Torben,
Thanks for the reply. I will look up the paper for sure. I am forwarding your
e-mail to the listserv.
thanks,
Appu
>===== Original Message From [log in to unmask] =====
Hi Appu
I can not post to the FSL mailing list from home (different email
address), so I will mail you directly, and you can forward it to the list
if you fancy :-) I am glad to hear that there are more sceptical people
around, I think that one should be concerned about cardiac noise even when
using a high-pass filter, because there is in general no specific
frequency band into which this aliasing happens. If you find the topic
interesting I have recently written up my views on noise in fMRI, the
paper is called "Non-white noise in fMRI: Does modelling have an impact?".
You can find them under "papers in press" on the NeuroImage homepage, or I
can send you the PDF if you want. I hope the paper will answer your
questions otherwise let me know.
Best
Torben
Torben E. Lund
Danish Research Centre for MR
Copenhagen University Hospital
Kettegaard Allé 30
2650 Hvidovre
Denmark
email: [log in to unmask]
webpage: http://www.drcmr.dk
> Hi,
>
> Our experimental task involves an ABAC pattern that is repeated four
> times.
> Each condition is 32s long so the total time for ABAC pattern is 128s. As
> recommended by the FSL software we chose a temporal filter is 128s i.e.,
> .007
> Hz. The FSL documentation states that a high pass temporal filter is
> effective
> in filtering low frequency scanner drift and physiological noise like
> cardiac
> and respiratory noise. Although cardiac and respiratory effects may occur
> at
> higher frequencies than the signal of interest, the temporal sampling
> frequency (the scanner’s TR setting) can interact with these high
> frequencies
> resulting in an aliasing of the signal to a much lower frequency. It is
> expected that this aliased lower frequency will be filtered out by the
> high
> pass filter along with the low frequency scanner drifts. I calculated that
> hypothetical cardiac rates of 1 or 2 Hz, with a sampling rate of 2 sec
> (TR=2sec) will alias down to 0 Hz resulting in nothing really to filter
> out.
> However, if the cardiac rate is around 1.2 Hz, with a TR of 2, it aliases
> down
> to a frequency of .2Hz. In such a case our high pass temporal filter of
> .007
> Hz will not be effective in filtering out cardiac noise. If this is the
> case,
> should we be concerned about cardiac noise aliasing into our signal of
> interest?
>
> Thanks,
> Appu
>
|