I would like to know what is the best way to train mice in a virtual environment to perform behavioural tasks (to run consistently, to train them to reach a specific target without stops before to reach this target, and to navigate in 2D).
Yes, We can't perform that in humans. The idea is to perform a cognitive task under specific conditions involving genetic manipulations and to record simultaneously the neural activity. That's why we need an animal models .
First you train the mice to do a spesific behavior that produces a reinforcer. When this is established, you can shape the behavior. For example expand the time or distance the mouse have to run to get the reinforcer. The virtual environment may eventually function as stimuli controlling the behavior. That is that some images in the VE may serve as discriminative stimuli that signals the reinforcer that can be produced by the "correct" behavior, and other images signals that the behavior will not produce reinforcers. There are different studies on how to train rodents. And it is difficult to control for these genetic manipulations if the animal models have different learning history. I am not sure if this answer his helpful, but if you need to get in touch with researcher that has experience in training rodents I'll be happy to give you some names. https://www.youtube.com/watch?v=fyHJxZB3pMs
Please explain how virtual reality is provided for rodents? Can mice see a 3D digital images? Have you developed stereo glasses for mice or are you using an analogue of CAVE? How do you prove that mice really see and respond to 3D digital images? Perhaps they act on some other motives?