Be careful with not over-interpreting source results. Anything you see only using one specific source model is suspicious. For deep structures, compare what you obtain with fully unconstrained models (volume source models: https://neuroimage.usc.edu/brainstorm/Tutorials/TutVolSource).
You should also always find a way to test the significance of the results you obtain (comparison with a baseline or between conditions). Here you could maybe compare between the brain responses to stimulations at different sites?