Do dreams rehearse danger? The threat simulation theory

Being chased, falling, arriving late for something that matters: many of the most common bad dreams share a theme. Something is wrong, and you need to act. In 2000, the Finnish consciousness researcher Antti Revonsuo asked a bold question about this. What if that is the point?

The theory in brief

Revonsuo’s threat simulation theory, published with peer commentaries in Behavioral and Brain Sciences, starts from the observation that dreams are not random. Compared with waking life, he argued, they over-represent some kinds of experience and under-represent others, and they are strongly shaped by frightening waking events. His proposal was that the biological function of dreaming is to simulate threatening events, and so to rehearse noticing threats and avoiding them.^[1]

The argument is evolutionary. Ancestral human life was short and full of danger, so a brain that rehearsed escapes at night, in safety, might have gained a small advantage that was passed on. Revonsuo drew on ordinary dream content, children’s dreams, recurrent dreams, nightmares, post-traumatic dreams and the dreams of hunter-gatherers to support the idea.^[1]

Evidence that fits

One prediction is that real danger should switch the system on. A 2005 study compared dream reports from severely traumatised and less traumatised Kurdish children with those of Finnish children growing up in safety. The severely traumatised children reported more dreams, their dreams contained more threatening events, and those threats were more severe.^[2]

In a 2009 review, Katja Valli and Revonsuo gathered studies that scored dreams with a tool called the Dream Threat Scale, using reports from Finnish and Swedish students, from traumatised and non-traumatised children, and from collections of recurrent dreams and nightmares. They concluded that the evidence gave strong support to the theory and that its strengths outweighed its weaknesses.^[3] It is worth remembering that the review was written by the theory’s own authors.

Evidence that fits less neatly

Independent tests have been more mixed. Antonio Zadra and colleagues in Montreal scored 212 recurrent dreams and found support for six of their eight predictions: 66% of the dreams contained at least one threat, the threats were usually aimed at the dreamer, and dreamers tended to respond in sensible ways. But fewer than 15% showed realistic situations critical for survival or reproduction, and dreamers rarely managed to escape, despite real effort.^[4] If dreams are a training ground, they are an odd one: unlikely scenarios, and few successful runs.

Researchers in Cape Town, including the neuropsychologist Mark Solms, had earlier found biologically threatening content in around a fifth of dreams, with successful escape in roughly a fifth of those. When they compared “approach” dreams, in which the dreamer moves towards something, with “avoidance” dreams, approach dreams were significantly more common, in both high-threat and low-threat environments. They suggested that dreaming may be driven more by the brain’s reward-seeking systems than by fear.^[5]

There is also a practical difficulty. The theory is about what dreams did for our distant ancestors, and that is very hard to test directly today.

A social twist

Dreams are crowded with other people, and in 2015 Revonsuo, Jarno Tuominen and Valli set out a broader idea: the social simulation theory. On this view, dreaming rehearses the social skills, bonds and networks that humans depend on. The authors framed it as testable and argued that it should be tested directly against threat simulation and against the continuity hypothesis, the view that dreams largely mirror waking life without any special function.^[6]

An early test, published in 2019, compared 232 waking reports with 116 REM and 116 non-REM home dream reports from 15 students. Dreams contained more social events than waking reports, as predicted, but a second key prediction was not supported. The authors described the support as partial.^[7]

A theory, not a settled fact

Threat simulation remains one of the most carefully argued ideas about why we dream, and it has prompted useful research. But it competes with other explanations, including social simulation, the continuity hypothesis, and theories about memory and emotion, and none of them has won. Dreaming may do several things at once, or its content may mostly reflect how the sleeping brain works rather than a single purpose.

Why nightmares still deserve care

It can be tempting to read this theory as a reason to put up with nightmares: if they are training, perhaps they are good for us. That does not follow. Even on its own terms, the theory describes a possible function in the ancestral past, not a reason to endure distress now. An alarm that rings every night is not keeping anyone safer. Frequent nightmares that disrupt sleep, make you dread bedtime or spill into the day are a recognised condition, nightmare disorder, and many people find they respond well to treatment.

Imagery rehearsal, the approach sleep medicine recommends for nightmare disorder, works with exactly this kind of recurring threat dream, such as being chased: you change the story while awake and rehearse the new version. Here is how it works. If your nightmares are tied to trauma, please see a professional; help works. And if nightmares come with thoughts of harming yourself, please reach out to a crisis line or call your local emergency number today.

Baku’s Rewrite studio is built on the same idea: the threat can stay in the story, but the ending is yours to write.