Does the MRSI toolbox support multi-peak fitting?

[SuhaoLiu]

Hi,

I would like to ask whether the toolbox supports multi-peak fitting, rather than only single-peak Voigt line shape fitting. If multi-peak fitting is not currently implemented, do you have any recommendations for extending the toolbox to support it?

Looking forward to your response. Thanks in advance!

Best,
Suhao Liu, Student from UIUC

[jamienear]

Dear Suhao, thank you for your question.

Until now FID-A did not have tools for multi-peak fitting. The reason for this is because FID-A is more geared towards data processing. For fitting MRS data, I typically recommend users export their processed data to one of the many MRS fitting softwares available (LCModel, Osprey, FSL-MRS, AB-fit, jMRUI, Vespa, etc.).

However, in response to your question I have now added a new function called 'op_HSVDfit' (https://github.com/CIC-methods/FID-A/blob/master/processingTools/op_HSVDfit.m), which uses the HSVD fitting method proposed by Barkhuijsen et al (JMR 1987) to fit a spectrum to a series of Lorentzian peaks. This function is closely related to the 'op_removeWater' function, which has been part of FID-A for a long time and uses the HSVD method to remove residual water peaks from a spectrum. In the past, I have often used the 'op_removeWater.m' function in an "off-label" way to fit spectra to a series of Lorentizan peaks, but I realize that most users will not be aware of this trick, especially because the name of the 'op_removeWater' and its function header do not hint at this functionality. SO I have now made a new version that is specifically for fitting spectra. The new function will return the fitted model spectrum as well as the fit residual. I hope you find this useful.

If you are interested in other fitting tools, I would be happy to discuss.

Best wishes,
-Jamie