Best fitting sphere estimation

As explained in the head modeling tutorial (neuroimage.usc.edu/brainstorm/Tutorials/TutHeadModel), these spherical models approximate the head with one unique sphere.
But the head is not a sphere, there is no way around it. Those different options are just different suggestions for the sphere, there is not necessarily one better than the other. In the end they are all bad, in different ways.

Use instead the head models suggested in the same tutorial: Overlapping spheres (MEG) and OpenMEEG BEM (EEG).

Cheers,
Francois