Interesting, I tried setting my PCIe slot to GEN 3 but the problem occurred again. I might play around with the SOC voltages.
I've been doing some other troubleshooting:
When using the primary monitor (PG35VQ) in deep sleep mode or turned off with the Index plugged in, the issue occurs.
When using the primary monitor (PG35VQ) with deep sleep disabled and the Index plugged in, the issue does not occur.
When using only the Index plugged in, I think the issue occurs judging from the Q-Code and LEDs - although obviously no way to see if there is an error message on POST.
When using the primary monitor (PG35VQ) in deep sleep mode and a secondary monitor (PG279Q) instead of the Index, the issue does not occur.
When using the PG279Q as the primary monitor in deep sleep with the Index plugged in, the issue does not occur.
So my conclusion is this is occuring because the Index is being interpreted during POST as the primary monitor. Unfortunately, the PG35VQ has a very deep sleep mode so it is unable to turn on quickly enough for the BIOS to see it as the primary monitor - instead it sees the Index and flips out. With the PG279Q with a faster activation from sleep the problem does not happen.
Now the question is why does the display type have any bearing on the motherboard correctly or incorrectly interpreting the presence of UEFI/GOP on the graphics card? Surely this is on the card or not, and the type of display should not impact boot.
This behaviour is unique to this motherboard though, as I said the exact configuration worked okay with the previous motherboard (Maximus VIII Hero).