Hi Carles, Susan,
I think, in Susan's case, hrf convolution would still make sense, since 1)
TR is far below hrf time resolution and 2) the time between two scans is
relatively short.
Volkmar
On Wed, 26 Sep 2007, CARLES M. FALCON FALCON wrote:
> Hi Susan,
>
> if acquisition is not continuous, it makes no sense to convolve with a HRF.
> What I did to analyze an auditory fMRI study with discontinous acquisition (8
> seconds delay between scans)was to set the stimuli as regressors instead as
> conditions. The regressors were set to 1 o 0 according if haemodinamic
> response to the stimuli was expected or not during each scan acquisition. It
> worked fine.
>
> Hope this helps
>
> Carles Falcon
> IDIBAPS-Hospital Clinic de Barcelona
> University of Barcelona
>
> Missatge de Susan Shuai <[log in to unmask]>:
>
>>
>> Dear Volkmar,
>>
>> My experiment is still a block design. Actually I have tried the methods both
>> with and without slice timing. And the parameter setting is exactly what you
>> suggested below. But the result is quite different from the original 2sec TR
>> (continuous acquisition) analyzed result. For the original one, I got
>> significant differences in contrasts I want. However, for the second
>> experiment (cluster acquisition), both analysis methods did not show
>> differences.
>>
>> I am not sure whether it is because of the individual differences,
>> hemodynamic changes caused by introducing TD or wrong analyzing procedure. I
>> followed SPM manual, the first two examples' procedure in chapter "data sets"
>> to do the analysis with my own parameters.
>>
>> Should I send you a more detailed SPMPrint file?
>>
>> Thank you and Best regards,
>> Susan> Date: Mon, 24 Sep 2007 10:56:42 +0200> From:
>> [log in to unmask]> Subject: Re: [SPM] Feel desperate in
>> analyzing the discontinuous acquisition data> To: [log in to unmask]> > Dear
>> Susan,> > your setup with silence periods between each scan is indeed tricky.
>> If > your experiment can be modelled as a block design, I would go for> - no
>> slice timing correction> - model TR as 3000 ms> - model blocks with onsets
>> and duration> > If you really need to model the design event-related, then
>> you should > enter TA as 1.55 secs and TR as 3.0 secs. For modelling, your TR
>> would be > 3 seconds, and you have to set your reference time bin in a way
>> that it > corresponds to the time when your reference slice was taken.> >
>> hope this helps,> > Volkmar> > On Fri, 21 Sep 2007, Susan L. SHUAI wrote:> >
>>> Dear SPMers,> >> > I am just a new comer to use SPM and doing fMRI
>> experiment. I have> > encountered much difficulty in analyzing the data now.
>> My experiment> > design is a !
>> block design and the acquisition is a cluster acquisition> > because it will
>> allow for silence to present auditory stimulus.> >> > Should I follow the
>> Event-related analysis? In slice timing, what should I> > put as TA? Is it
>> the start time of the last slice acquisition? For> > example, I have 32
>> slices, and they are acquired in 1.6s. Is the TA 1.55?> >> > In the original
>> experiment, I did the continuous acquisition and the TR is> > 2000ms. In the
>> cluster acquisition, the TR is 3000ms and TD is 1400ms> > allowing for
>> silence. But the analyzed results are very different. In the> > first case,
>> most contrasts show significant differences, while in the> > second case,
>> little contrasts show differences.> >> > I think the slice order and
>> reference slice are all set right. What would> > be the reason for such large
>> differences? How can I check my analyze> > result? I am feeling desperate
>> about the situation now.> >> > Thank you for your help!!> >> > Regards,> >
>> Susan> >> > -- > Vo!
>> lkmar Glauche> -> Department of Neurology volkmar.glauche@uniklinik-fr
>> eiburg.de> Universitaetsklinikum Freiburg Phone 49(0)761-270-5331> Breisacher
>> Str. 64 Fax 49(0)761-270-5416> 79106 Freiburg
>> http://fbi.uniklinik-freiburg.de/
>> _________________________________________________________________
>> Discover the new Windows Vista
>> http://search.msn.com/results.aspx?q=windows+vista&mkt=en-US&form=QBRE
>
>
>
>
> -------------------------------------------------
> This mail sent through IMP: http://horde.org/imp/
>
--
Volkmar Glauche
-
Department of Neurology [log in to unmask]
Universitaetsklinikum Freiburg Phone 49(0)761-270-5331
Breisacher Str. 64 Fax 49(0)761-270-5416
79106 Freiburg http://fbi.uniklinik-freiburg.de/
|