Failed import of Bruker ISIS 31P MRS data into jMRUI

Ask Anything!
1

Hi all,

I am working with in vivo mouse cardiac 31P MRS data acquired on a Bruker 9.4T system using an ISIS spectroscopy sequence.

The problem is that Bruker ParaVision 360 3.6 clearly shows multiple 31P spectral peaks, but after importing the raw data into jMRUI 5.0 using SERload, the spectrum appears essentially flat, with no meaningful peaks.

What I tried:

  • jMRUI SERload / Bruker SER format
  • importing rawdata.job0
  • checking method and acqp parameters manually
  • attempting Python-based conversion of the raw spectroscopy data into a jMRUI-readable text format

My main question:

  1. Why does ParaVision show clear reconstructed ³¹P real peaks, while jMRUI shows a flat spectrum after SERload import?
  2. What is the correct Bruker PV360 ISIS → jMRUI import workflow? Must the ISIS combination be performed before the jMRUI import?
  3. Whether there is a Python workflow to convert the PV360 ISIS raw data directly into a jMRUI-readable text file. While keeping the spectrum consistent with the real peaks shown in ParaVision, so that phase correction can later be performed in jMRUI.
  4. I would also appreciate suggestions for how such a conversion prompt/workflow should be formulated in Codex or other AI coding tools to generate the correct text-conversion script.

My current hypotheses are:

  1. ParaVision may perform ISIS combination/reconstruction internally before displaying the spectrum. but the rawdata.job0 file may still contain separate ISIS acquisitions/subscans that need to be combined manually.
  2. jMRUI SERload may not correctly interpret the PV360 ISIS raw data structure. The issue may be related to complex data dealing. Then we need to use Python to convert the data into text format.

System:

  • Scanner: Bruker 9.4T
  • ParaVision: PV360 3.6
  • jMRUI: 5.0
  • Sequence: ISIS spectroscopy
  • Nucleus: 31P
  • OS: Windows

Dataset / key parameters:

  • Single voxel ISIS spectroscopy
  • Voxel size: 6 × 6 × 6 mm
  • Spectral points: 2048
  • Bandwidth: 7936.5079 Hz
  • TR: 4000 ms
  • Averages: 78
  • ISIS averages reported: 624
  • BYTORDA: little
  • BF1 / SFO1: 162.04141926086106 MHz
  • ACQ_size: 1024
  • NI: 1
  • NR: 1

Files available:

  • method
  • acqp
  • rawdata.job0
  • pulseprogram.precomp
  • shimcondition

Attached screenshots:
jMRUI SERload result showing an essentially flat spectrum after importing the same dataset

Jmrui_SER
Jmrui_SER2526×1657 268 KB

Thanks!