Call it a nerd crossover, but I've always been curious about how we play games on a mechanistic level. However, even coming from a rational viewpoint video games seem like a fruitful domain of study, since data collection is simple and complete, environmental variables can be easily manipulated, and the task is far richer than typically used in cognitive psychology experiments.

So, have any psychological theories been formed by/applied to in this domain? It seems like eye-tracking and cognitive modeling approaches are both particularly well suited to this domain, but any work in cognitive science on the phenomena involved here would be interesting to see.

The only things I've found are the interesting experimental work on semantic actions in Tetris, and some modeling work done on unreal tournament within the ACT-R community. It should be noted however, that the latter example uses a video game environment as decent simulation of reality and thus models an agent inside the game, whilst the former focuses on the game from the player's point of view. This might be a personal bias, but the latter seems much more fruitful and more in line with the spirit of my question.