Electro- and Magneto- encephalography (EEG and MEG) are measuring non-invasively the post- synaptic brain electrical activity. At each time instant, EEG and MEG provide a few hundreds of measurements at the scalp (or near) level, and these measurements can be obtained at about a millisecond time resolution. In the frequency domain of interest for brain signals, the electromagnetic propagation over the head is governed by the quasi-static approximation. The inverse problem of source localisation seeks to identify the cortical electrical sources from the measurements. As the cortex is a very convoluted surface, it is usually modeled with several thousand variables, which make the source localisation an ill-posed problem. We will explore several strategies to solve this inverse problem with different strategies to regularize this ill-posed problem.