source localization & synthetic data

source localization methods in general aim at estimating brain activity inside the skull based on a signal measured outside the skull, the EEG or MEG. They have to solve the so called 'inverse problem': with a given surface signal and with the brain activity unknown. The other way around, the 'forward problem', with the brain activity given, and the surface signal unknown, is a lot easier, but simply the other side of the coin. Thus if a model can give you an inverse solution, it can also give you a forward solution, for instance for simulation purposes. That is the reason, why source localization and generation of synthetic data is presented together in the present section.

Brain activity is most often estimated in terms of 'dipole activations'. That is because electrical dipoles ( generators that have one Plus and one Minus pole ) are the main generators of a measured EEG and MEG, as the have a considerably spread electrical and magnetic field. Other electrical sources (e.g. a circular set of Plus and a central Minus Pole ) have a much more closed electrical and magnetic field, and therefore will not contribute much to the EEG and MEG signal.  Moreover,  postsynaptic potentials on pyramidal cells of the cortex, the main source of the EEG and MEG signal, are classical dipoles in this sense.

Different source localization methods differ concerning the number of dipoles that are used, concerning the degree of realism of the used head model (it's shape, it's conductivity properties etc. ), and concerning the additional assumptions, that are used to solve the inverse problem.

Emegs offers a source localization method that works with multiple distributed sources (dipoles) and a spherical head model.  This method solves the (in this case underdetermined)  inverse problem, by using the additional constraint of preferring the solution with the least overall energy, and therefore is called as this constraint  minimum norm solution.

Using a forward solution with  the same head model, emegs offers a tool for generating synthetic data based on user defined dipoles.

Emegs also lets you perform a related method, the curent source density estimation, which aims at improving the spatial resolution of scalp EEG without making any assumptions about a head model and dipole locations. Moreover, it allows to estimate the potential distributiont on the cortical surface.