Dear Aser,
On 14/12/16 09:36, fmri2013 wrote:
> Dear Guillaume
>
> Many thanks for the response. In the manual it suggests using an F test
> using a factorial design. I have two questions here
>
> when using both basis the temp and the disp spm specify first the temp
> regressors and then the disp?
With the so-called informed basis set, the order of regressors is indeed:
[ canonical_response temporal_derivative dispersion_derivative ]
> Also if I would like to see the earlier response or the wider response
> is it correct to use a t test in the factorial design by using contrast
> of (0 1 0) for the early response and (0 0 1) for the wider response?
> And also using a a negative t test (0 -1 0) to see the later response?
> or should I perform a one sample t tests on the contrast maps for each?
You have to be careful with the interpretation of these tests as you are
looking for the consistency of an effect over subjects. Concerning the
temporal derivative, have a look at the recent email thread of "the
calculation of the latency of the BOLD response" and Cyril's insightful
comments.
Best regards,
Guillaume.
> Thanks a lot
>
> Aser
>
>
>
> Sent from my Samsung Galaxy smartphone.
>
> -------- Original message --------
> From: Guillaume Flandin <[log in to unmask]>
> Date: 12/12/2016 14:25 (GMT+00:00)
> To: Aser A <[log in to unmask]>
> Cc: [log in to unmask]
> Subject: Re: [SPM] Fwd: effects of PST
>
> Dear Aser,
>
> PSTH plots are created using a Finite Impulse Response (FIR) basis set
> so this is the basis set you should select for each of your subjects.
> See chapter 32 of the SPM manual for a detailed analysis at the group level:
> http://www.fil.ion.ucl.ac.uk/spm/data/face_rfx/
>
> Best regards,
> Guillaume.
>
>
> On 12/12/16 13:09, Aser A wrote:
>> Hi all,
>>
>> I am forwarding this again as I could not find an answer.
>> ---------- Forwarded message ----------
>> From: *fmri2013* <[log in to unmask] <mailto:[log in to unmask]>>
>> Date: Wed, Nov 23, 2016 at 11:07 AM
>> Subject: effects of PST
>> To: SPM <[log in to unmask] <mailto:[log in to unmask]>>
>>
>>
>> Hello all,
>>
>> Is it possibke in SPM to see the effect or changes on the Post stimulus
>> time (PST) (or perstimulus time - I am not sure if they are the same)
>> for each subject and make a groupe statistical inferences on the PST
>> changes? I used though the standard canonical HRF function provided by
>> SPM. The task is an event related design and stimuli are jittered.
>>
>> Thanks in advance
>>
>> Aser
>>
>>
>>
>
> --
> Guillaume Flandin, PhD
> Wellcome Trust Centre for Neuroimaging
> University College London
> 12 Queen Square
> London WC1N 3BG
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|