PASTIS stands for (P)rocessing (AS)sessment (T)echnique for (I)mproved (S)pectroscopy 😉

PASTIS is a Python package that can be used to process and quantify single-voxel Magnetic Resonance SPectroscopy (MRS) data. It can also simulate MRS data using various MRS sequences for different B0 fields. It was originally developed to reconstruct, process and quantify spinal cord MRS data at 7 T and has therefore special features related to motion detection and compensation. For these reasons, it was later on used for cardiac MRS but could work for any organ really 🙂

PASTIS relies a lot on the suspect package. More info here:


PASTIS was originally written by Tangi Roussel. If you are using PASTIS or part of it in your work, please cite the original paper:

Respiratory-triggered quantitative MR spectroscopy of the human cervical spinal cord at 7 T.
Roussel T, Le Fur Y, Guye M, Viout P, Ranjeva JP and Callot V. Magn Reson Med. 2022

Main features

  • Read raw data from Siemens (TWIX MR Syngo VB17 & VE11)
  • Read raw data from Bruker (fid files)
  • Read dicom data (standard dicom, Siemens MR Syngo VB17, Siemens MR Syngo XA20)
  • Read and write NIFTI MRS files
  • VOI overlay on anatomical image
  • Data processing
    • Automatic phasing
    • Automatic channel combination
    • Zero-filling
    • Fully automatic data discard for SNR and FWHM enhancement
    • Automatic frequency realignment
    • Apodization
    • Peak HSVD removal
    • Signal and noise estimation
    • Linewidth estimation
    • Spectral display
  • Data simulation
    • Based on GAMMA library
    • PRESS, STEAM, sLASER sequences
    • Including real RF pulse shapes for sLASER
    • Fully editable metabolite basis set
    • Macromolecular baseline modelization
    • Linear combination time-domain model
    • Possible to save/load metabolite simulation signals
  • Quantification
    • Based on the previous model
    • Dynamic model options
      • Parameters can be freezed
      • Parameters can be linked to each other
    • Jacobian matrix information
    • Cramér-Rao Bounds estimations
  • Quantification using suspect’s LCModel wrapper
  • Dataframe-based storage in pkl files
  • Reconstruction and quantification of diffusion-weighted MRS data (experimental)


PASTIS is still under development and is available at:

See the file for more details about installation and usage.