My hypothesis of them communicating via the main processor may be incorrect. I can almost confirm the link is controlled by the pals at u111 and u38 and uses a parallel type connection bus. It was a nice find but I had to get someone angry in the process.
Some good news is I did end up fixing the linking problem on both Cruis'n USA machines as the ribbon connector to both pcbs were backwards so reversing them fixed the link itself but not the slave monitor issue (at least that is not directly related to the link so connection issues shouldn't be a problem).