Dreams are full of possibilities; by drifting into the world beyond our waking realities, we can visit magical lands, travel through time and interact with long-lost family and friends. The notion of communicating in real time with someone outside of our dreamscapes, however, sounds like science fiction. A recent study demonstrates that, to some extent, this seeming fantasy can be made real.

Scientists already knew that one-way contact is attainable. Previous studies have demonstrated that people can process external cues, such as sounds and smells, while asleep. There is also evidence that people are able to send messages in the other direction: lucid dreamers—those who can become aware they are in a dream—can be trained to signal, using eye movements, that they are in the midst of a dream.

Two-way communication, however, is more complex. It requires a person who is asleep to actually understand what they hear from the outside and think about it logically enough to generate an answer, explains Ken Paller, a cognitive neuroscientist at Northwestern University. “We believed that it was going to be possible—but until we actually demonstrated it, we weren’t sure.”

For this study, Paller and his colleagues recruited volunteers who said they remembered at least one dream per week and provided them with guidance on how to lucid dream. They were also trained to respond to simple math problems by moving their eyes back and forth—for example, the correct response to “eight minus six” would be to move your eyes to the left and right twice. While the participants slept, electrodes attached to their faces picked up their eye movements, and electroencephalography (EEG)—a method of monitoring brain activity—kept track of what stage of sleep they were in.

As Paller’s team was conducting these experiments, the researchers discovered three groups in Germany, France and the Netherlands who were trying to accomplish the same thing. Instead of competing, the groups decided to collaborate. They carried out similar experiments, though with slightly different methods of answering questions and receiving responses. The German group, for example, transmitted its math problems using Morse code, and the French group asked its participant— a person with narcolepsy who had expert lucid-dreaming abilities—to answer yes-or-no questions with facial muscle contractions rather than eye movements.

Across the four studies, there were a total of 36 participants and 158 trials during which the researchers could verify lucid dreaming and attempted to establish contact. Answers were considered correct if three of four raters were in agreement on whether the responses, sometimes very subtle movements, were accurate. Correct responses were given in 18 percent of trials; another 18 percent were classified as ambiguous because raters could not come to a consensus about whether participants gave a correct response or whether they had responded at all. Incorrect responses were given in 3 percent of the trials. Overall there was no response in 60 percent.

One of the co-authors, Karen Konkoly, a graduate student in Paller’s lab, speculates that participants failed to respond in 60 percent of the trials because they simply did not perceive the incoming communication. In those cases, they rarely reported any incorporation of the questions into their dreams after waking up. But she adds that it is also possible that dreamers perceived the inputs but paid little attention and forgot before awakening. The proportion of people who respond could potentially be improved with more training or by presenting questions when individuals are in specific sleeping brain states, Konkoly says.

After establishing successful two-way communication, participants were woken up and asked to recount their dreams. In most cases, they could remember receiving the experimenters’ questions while asleep; in some cases, the questions appeared to be coming from outside the dream, whereas other times they were integrated into the dream. (One participant reported that the lights in their dream started flickering, which they were able to recognize as the Morse-coded math problem.)

There were instances, however, when people either did not recall the interactions or had a distorted account. For example, there were trials in which individuals answered a math problem correctly while asleep but did not remember the question correctly after waking up. These findings were published in February 2021 in Current Biology.

The findings “challenge our ideas about what sleep is,” says Benjamin Baird, a researcher who studies dreams at the University of Wisconsin–Madison and was not involved in this study. Sleep has classically been defined as unresponsiveness to external environmental stimuli—and that feature is still typically part of the definition today, Baird explains. “This work pushes us to think carefully—rethink, maybe—about some of those fundamental definitions about the nature of sleep itself and what’s possible in sleep.“

This kind of two-way communication with dreamers could be used as a tool to better study dreams, according to Paller. In particular, he says, the observation that the responses some people gave during dreams did not match their reports after waking provides evidence that such real-time techniques will help researchers get more accurate accounts of dreams—and address whether dreams play a useful role in processes such as memory. Paller and his colleagues also suggest this technique could be used by people to enhance problem-solving and creativity, by providing a new way to process content in their dreams.

“I really liked this study,” says Christine Blume, a sleep scientist at the Center for Chronobiology in Basel, Switzerland, who was not involved in this work. “The extent to which information can be processed and responded to surprised me.” But she adds that it is important to keep in mind that the findings relate specifically to lucid dreaming, which is a special type of dreaming that not many people are able to experience.

Blume notes that even with lucid dreamers, in most trials, the researchers were not able to establish communication. Therefore, how applicable this technique would be to learning or creativity remains an open question, she says.

Paller and his colleagues are now exploring what other types of questions can be asked during sleep, as well as other ways of receiving messages from sleepers, such as sniffing. “We are hopefully going to get better at doing this kind of experiment,” he observes. “Then [we can] ask new questions about what’s happening during dreams.”