The body swap illusion

Body ownership - the sense that one's body belongs to one's self - is central to self-awareness, and yet is something that most of us take completely for granted. We experience our bodies as being an integral part of ourselves, without ever questioning how we know that our hands belong to us, or how we can distinguish our body from its surroundings.

These issues have long intrigued philosophers and psychologists, but had not been investigated by neuroscientists until recently. Now researchers from the Karolinska Institute report that they have induced a "body-swap" illusion, whereby subjects perceived the body of another person as belonging to themselves. Their findings are published today in the open access journal PLoS One.

Our sense of ownership of our own body depends largely upon the brain's representation of the body, which is constructed largely from various types sensory inputs. For example, visual inputs provide us with information about body posture; tactile inputs provide information about objects which we come into contact with, and proprioceptive inputs inform us about the position of the body within its surroundings. Normally, the position of one's body corresponds directly to the way it feels and the way it looks and, as a result, one can identify their own body as belonging to themselves, and also distinguish very easily between the "self" and "non-self" (or objects in the external world).

Under certain circumstances, however, the sense of body ownership can be perturbed, and sometimes this has bizarre consequences. Take, for example, somatoparaphrenia, a condition in which ownership of the left hand or leg is denied, following damage to the right parietal lobe. Often, patients with this condition perceive their limb to belong to, and to be controlled by, another person. Or take hemispatial neglect, a related disorder, which also occurs as a result of brain damage. This causes a complete loss of awareness of half of the body, and causes the patient to act as if it didn't exist - she will apply make-up to one half of the face while ignoring the other, and eat food from only one half of a plate that is placed in front of her. If the plate is then rotated by 90°, the patient will continue to eat half of the remaining food!

As well as going awry in various neuropsychiatric conditions, the sense of body ownership can be distorted experimentally in healthy individuals, by manipulating the sensory inputs entering the brain. Henrick Ehrsson, who is now at the Stockholm Brain Institute, was one of the first to demonstrate this. In a 2004 study, Ehrsson and his colleagues induced in their subjects what is referred to as the rubber hand illusion. This was done by seating the subjects at a desk, with one hand hidden from view. In front of them on the desk was an artificial rubber hand. When the hidden and rubber hands are stroked simultaneously and in the same way, the visual information overrides the proprioceptive inputs, evoking the illusion that the sensation originates in the rubber hand, so that the subject perceives it as belonging to them.


 The rubber hand illusion activates neurons in the premotor cortex. The felt touch is represented in green, the seen touch and illusory position of the arm in blue, and the visual field in yellow.
(From Botvinick, 2004)

This illusion is the exact opposite of somatoparaphrenia - the artificial limb is perceived as being part of one's own body. Ehrsson's subjects underwent functional neuroimaging whilst the illusion was induced; this showed that the illusion was correlated with activity in the premotor cortex, which contains neurons that are known to respond to both visual and tactile stimuli, and to integrate these stimuli. These neurons therefore register and process multisensory stimuli which are associated with the "self", and their activity is somehow  correlated with body ownership.

Last year, Ehrsson went one step further, and succesfully induced out-of-body experiences in the healthy participants of another study. The participants sat on a chair wearing a head-mounted display that was connected to video cameras placed behind their back. They could therefore see themselves from the perspective of another person sitting behind them, such that their "self", or centre of awareness, was located outside of their physical body. While standing next to the chair, Ehrsson used plastic rods to simultaneously touch the each participants' chest and the chest of an illusory body located just ahead of the cameras. When the movements of both rods corresponded to each other, the participants reported the experience of sitting behind their actual bodies

Because of the position of the eyes, we always perceive our bodies in exactly the same way,  from the first-person perspective, with our limbs appearing in a number of limited orientations beneath us. And this earlier work suggested that this perspective is critical for the sense of body ownership. On the basis of these earlier findings, Ehrsson and his colleague Valeria Petkova reasoned that they could induce the illusion that one "owns" the body of another, by manipulating the visual perspective so that the other body is viewed from the first-person perspective and providing the corresponding tactile stimuli.

In the first set of experiments, they used a life-sized mannequin fitted with two CCTV cameras, which were positioned in such a way that the images from each of them corresponded to the mannequin's eyes. The 32 participants wore a set of head-mounted visual displays, which were connected to the cameras in such a way that the images from the left and right video cameras were presented to the participants' left and right eyes, respectively. The participants therefore  recognized the mannequin clearly, but when they tilted their heads downwards, they viewed its body from the first-person perspective, as if it was their own.

The researchers then used short rods to stroke each participant's abdomen and that of the mannequin. In the "synchronous" condition, the number, length and location of the strokes applied to the participants and the mannequin were identical; in the "asynchronous" condition, the same strokes were applied to a dfferent part of the mannequin. After doing this for two minutes, each participant was asked to complete a questionnaire about the perceptual effects they had experienced. From the answers they provided, it was clear that the participants perceived the mannequin's body as their own in the synchronous, but not in the asynchronous condition, even though all the strokes were applied in full view.

In a second set of experiments, the researchers sought to obtain objective evidence of these initial findings. As before, they used rods to stroke the participants' abdomens, both in and out of synchrony with the strokes applied to the mannequin, for one minute. This time though, they simulated a threatening situation by "cutting" the mannequin's body with a knife, while at the same time recording the participants' skin conductance response, as a measure of anxiety. This physiological response was found to be significantly greater when the mannequin was threatened with the knife in the synchronous condition than in the asynchronous condition, or when the mannequin was approached with a spoon in either condition.

This demonstrated that, in the synchronous condition with the knife, the participants responded as if it was their own body being threatened. But these responses were not limited to just one part of the body - the same results were obtained when synchronous tactile stimulation was applied to the hands. However, when the mannequin was replaced with a rectangular box of the same size, the threat-evoked response in the synchronous condition was significantly weaker. Thus, synchronous stimulation of any part of the body (or at least the abdomen and hands) is sufficient for the mannequin's artificial body to be experienced as one's own. However, this illusion only works with objects that resemble one's own body. 


Experimental set-up for inducing the body swap illusion, in which the participant views the mannequin's body in the first-person perspective. (From Petkova & Ehrsson, 2008)

Finally, Ehrsson and Petkova sought to determine whether the illusion could be extended so that the participants would perceive themselves to be localized within someone else's body, or to literally swap bodies with another person. In this experiment, one of the researchers wore the cameras that had previously been mounted on the mannequin's head, so that their visual perspective was conveyed to the head-mounted display worn by the participants. In this way,  they viewed the upper part of their own bodies from the third-person perspective. They were then asked to stand opposite the researcher and to take hold of their hand and squeeze it. In the synchronous condition, the participant and researcher squeezed each other's hands simultaneously, while in the asynchronous condition, they squeezed each other's hands in an alternating rhythm.

Afterwards, the participants were interviewed, so that their perceptual experiences could be established. They reported that in the synchronous condition they had perceived the researcher's arm as their own, and that they sensed their entire body behind it.  Some even spontaneously remarked with comments such as "Your arm felt like it was my arm, and I was behind it", "I felt  that my own body was someone else" or "I was shaking hands with myself!". Remarkably, they also reported that the sensations evoked when the researcher squeezed their hands seemed to originate from the researcher's hand and not from their own. This illusion was vivid  and robust - it persisted even though their own body was in full view, and  regardless of either the sex of the researcher or the shape of their body. Furthermore, the participants exhibited anxiety when a knife was placed just above the researcher's wrist, but not when it was placed near their own.

Thus healthy volunteers can easily be made to perceive themselves as "owning" the body of both a humanoid mannequin and another person. This body-swap illusion can only be induced with another person or something resembling a person; this must be viewed from a first-person perspective, and there must be a continuous match between the visual and tactile information about the state of the body. These findings support the view that the sense of owning one's own body depends on multisensory cues - visual and tactile - from all parts of the body and that these cues are integrated and processed together. They also demonstrate that visual information from the first-person perspective is critical for the experience of owning a body.

Why can the sense of body ownership be manipulated so easily? In a real life situation, identifying one's own body in space quickly and accurately can be of vital importance. It may therefore rely on memory during any decision-making process, because we have lifelong experience of seeing our own bodies from the first-person perspective, and because our bodies generate typical patterns of sensory signals in from the different modalities (vision, proprioception, and so on). Thus, information from memory is used to generate an estimate of the body's location in space, and this estimate is continuously refined by the incoming sensory information.

This refined information is then processed by populations of neurons in the premotor cortex which integrate multisensory cues and map them onto co-ordinates within which the body is at the centre. In the body-swap illusion, these same neuronal populations encode and integrate visual information from a new visual perspective, but because of their egocentric co-ordinates, the individual's centre of awareness is shifted, so that she perceives her "self" to be located at the new position in space. Consequently, the sense of ownership of one's own body can be modified simply by altering the perspective from which the body is viewed.    

As well as providing insights into how the brain constructs its representation of the body and of the "self", these fascinating findings could have useful applications in the clinical and industrial settings. For example, body swapping could prove to be a valuable research tool for researchers investigating self-identity or body image disorders. The findings could  provide improved control within virtual reality situations, and could also be applied to the field of tele-robotics, which is concerned with the control of robots from a distance. Remote-controlled robots are routinely used in the exploration of space and the deep sea; those operating such devices could possibly gain better control of them if they swapped bodies with them.


Petkova, V.I. & Ehrsson, H.H. (2008). If I Were You: Perceptual Illusion of Body Swapping. PLoS One 3 (12): e3832. doi: 10.1371/journal.pone.0003832.

Ehrsson, H.H. (2007). The experimental induction of out-of-body experiences. Science 317: 1048.

Ehrsson, H.H. et al (2004). That's my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science 305: 875-877.

Botvinick, M. (2004). Probing the neural basis of body ownership. Science 305: 782-783.

More like this

A ridiculous number of science-fiction TV series and films have moments where characters exchange minds, from the brilliance of Quantum Leap to the latest season of Heroes. Body-swapping is such a staple of imaginative fiction that it's tempting to think that it has no place being scientifically…
SECOND LIFE is an online "virtual world" which enables users to create a customised avatar, or digital persona, with which they can interact with each other. It has become incredibly popular since its launch just over 6 years ago, with millions of "residents" now using it regularly to meet others,…
One of the bigger challenges facing researchers who are developing artificial limbs is to create prostheses that not only act but also feel like real limbs. This is especially true for the hand, which is one of the most sensitive parts of the human body, and although advanced prosthetic hands with…
Your face is a major component of your self-identity, but when you look into a mirror, how do you know that the person you are seeing is really you? Is it because the person in the reflection looks just like you? Or because the reflection moves when you move? Or perhaps because you see the face in…

"Call me Joe" (1957) ... by Poul Anderson. ....

Joe is awakened in his den, when a pack of predators is attacking him. Using his great strength, and weapons made from sculpted ice, he kills the animals and, exultant, bays at the moon above him. A vital component shorts out, and Joe reverts to being a human, Ed Anglesey, wearing a special headset on a space station orbiting Jupiter. ...

By Hank Roberts (not verified) on 02 Dec 2008 #permalink

This is fascinating. The whole issue of 'self' comes into play here.

I distinctly remember trying to wrestle with the notion of physical self and id/psyche/ego (whatever the correct term is as identified by clever people) when I was less than 5 years old. I'm sure I didn't have the vocabulary to express it properly, but I was trying to ask my mother how I could be sure that all of 'me' was in this body. I was fascinated with the idea of selves being siloed so neatly into physical entities without spillage or overlap. I asked her how it worked and how we could be sure that it worked perfectly. I was also curious as to how it had transpired that my 'self' should be in this body as opposed to any other.

I have a perfect snapshot memory of the scenario, with my Mom standing in the door to the bedroom as I questioned her. She has no particular memory of the occasion, but does remember that I had a tendency to a lot of 'very odd' questions as a child!

My point is that, while I have some to realise that I was quite a bright child, I was no prodigy, so I wonder if children don't have a better grasp of these things that we realise. Which in turn makes me wonder how the experiment would play out if conducted with children as the subjects. No doubt ethics would not permit it, since we have no guarantee that children would not be profoundly disturbed by the perceptions generated, but it would be interesting to hear their views!

Has anyone looked at the phenomenon I have experienced and others have mentioned to me, that when driving the car can seem to be an extension of your own body? If a rock is seen on a road too late to take avoiding reaction, many drivers automatically cringe before hitting it, for example.

By Richard Simons (not verified) on 03 Dec 2008 #permalink

@Richard: What you're referring to is what neuroscientists would call a projection of the peripersonal space. I think most studies look into how tool use alters the somatosensory cortical representation of the body.

Psychologists Michael Graziano and Charles Cross are among the researchers who have investigated this phenomenon, and there is a very good discussion of it in a recent book called The Body Has a Mind of Its Own, by Sandra Blakeslee and Matthew Blakeslee.

Very interesting. We now need to find out for how long we can make it persist and whether subjects can be made to feel this way (interchangeable) bilaterally, between 2 human subjects. Seems like parapsycology and black magic may have some base. I am serious.

Interesting info to peruse when designing rehab techniques for stroke victims.

this illusion only works with objects that resemble one's own body.

If a subject were to spend a month in a setup like the one described, but with (say) an insect-like mannequin, would they start to feel sensations in the mannequin's extra, nonhuman sets of legs?

If the mannequin were humanoid but ten feet tall, how long would it take the subject to start feeling like they were a giant?

It got me thinking about compassion, apathy and voodoo. The expression of cringing that comes on people's faces when the goal keeper gets hit on the balls...the smile that comes on people's face when they see a child lapping about in a pool....isnt it the overspilling of self? What are you neuro guys doing to concentrate the experience? Meditation is the first word that comes to my mind...Not that i ever did it.

Is a blind person unable to do body-swap? Or more importantly, unable to own their body given the lack of vision as part of the multi-sensory experience necessary in these trials? I suggest that higher levels of proprioceptive awareness may partially substitute for visual cues in a blind subject's development of body ownership. The alternative premise, i.e. that an individual without sight cannot own their own body, is unacceptable.

I'd think that blind people would be able to experience the body swap illusion, but would still have a representation of their body, constructed from the other senses, and a feeling of ownership over it.

Kapitano (at #10), I think that feeling like a giant would be rather quick, based on the experiences of technicians using waldos.

However, there is no brain mapping for more than four limbs, and I don't believe we are equipped to acquire one. The grown-up thalidomide babies who don't have arms, for example, can do nothing with artificial arms, for they have no body image for them. In effect, their legs are both legs and arms, since they have always used their toes to manipulate things (as we all did when we were babies, but then more or less gave up on).

Could you perhaps comment more on how blind people would experience and develop body ownership? I think the article's conclusion that a visual first-person perspective develops body ownership barely penetrates the surface of what's going on. From what I've read about Harold Burr's studies and the morphogenetic field, it seems that body ownership begins to develop in the womb, long before the tissues materialize.

By Kristina Lumme… (not verified) on 13 Oct 2009 #permalink

I usually use my prosthetic steel hook for cooking. *It*. Not to hold a tool but directly. So I handle meat, I turn it on the grill, I lift hot stuff out of the oven with it, I scrub it down regularly, I grab pasta out of boiling water to see when it'll be al dente and so on. But with my prosthetic hand (tips made from silicone, fingers from steel, it'll be alright up to some 250 degrees C) I have this same block that I have with my "real" hand or any other "real" body part - before I could start to risk pain I'd get nervous, I'd get sweaty palms and I really cannot bring myself to dip it into boiling water. So, weird weird weird: I cannot *emotionally* get myself to using my Becker hand as a tool similarly to how I'd use the hook.

But then, I feel naked when I leave the house without keys, when I don't have my arm on, and lo and behold it turns out I have a lot of quasi-prosthetic items that extend my body in sort of ways. But this hook hand thing, that's quite peculiar.

Itâs well-known that some people experience an out-of-body sensation while in surgery, viewing the action from outside their bodies. The chemical changes in their brains from the drugs used in the surgery are the obvious culprits. I experienced the same phenomenon as a result of antipsychotic and mood stabilizer drugs I took. It's easy to discount the out-of-body experience as not legitimate. And yet let us not discount the reality of the experience by saying that it isnât ârealâ because a "self" can't really be outside someone's body. The perspective of seeing life from outside the physical confines of your body is as legitimate as any other perspective. Your article gave me pause to think, and I wrote a blog about it. Thank you for the idea.