Inconsistency of cerebellum's forward grid and inverse source grid

Dear editors,
Recently I tried to start a reasearch aiming on the ERD in cerebellum areas so I computed volume singlesphere forward model using a group template, and the grids of the cerebellum areas seems to be good.

Then I used a lcmv beamformer to reconstruct the source, but the cerebellum's data didn't show up even if I adjusted the amplitude to zero and it went like this:

I also tried the OpenMEEG volume forward model and it turns out the same result that seems like cerebellum's grid didn't get involved into the inverse computation.

It seems that problem happended in the case that I tried to warp a template grid to individual space. I checked that my template grid already included the cerebellum part like this:

While not using the template warping method aiming on group analysis but building on individuals' own regular grid turns out good.

Now it seems like what kind of set did I miss of the template warping method?

Can you check if your forward model includes all the grids?
Right-click on the "head Model" and then "check source grid".
Check both options "cortex" and "MRI"