I think each rafter cavity should be vented and well vented at that.
My thinking is that in cold weather moisture from the living space will enter the edge of each rafter and migrate inside the rafters toward the cold outside edges. The rafters must have a place for the moisture to be extracted.
Without ventilation on the outer edges the moisture level would rise in the cold outer edges of the rafters. These should have air flowing over the sides of the outer edges to dry them out.
In the a/c season the interior humidity should be low and no moisture from the interior should enter the edges of the rafters. If moisture would enter from the outside due to humidity or a roof leak, a ventilated air gap under the roof decking would be needed and should be effective in drying out the rafters.
You should Google Joseph Lstiburek building science. Lstiburek's videos are painful to sit through but he is considered a guru in managing moisture in highly insulated buildings. In one video he describes how hyperinsulation has been the cause of rotting of roof structure.
In particular, Lstiburek maintains that moisture moves up in buildings and will collect at the top unless there is a means of drying to the outside at the top. He stresses that the building materials must have a path to allow drying. I think it would be wrong to fill up the remaining 3" under the roof decking with anything. Use these spaces as paths for air flow for drying of the rafters and the underside of the decking.
https://youtu.be/pWYjwZq_1Ao?feature=shared
https://youtu.be/zhkGcklWB_Q?feature=shared
EDIT
After watching some of these videos I am confused. Lstiburek indicates that introducing warm moist air into the spaces under the decking between the rafters would add moisture to the rafters rather than drying them out. In the a/c season the rafters might get cold all the way from the indside to the roof deck. Warm moist outside air flowing over them might condense water onto the top 3" of the rafters. Of course in Colorado the air in the summer might be very low in humidity so this air would not be carrying moisture.
Lstiburek is telling Rissinger that dehumidifed air from the conditioned space must be injected into Rissinger's conditioned attic.
In the present case the equivalent would be accomplished by dehumidified air being carried in tubing (I guess from the conditioned space) into each rafter cavity on one end and allowed to flow back into the conditioned space on the other end.
Of course, the present case is different in that the open space is between the decking and the outside (top) of insulation. Air circulated into that space would pick up a lot of heat in the summer. So the best thing would be to vent it to the outside rather than recirculate back into the conditioned space. This means you would have vents to the outside on only one end of each rafter cavity.
However, the practical problems of boring through the drywall ceiling and through 8" of closed cell foam in each rafter cavity to pump in dry air are so great that the idea seems ridiculous.
