Improve support for yokogawa (and possible future OPM systems)

[contsili]

ft_sensorrealign is obsolete. ft_electrorealign can be used instead.

However, cfg.method = 'fiducial' can be used only for elec structures and not for grad structures. So instead we use cfg.method = 'template'.

[contsili]

Hi @robertoostenveld @schoffelen,

In my Yokogawa dataset, using cfg.method = 'template' to align polhemus and sensors based on three fiducials didn't perform well. For >3 fiducials it seems to work better; see FT coreg tutorial.

To align Polhemus data with Yokogawa sensors (+ newer OPM systems will benefit from that!), I suggest we expand ft_electroderealign to allow for cfg.method = 'fiducial'to be used also for grad structures, or revive ft_sensorrealign or create a new function ft_gradrealign

[contsili]

OPMs have the flexibility to be on the scalp as EEG, so many EEG methods can be extended to the OPMs e.g., via the ft_gradrealign. For example I needed to use the cfg.method = 'moveinward' for my thesis and there was no FT code to that for grad but only for elec. So I wrote my own code in the end.