We study the inverse problems of source recovery and cortical mapping for sEEG, EEG and MEG under the quasi-static approximation of Maxwell's equations. These problems are known to be ill-posed due to the existence of silent source, i.e, those non-zero sources that generate null field. In addition the collected data are corrupted with noise that has to be corrected for, hence regularised Tikhonov problems will be solved. We make use of single and double layer potentials to express the electromagnetic (EM) fields associated with a source which allows to simultaneously solve the source recovery and cortical mapping problems for a particular modality or coupled data. The expression of the electromagnetic fields we use make the different couplings of sEEG, EEG and MEG data direct and complementary. Numerical results of experiments performed using meshes of realistic head geometries will be presented.