Model of disk surface density distribution which invokes both MRI and GI. The MRI is efficient only in the innermost disk, where thermal ionization exists, or at large radii where cosmic and X-rays can ionize enough of the disk. The model is compared with recent results from submm interferometry at large radii.
Research by former graduate student Zhaohuan Zhu: Long term evolution of protoplanetary disks

As described on the research page on forming stars with disks, the large factor by which protostellar clouds must collapse to form stars means that any finite angular momentum results in collapse initially to a rotating disk. The further implication is that much if not most of the mass of a typical star must be accreted through the disk. The mechanisms of angular momentum transport, along with the initial sizes and masses of disks, then set the mass distributions within these disks that are the starting point for planet formation.

While it appears that gravitational instabilities (GI) are adequate to transport mass inwards so that most of the material ends up in the star, GI transport will stop when the disk is still quite massive. The open question then is how efficient other mechanisms of transport are. The magnetorotational instability (MRI) can be quite effective, but only in regions where there is sufficient ionization to couple the magnetic field to the gas. Protostellar/planetary disks are likely to be quite cold and low-ionization objects over wide ranges of radii; this will result in a “bulge” of mass where transport is inefficient once GI has ceased. In our schematic models of disk evolution, we find that large mass bulges can be present at scales of a few AU, as shown in the figure on the left. This is actually desirable from the point of view of explaining FU Ori outbursts, where we need to accrete large amounts of mass over short timescales – putting the mass at large radii won’t work. The result obviously has major implications for planet formation and migration.