[Eeglablist] Question regarding Cleanline

[Tim Mullen]

Hi Michael,

Power grid fluctuations and changes in EMI from movement or other factors
can cause line noise frequency and complex amplitude to vary over time.
Cleanline subtracts from the data a time-domain sinusoid representing the
best estimate of the deterministic line noise signal (based on a
frequency-domain regression). If a single sinusoid estimate is obtained for
all the data, but the line noise frequency or amplitude drifts over time
this may result in suboptimal correction. For this reason, using a shorter
sliding window for noise correction is generally the best approach. There
are options for this in Cleanline. I usually recommend starting with a 4
sec window with 4 sec step size (no overlap), but you may want to start
with 10 seconds since you've already had good results there.

Tim


On Sat, Dec 14, 2013 at 11:43 AM, Michael Schubert <mschuber at mail.upb.de>wrote:

> Hi Tim,
>
>
>
> we recorded EEG data outdoor and unfortunately have a lot of line noise
> (plus its harmonics) in it. I tried Cleanline to get rid of the noise but
> wasn’t very happy with the result when I used the whole recording (approx.
> 500s). However when I only use the first 10 or 20 seconds, Cleanline
> performs great.
>
> From this, can you guess what the problem might be? Its continuous data
> with no boundaries in it, so I’m guessing phase changes are not the
> problem?!
>
> The plots I attached show the effect and were created using Cleanline with
> default settings.
>
>
>
> Thanks,
>
>
>
> Michael
>
>
>



-- 
---------  αντίληψη -----------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20131214/bd328d02/attachment.html>