The edge of chaos can be thought of as a zone in the ratio between parameters where evolution or optimization of a system (in this case, maximizing information processing) can occur with access to the most dynamics (needed to explore optimization space) without the system becoming unstable.

It's traditionally applied to order/disorder, but in this case the order and disorder are more abstract as the control mechanisms in question are the excitation or inhibition of a stimulus. Excitation and inhibition are a way of controlling the dynamics of the brain, and in general are two basic controls for maintaining stability in any embedded dynamic physical system, and according to the good regulator theorem[0], a perfect inhibition/excitation control system contains a perfect model of both the embedded system and its local environment, so even learning and understanding are rooted in these two basic ways of interacting with information, and thus we see the edge of chaos appear all over in biology, neuroscience and physics if you squint a bit.

[0] https://en.wikipedia.org/wiki/Good_regulator