General workflow#

Prerequisites#

MEEGsim re-uses the forward and inverse modeling functionality provided by MNE-Python, so to perform simulations with MEEGsim, you will need:

A source space (mne.SourceSpaces) that contains all possible locations for simulated sources. The source space is required for every simulation.
A forward model (mne.Forward) that contains all sources in the source space mentioned above. The forward model is required only if you wish to obtain sensor-space data or adjust the sensor-space SNR of simulated sources.
An mne.Info structure describing the channel layout is required for projecting the simulated source activity to sensor space.

MEEGsim provides two main classes that should be used for performing the simulations:

With the SourceSimulator class, you can add sources of activity to the simulation, configure their signal-to-noise ratio and set up coupling between sources. The class does not contain the simulated data. Instead, it should be used to generate different configurations of sources (see below) as instances of the simulation (e.g., if some of the sources are placed in random locations).

The SourceConfiguration class actually contains the simulated data. We store the waveforms of all sources separately to ensure that they are always accessible (e.g., in case two sources overlap in space) but also provide methods for converting the configuration to commonly used stc and raw objects for source and sensor spaces, respectively.

Learn how to describe the simulation in the next section.