Here is my understanding to noise in fmri experiments. Please correct me if i'm
wrong.
If you take an roi (could be one voxel) under any condition and plot the density
power spectrum using the autocorrelation function, you will get the "1/f" graph
and possibly some "squiggles".
The "1/f" graph is the so called "DC" component of the spectrum while the
"squiggles" are the "ac" components.
A better view for the "ac" components may be obtained by removing the "DC"
component, which hides lower "ac" freqeuncy "squiggles". [Simply subtract the
mean of spectral density]
In our lab, we reduce the noise by applying a bandpass filter (hamming for
example), with the cuttoff frequencies based on the experimental design.
Example1: For a block design 30secON/30secOFF, we use (4*30sec, 30sec) as cuttof
frequencies.
Example 2: For an experimental design which has 30secON/30secOFF &
15secON/15secOFF, we use (4*30sec,15sec) cutoff frequencies.
Our statistics improves (for our data where physiologic noise density is
comparable to our signal) by applying a bandpass filter compared to modeling the
the "i/f" noise. However, we didn't test for zero mean, and guassian distribution
of the noise after applying the bandpass filter.
-s madi
Drexel university
> John,
>
> This is a very interesting result. Were these data motion-
> corrected prior to your analysis?
> We have done a different sort of test to address the same
> question (i.e., "that motion artifact is a contributor to the 1/f
> portion of the spectrum.") We compared the voxel-averaged power spectra
> (from the whole brain, not selected based on activation in this case)
> of several subjects before and after motion correction. In contrast to
> your results, we found that the 1/f component of the spectrum was reduced
> by motion correction (Fig. 3D, Neuroimage, 5:179-197). Possible explanations
> for the difference between our result and yours are 1) that perhaps your
> data were analyzed after motion correction which could have eliminated the
> differences due to motion 2) different sets of voxels were examined, or 3)
> the more fidgety subject would have actually had less 1/f noise than
> the more stationary one if motion was a controlled variable.
>
> Sincerely,
> Eric
>
> >
> > I don't have any answers, but here is a small tidbit of information.
> > We acquired 1536 frames of fixation-only data in
> > two subjects. One of these subjects (our favorite) is blessed with the
> > ability to remain utterly motionless (as far as our realignments programs
> > can tell). The other squirmed more than my 2-year old nephew, yielding data
> > that we considered unusable because of motion artifact. I tried to
> > test the hypothesis that motion artifact is a contributor to the 1/f
> > portion of the spectrum. To this end, I computed a noise spectrum for each
> > of the 12 runs (128 frames per run) for each subject. I then averaged the
> > spectrum across runs, and then averaged the spectra across voxels
> > in visual cortex activated by a flickering checkerboard (activated voxels
> > passed a Bonferroni correction). This should give an average spectrum
> > across voxels that are primarily gray matter, and excludes ventricles and
> > large white-matter tracts. I expected the squirmy subject to be noisier.
> > In fact, the spectra lie on top of each other with an uncanny degree of
> > precision. Both the lowest and higest frequencies matched very well, with
> > the squirmy subject having a few more squiggles in between.
> >
> > Apparently noise due to motion largely cancels when averaging across
> > voxels, and the residual 1/f noise is due to some other effect.
> >
> >
> > John
> >
> > --------------------------------------------------------------
> >
> > John Ollinger
> > Washington University
> > Neuro-imaging Laboratory
> > Campus Box 8225
> > St. Louis, MO 63110
> > http://imaging.wustl.edu/Ollinger
> >
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|