I was lucky enough to inherit an xbr4 with this fault a while ago, and so I have been spending a bit of time looking into the cause of the fault.
After some rather in depth analysis on the test bench, it looks like the problem is not with the flex interconnect's bonding to the panel, but rather with the connecting bonds between the flex interconnects and the long, narrow semiconductor chip that is mounted directly on the flex interconnect, about halfway between the LCD end and the PCB end. This chip looks like it might be a signal line termination or impedance matching network, and gets quite hot during operation (the design of the big metal LCD retaining frame includes a series of long depressions along the top section that push against the flex interconnects where these chips are located, to provide some not very effective heatsinking).
The heat from these chips, combined with poor heatsinking, and the structure of the LCD frame appear to be the root cause of these failures.
I am working on a repair method, though it will likely be something ad-hoc, and not something that could be classed as a "permament" fix. Still, if it can resurrect even some of these panels, then I think it is worth a shot.