Vespa – Versatile Simulation, Pulses and Analysis Created via an NIH software support/maintenance grant Derived from three existing applications on the ‘sneakernet’ MatPulse – MatLab GAMMA – C++ SITools/FITT – IDL/C/Fortran URL: http://scion.duhs.duke.edu/vespa

Vespa – Versatile Simulation, Pulses and Analysis Interconnected, Python-based, applications for MRS spectral analysis Pulse designs, spectral simulations and metabolite fits all shareable via database/XML Full provenance/versioning for all design, processing and analysis steps Open source, runs on Windows, OS X and Linux Vespa Package Simulation Application Pulse Analysis Optimized RF pulse waveforms Spectral simulation metabolite basis set PRESS TE=30ms 1.5T brian.soher@duke.edu

Vespa – Pulse Application Pulse allows users to create, compare and analyze RF pulses for MR applications. The Python language allows for easy prototyping and cross platform compatibility. The GUI interface displays the RF waveform at all creation/modification steps. The integrated design/test GUI lets users contribute their own algorithms. Side by side display of two pulse designs … an SLR 90 and a BASSI hyperbolic inversion pulse

Vespa – Pulse Application Standard Results Display Interactive Algorithm Test Windows Create new pulses from a list of user defined algorithms Store/re-load pulse designs and their design parameters into a database Display results for each design step and compare side-by-side results from one or more designs Output results graphically or in MR manufacturer platform formats Import pulse designs to/from other users – or from manufacturer formats

Vespa – Simulation Application Simulation allows users to create and analyze spectral simulations. Pulse sequence simulations can use ‘real’ RF pulses from Vespa-Pulse results. The GUI interface displays the simulated spectral results and 1D and 2D analyses. Provides a user friendly front end to the GAMMA/PyGAMMA NMR library. Side by side display of two Experiments … Ideal PRESS for multiple metabolites for one TE value, and Ideal PRESS for multiple TE1 and TE2 values showing integral variations through time

Vespa – Simulation Application Metabolite Controls Interactive Pulse Sequence Test Windows Run simulated Experiment from lists of metabolites and pulse sequences Store Experiment results in a database Display the results in a plotting/graphing tool Compare side-by-side results from Experiments Export/Import results from other users Design and test your own PyGAMMA pulse sequences for addition to the list of pulse sequences available for use in Experiments

Vespa – Simulation Example Code Ideal PRESS import pygamma as pg def run(sim_desc): # This is an example PyGAMMA pulse sequence spin_system = sim_desc.spin_system obs_iso = sim_desc.observe_isotope te1 = sim_desc.dims[1] / 1000.0 te2 = sim_desc.dims[2] / 1000.0 H = pg.Hcs(spin_system) + pg.HJ(spin_system) D = pg.Fm(spin_system, obs_iso) ac = pg.acquire1D(pg.gen_op(D), H, 0.000001) ACQ = ac sigma0 = pg.sigma_eq(spin_system) # excite, propagate, refocus and acquire the data sigma1 = pg.Iypuls(spin_system, sigma0, obs_iso, 90.0) Udelay = pg.prop(H, te1*0.5) sigma0 = pg.evolve(sigma1, Udelay) sigma1 = pg.Iypuls(spin_system, sigma0, obs_iso, 180.0) Udelay = pg.prop(H, (te1+te2)*0.5) Udelay = pg.prop(H, te2*0.5) sim_desc.mx = pg.TTable1D(ACQ.table(sigma0))

Vespa-Analysis: MRS Data Processing and Quantitation Can be run interactively in GUI mode or as batch job from command line Highlights Multiple data sets can be loaded and compared Native read for Siemens, Bruker, GE, Philips and Varian MRS data User extensible via Python scripts HTML and CSV output Individual FID B0 and Phase 0 correction. (future) Inline processing of Siemens MRS data via embedded Python/Analysis within the ICE pipeline Example: Siemens Trio SVS-EDIT, raw file (twix) read/process, fitted data for both Sum and Difference spectra brian.soher@duke.edu

Vespa-Analysis: MRS Data Processing and Quantitation brian.soher@duke.edu

The Future and my Wish List Inline processing of Siemens MRS data via embedded Python/Analysis within the ICE pipeline Output to bitmap image of results Standardized semi-LASER sequence and processing on Siemens and Philips Want to wrap results into DICOM Structured Report More centralized processing location – Syngo Via.Frontier server? Also able to embed Python code NiPype – Neuroimaging in Python workflow package

Thank you!