Source analysis of EEG oscillations using high-resolution EEG and MEG
Ramesh Srinivasan1, William R Winter, Paul L Nunez
1Department of Cognitive Sciences, University of California, Irvine, USA. r.srinivasan@uci.edu
Progress in Brain Research
|October 31, 2006
Summary
This study reveals that simple "equivalent dipole" models are insufficient for understanding brain oscillations. Comparing electroencephalography (EEG) and magnetoencephalography (MEG) provides new insights into the spatial distribution of neural sources.
Area of Science:
- Neuroscience
- Biophysics
- Signal Processing
Background:
- Scalp electroencephalographic (EEG) oscillations exhibit complex spatio-temporal dynamics.
- Previous models simplifying neural sources to a few equivalent dipoles are inadequate for capturing this complexity.
- Understanding the spatial properties of neural sources is crucial for accurate interpretation of EEG and magnetoencephalography (MEG) data.
Purpose of the Study:
- To develop and apply an approach using volume conduction models to characterize the spatial filtering of cortical source activity by different neuroimaging techniques.
- To infer the spatial properties of EEG oscillations without making prior assumptions about the neural sources.
- To compare EEG and MEG recordings to gain insights into the underlying source distributions.
Main Methods:
- Utilized volume conduction models to simulate and analyze the spatial filtering effects of average reference EEG, high-resolution EEG, and MEG.
- Applied the approach to spontaneous EEG oscillations recorded during eyes-closed rest.
- Compared the spatial characteristics and frequency content of alpha and theta rhythms between EEG and MEG.
Main Results:
- Both EEG and MEG detected robust posterior alpha rhythms, but with differing dominant frequencies.
- Frontal alpha and theta rhythms were primarily generated by superficial radial sources, yielding strong EEG but weak MEG signals, likely originating from frontal gyral crowns.
- MEG and high-resolution EEG suggested local tangential and radial sources in posterior cortex, on both sulcal and gyral surfaces, while also indicating contributions from widespread, non-local sources.
Conclusions:
- The spatial filtering properties of EEG and MEG, when analyzed with volume conduction models, allow for inferences about neural source distributions.
- Simple equivalent dipole models are insufficient; complex source distributions, including superficial radial and local tangential/radial sources, contribute to observed oscillations.
- Alpha and theta rhythms involve contributions from both local cortical sources and potentially larger-scale, non-local activity.
