[Eeglablist] Question about Cleanline

[Makoto Miyakoshi]

Dear Tim and Eric,

Thanks for your comments and corrections!

> Cleanline uses multitaper regression, with Thompson's F-statistic for
determining the probability of sinusoidal artifact at a given frequency.
There is no "Welch FFT" nor any need to high pass filter the data (this
would have no effect on Welch's method either).

Ah that's true, I was wrong. Hence the flat peak and two inputs for
determining the multitaper parameters (which is fixed to [3 5] in GUI).
Thank you for correction.

I referred to the PREP paper to show evidence that high-pass filter helps
CleanLine, so the data should be treated with high-pass before CleanLine. I
did not intend to refer to the paper to explain what CleanLine is. Sorry if
my explanation was misleading.

Makoto



On Wed, Apr 6, 2016 at 3:34 AM, Tim Mullen <mullen.tim at gmail.com> wrote:

> Actually, that is not correct.
>
> Cleanline uses multitaper regression, with Thompson's F-statistic for
> determining the probability of sinusoidal artifact at a given frequency.
> There is no "Welch FFT" nor any need to high pass filter the data (this
> would have no effect on Welch's method either).
>
> Please see the documentation and references I have provided for Cleanline,
> in the toolbox Readme file on NITRC, paying attention to the "Tips on
> Running Cleanline" section:
> https://www.nitrc.org/projects/cleanline/
>
> Eric, you may try running Cleanline iteratively on the data, gradually
> increasing the significance threshold (LineAlpha). You may also try
> different window sizes, as suggested in the tips. Other tips there may also
> apply. If you have multiple epochs, ensure your window size and step size
> equal (or divide) the epoch length.
>
> You may also consult our PREP paper/toolbox, which wraps around Cleanline
> and local detrending tools I have previously discussed on this list, as
> well as other artifact rejection tools developed by Christian Kothe.
>
> Best,
> Tim Mullen
> On Apr 5, 2016 2:09 PM, "Makoto Miyakoshi" <mmiyakoshi at ucsd.edu> wrote:
>
>> Dear Eric,
>>
>> First of all, did you apply it on the high-pass filtered data? CleanLine
>> uses EEGLAB's Welch's method for FFT (if I remember correctly) to evaluate
>> the spectrum, so the data needs to be high-pass filtered. See Nima
>> Bigdely-Shamlo's PREP pipeline paper for detail.
>>
>> Did you average reference the data? It may help, especially if your data
>> are not re-referenced at all (like Biosemi's data).
>>
>> Another thing is that you can expect a small improvement by doing this
>>
>>  'SlidingWinLength',4,'SlidingWinStep',4
>>
>> If these does not work, I recommend you band-pass filter your data
>> 45-55Hz to see stationarity of your 'line noise'-- is it really it??
>>
>> Makoto
>>
>>
>> On Fri, Mar 4, 2016 at 2:14 AM, Eric HG <erichg2013 at gmail.com> wrote:
>>
>>> Hi everybody,
>>>
>>> I've been trying to apply Cleanline instead of a notch filter but there
>>> seems to be a problem with some of the subjects. Cleanline doesn't seem to
>>> remove all the line noise (at 50 Hz) (see picture). Before I use cleanline
>>> I use the function pop_firwsord to high- and low-pass the EEG.
>>>
>>> I use cleanline with the following settings:
>>>
>>> [EEG, Sorig, Sclean, f, amps, freqs, g] = pop_cleanline(EEG, 'Bandwidth'
>>> ,2,'ChanCompIndices',[1:EEG.nbchan], ...
>>>
>>>                           'SignalType','Channels','ComputeSpectralPower',true,
>>> ...
>>>
>>>                           'LineFrequencies',[50] ,'NormalizeSpectrum',false,
>>> ...
>>>
>>>                           'LineAlpha',0.01,'PaddingFactor',2,
>>> 'PlotFigures',false,          ...
>>>
>>>                           'ScanForLines',true,'SmoothingFactor',100,
>>> 'VerboseOutput',1,    ...
>>>
>>>                            'SlidingWinLength',4,'SlidingWinStep',2);
>>> Does anyone know what might need to be changed to fix this?
>>>
>>> Best,
>>>
>>> Eric
>>>
>>>
>>> [image: Inline image 1]
>>>
>>> _______________________________________________
>>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>>> To unsubscribe, send an empty email to
>>> eeglablist-unsubscribe at sccn.ucsd.edu
>>> For digest mode, send an email with the subject "set digest mime" to
>>> eeglablist-request at sccn.ucsd.edu
>>>
>>
>>
>>
>> --
>> Makoto Miyakoshi
>> Swartz Center for Computational Neuroscience
>> Institute for Neural Computation, University of California San Diego
>>
>> _______________________________________________
>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>> To unsubscribe, send an empty email to
>> eeglablist-unsubscribe at sccn.ucsd.edu
>> For digest mode, send an email with the subject "set digest mime" to
>> eeglablist-request at sccn.ucsd.edu
>>
>


-- 
Makoto Miyakoshi
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160406/7856bbe4/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Cleanline example.jpg
Type: image/jpeg
Size: 204931 bytes
Desc: not available
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160406/7856bbe4/attachment.jpg>