Society for Neuroscience Washington, D.C.
Nov. 12-16, 2005
Julie Onton and Scott Makeig , Swartz Center for Computational Neuroscience, Institute for Neural Computation University of California San Diego, La Jolla CA, USA
Variability of peak alpha frequency within single subjects examined with independent component analysis
A proposed method for normalizing peak electroencephalographic (EEG) frequencies across subjects in an experimental pool is to determine the 'individual alpha frequency' (IAF) for each subject and create individualized theta and alpha bands according to this frequency (Doppelmayr, M., Klimesch, W., Pachinger, T., Ripper, B., 1998. Biol Cybern 79, 49-57). We show here that the mean peak alpha frequency across scalp electrodes represents an average of possibly widely varying frequency peaks of independent EEG processes. When mixed signals at scalp electrodes were decomposed using independent component analysis (ICA), independent sources with different peak alpha frequencies were revealed. Our results demonstrate not only considerable inter-subject variability in peak frequencies (as noted in the literature), but also that a variety of peak frequencies were commonly exhibited by independent sources in the same brain. This finding suggests that reports showing task-related power changes exclusive to 'lower' or 'upper' alpha frequency bands could in fact reflect power variation in peak alpha frequencies of different EEG sources.