From my point of view - power sequencing is poorely organised by driving ENABLE pins of low dropout linear regulators.
For n770 all power sequencing for LCD is being held by state of voltages vs loads.
It can lead to a failure.

I've had an opportunity to exchange display in my wsod'ed n770 and it DID work.

For LCD's manufactures give detailed recommended power sequencing diagrams.
Not knowing one it's hard for me to judge what would be optimal.
Yes - there is a possibility to monitor all signals with the scope and doing some tests. Then one probalby could notice the source of this failure.

But... is there anyone who wish to donate 3 x n770 ?
One for finding critical sequence, second to test other sequences except probable one and the third to make some normal vs. crash sequence benchmark?

After all that - one could develop some fix if applicable.
Yes - maybe it's just easy as solder few additional resistors, diodes and capacitors on critical paths but - without knowing for sure - I can only guess.