Dear Paul,
VBM8 is indeed applying the correction of Hayasaka. This is done by using local RESEL values and as result you obtain a corrected (and now valid) cluster statistic.
Best,
Christian
On Tue, 23 Oct 2012 15:03:25 +0200, Gabor Oederland <[log in to unmask]> wrote:
>Hello Paul,
>
>
>some clarification, the non-stationarity correction in SPM8 is based on
>RESEL density, for methods see Hayasaka et al. (2004, Neuroimage) - so it's
>not "simply" adjusting p-values. But the data itself seems not to be warped
>directly, at least you don't get any corresponding isotropic output. That's
>why I said a cluster threshold determined in another programme won't help.
>
>The non-stationarity correction in VBM8 should be based on the methods
>provided by Hayasaka et al. (2004, Neuroimage) as well. But this is
>something Christian Gaser is familar with. I set him on CC, as the title of
>the thread does not directly refer to VBM8 toolbox.
>
>
>Best,
>
>Gabor
>
>
>
>>
>> -------- Original-Nachricht --------
>> Datum: Tue, 23 Oct 2012 03:45:31 +0100
>> Von: Paul Klauser <[log in to unmask]>
>> An: [log in to unmask], Gabor Oederland <[log in to unmask]>
>> Betreff: Re: SPM8 AlphaSim non-stationarity correction
>>
>> Thanks Gabor!
>>
>> I am wondering if the answer also applies to non-stationarity correction
>> in VBM8. Could anybody from the VBM8 team confirm that the
>> non-stationarity correction applied through the VBM8 toolbox also
>> corrects cluster-extent estimates, such that it would be appropriate to
>> use AlphaSim in such a case?
>>
>> Otherwise, is the only other alternative to using cluster statistics in a
>> non-parametric framework to use SnPM?
>>
>> Thanks for your help
>>
>> Paul Klauser
>>
>>
>
>
|