[Eeglablist] Brain PSD thresholds by frequency band
mmiyakoshi at ucsd.edu
Thu Jan 26 14:47:20 PST 2017
> I get a lot of ICs that are very difficult to label so I also take in
If the spectra curve is not very smooth or does not follow 1/f, your ICA
decomposition may have a problem.
> However I still dont know if there are any thresholds which I could use
as a reference to evaluate if the IC could be EEG or not.
You cannot apply a simple threshold on one band to call something EEG or
noise. Particularly, the low frequency band has a very high power and
standard deviation as well.
Try my plugin to evaluate spectral fluctuation and cross-frequency
correlations if you are interested in.
On Thu, Jan 26, 2017 at 5:03 AM, Alberto Sainz <albertosainzc at gmail.com>
> Dear Makoto and EEGLAB list,
> Yes, I mean uV^2/Hz. I get a lot of ICs that are very difficult to label
> so I also take in account PSD. However I still dont know if there are any
> thresholds which I could use as a reference to evaluate if the IC could be
> EEG or not.
> 2017-01-25 0:45 GMT+01:00 Makoto Miyakoshi <mmiyakoshi at ucsd.edu>:
>> Dear Alberto,
>> After performing ICA, you evaluate each IC not only by their PSD but also
>> their spatial location etc. If an IC shows unmistakable EEG characteristics
>> in spatial and time domain, I would take it a brain EEG even if delta power
>> is 0-20 (??? what is 0-20 anyway? uV^2/Hz?)
>> On Wed, Jan 18, 2017 at 10:32 PM, Alberto Sainz <albertosainzc at gmail.com>
>>> I am struggling a bit with ICA rejection when measuring PSD. I am not
>>> sure if i am prunning too much data when i look at some results. For
>>> example, I saw references of delta power between 0-6 and others between
>>> 0-20. I am not sure when high delta power (20?) is an artifact or brain
>>> For alpha band right now i dont get power values larger than 1.
>>> Do we have any reference for (EEG-measured) PSD thresholds by frequency
>>> band? Also any for typical PSD artifacts like eye blinks?
>>> I hope I explained it correctly.
>>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>>> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.uc
>>> 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
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the eeglablist