Small particle reconstructions are "hollow"

I’m working with a fairly small particle size on the order of 100 kDa, and it seems like both the heterogeneous and homogenous refinement jobs are wanting to align really strongly on the edges of my particles, yielding what is essentially a “hollow” 3D volume. Example below.

Wondering if there are any advanced parameters I’m not aware of that could be helpful for working with this dataset and overcome this issue.