Hi,
You can find details about how we calculate smoothness in:
http://www.fmrib.ox.ac.uk/analysis/techrep/tr00mj3/tr00mj3/tr00mj3.html
Note that we do not average the sigma values (or even the sigma^2 values)
but take the average correlations to feed into the calculation.
I'm not sure what AFNI uses to do their estimation, but the above technical
report shows the difference between two commonly used methods (we use
the discrete version proposed by Forman et al).
It also is probably not worth obsessing too much over this as these are
fairly gross estimates, assuming that the residuals are a beautiful stationary
Gaussian field. In practice they are not, and these are approximations to
be able to employ things like GRF (or AlphaSim) which are known to be
relatively approximate, but good enough for most statistical purposes (erring
on the conservative side).
I hope this helps.
All the best,
Mark
On 14 Aug 2012, at 05:59, MCLAREN, Donald wrote:
> Interesting. I'd wait for some input from the FSL authors. I've always
> wondered about the differences in smoothness between programs and
> whether they converge or diverge across different studies.
>
> Best Regards, Donald McLaren
> =================
> D.G. McLaren, Ph.D.
> Research Fellow, Department of Neurology, Massachusetts General Hospital and
> Harvard Medical School
> Postdoctoral Research Fellow, GRECC, Bedford VA
> Website: http://www.martinos.org/~mclaren
> Office: (773) 406-2464
> =====================
> This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED
> HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is
> intended only for the use of the individual or entity named above. If the
> reader of the e-mail is not the intended recipient or the employee or agent
> responsible for delivering it to the intended recipient, you are hereby
> notified that you are in possession of confidential and privileged
> information. Any unauthorized use, disclosure, copying or the taking of any
> action in reliance on the contents of this information is strictly
> prohibited and may be unlawful. If you have received this e-mail
> unintentionally, please immediately notify the sender via telephone at (773)
> 406-2464 or email.
>
>
> On Tue, Aug 14, 2012 at 12:54 AM, Jessica Andrews-Hanna
> <[log in to unmask]> wrote:
>> Great suggestions. Thanks very much for your advice!
>>
>> In answer to your questions/suggestions, smoothest does seem to be applying a mask. While it will not run without inputting a mask, switching the mask changes the smoothness values.
>>
>> Second, neither 3dFWHMx nor smoothest seem to simply be averaging the smoothness values calculated across the individual volumes.
>>
>> Running 3dFWHMx on each volume separately (from fslsplit) and averaging across volumes for the x, y, and z dimensions yields:
>>
>> x = 9.886185862
>> y = 9.917113793
>> z = 9.207430345
>>
>> while doing the same thing using smoothest yields
>>
>> x = 11.15205172
>> y = 11.22923862
>> z = 10.41682207
>>
>> Best,
>> Jessica
>>
>
|