Dear Freesurfers,

I spent some time and answered my own question, which (after thinking about the problem) is a really stupid one. I only analyzed 15 subjects with the 160 slices MPRAGE. If I'll have some more time I'll analyze the 176 versions also, but after these results It will have extremely low priority.

Theoretically the SNR gain from the extra 16 slices in my case is 4%. Which is further reduced a bit due to some other effects like inhomogenity, etc, but let's say it is the maximal 4%.

The inter-subject variance of SNR (in %) (i.e. CV) (SNR calculated in 3 ways: signal avg/sd of noise; (signal avg-noise avg)/sd of noise; signal avg/noise avg, whichever you prefer) is 9.20%, 9.28% and 8.13% respectively. The biggest differences between the highest and lowest SNR among the 15 patients in % are 30.6%, 30.9% and 33.6% respectively.

The SNR is so different between two measurements (which is not surprising, as it is affected by dozens of factors (even head size, but if you think about the contrast, as the MPRAGE is a T1 image, it's easy to understand that the T1 time of the tissue severely affects your measured signal, and T1 time differs in people around 10% in healthy young ones), but if you repeat the MPRAGE in the same subject within the same session and let the scanner adjust everything again, you will still get quite different SNRs, within the same patient)

I think it is very unlikely that the 4% of extra SNR would cause any bias in the comparison.

Being honest, If it caused any bias, more serious problems would emerge: as you could not compare anybody, ever. As the age, level of hydration (how much he/she drunk in the last few days), sex, periods in women, many more can all affect the SNR more than the 4% mentioned above.

But as this problem can be discussed from many points of view, it would result in an endless ati/nvidia type thread, which would not beneficial and would be only the waste of time, so for me, the thread is closed.

Thank you for your replies and generous help.

Best Regards,

Gergely

2011/1/20 Michael Harms mharms@conte.wustl.edu

Hi Gergely, Was this collected on a Siemens scanner? If so, the Siemen's MPRAGE sequence is such that the "slices" form the "inner loop" of the acquisition -- that is "Number of slices" lines in k-space are acquired for each TR period. So, having a differing number of slices theoretically has an impact on the spatial filtering properties of the acquisition. My intuition suggests to me that this would be a minor effect that would be washed out by other noise and inter-subject differences. But perhaps others have quantitative results in this regard...? (e.g., collecting both sequences on a set of subjects and then looking for differences using an intra-subject design).

cheers, -MH

...

Hi Freesurfers,

I have a quick question:

I have two groups, both measured on the same scanner, with the same sequence (3D MPRAGE), and with almost the same parameters.

TR/TE/TI/BW/MATRIX/FOV/FOV-phase/Flip angle/Slice thickness/in-plane resolution... everything is the same, but the number of slices. Total measurement time was also the same.

One group was measured with 160 slices, the other with 176.

Can I compare the two groups in freesurfer?

Thanx,

-- Gergely Orsi biologist - research associate Diagnostic Center of Pécs H-7623 Pécs, Rét str. 2. Tel.: 0036-70-3174676 E-mail: gergo.orsi@gmail.com _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you

in

...

error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

Thanks for the speedy replies,

I think that in case of a 3D acquisition, the bigger the volume the better the SNR (until a certain point, of course), and in this case the volume is bigger with 9,9%.

The parameters of the MPRAGE are:

TR/TE/TI: 1900/3.41/900 ms; FA: 9°; FOV: 210x240 mm2; 224x256 matrix; slice thickness: 0.94 mm;

160/176 slices; voxel size: 0.94x0.94x0.94 mm3; 180 Hz/pixel receiver bandwidth

But, (i know it can not be answered quantitatively, or at least hard) how big the influence of this effect in such comparison after all the preprocessing steps like intensity norm.

Ok, my last question was hypothetic,

I'm waiting for the answer from Andre van der Kouwe, for further help.

Thankx,

2011/1/20 Bruce Fischl fischl@nmr.mgh.harvard.edu

Hi Gergely,

I think this will be trouble. I'll cc Andre van der Kouwe, but for the 3D acquisitions, if you have a bigger field of view is has an effect on image SNR, so that will be a bias between your groups.

cheers Bruce

On Thu, 20 Jan 2011, Gergely Orsi wrote:

...

Hi Freesurfers,

I have a quick question:

I have two groups, both measured on the same scanner, with the same

sequence

...

(3D MPRAGE), and with almost the same parameters.

TR/TE/TI/BW/MATRIX/FOV/FOV-phase/Flip angle/Slice thickness/in-plane resolution... everything is the same, but the number of slices. Total measurement time was also the same.

One group was measured with 160 slices, the other with 176.

Can I compare the two groups in freesurfer?

Thanx,

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.