That really surprises me, since the disk is also orbiting and it is much easier to see how disk interactions could cause circularization than it is to see how they cause the "death spiral", and so quickly. It sounds like this is telling us that circularization happens very fast, and is due to interactions with disk material at the same distance to the Sun as the planet, whereas there is a slower (but not a lot slower!) interaction with disk material that is preferetially farther out (which is what you need to extract angular momentum from the planet). Note these "death spiral" models need the disk to continue down toward the Sun, whereas the outward migrations need the disk to extend outward away from the Sun (you have to keep the planet immersed in the disk to keep circularizing the orbit). So the Neptune/Uranus conjuctions must not do that much since they can't make both planets migrate outward (perhaps it helps separate the two orbits, though) whereas the resonant interactions with Saturn and Jupiter can cause an outward spiral even in the presence of the circularization.