THE RUBBER HAND DELUSION   ⏱39:59

yeah, i'm Pete Lush. i'm a postdoc at the University of Sussex

cup of tea

and a biscuit

um, you invited me!

yeah, sure, so the rubber hand illusion is an effect that was first written about 20 years ago now. you're sitting at a, at a desk, there's a box in front of you that you put your hand inside, so that you can't see it. you put your right hand inside there. there is a fake hand - a rubber hand - that would be placed away from where your hand is. you can't see your own hand, you can see a fake hand. the experimenter then takes two brushes and begins to stroke the fingers of your hand and the rubber hand, at the same time. this is usually referred to as the synchronous condition

so what you're seeing on the rubber hand is matching what you're feeling on your own hand. this is talked of as a multisensory illusion, and by multisensory it just means that there's two signals going on: you've got er vision - you can see the hand, the rubber hand - and you've got this tactile, this, this sense of touch - you can feel the stroking of the brush on the other hand

after this procedure people will then report an experience of touch felt on the rubber hand, rather than on their own hand, and an experience of ownership of the rubber hand. that's your, your basic set up. there is a control condition where the brushstrokes happen at different times. this is usually called the asynchronous condition

if it's going to be about this, this experience of ownership and touch transferred or referred to the fake hand, we need to ask people what their experience is, and the way that that's typically done is to give a set of statements that describe experience of the illusion. for the referral of touch this would be something like 'i felt the touch of the paintbrush on the rubber hand', and for ownership it would be the, er this, 'i felt as if the rubber hand were my hand'

there are other measures. the most popular of those is 'proprioceptive drift' - a measure of where participants perceive their own hidden hand to be before and after the rubber hand stroking. you might typically set up a ruler in front of you and ask you to point or to, to say where on this ruler you felt your hand was positioned before they start doing the rubber hand illusion procedure, and then afterwards they ask you again. and you'd see typically a shift in the direction of the rubber hand, a shift in the reported perception of the position of your own hand towards that of the rubber hand

i think it ultimately comes down to, if we're going to be measuring experience, everyone accepts that it has to be tied to report at some stage, the report of where you feel like your hand is. all this has built into a huge body of literature which claims to show us, i think it's seen as, it's been very very popular. there's this, this idea that this experience of body ownership, ownership of our bodies is so fundamental, the idea that it can be disrupted in this way, just through a very simple multisensory stimulus, for a very brief period of time, is really surprising, because it seems to tell us a lot about how our experience of body ownership comes about

yeah, but i think, i think that the key thing is that it's really exciting because, because it's really surprising right? it might be somewhat maladaptive to be so easily mistaken

yeah, or if you imagine a situation, if we imagine a sort of, real life situation, you're sort of, you know, you're at a bar or something. the person standing next to you at the bar, their hand happens to be touched in the same way as yours at the time when you are standing there, you don't, i mean maybe people do - it's an interesting question - hang on, i mean is that my hand or is that my hand?

erm, no. to be honest i'm not sure what the ownership question really means, i'm not sure how to interpret it, erm so i don't know if i've experienced that. the referred touch, the feeling of touch on the hand, that hasn't happened to me spontaneously, but i feel that with a sort of, effort of imagination, i could imagine as though the touch was on the rubber hand. i think i get some insight into the experience then, perhaps

let's take a segue into, into another area of psychological research, and this is the history of, er, imaginative suggestion. it's been primarily studied within the context of hypnosis. the way that hypnosis is usually measured in the, in the laboratory is through scales, these standardised hypnotisability scales. these would typically be preceded by a hypnotic induction which can take virtually any form. usually in these scales it will be some sort of relaxation procedure, counting down, and being told that you're in a state of hypnosis. um, but a hypnotic induction can also be riding on an exercise bike, as long as you've been told that it's, er, hypnosis. it can be just 'now you are hypnotised'. it can be anything, pretty much, as long as it establishes that this situation is a hypnotic situation and things are going to happen. in other words, people's expectancies - what they're expecting to happen - is kind of primed, primed for something to take place

yes. and, and the important point is that they will vary in their ability to meet those expectancies by generating some experience, by imagining something, but probably not being aware that they are, that that is what they're doing. it means that every time you run an experiment, you don't know whether they're the sort of person that reports nothing for all these imaginative suggestions, they don't experience any paralysis, they don't hear any mosquitoes, nothing. or whether they're somebody who gets all of it, and they're completely blown away by it, and it's like a life-changing experience for them, just sort of: 'wow, this is the most amazing!'

and i should say as well that there's a lot of evidence that there are changes in experience in imaginative suggestion. a hypnotic suggestion for analgesia, for example, people will repeatedly choose that other, over other forms of anaesthesia for dental surgery. if there's not any experience involved, you'd see, it would be quite a commitment to pleasing the experimenter or pleasing the dentist. you'd have to, you know, gritting your teeth, go: 'i'm going to pretend this is working'

so i think, i think probably here's a good place to talk about the concept of demand characteristics. this is a, a term that was introduced about sixty years ago by Martin Orne, who was a hypnosis researcher, it grew out of hypnosis research. he defined it as any aspect of the experimental situation, or any pre-existing beliefs of participants, that could affect expectancies, affect expectations in an experimental situation. he pointed out, or he argued, that, when you're studying humans, if you're running an experiment with humans and they turn up, we're not just blank slates, we don't just sit there absorbing, you know, or just performing the tasks that we've been instructed to do. we, we can't help ourselves, we're constantly trying to work out what's going on, it's an effort to stop yourself guessing what it is that you're supposed to be doing

Orne argued that, therefore, that what people expect in an experiment can bias the responses. some people might want to try and be a 'good participant' and give the experimenter what they believe the experiment wants. that is the key point that we're making when we talk about the rubber hand illusion, and when we talk about a range of other effects that we've been looking at as well. and this is that, what you expect in an experimental situation may lead to the generation of experience, by the same mechanisms as people respond to imaginative suggestion. so you expect a particular experience and you generate that experience

if you take a situation like the rubber hand illusion, some people will, they really seem to have had an experience, they're really wowed by it, it doesn't seem like they're putting it on, they go 'woah that was amazing!' but if you have run imaginative suggestion screenings, you get the same kind of responses. you get, there will be some people who are just really blown away: 'wow that was amazing!', 'how did you do that?', 'i could really, you know, it was like there was a mosquito walking on me!', 'i couldn't move my arm!'

if we return to the demand characteristics idea, we can think about what people might expect in the rubber hand illusion. you're sat there, you've got a rubber hand there, you've got your own hand. the experimenter starts to stroke your hand at the same time, what was that experimenter expecting to happen? people are going to be trying to work that out. if people can work out that they're expected to feel the touch over there on the fake hand, and if they can work out that they're supposed to feel like that hand was their own hand, then this could act like an implicit imaginative suggestion

hypnosis grew out of um, Mesmerism, it developed from Mesmerism, a sort of eighteenth century suggestion effect, which was attributed to the manipulation of magnetic fluid in the, in the air. at the time that was plausible enough scientifically. a hundred years later it wasn't so plausible and that's when hypnosis came along. in Mesmerism, typically there would be a sort of, something like waving hands over the body, or iron rods over the body. Mesmer used iron rods um at first, this was, you know they manipulated magnetic fluid, then he fell out with his iron rod maker, then he discovered he could do it with his hands... waving hands over the, over somebody, you don't tell them what they're going to experience, instead they just respond in some way. and there were people who had convulsions, or they would be cured of things, or you know, just this huge range of, there's a wide range - anything they could come up with, really, from that situation

well, the degree to which someone reports experience on a hypnosis scale predicts the degree to which they will report experience of the rubber hand illusion, given this particular context. and i'd say also, i'd point out that the scientific context perhaps has quite a lot in common with the hypnotic context. i mean the scientist wouldn't have got you in there unless they expected something to happen. if this sort of authority figure, the scientist, taking the place of the hypnotist, your expectancies are raised

what's important in this situation is the controls for demand characteristics, that is the controls for what people expect. the control for demand characteristics in rubber hand illusion studies, chiefly is this asynchronous stroking condition we discussed earlier. so the key assumption there for this to be used as a control is that people expect experience similarly for synchronous and asynchronous conditions

i think if you just introspect about that for a moment, imagine yourself receiving these two conditions. would you expect the same experience in both? i wouldn't. but, i tested it, just used a very simple procedure - this is based on ideas of Martin Orne's again from the sixties - um just basically describe the procedure to people, there was no, they didn't take part in a rubber hand illusion. showed them the video, showed the procedure, didn't say it was anything to do with illusions, or rubber hands etcetera, described the synchronous and asynchronous conditions, and showed them the synchronous and asynchronous conditions. and then just gave them the questionnaire, the standard questionnaire for the rubber hand illusion, but for what they would expect to experience, rather than what they did experience

and we find that what they expect in the synchronous condition and what they expect in the asynchronous conditions differs. that is, they expect the experience for synchronous stroking, they don't expect it for the asynchronous condition. people, on average, know that they are supposed to experience something in the synchronous condition, not the asynchronous condition

which means that there's the possibility that the rubber hand illusion, these reports and these objective measures, are compliance, bias. this is the key step we take. people responding to their ability to meet their expectancies for experience [see also the second paper]

it hasn't gone down well, there's complete disbelief. the thing that struck me the most is the lack of interest in the demand characteristics control issue. nobody's argued with me about that, about why that could be a valid measure for demand characteristics. erm, so i find that quite strange and surprising. but, i don't know, maybe i shouldn't be surprised and this is just how these things go, you can't expect people to open their arms to something like that

for me, i just, I focus on: we've got an alternative theory, we've got evidence consistent with that theory, and we've got evidence that the existing theory isn't as solid as was, as was thought. to me that's just normal science. if i'm wrong, then it will strengthen the rubber hand illusion, because stronger, valid controls will be developed

i don't think though, in the long term, there's nothing to lose from it. well i, i get the impression that people feel, people feel that they're losing something. i don't, i don't see it in that way, because every time we discard a particular hypothesis or find some new evidence, i mean, that's, that's science. if you go into science, you know there's going to be change, disagreement, alternative theories

the first thing is, the way that it's been done with the synchronous and asynchronous conditions, trying to show that the score is higher in the one than the other, that's fine, we can retain that. but just that they need to be matched in expectancies. you need to make sure that people expect experience in the control condition the same as they do in the illusion condition

i don't, i don't, i really didn't intend to be the rubber hand guy. and i'm gonna, i'm going to do my utmost to not be the rubber hand guy. i mean, i'm gonna, i'm doing it with other things, one after another. so i'll probably just be the guy everybody hates

i think i would have jumped to the end point more quickly. i think that the response, i think i thought that when i raised it in the first place, people would go: 'oh wow, that's a concern' and start thinking: 'we should have a go at sorting out these control methods'. and that isn't what happened, i had just absolute opposition, really

actually, that's mostly form the scientists. from the, from the philosophers that work in, or are peripheral to, the rubber hand illusion, who use it in their arguments, they have, there's been a lot more acceptance and even support. maybe that's because, i don't know, philosophers are more used to sort of, contentious argument in that way. that they're, kind of, that's a philosopher's job, basically, to come along and say 'i've got a different take on this, [i] disagree with you'

having more of that in science would be helpful. it would, it would, not getting so, sort of, attached maybe? not be personally invested in your data as much as possible; be aware that you're gonna be fooling yourself, and always be trying to. i mean, i'm going to be making the same mistakes, and someone will come along and, you know, if this thing manages to stand up for ten years, then someone will probably tear it apart in ten years, then i'll be angry at them. i'd like to think that i'd be able to... i'll try to, i will you know, i will certainly be trying to er, to think 'alright, well, what did you expect?'

yeah, i think that, that doing this has really kind of opened my eyes to demand characteristics. and now i'm seeing them everywhere, and it feels like, there's this situation in psychology where, to me it feels a bit like if, you know, after Fleming, and penicillin, we realised from that that you could contaminate your Petri dishes if you didn't wash them. like ten years later, twenty years later, everyone had just stopped washing their Petri dishes again. it feels a bit like, um, like it's sort of, hiding in plain sight. and now when i read papers, i think, i'm just seeing demand characteristics all over the place. and that of course applies to my own earlier work

maybe i should say that again in a way that would actually not take five minutes?

otherwise you're going to be stuck there with the edit tool

um, yeah, i would say um, looking back at my own earlier work, now that i've been looking at demand characteristics, i'm starting to see them, er, in other places, i'm seeing them, you know, you become aware of this thing of will, of thinking what will the participants [be] expecting? is there any way that this could be demand characteristics? and when i look at some of my earlier studies, then, yes, there is. and, erm, i'm testing that at the moment. i hope not to be publishing a paper in a short while saying: 'yeah, my studies were demand characteristics', but, you know, we'll see

yeah hello, my name is Regine, Regine Zopf. and, er i'm currently in Sydney, Australia. and i study, i study body perception, so how we perceive our own body. and i study that in healthy participants, but also in disorders, like the eating disorders. and for that i'm particularly interested [in] how different signals of the body sort of are combined and integrated and how that influences our, our body perception, and then also how we use these signals to then interact with the world around us

yeah, sure

i might have a, erm, some red wine, maybe a Merlot?

yeah, i must say that i initially found them quite interesting. i thought, 'oh there's some interesting new data', and also some interesting criticism and findings of that, are, i think interesting for the rubber hand illusion researcher

yes, i think, i guess what crossed me more was, sort of, this strong claim based on these data, that the illusion, that the rubber hand illusion can be explained by suggestibility. i sort of can distinguish between a weak version of the data that there's something not quite right with the methods, and that response bias can influence our methods, and the sort of strong version that the whole rubber hand illusion can be explained by suggestibility, which is more the idea this implicit suggestion can control what we actually experience in the illusion. and i think it's that second sort of idea that i think has been sort of pushed quite a bit by the authors and also by the media and maybe on Twitter, and that's more what i have a problem, not a problem, but i don't agree with that. and i don't find that they actually have strong evidence for that. i'm not convinced

well i think there are other ways, to use measurements that are not as prone to these expectations or possible response biases that people can guess what the experimenter wants you to do, and, which might even underlie these correlations that they then found in the second study. and we always knew that, i mean it's not a new thing that we think that these subjective experiences that we sort of measure using these questionnaires could be biased by expectations and bias, and so i think um, one way is to use measures that are maybe not so prone and not so subjective. and there are quite a few as well

like i have two parts to this, two answers, so maybe first i'll talk about why i think, why i don't really believe there's strong evidence for the suggestibility idea, although i'm, i'm intrigued by it as well and especially when you think about the rubber hand illusion, you put this hand, and then you also have the stroking and it's quite rhythmic, and i also think maybe there is something hypnotic about this. it's often dark and things like that

i don't quite, based on the data i'm not a hundred, i'm not convinced. we know there's a lot of individual differences, right? so we have often eighty percent who get the illusion and twenty percent who don't get the illusion. so, this, i think this is very intriguing, we really don't know why some people don't get the illusion. but what Lush et al found is some interesting like correlation in the synchronous condition, the measures, the subjective measures and also the drift correlates with this er suggestibility or hypnotisability

like, my second point is that these correlations can also be, be kind of explained by two types, like first it can be this response bias, or it can be explained by the implicit suggestions and then they control sort of the phenomenology to kind of respond to these expectations, so kind of the strong claim that something that happens in the rubber hand illusion by suggestion just changes the experience - doesn't have to be the case

theoretically, i think what i also have missed in this whole discussion is, i kind of think if they want to make this strong claim that what you change in the rubber hand illusion is you make these you have this implicit suggestion and it changes the phenomenology, and there's no, it has nothing to do with kind of multisensory processes or something like that, but is purely a cognitive, top-down effect, then i really believe that they also need to develop a stronger theory, and i think this theory also has to then explain some other previously-established and well-established rubber hand illusion effects

and i'm not quite sure if they can and i, and i'll give you an example, for example in the rubber hand illusion we know some factors have a strong effect and some don't. for example when i see, er, like, a hand then i can experience the rubber hand illusion, but if it's just a block of wood it's not possible to induce ownership or embodiment for that block of wood, it has to be kind of a hand, hand form is really important, that it's sort of body-like. but other factors are really not so important for the illusion, like um, especially er, the skin tone, or the details of the hands

but let's take the example skin tone. here we have basically the idea that, we know that somebody who has maybe a dark skin tone can experience the illusion for a light skin tone, and the other way around. and that's, for example, a finding that i don't really think this sort of suggestibility, er, theory can explain, because, in theory when, when i see a hand that has a different skin tone to mine i would, i would really expect that this is not my hand. but at the same time when you induce the rubber hand illusion it doesn't seem to make a difference

for me, this is something that really can't be explained by this suggestibility idea, but we, at the same time we have a lot of these multisensory theories, or multisensory theories on the rubber hand illusion that are based on multisensory processes and they often are very much based on, sort of, how we perceive something in time and space, sort of binds signals from multiple senses together, so if you feel something at the same time as we see it, this seems to be a signal that these things belong together. or if we feel, experience it in the same space, the two different signals and we are also combining them together. so these signals that are important for multisensory, er binding are really sort of important for the rubber hand illusion, whereas signals that are, maybe like skin tone is a purely visual one, we can't feel skin tone, so these don't seem to be important

so that really fits nicely with these multisensory ideas and i think a challenge is then for this other theory to really explain it

yes, quite strongly, yes

oh, that's an interesting question, i never thought about it. maybe it, it can influence it, but i don't know, yeah maybe. i don't know, we just have an N of one and one

oh, there might be a study there

yeah, i completely agree. it can be very messy, or has been very messy, or seems very messy. but i think one of the, sort of, reasons for this, i think that the illusion itself is so complex, that body perception is so complex. when we er have the illusion it's, sort of, it's not just that we experience ownership for the rubber hand, but we also feel embodied in this rubber hand, that the rubber hand is in that location, that the touch we feel is coming from the rubber hand being touched, and we also have a bit this feeling that we can control the rubber hand

it's very complex, and there are a lot of different aspects and it really has these multiple components, like Matthew Longo has, and his colleagues, they have um investigated a long time ago, and studied these different components like ownership, location, agency all seem to play a role in, in the rubber hand illusion and this embodiment feeling. and i think that's sort of one answer why it is so confusing what it the illusion because it really has all these aspects. um in addition to these sort of subjective measures we also have these other measures where we can observe behaviour of people in the rubber hand illusion, like how somebody reaches or how somebody responds to touches, or how somebody protects their hand or

so there are all these different aspects as well of the body protection and, and, and action and multisensory perception. um, and i think all of this is can really, is really part of the illusion and i think if one like decides on a measure it's really important that one kind of thinks about what, what the question is and what measure and what aspects one is interested in. is it like the complexity of the illusion or is it just a subset?

yeah, so i think it always depends on the question what measure's best and with each measure we should really carefully think about what potentially problems there is and how good this measure is. as long as one is clear why one does that, i think it seems OK for, like, in my opinion

i think we just have to think more about the complexity, and measure, yeah measure it more more, and be, try and get the best, the best measurement we can get, and try and then tease it apart and think more about it

yes, definitely. i don't think we can explain it away. i mean the illusion still exists in both theories, and, but i don't think we can explain away the er, like these multisensory um effects and multisensory interactions that seem to be quite important

mmm, i think what was was really funny, i had a lot of people sending me the media reports and a lot of colleagues sending me 'oh, what do you think about this?' and 'is that the end of the rubber hand illusion?' and i thought i had to explain um, had to explain what's, what's sort of happening. so there was a lot, i felt like, it came up all the time somehow last year, and also in a lot of discussions i had, it somehow came up. and also yeah on Twitter i, i saw it. i don't really think it changed because i think i read it before, i don't think i really changed my view on it

yeah i think, for me, the most important thing, is that our methods, like especially the subjective measurements, can be influenced by er expectations and then response bias, and but i think that was, to me at least, clear before. i don't think that was really, er, really a new point, but it just highlights it again, and i think it's an important, something important to take away, and it should influence us, maybe we should try and maybe think about the questionnaires a bit more that we're using, and develop maybe better control questions, if that's possible, or try and, yeah, kind of discuss how we can improve what we are doing there

for the theoretical work, i think, like we discussed, i think that needs just, that needs more work for me, i'm not convinced. and at the same time, maybe it's important that people in the field who've worked on the rubber hand illusion maybe pitch different theories against each other, and maybe that's, and maybe that will be interesting to improve different theories and how we talk about them, er, and maybe yeah maybe we can improve each other's models as well

yes, i will be, i think there are still some open questions, and um, i mean i don't have any direct plans at this moment, i'll, i'll be honest, but um, down the track i think if something interesting comes up, i think it's um really a great tool, because i'm interested in different signals and multisensory perception of the body, it's a great tool to introduce different conflicts in space and in time, and i think it's very very useful to get to, yeah, questions about the body, because it really allows us to manipulate something we, we were unable to manipulate before, the rubber hand illusion

ah, yes, yeah. we talked um, we talked a bit about the complexity of the rubber hand illusion and the different questions that we can use to measure the rubber hand illusion, and i must say that early in my career, actually, because, when we analyse, we can, we can look at different components, we can look at individual questions or we could, for example, also just calculate a mean score across all the questions

erm, and one sin i, i did like is that, probably when i had my, questionnaires of my participants that i actually played around a bit and explored what might be a good way to analyse the data. and maybe tried out different options, which i don't think i would do any more like that. then later, er, became much more aware of the problem that this might create in terms of bias. and i have pre-registered a lot of my data or my analysis since, and and that's probably a sin of the past hopefully for me, that i am not really proud of, looking back

yeah, i just, experiencing a pre-registration process really showed me how important it is