Put a rubber hand on a table.
Hide the participant's real hand. Stroke both the hidden hand and the visible rubber hand in synchrony. After a surprisingly short time, many people begin to experience the rubber hand as if it belongs to their body.
Nothing has changed about anatomy.
Something has changed about ownership.
Henrik Ehrsson's research programme has used this kind of controlled illusion, alongside neuroimaging, brain stimulation, psychophysics and computational modeling, to ask a deceptively simple question: how does the brain decide which body is mine?
The work matters because body ownership feels immediate and unquestionable in ordinary life. Experimental manipulation reveals that the feeling is constructed from sensory evidence.
Constructed does not mean unreal.
It means generated by mechanisms that can be studied.
Current role and research programme
As of 23 September 2026, Karolinska Institutet lists H. Henrik Ehrsson as Professor in the Department of Neuroscience and leader of the Cognitive Neuroscience of Body and Self research group in Stockholm.
KI describes the programme as investigating how the brain integrates visual, tactile, proprioceptive and interoceptive signals into a coherent representation of the body. The group combines behavioral experiments, neuroimaging, electrophysiology, brain stimulation, robotic multisensory stimulation and computational modeling.
The current programme extends from limb ownership to full-body self-location, memory, self-concept and applications in prosthetics, virtual embodiment and human-technology interaction.
This breadth makes the profile relevant beyond one famous illusion.
The rubber hand illusion became a causal tool
The rubber hand illusion predates Ehrsson's work, but his programme helped turn it into a precise neuroscientific instrument.
In the illusion, spatially plausible and temporally synchronized visual and tactile signals can lead participants to experience an artificial hand as part of their body.
Ehrsson, Spence and Passingham's 2004 Science paper reported premotor-cortex activity associated with the feeling of ownership. Ehrsson, Holmes and Passingham's 2005 work further linked ownership-related experience to multisensory brain areas.
The important contribution is not "the premotor cortex is where ownership lives."
Modern body-ownership research describes distributed networks and computations.
The experiments showed that ownership could be manipulated while sensory evidence was controlled, allowing researchers to connect subjective experience with measurable changes in behavior and neural activity.
Ownership depends on multisensory integration
The brain receives several kinds of information about the body.
Vision shows where a limb appears to be. Proprioception provides information about limb position. Touch provides contact events. Motor predictions accompany voluntary action. Interoceptive signals contribute information about internal bodily state.
For a visible hand to be experienced as one's own, these streams usually need to fit within plausible spatial and temporal relationships.
If the rubber hand is positioned impossibly or the strokes are strongly asynchronous, the illusion weakens.
This makes body ownership a multisensory inference problem rather than a fixed tag attached to biological tissue.
The brain has to determine which sensory events have a common cause.
Causal inference gives the field a computational language
Ehrsson's current research group explicitly frames body ownership in terms of causal inference.
The nervous system estimates whether visual, tactile and proprioceptive signals likely arise from the same body. Temporal synchrony, spatial correspondence and prior expectations constrain the estimate.
This framework helps explain why ownership illusions are graded rather than all-or-none.
A stimulus can become more or less plausible as evidence for "my body" depending on timing and configuration.
It also gives researchers quantitative predictions.
Instead of asking only whether a participant reports ownership, experiments can estimate perceptual sensitivity, temporal binding windows and uncertainty.
That move from demonstration to computation is one reason the programme deserves independent treatment.
The sense of self-location can also be manipulated
Ehrsson's work has extended beyond hands.
Full-body paradigms use visual perspective and synchronized touch to alter where participants experience themselves as located or which body they experience as their own. His 2007 Science report on experimentally induced out-of-body experience became widely known because it showed that aspects of bodily self-location could be shifted in healthy participants under controlled conditions.
The media language around such experiments can become dramatic.
The scientific claim is narrower.
The work does not show that consciousness literally leaves the body. It shows that multisensory manipulations can alter the experienced spatial relation between self and body.
That distinction should remain explicit.
Threat makes ownership behaviorally consequential
If a rubber hand feels like part of the body, threatening it should matter.
Ehrsson and colleagues showed that threatening an embodied rubber hand can elicit anxiety-related responses. This provided converging evidence that ownership reports were not merely participants agreeing with an experimenter's suggestion.
Again, the measure needs care.
Physiological response does not prove full equivalence between artificial and biological limbs. Demand characteristics and expectation remain relevant in illusion research. Strong designs therefore use multiple measures: subjective reports, proprioceptive drift, threat response, perceptual discrimination and neural data.
No single indicator is the ownership construct itself.
Recent work makes timing causal
The programme continues to develop rather than resting on classic illusions.
D'Angelo, Lanfranco, Chancel and Ehrsson's 2026 Nature Communications study examined whether parietal alpha frequency shapes temporal integration in body ownership. Individual alpha frequency predicted temporal binding windows and perceptual sensitivity, and brain stimulation that modulated alpha frequency altered those measures.
The work links an oscillatory property of cortical processing to the timing window in which bodily signals are integrated.
That is stronger than a simple correlation because the stimulation manipulation supports a causal interpretation within the experimental paradigm.
It still does not mean one frequency "creates the self."
A causal contribution to one temporal-integration mechanism is not a total theory of bodily identity.
Nociception and ownership remain separable questions
Current work from the Ehrsson group also probes relationships among body ownership, nociception and pain.
Coppi, D'Angelo and Ehrsson's 2026 Neuropsychologia paper examined visuonociceptive facilitation in peripersonal space and its relationship to body ownership. The group's 2026 publication list also includes a systematic review and meta-analysis concluding that pain perception is not straightforwardly modulated by the rubber hand illusion in healthy participants.
This is scientifically healthy.
An influential paradigm should not expand by assumption into every bodily domain.
Ownership, peripersonal threat, nociception and pain can interact while remaining distinct constructs.
Body ownership is not body image
Body ownership answers a basic question: does this body or body part feel like mine?
Body image concerns beliefs, attitudes and representations about appearance, shape, size and evaluation. Body schema concerns sensorimotor representations used for action. Sense of agency concerns the experience of causing an action.
These constructs can influence one another, but they are not interchangeable.
Ehrsson's programme is especially useful because experimental illusions let researchers alter ownership while holding many other features relatively stable.
That separation improves the entire ontology of embodied selfhood.
Ownership and agency can dissociate
A person can experience a limb as theirs without moving it.
They can also produce an action while ownership and control signals are manipulated independently in virtual or experimental settings.
Ehrsson's current group explicitly studies the relation between voluntary movement, sensory prediction, agency and ownership.
This matters for prosthetics and teleoperation.
A device may respond accurately to a user's command while still feeling external. Conversely, an avatar can feel bodily without providing full causal control.
Human-technology integration therefore has at least two problems: making action controllable and making the controlled system perceptually incorporable.
The two should not be confused.
Prosthetics and virtual embodiment are a natural application
If multisensory consistency helps generate ownership, artificial limbs and virtual bodies can be designed to provide more coherent sensory evidence.
This has obvious implications for prosthetics, rehabilitation, virtual reality and teleoperated systems.
But laboratory ownership effects do not guarantee long-term adoption, function or wellbeing.
Real prostheses involve comfort, motor control, appearance, maintenance, social experience and years of use. Virtual embodiment can produce temporary ownership without durable changes in identity.
The laboratory can isolate mechanisms.
Applications have to put the complexity back.
The work also raises questions about self-concept
Ehrsson's group has investigated whether changing bodily self-perception can alter higher-order self-related processes including memory and self-concept.
This direction is theoretically important because it tests whether low-level bodily inference can propagate into more conceptual aspects of selfhood.
The evidence should be read proportionately.
Short-term experimental manipulations can show causal effects on a measured self-related outcome under controlled conditions. They do not establish that identity as a whole is reducible to multisensory body representation.
The self has biological, autobiographical, relational, linguistic and cultural dimensions.
Bodily ownership is foundational without being exhaustive.
Measuring an illusion requires converging evidence
Body-ownership studies often combine several outcomes because no single measure is sufficient. Participants can rate statements about ownership. Researchers can measure proprioceptive drift, asking whether the perceived location of the real hand shifts toward the artificial one. Threat responses can test whether an embodied object acquires defensive significance. Psychophysical tasks can estimate temporal or spatial sensitivity. Neuroimaging can characterize associated systems.
These measures do not always move together. A participant can report ownership without a large proprioceptive drift, or show one physiological response without endorsing every questionnaire item.
That divergence is scientifically useful. It suggests that "the illusion" contains multiple processes rather than one hidden state perfectly observed by any single proxy.
The strongest experiments therefore define in advance which outcome indexes which theoretical component.
Demand characteristics are a real methodological challenge
Bodily illusions are vivid and easy to explain to participants. That creates a risk: people may infer what the experiment expects and change their reports accordingly. Researchers have debated how much demand characteristics contribute to classic ownership findings.
The appropriate response is neither to dismiss subjective reports nor to treat them as self-validating. Control conditions, implicit measures, forced-choice psychophysics, threat responses and manipulations with specific temporal predictions can constrain expectancy explanations.
Recent computational and stimulation studies are valuable partly because they generate quantitative predictions that are harder to explain by simple compliance alone.
Subjective experience remains central because ownership is a subjective phenomenon. Scientific rigor comes from surrounding the report with discriminating tests.
Individual differences are not noise to be erased
Not everyone experiences the rubber hand illusion with the same strength. Differences can reflect sensory precision, prior expectations, body representation, attention and task interpretation. Some participants may experience little ownership at all.
That variability matters for application. A prosthetic or virtual embodiment system optimized around an average user may fail for people whose multisensory integration differs. Clinical populations may also show altered ownership in ways that cannot be reduced to a single continuum of "more" or "less" embodiment.
The future of the field therefore includes personalized models of integration without turning variation into pathology by default.
Long-term embodiment is a different research problem
A laboratory illusion develops over minutes. A prosthesis may be used for years. An avatar may become familiar over hundreds of hours. Repeated tool use can change motor and perceptual representations in ways that a brief experiment cannot capture.
Short-term ownership paradigms reveal mechanisms of integration. They should not be assumed to predict long-term attachment, identity or functional adoption without longitudinal evidence.
Time is not a detail added after the mechanism. It may change the mechanism itself through learning.
What the evidence does not show
Ehrsson's research does not show that body ownership is an illusion in ordinary life.
It does not show that the premotor cortex is a single ownership center.
It does not show that consciousness can literally leave the body.
It does not show that a rubber hand becomes physiologically part of the participant.
It does not show that every participant experiences bodily illusions in the same way.
It does not show that body ownership, body image, agency and self-concept are one construct.
It does not show that inducing ownership over an avatar produces durable identity change.
It does not show that pain can be reliably reduced merely by manipulating ownership.
And it does not justify inferring a person's stable sense of self from one laboratory illusion.
Current relevance to AI and human-technology integration
As artificial agents, avatars, prostheses and telepresence systems become more capable, the question "what can I control?" is increasingly joined by "what can feel like part of me?"
Ehrsson's work provides mechanisms for thinking about that second question.
Temporal synchrony, multisensory congruence, spatial plausibility and causal inference can shape embodiment.
This creates design opportunities and ethical questions.
A system that feels incorporated may change behavior more strongly than one experienced as an external tool. Virtual bodies can alter social perception and self-related processing. Prosthetic or robotic extension can become experientially intimate.
None of this requires claiming that technology literally becomes the biological self.
It requires taking perceptual ownership seriously as a design variable.
Institute synthesis: ownership is an inference with consequences
The Institute uses ownership calibration as a synthesis phrase for understanding how easily the feeling of "mine" can be shaped by coherent sensory evidence.
This phrase is Institute language, not Ehrsson terminology.
The implication is broader than bodily illusions.
Humans can experience tools, roles, outputs and representations as extensions of self. The analogy should not be confused with neural equivalence, but body-ownership research offers a disciplined reminder: familiarity and integration can generate intimacy before reflective judgment catches up.
In AI systems, that raises a question worth testing rather than assuming: when does a tool begin to feel less like an external aid and more like part of one's agency or identity?
Why this profile belongs independently
Ehrsson's programme has supplied both experimental paradigms and computational language for a central element of embodied selfhood.
The work links psychophysics, neuroscience, body representation and technology while remaining unusually testable. It also provides direct Nordic authority in the Knowledge Library through a current Karolinska Institutet research programme.
The profile is therefore not a biography page.
It is an entry point into a coherent research architecture about how the nervous system builds the experience of owning a body.
Editorial limitation note
This profile maps a research programme rather than elevating laboratory illusions into a total theory of selfhood. Current institutional status is dated; experimental effects are kept within their populations, paradigms and measures.
Research references
- Karolinska Institutet. Henrik Ehrsson: current institutional profile. Accessed 23 September 2026. https://ki.se/en/people/henrik-ehrsson
- Karolinska Institutet. Cognitive Neuroscience of Body and Self – Henrik Ehrsson research group. Accessed 23 September 2026. https://ki.se/en/research/research-areas-centres-and-networks/research-groups/cognitive-neuroscience-of-body-and-self-henrik-ehrsson-research-group
- Ehrsson HH, Spence C, Passingham RE. That's my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science. 2004;305:875-877. DOI: 10.1126/science.1097011. https://doi.org/10.1126/science.1097011
- Ehrsson HH, Holmes NP, Passingham RE. Touching a rubber hand: feeling of body ownership is associated with activity in multisensory brain areas. Journal of Neuroscience. 2005;25:10564-10573. DOI: 10.1523/JNEUROSCI.0800-05.2005. https://doi.org/10.1523/JNEUROSCI.0800-05.2005
- Ehrsson HH. The experimental induction of out-of-body experiences. Science. 2007;317:1048. DOI: 10.1126/science.1142175. https://pubmed.ncbi.nlm.nih.gov/17717177/
- D'Angelo M, Lanfranco RC, Chancel M, Ehrsson HH. Parietal alpha frequency shapes own-body perception by modulating the temporal integration of bodily signals. Nature Communications. 2026;17:53. DOI: 10.1038/s41467-025-67657-w. https://doi.org/10.1038/s41467-025-67657-w
- Coppi S, D'Angelo M, Ehrsson HH. Visuonociceptive facilitation in peripersonal space and its relationship to body ownership. Neuropsychologia. 2026;233:109580. DOI: 10.1016/j.neuropsychologia.2026.109580. https://pubmed.ncbi.nlm.nih.gov/42665226/