Hi Michael and the SPM list,

Many thanks for you response. What I meant by time window was the window
length. I also specified the order as 1.

Could you please let me know:

1) Why is the standard time window value for FIR 20-24s? Are there any
references on this in relation to FIR?
2) In my design, the volume acquisition occurs 4s after stimulus
presentation. When specifying a window length, do I need to take into
account the delay in scan time?

My understanding is that the FIR would allow better modelling of sparse
sampling design. However, I'm not sure how we can enter timing

thank again.


> Hi Catherine,
> If the time window prompt you are talking about is the window length
(xBF.length), then I believe the window is the time it takes for your
function to return to baseline in seconds. I believe a standard value
for FIR is 20-24. However, if you look in the SPM functions
> (spm_fmri_spm_ui or spm_fMRI_design) you will see that window length is
overwritten automatically by SPM. If you want to change the decay
function you can try changing the time bin length(xBF.dt), which is set
by SPM as TR/#ofTimeBins. If you want to experiment with this the
easiest way is to change the number of time bins per scan. Careful when
messing with these attributes as it can have a serious effect on your
parameter estimates. If the time window you were talking about was the
duration of the covariate onset, then it seems logical for you to model
the stimulus as an event (duration of 0).
> Hope this was helpful,
> Michael Rubens
> --
> Research Associate
> Gazzaley Lab
> Department of Neurology
> University of California, San Francisco
> Lab Phone # (415) 476-2164
> NIC Phone # (415) 476-4232
> Cell Phone # (415) 902-3138
> UCSF, Mission Bay
> MC2512, 1700 4th st.
> Byers Hall, Room 102C
> San Francisco, CA 94158
> On Sun, Apr 6, 2008 at 6:30 PM, Catherine Wan
> <[log in to unmask]> wrote:
>> Dear SPM -list,
>>  I'm writing to seek clarfication on FIR analysis using SPM5. I'm
>> running an
>>  auditory experiment using a sparse sampling design.  The design
>> incorporates
>>  a silent period, such that the actual interscan interval is 14s, and
>> the
>>  single volume image acquisition occurs 4s after the stimulus
>> presentation to
>>  capture peak of HRF (TR = 4.69s).
>>  The basis function I'm using is FIR. In the time window  prompt, I've
entered 4s to allow for the delay between stimulus and image
>> acquisition. Is
>>  this correct? I've also tried it with other values and the outputs turn
>> out
>>  to be identical.
>>  Many thanks - any assistance will be much appreciated.
>>  Regards,
>>  Catherine