optimize fit: residual issue

[Wu Chen]

Dear Martin

Dear all, 

I have problems getting an optimized fit for my Siemens 3T 31P data. We use  3D CSI  for the acquisition and so far Tarquin has been working nicely, but I need help optimizing the fits w.r.t residual signal. I've attached the output from one of my voxels. My question is why I lose so much signal as residual(as shown in the figure in red). Shouldn't it be flat (as shown in the figure in green)?? Is there a way to tweak some parameters so that I can fit the data optimally with the basis set and flatten the residual signal? 

Data Parameters:

Dynamic  average : default

Dynamic corr. Ref signals = 31P PCr

Preprocessing:

Water cutoff = 0

Reference offset = 0

Reference signals = 31P PCr

Max dref = 0.1

Fitting:

Start point = 17

End point = 1024

Initial beta 1000

Water concentration = 0

Water attenuation = 0

Max metabolite shift = 3

Max broad shift = 3

Internal Basis Set = 31P brain, 1H decoupled

Postprocessing:

Right PPM limit = -20

Left PPM limit = 20

Smoothing points = 25

Line-broadening = 15

External basis: the one you uploaded here in this forum. Everything else was set to default. 

I would appreciate your efforts if you could help me. 

Many thanks!

Best wishes,

WC

[Wu Chen]

Sorry. Please find attached the corresponding figure with this message

[harshal patel]

Dear Wu,

Please try Initial beta 500, 1000, 1500 &2000 and see if it helps. 

Best wishes

HJP