I am a late adapter of DCS having bought into the equipment in the last 3 years so I'm unfamiliar with the recommendation to also isolate the ground rails. Seems like it would work, I think it is maybe extra work for little gain as the signal will follow the shortest path, for the most part. It could be part of a solution in some cases.
The main goal is to avoid any duplication (reflections) of the DCS packets on the power rail so in theory we isolate the track at length short of where signal propagation might result in the ghosting of packets, and degrade signal integrity. I believe this is why the DSC Companion recommends blocks no longer that 100 feet.
But we also have to contend with electrically noisy track connections which negatively impact packet detection by increasing the noise, so DSC Companion recommends 5 or 6 track joints as a limit for passing the DSC signal through. The DCS companion also recommends keeping blocks to not more than 12 track sections. My interpretation of that spec is that it applies in two directions from the source, the power rail drop in the block.
If there is a terminal strip on the output of the TIU channel, as it seems most of us do, then that is a split in the signal that affects every block attached to it. That seems like the nucleus of the star to me.
These are my thoughts as to how I follow the DCS recommendations. It is a conservative approach. I will not claim this is the only way to get things to work.
BTW, I love seeing pictures of your work.