I've spent quite a bit of time over the last year drawing blocks (microplates) that are partially bounded by faults, to make block models that I can use to invert geodetic velocities and geologic slip rate data to get internally-consistent slip rates for all faults in a deforming region.
One of the challenges of block modeling (above simple mapping of fault traces) is that faults are pretty clearly discontinuous in the crust, and it's generally quite ambiguous how they are joined. With most fault-oriented studies it's easy to ignore areas without obvious faults, but when you're thinking about the continuous strain field in the crust, you can't think of faults just ending. Something has to happen in between the faults. If the blocks between faults are not very rigid, then the strain field may be pretty consistent away from the faults. Alternatively, if the blocks are actually fairly rigid, then there may be somewhat narrow zones of diffuse deformation that join faults.
There are a couple of different ways of addressing this. One of them is through high resolution bedrock (and perhaps Quaternary) geologic mapping, to try to estimate the strain of the crust in between known faults. A big issue with this approach is that plastic bedrock deformation is cumulative so it may not be easy to separate strain from the current deformational regime from earlier strain, but this is going to vary regionally quite a bit. However you basically get a maximum amount of strain (assuming you can actually quantify anything) which may still be quite useful.
A second approach, that might be a bit quicker, is to look at the distribution of microseismicity (and likely moderate seismicity as well). If strain away from faults is concentrated in fairly narrow bands, there should be more seismicity in the bands. Furthermore, if the total strain rates across narrow strain zones are comparable to slip rates on the faults they connect (and in a block model context, they are) then there should be much more low-magnitude seismicity in these zones than around the faults themselves, on average, as faults should host larger earthquakes which soak the majority of seismic moment, while the diffuse strain zones are probably less likely to have regular large-magnitude seismicity and therefore expend much more of the moment on the little events.
As always with geology, it's unlikely that all deforming regions will behave similarly, so any studies of this will have to proceed on a region by region level until broader patterns emerge. It's quite possible that this has been addressed in some regions, but I haven't come across the papers. However, now that the GEM Global Active Faults database is published, global-scope and local-resolution studies should be possible, though a compatible global block model would make it quite a bit easier.