One way I know this is possible, from prior research, is by implanting electrodes into the rat's hippocampi to detect activation in the place cells. In a nutsh

One way I know this is possible, from prior research, is by implanting electrodes into the rat's hippocampi to detect activation in the place cells. In a nutshell, place cells are cells in the hippocampus that are associated with particular regions of the environment (their 'place field'). They activate when the animal visits a place, thinks about a place or, critically, dreams about a place. The theory is that they're used by your hippocampus to tie together memories gathered in a particular place during memory consolidation, to help you recall details about a familiar environment, and to help build navigation maps of the environment (by linking place cells' fields together where their fields are close or overlap). In one study I read a few years ago, neuroscientists implanted electrodes into a rat's brain and had them run the maze repeatedly, tracking where they were in the maze and what cells were activated. Later, when the rats were asleep, they recorded the activation again, and found that they activated in the same sequence that they did when the rats were actually in the maze. This strongly implies that they were dreaming about running the maze. This study seems to be an extension of that where they showed they could direct the sleeping rats through the maze with 'turn left' and 'turn right' cues the rats learned while awake. http://en.wikipedia.org/wiki/Place_cell http://en.wikipedia.org/wiki/Place_cell Edit: To be clear, the hypothesis that this replay happens because the rats are dreaming is pure speculation. My personal hunch is that's what's going on, but that's all it is: a hunch, an assumption. There's no way to be sure whether they are experiencing what we call 'dreaming'. All we know is that the place cells in the rats' hippocampi 'replays' the place cell activation in the same order that they activated while awake.