spm_opm_headmodel: Solution to fiducial issue

[georgeoneill]

Now with less bugs. On test script this fails in two places.

S               = [];
S.data          = 'meg.bin';
S.coordystem    = 'coordsystem.json';
S.positions     = 'positions.tsv';
S.channels      = 'channels.tsv';
S.template      = 1;
S.meshres       = 3;
D               = spm_opm_create(S);

The registration is all wrong, but looking at the coordsys file the fiducials in the test data are nonsensical. it would explain why the registration is off. (and also why there are no marked fiducials in the above plot as the fiducials are in the centre of the brain! I propose updating the coordsys file going forward with the true locations.

Test code also fails at

S =[];
D = spm_opm_create(S);

I think this is due to the fact all sim code has been removed from spm_opm_create a while ago so this will no longer work as expected. Propse this test is removed from the script entirely.