Can you modify the circuits to accommodate multi-colour LEDs (4 leads)?

The circuit boards could be redesigned to take LEDs like that, but it would be a major change. Even more of a major change if it is to drive the different color LEDs to make something interesting happen.

What if all they did was cycle through their full colour range?

If they all started from the same colour and the signal is propagated to the adjacent LEDs, it would emphasize the ripple effect more.

Perhaps you can use bi-, or tri- coloured LEDs, though I don't think it would look as good.

Being able to adjust the change rate would be nice, but you could link it to the "ripple rate."

Right now we do not have separate "color control" signals going to each LED-- they would have to be added, which would require redesigning the circuits. Not impossible, but not trivial either.

I don't know if using the two-terminal color-changing LEDs would work or not, with the way that we are driving them; I will investigate at some point.

What if you attached half the LED bank per module to one color (blue maybe, the other half to green) so the area oscillates between blue and green. Take signals from the adjacent modules and attach those to the red leads of the LEDs closest the the adjacent module. You'd start with blue green oscillations locally, and the color would shift further toward red as the patterns ripple outward.

That works, assuming all the LEDs feature a common anode or a common cathode. I don't have a board or schematics, so I couldn't tell you for absolute certain. There is a possibility that you would have to reduce the number of LEDs per module, or duplicate the LED driving circuits and beef up the power supply to support more LEDs. Each color will require a different resistor to level out the current draw and brightness.

What if you attached half the LED bank per module to one color (blue maybe, the other half to green) so the area oscillates between blue and green. Take signals from the adjacent modules and attach those to the red leads of the LEDs closest the the adjacent module. You'd start with blue green oscillations locally, and the color would shift further toward red as the patterns ripple outward.

Should work... Our tables really work best with blue/green/white LEDs, but blue-green oscillations would be cool.

If the Red had a higher resistor than would normally be installed, it'd be naturally dimmer. I should have added that before. By combining the red with the blue green oscillations, you would get something of a yellow, orange, violet mix spaced in with the blue and green. since each LED has either blue green, or a varied mix of the two on at any given time, your color would shift toward red, but never reach red, as it would always mix with the blue and green. It doesn't give you full RGB range, but I find red to be slightly annoying, so I think maybe others might not mind the lack of pure red. There might be some dim red on the leading edge of a propagating ripple, providing that the outward ripple starts dim.

This is all theory to me, since I have not actually had the opportunity to build one of these yet.

I want one though... for my cat!