[Eeglablist] Separating out Alpha from mu oscillations

Wed Mar 26 18:46:09 PDT 2014

Dear Bala and Colleagues,

To Scott Makeig's statement ("Of course, at the scalp channels they are
highly mixed by volume conduction and can only be separated by ICA, so far
as I am aware."), I would like to add that mu and alpha are further
separable by frequencies and locations (with the understanding of course
that this explains that).

This can be achieved in most subjects under three conditions of (1) very
high spectral resolution (2) sufficiently high spatial density, and (3)
respect of interindividual idiosyncrasies against the tendencies to map
statistical strategies to the imperative of group analyses. Alphas and Mus
blend spatially, spectrally or functionally when one or more of those
conditions are not met. One's experimental paradigm of course matters as
well. Mu and alpha might covariate due to task demand, nevermind their
genuine independency.

May I point your attention to:
http://arxiv-web3.library.cornell.edu/abs/1310.7662
This paper makes the specific argument that left and right mu (or left and
right alpha) can be dissociated from each other. I hope that the simpler
alpha-mu dissociation will be obvious enough from the material exposed in
figures.
---
To return another specific suggestion to your initial question dear Bala,
the frontal aspects of occipital alpha originate from a strongly dipolar
presentation of this wave, depending on montage and reference electrodes,
the frontal alpha antipole would often be much more anterior than mu's
Rolandic Fissure (examine bandpassed oscillations in time domain,
emphasizing on maps). This fortunate spatial organization is favorable to
separation of the frontal part of alpha and the (fronto)central part of
mu. One way to achieve this separation is to adjust the frequency band of
each subject to neuromarkers-of-interest (my mu might be 11.4-12Hz, no
need to include the 10-11.4Hz; yours might be 10.3-10.9Hz, no need for
your own irrelevant frequencies) - and to proceed intelligently about
where in space to sample information about the wave of interest (e.g.
algorithm or well-reasoned, empirically-verified electrodes of interest -
I use control task "localizers" for alpha and mu in all my subjects).

With kind regards,
_________________________________________________

Emmanuelle Tognoli - PhD
The Human Brain and Behavior Laboratory
Center for Complex Systems and Brain Sciences
Florida Atlantic University
777 Glades Road, Boca Raton, FL-33431
phone: (int+1) 561-297-0110
http://www.ccs.fau.edu/~tognoli

> Thanks Dr.Makeig and Jason for your valuable comments.
>
> I will try ICA on our dataset to see if I can successfully separate out
> the
> influence of occipital alpha from central/frontal regions.
>
> Bala
>
>
> On Tue, Mar 25, 2014 at 5:38 PM, Scott Makeig <smakeig at ucsd.edu> wrote:
>
>> Bala,
>>
>> The subject of the generators of EEG alpha/mu oscillations is a complex
>> one.  Julie Onton made a poster on what ICA tells us
>> here<http://sccn.ucsd.edu/~scott/pdf/Onton_SfN05_AlphaPosterMini.pdf>.
>>  Basically, we have found no dramatic difference between motor mu
>> oscillations and occipital/parietal alpha oscillations:
>>
>> They both have a peak near 10 Hz (in adults) - by convention in the
>> range
>> 8-12 Hz.  They both are non-sinusoidal (i.e., they have harmonics in
>> their
>> power spectra, for mu a bit more pronounced). They both separate into
>> lower
>> and higher frequency alpha processes under Independent Modulator
>> analysis
>> (as discussed
>> here<http://sccn.ucsd.edu/~scott/pdf/Onton_CNS09_AlphaIMPoster.pdf>
>> ).
>>
>> They are both associated with inattention (positive disregard for the
>> 'receptive focus' of the cortical area --> alpha/mu flooding;  positive
>> attention to the 'receptive focus' --> alpha blocking). They differ
>> principally in location (alpha, across the occipital and parietal
>> cortices;
>> mu, near the somatomotor strip).
>>
>> Of course, at the scalp channels they are highly mixed by volume
>> conduction and can only be separated by ICA, so far as I am aware.
>>
>> Some subsets of IC sources generating alpha band oscillations form
>> (somewhat) dependent subspaces (as confirmed by pairwise mutual
>> information
>> measures, for which Jason Palmer has contributed a function in the
>> EEGLAB
>> miscfunc folder). These dependencies may index mutual influences of
>> various
>> types including (for alpha ICs with adjacent equivalent dipole
>> locations)
>> the possibility of (local) alpha traveling wave phenomena.
>>
>> Scott Makeig
>>
>>
>> On Tue, Mar 25, 2014 at 6:58 AM, B L
>> <thirstyforknowledge123 at gmail.com>wrote:
>>
>>> Hi,
>>>
>>> I guess this topic has been discussed in the past but I cant find an
>>> effective solution.
>>>
>>> We know that, it  is very common for the alpha oscillation to have
>>> influence up to almost the frontal electrodes. My question is - If we
>>> are
>>> interested only in mu rhythm in a visual-motor task, what is the best
>>> way
>>> to tease out the alpha activity from the motor regions so the
>>> comparison
>>> across groups is more effective?
>>> Will ICA work better for this(just identifying the alpha component and
>>> removing it)? Is there any other method researchers commonly use for
>>> this
>>> purpose?
>>> Any help is much appreciated.
>>>
>>> Thanks!
>>> Bala
>>>
>>> _______________________________________________
>>> 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
>>>
>>
>>
>>
>> --
>> Scott Makeig, Research Scientist and Director, Swartz Center for
>> Computational Neuroscience, Institute for Neural Computation, University
>> of
>> California San Diego, La Jolla CA 92093-0961,
>> http://sccn.ucsd.edu/~scott
>>
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
> _______________________________________________
> 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

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.