FW: [Eeglablist] fMRI artifact removal
Rami K. Niazy
rami at fmrib.ox.ac.uk
Thu Sep 29 11:51:27 PDT 2005
For the fmrib plug-in, the sampling requirement is that you sample high
enough so that you do not have any aliasing in the data. So that
depends on what filters your hardware has. As a minimum, your hardware
should completely remove any frequencies above 400Hz or so to avoid
aliasing at 1kHz (theoretically, you should always sample at twice the
highest frequency in your data. Practically, that should be 3-4 times
the highest freq.)
Negishi et. al. (2004) applied similar techniques to those used in the
fmrib plug-in to data sampled at 1kHz and had good results. We have
tried the fmrib plug-in on data collected at 1kHz and had really bad
results because high frequencies were aliasing. At those low sampling
frequencies (1kHz - 2kHz), even though you avoid aliasing, you get some
shape differences in the artifact as Marcel has pointed out. However,
our program is designed to approximate these differences and deal with
them so that you don't have to sample at 5kHz. You just need to avoid
I hope that helps.
Yuko Yotsumoto wrote:
> Hi Marcel,
> According to their publication and web page, fMRIb plug-in works with
> as low as 2048 Hz sampling rate, which I think makes this plug-in very
> But 957 Hz is probably too low... I am wondering if anybody knows the
> minimum requirement for sampling rate to apply fMRIb.
> Marcel Bastiaansen wrote:
>> Dear Yuko,
>> I have no experience with the fMRIb tool, but:
>> It is my experience that everything below 1 KHz is way too low to
>> adequately sample the gradient artefacts produced by the EPI
>> This will result in a shape of the gradient artefact that varies from
>> volume to volume. This makes it impossible to compute anaverage
>> gradient artefact for the subtraction.
>> You should sample at 5 KHz or higher to get rid of this problem.
>> Marcel Bastiaansen
>> Yuko Yotsumoto wrote:
>>> Hi all,
>>> I have EEG data recorded in the fMRI scanner.
>>> Our EEGs were recorded with 957 Hz, while fMRI was operating at 4TR
>>> with 35 slices. Recording time was about 160 sec. Trigger signals
>>> were obtained at every image acquisition (=35 triggers for 4 TR).
>>> I'm trying to remove EPI noise artifacts with fMRIb plug-in, but
>>> have been unable to get clean results.
>>> With fMRIb plug-in, I used parameter sets such as follows.
>>> Low-pass filter cut off : 30
>>> Interpolation folds : 21
>>> Artifact-timing event : slice
>>> ANC : on
>>> Correct for missing triggers : off
>>> Relative trigger location : 0
>>> # of residuals PC to remove : auto
>>> I am wondering if our sampling rate of 957 Hz might not be high
>>> enough to use fMRIb plug-in. Have anybody in this mailing list tried
>>> this plug in with a similar sampling rate? Did it work?
>>> eeglablist mailing list eeglablist at sccn.ucsd.edu
>>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>>> To unsubscribe, send an empty email to
>>> eeglablist-unsubscribe at sccn.ucsd.edu
> eeglablist mailing list eeglablist at sccn.ucsd.edu
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
Rami K. Niazy, DPhil Candidate
University of Oxford, Centre for Functional MRI of the Brain (FMRIB)
John Radcliffe Hospital, Headington, Oxford OX3 9DU, United Kingdom
Tel: +44 (0) 1865 222739 / Fax: +44 (0) 1865 222717
e-mail: rami at fmrib.ox.ac.uk / URL: http://www.fmrib.ox.ac.uk/~rami
More information about the eeglablist