GANNET-Quality measures

#1

Hi all

I’m new in the spectroscopy field and I just analysed my data with Gannet for the first time. I would like to check the quality of my data, but I can’t seem to find the quality metrics I’m looking for.

  1. Regarding the residual water frequency, I can see the graph, but I was wondering if there are also values available to use for quantification (such as mean standard deviation).

  2. For the SNR, I get values for both Cr SNR and water SNR. Is there any way to calculate the SNR for Glx (I’m analyzing Glx)? Or is Cr/water SNR more frequently used to check for quality?

Any help would be really appreciated!

Kind regards
Marte

#2

Hi @Marte_VdS,

  1. I’m not quite sure what you mean regarding the use of the residual water frequency for use with quantification. Do you mean a measure of the frequency shift/drift based on the residual water frequency?

  2. Glx SNR is calculated. Assuming you are using the latest version of Gannet, it is stored in the MATLAB output structure (e.g., MRS_struct.out.vox1.Glx.SNR).

Mark

2 Likes
#3

Hi Mark

  1. Yes, I meant a measure for frequency drift based on the residual water frequency. I want to use this in my statistics as a covariate, but I don’t know which value I can use that represents this residual water frequency.

  2. I will check this out!

Thank you for your help!

Marte

#4

The variable MRS_struct.out.AvgDeltaF0 is the variable to use. It is defined as the mean difference between the observed residual water frequency in each average and the nominal water frequency (4.68 ppm). For HERMES acquisitions, we use the 3.02 ppm Cr signal instead.

2 Likes
#5

It works!

Thanks again for your help, I really appreciate it.

Kind regards
Marte

1 Like
#6

Glad to be of help.

Note that if your (non-HERMES) sequence is using very strong water suppression, AvgDeltaF0 is not reliable. Please let me know if this is the case for your data.

#7

Hi Mark

For data acquisition, we used a regular Johns Hopkins MEGA-PRESS sequence (TE = 68 ms, TR = 2 s, 2-kHz spectralwidth, 112 averages). Is AvgDeltaF0 reliable in this case?

#8

Yes, with that sequence, it should be.

1 Like