE1, STEAM TM and CPMG pulse number parameters. Bug fix for incorrect DC offset correction for Varian data.
· Crash fix for DCM MRSI data that also contains water unsuppressed data.
· Added options to perform dynamic frequency correction.
· Bug fix for phasing water unsuppressed data.
· Added water unsuppressed freq to output.
· GE SVS quot;framesquot; are now treated as dynamic scans and can be fit individually or just averaged and fit as normal.
· When there is a mismatch between the number of water suppressed and water unsuppressed FIDs the water unsuppressed FIDs are averaged and then duplicated to match the number of water suppressed FIDs.

· init_beta estimate has been improved which should result in slightly improved fitting for 1H brain
· Initial shift search resolution has been increased
· Initial shift is now plotted in the gui
· Zero-filling factor can now be changed in the GUI (defaults to no zero-filling)
· x and y scales are preserved when moving between voxels and other options, press 'x' or 'y' to reset to match the data
· Zero-filling can now be applied to k-space for 2D CSI
· Fixed a bug where init_beta was not optimized for each voxel for CSI data
· Text format spectra can now be exported prior to fitting
· Water unsuppressed linewidth is now measured where possible and exported in the txt and csv output