[Eeglablist] Modeling alpha activity in EEG data

[Scott Makeig]

Jinwon -

Julie Onton made an informative 2009 poster
<https://sccn.ucsd.edu/~julie/AlphaPosterMini.pdf > exploring the nature of
alpha activity in our EEG cognitive experiment data.

1. Using the ICA decomposition approach, she showed that peak alpha
frequency of different effective cortical sources isolated by ICA
decomposition can vary within individuals.

2. Using a subsequent ICA decompositions of the log spectrum of such 'brain
ICs', we found that peak alpha frequency does typically vary within
individuals across time - even within single task performance periods. But
no simple rules would apply to every individual or to every data
recording - to deal with our highly complex human 'eco-niche', after all,
requires high-dimensional flexibility in brain processing.

These 'Independent Modulator Analysis (IMA)' applications of ICA often
separated ongoing variability in the source spectra into maximally distinct
frequency modes (IMs), something we showed first (relative to frontal theta
activity) in this 2005 paper
<https://urldefense.com/v3/__https://www.sciencedirect.com/science/article/pii/S1053811905002673?casa_token=UibPU--axWAAAAAA:J-0CTcHpJV8BKBTD-MPfnyE1659Q527qXGY5w4P-Kg5sYBdAq38wNgTEdufjHjEZIQ47W2iVfA__;!!Mih3wA!BliQLUCUjvX_OwMbwzt8CMbE7Q9URcT0MsxyVabuF7BIcWuCj2sWMO1mUUyHpYAmc7L34h2XWs8OXWUVVfZQ$ >.
The Matlab code Julie developed for this and subsequent analysis projects
was formed into an available EEGLAB plug-in, the Independent Modulator
Analysis Toolbox (IMAT) <https://urldefense.com/v3/__https://eeglab.org/plugins/imat/__;!!Mih3wA!BliQLUCUjvX_OwMbwzt8CMbE7Q9URcT0MsxyVabuF7BIcWuCj2sWMO1mUUyHpYAmc7L34h2XWs8OXQP8JTzx$ > by Johanna
Wagner and Ramon Martinez-Cancino (though note that the 2005 paper had used
a somewhat different formulation of the log spectral decomposition
approach).

 IMA takes an approach orthogonal to FOOOF.  Wherease FOOOF considers the *form
of the log spectrum* of a data source - in itself - IMA pays no attention
to the mean log spectrum (removing it from consideration first of all). IMA
then considers the following question: What maximally distinct modes of *log
spectral variability* does the data source exhibit across time?

IMA, for example, could possibly isolate multiple modes whose summed
activities across time (i.e., in the grand mean spectrum) happened to
cancel each other out at frequencies of interest. Here, FOOOF would not
find any evidence of them.

More typically, IMA can neatly separate beta-specific activity from
alpha-harmonic activity, or at higher frequencies, can separate narrow band
gamma activity from high-frequency broadband.

Our paper best explaining IMA is this 2009 report
<https://urldefense.com/v3/__https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/neuro.09.061.2009/full__;!!Mih3wA!BliQLUCUjvX_OwMbwzt8CMbE7Q9URcT0MsxyVabuF7BIcWuCj2sWMO1mUUyHpYAmc7L34h2XWs8OXSJ7Y9Cf$ >
on high-frequency broadband activity during imagination exercises - though
our also-interesting results on alpha IMs in those data did not fit into
the paper as published.

Scott

-- 
Scott Makeig, Research Scientist and Director, Swartz Center for
Computational Neuroscience, Institute for Neural Computation, University of
California San Diego, La Jolla CA 92093-0559, http://sccn.ucsd.edu/~scott