Could be an IC issue, could be the impedance control of the connector, could be the impedance control of the traces on the circuit board, could be poorly chosen passives, could be interference from nearby circuits, etc.
I wish I could give you a better answer but high speed signal integrity is one of the most obtuse and most black art aspects of electronics design, there's a fair amount of non-intuitive physics going on, e.g. signals going down a wire don't just go, they 'ricochet' off impedance differences. It's far too large a topic to treat in a HN comment and its whole system encompassing so 'what it could be' is a large set.
These types of things are not uncommon in hardware development, especially with high speed signaling. The interactivity of such systems can be high and hard to predict fully at times.
EDIT: Simply put - display outputs scan top to bottom, left to right. The fact that the pattern that triggers the issue appears to be horizontal means it would occur in the raw bit stream of the output in a periodic way. That means there is some low order frequency in the system caused by this pattern (maybe it repeats every 500khz or something like that). If there is something in the electronics of the output of the RMBp that resonates at that frequency (easier than you think), then it could possibly cause the described failure.
