With plain old standard (including rottendog) hardware, this is a bad connection between MPU and display board. Primary suspects are the connectors, and R95 as mentioned by astyy. It's a hard-working resistor so look up the band colour code and measure its resistance. If it doesn't match the bands, replace it. (Manual says it's a 33K 1/2 Watt resistor.) One of those rare cases where you do have a legitimate case to start looking at passives like resistors...
For the connectors, clean aggressively with solvent like IPA and try again. If there's corrosion on the pins, replace the connectors.
After both of these you will be 100% right as rain. If not then the machine's power supply is starting to go - that's the only remaining candidate and it's not my first suggestion as other effects would likely show up, as well.
Nothing wrong with the above mentioned general maintenance checks (e.g. reseating display ROMs) but that command overrun message has a specific meaning, and it is a breakdown in signal timings between MPU and display board - on these machines the MPU and the display board have separate CPUs working independently, and this is the display's CPU telling you that it's getting crappy inconsistent messaging from the MPU. If the game works in the slightest (proving that the MPU is sane) then the only candidate is the copper between MPU and display board.
(Just for entertainment's sake, it's worth mentioning that the MPU in these Data East/Sega games will usually not boot on the bench unless the display board is connected and powered on - the MPU tries to send ROM version data for the display board to show, and it waits for the display board to give the go-ahead to send that info. Without the handshake the MPU will sit and do nothing useful. They're not like Bally/Williams games that will at least have a fair shake at running 'headless'.)