Next experiment idea: simultaneously send different lasers to each cone.

Pretty neat. tl;dr: they produce impossible colors by selectively activating one specific type of photoreceptors in the eye with a laser. Which is normally impossible as there's significant overlap in spectral response in different types of cones. They map the retina in a pretty elaborate way and send the pulse to M-cones only, driving it with a FPGA to achieve real-time performance.