Monte-Carlo Analysis and Confidence Region Ellipsoids for Equivalent Current Dipole Solutions to EEG/MEG data

Although equivalent current dipoles (ECDs) are point source representations of the underlying neuronal currents they are in reality approximations to the true focus or foci of activity. The actual neuronal distribution is not point-source in nature, and may be distributed over a significant portion of the cortex. If an equivalent current dipole is a successful model for this distribution, either the neuronal current is localised or the spatial distribution is such that it is amenable to representation as an ECD. However, noise of all kinds will lead to an indeterminacy in the localisation of the ECD. Noise in the case of Magnetoencephalography (MEG) or Electroencephalography (EEG) refers to brain noise, signal noise, and experimental errors such as mis-registration of the detectors with the head. The estimation and display of positional uncertainties due to noise is the subject of this paper.