Stephen LaBerge has been a natural lucid dreamer every since he was a child, but his interest in the subject blossomed in the 1970s when, as a graduate student at Stanford, he began working with pioneering sleep researcher William Dement.
Back then, scientists were aware of lucid dreaming, but they dismissed the phenomenon as a hallucination having nothing to do with sleep. In a series of landmark experiments, LaBerge proved that assumption to be false.
In his studies, lucid dreamers were wired to polysomnogram machines and electronically monitored while they slept. When the subjects entered REM sleep, they would try to perform tasks agreed on before they went to sleep: They'd breath rapidly, for example, or move their eyes in a steady left-right patterns. When these actions were detected and recorded by the polysomnogram, LaBerge had proof it was possible to act consciously while sleeping.
In another fascinating experiment, LaBerge electronically charted the eye movements of subjects as they followed a moving fingertip with their gaze. The resulting graph was a smooth, gracefully undulating line.
Next, the subjects were asked to close their eyes and follow an imaginary fingertip. This produced a halting, jagged line on the graph.
Finally, the subjects went to sleep and were asked to follow a moving fingertip in their dreams. When the subjects entered REM phase, they would signal with several steady eye movements (visible on the polysomnogram) that the dream had begun, then they'd follow the moving finger. This yielded a smooth, graceful line virtually identical to the graph produced during the waking phase of the experiment. According to LaBerge, this is evidence that dream activity is more similar to waking experience than it is to imagination.