Moving rubber hand illusion: a switched perspective

(1)

1

Moving rubber hand illusion:

a switched perspective

44 EC

01-09-2011 –04-04-2012

Koen Knip 0446440

MSc in Brain and Cognitive Sciences, University of Amsterdam Track: Cognitive Neuroscience

Supervision: Mark Schram Christensen / Anke Karabanov University of Copenhagen, Copenhagen Neural Control of Movement UvA representative: Martijn Wokke

(2)

2

When we move or see our hands we know that they are part of our body. This feeling of ownership has been studied repeatedly, mostly using multisensory paradigms. In the rubber hand illusion stroking an unseen real hand and a fake hand in sight simultaneously causes attribution of the fake hand (Botvinick & Cohen, 1998). Kalckert and Ehrsson (2012) showed that the illusion of ownership can also be elicited by movement. Here we replicate their study and introduce a two-finger moving rubber hand and a switched finger condition. We showed that the moving rubber hand illusion is successful in eliciting the feeling of ownership over a model hand. Moreover with the help of subjective report and proprioceptive drift data we showed, against our expectations, that when the index finger and the middle finger responses are switched (meaning the real index finger is attached to the rubber middle finger and vice versa) a sense of ownership over the fake model hand is being established. This study sheds new light on the boundaries of ownership over body parts. By unraveling these boundaries, body ownership and agency studies like this will benefit potential clinical applications.

(5)

5 Introduction

Everyone knows where their hands are and that they are part of their own body. We use them daily, and almost continuously receive multi-sensory information about their shape, movement, the position they are in and whether or not they are touching an object. These bodily sensations are automatic and we are not constantly aware of them. The hands have been used many times to investigate the sense of ownership over body parts. Ownership over body parts and ownership over the body in general are important central issues in the psychologically and philosophically based discussion on the relationship between our body and the sense of ‘self’ (Gallagher, 2000; Haggard, 2005). Identifying and localizing oneself in the sensory

surroundings of one’s body is most likely an adaptive trait shaped by evolution (Graziano & Botvinick, 2002; (Makin, Holmes, & Ehrsson, 2008). The last couple of decades the study of perception of one’s own body has taken flight in cognitive neuroscience, among other disciplines. Recently de Vignemont (2011) stated there are two main pathways to study the sense of body ownership. The first pathway is to study

embodiment of an object that is experienced as part of one’s own body while the second pathway investigates patients who perceive a part of their body as alien. For practical reasons the first pathway is the main focus of most studies. The range of studies on embodiment extent from tool use to controlling an avatar and from prosthesis to full body and rubber hand illusion (for a full overview of embodiment and disembodiment studies see table 2 in de Vignemont, 2011).

The sensation of ownership over a rubber hand is probably the most studied paradigm in exploring the representation of the bodily self. Here, simultaneously stroking of an unseen real hand and a visible rubber hand within the boundaries of (peri)personal space and congruent with body orientation can result in attribution of the fake hand to one’s own body. This experimental set up, referred to as the rubber hand illusion (RHI), combines multisensory information necessary for ownership sensation of body parts and was experimentally introduced by Botvinick & Cohen (1998) after a similar sensation has been reported anecdotally in 1937 by Tastevin. The study so far had several replications (Armel & Ramachandran, 2003; Ehrsson et al., 2004; Tsakiris & Haggard, 2005) and modifications (Ehrsson, 2007; Ehrsson et al., 2008, Kammers, et al., 2009; Petkova & Ehrsson, 2008, and many others, for a more elaborate list see Tsakiris (2010, p. 704)). The illusion takes about 10-30 seconds to build up in most cases and approximately 30% of the population seems to be resistant to the illusion, they remain unaffected after its induction.

Different modifications of the paradigm revealed the boundaries of the illusion and thereby the boundaries of ownership over a body part. Placing the rubber hand in an incongruent anatomical position abolishes the RHI (Constantini & Haggard, 2007; Tsakiris & Haggard, 2005; Ehrsson, Spence &

(6)

6 Passingham, 2004). Furthermore, clever studies show that over exceeding the distance between the rubber hand and the participants’ real hand eliminates the illusion (Lloyd, 2007), as does placing the artificial hand in a position too much in front of the expected position of the rubber hand (Armel & Ramachandran, 2003). These findings support the view that the illusion works only within peri-personal space, mostly described as the space within reach of any limb of an individual. (Maravita, Spence, & Driver, 2003; Maravita & Iriki, 2004). Tsakiris & Hagard (2005) showed that resemblance of a model hand to a real human hand is of great significance. Longo, Schuur, Kammer, Tsakiris and Haggard (2009) even showed that participants perceive their own hand as more physically similar to the rubber hand when experiencing the rubber hand illusion. Measuring the experience that the RHI elicits has mostly been with the help of subjective reports and questionnaires, as the one in Botvninick and Cohen (1998). They combined these subjective measures with an objective measure of proprioceptive drift, the change in distance between the position of the real hand and where the participants perceive their hand to be. Botvinick and Cohen (1998) and Longo et al. (2008) found that there is a correlation between proprioceptive drift and the subjective experience rating of ownership over the model hand, high ratings correlate with a larger drift. Skin conductance response (SCR) was used by Armel and Ramachandran (2003) and Petkova and Ehrsson (2009) to test whether or not subjects would show an emotional response when the rubber hand was threatened. Ehrsson, Wiech, Weiskopf, Dolan, & Passingham (2007) showed that the level of activity in the insula and anterior cingulate cortex, brain areas associated with anxiety responses, is similar when a rubber hand is threatened compared to when a person’s real hand is threatened. Moreover they show that the activity in these brain areas rises when a subject is experiencing a stronger illusion. These studies show that

threatening a model hand leads to an emotional response, thus increases skin response, and changes in the interoceptive systems when subjects are having an ownership illusion compared with the suitable control conditions. A more recent objective autonomic measure is registration of a temperature drop in the real hand during the RHI. Moseley et al. (2008) showed that skin temperature of the real hand decreases when participants experience ownership over the rubber hand, adding that the stronger the illusion is

experienced the more the temperature decreases. Barnsley et al. (2011) showed that the rubber hand illusion increases histamine activity in the real arm, suggesting a role of the immune system in

discriminating between bodily-self and the ‘non-self. ’

A problem with the classical set up of the rubber hand illusion is that is a static illusion, there is no movement involved. The existing cognitive neuroscience research framework strongly points out that a sense of being in control of one’s own movements, also referred to as authorship over bodily movements or in short agency, plays an essential role in being conscious of the ‘self’ (Blakemore & Frith, 2003; Haggard,

(7)

7 2005; David et al. 2008). Several studies tried to replace visuo-tactile stimulation with movement in the classical RHI set up. By measuring drift experiences in participants towards a video-projected image of their own hand with synchronous visual feedback, Tsakiris et al. (2006) were among the first to involve

movement in the RHI paradigm. However, IJsselsteijn et al. (2005) showed that ratings of ownership over 2D images of hands projected onto a screen are weaker than with the physical classic model of the RHI set up. Dummer, Picot-Annand, Neal, & Moore (2009) build a mechanical set up that was able to synchronize visual with experienced movement. This in contrast with asynchronous movement shows stronger reports of the RHI. The contribution of proprioceptive signals in the RHI was studied by placing an anaesthetized hand with no cutaneous feedback in a moving mechanical set up (Walsh, Moseley, Taylor, & Gandevia, 2011). Even the possibility to elicit ownership over a virtually displayed, moving hand has been addressed. To prove that motor activity in synchrony with proprioception and visual information is able to induce a feeling of ownership, Sanchez-Vives et al. (2010) successfully used a data-glove that measured and

transmitted the position of the real hand, resulting in a visual image of a virtual projected hand to create a synchronous movement condition. In a reaching task modification of the classic RHI paradigm Kammers et al. (2009) showed that the perceptual judgment (pointing) of subjects is still sensitive to the illusion of ownership over the rubber hand after moving their real hand. This challenges the view that active

movement of the stimulated limb, causing a proprioceptive update, erases the illusion. A recent study done by Newport et al. (2010) presented their participants with two moving left model hands in a computer and mirror set up called MIRAGE. When both hands and the real left hand are being stroked synchronously, participants experience ownership over both fake hands. When one fake hand is being stroked

synchronously, then only ownership over that hand is claimed but not over the asynchronously stroked fake hand. Hereby showing that the body image of ownership can incorporate multiple limbs, but agency can only be addressed to one limb within this body image.

In contrast to previous experiments studying the moving rubber hand illusion one of the latest studies in the field tried to completely dissociate ownership and agency in one paradigm using mechanically controlled index-finger movement of a physical wooden model hand. In a 2x2 design with synchronized movements Kalckert and Ehrsson (2012) used two different modes of movement (active; participants moving the finger their self by mechanically attached rod vs. passive; experimenter unseenly moved the finger with the maintained mechanical connection) and two different hand positions (congruent; in line with position of the body vs. incongruent; turned around 180 degrees). They show that passive movements do not influence the sense of ownership over the moving rubber hand, while it does abolish the sense of agency. Whereas incongruent positioning of the model hand keeps the sense of agency intact, it eliminates ownership. An experimental condition with asynchronous movement eliminates both agency

(8)

8 and ownership feelings over the model hand. Kalckert and Ehrsson (2012) indeed show a double

dissociation of ownership and agency, which opens possibilities to study both cognitive functions separately in future research.

In the present study, we replicated and modified the moving rubber hand illusion set up used by Kalckert and Ehrsson (2012). We used a wooden model hand mechanically connected to the index and middle finger of the participant’s real hand that is resting out of sight in a rectangular box.

Consequently, in contrast to the study of Kalckert and Ehrsson we used the movement of two fingers, with the help of two separate rods. Furthermore we attached movement sensors to both fingers in order to measure response time, and introduced a switched finger response condition. In this condition the index finger of the model hand was attached to the middle finger, while the middle finger of the model hand was attached to the index finger of the participant. In the switched conditions participants will experience asynchrony in their visual feedback; when moving their index finger the rubber middle finger will move. This key manipulation is hypothesized to abolish the illusion. Like stated before, impossible manipulations, like incongruent hand position (Constantini & Haggard, 2007; Tsakiris & Haggard, 2005; Ehrsson, Spence & Passingham, 2004) or placing the fake hand far from the body (Armel & Ramachandran, 2003; Lloyd, 2007), are not able to establish an experience of ownership over a model hand. Therefore we hypothesized that ownership over the wooden model hand will only be claimed in the non-switched conditions. All in all we used three conditions (respectively non-switched, switched, and non-switched) and divided each over 3 blocks in order to test time-effects. Participants were asked to react to a red LED light with the

corresponding finger of the model hand. We hypothesize that the sense of agency will be maintained during the entire study, although needs building up in the switched condition. Also we hypothesized that reaction times will be slower in the switched condition but will decrease over time when subjects will get acquainted with the switched condition. The acquaintance of the switched condition can be seen as an adaptation process. Therefore we expect to see after- effects when turning back to the normal condition. Furthermore illusion responders in the switched condition will have a harder time in the switched condition than non-responders; it might be so that it costs them more effort to disentangle the motor response and the visual feedback.

We found that against our expectations a large part of the illusion responders were able to build up a sense of ownership over a fake model hand when finger responses are switched. In the switched condition correlation between ownership subjective measures and proprioceptive drift was positive and got stronger over time.

(9)

9 Methods

Participants

In this study 15 right-handed healthy, paid volunteers participated (8 male and 7 female, mean age: 23.5 years, ±SD 1.5 years). All of them had normal or corrected-to-normal vision. Right Handedness was tested with the Edinburgh handedness inventory (Oldfield, 1971). Subjects were recruited via leaflets on the local University of Copenhagen and via Facebook. Before participating, all subjects read and signed an informed consent. The study protocol was approved by the local ethics committee of the Capital Region Copenhagen.

Apparatus and procedures

During the reaction time task the participants were seated in front of a table with their right hand placed inside a wooden box. This set up is based on and similar to the one described by Kalckert & Ehrsson (2012). The box was placed approximately 30 cm in front of them at a position where the rubber hand would be in a lifelike position and in line with the body. Participants wore a blue rubber glove on their left hand (in order to have a similar visual impression as the model hand) and would let it rest on a resting platform that had the same height as the wooden box. On top of the wooden box a life size wooden model hand wearing a blue rubber glove was placed. First two small movement sensors (s720, Measurand) were attached to the participant’s fingers, one to the index, and the other to the middle finger. This was in order to see which finger moved in the separate trials and importantly at what time the fingers were moving. The index and the middle finger of the participants hand were mechanically attached to the model hand with small rods. These rods were sticking through a small hole in the wooden box. They were constructed out of iron and guitar fingerpicks. Two straight rods were used for the non-switched conditions and two bended rods were used for the switched condition. Two red LED lights were placed on top of the box, one in front of the model hand’s index finger the other in front of the middle finger of that hand. The lights were controlled by a 1401 Micro AD-converter (Cambridge Electronic Design) and the software package Signal (Cambridge Electronic Design) running on a desktop PC. This was also used to record the movement signals. The reaction times were measured with the help of a script in Signal that saved the time of each finger movement onset. The position of the movement onset was manually determined for each frame. The reaction time was defined by subtracting the onset of the light from the onset of the movement.

Like other rubber hand settings a blanket was used to cover the space between the wrist of the model hand and the shoulder of the participant, hereby making it visually more likely that the model hand could be part of the participant’s body. During the attachment of the rods to the index and middle finger the set up was covered with a cardboard box, the back of this box was cut out so the experimenter was able to adjust

(10)

10 without the subject being able to see the adjustment. On the left side of this cardboard box a piece of millimeter paper was placed. This sheet of paper was used to measure proprioceptive drift before and after every block of the reaction time task. Before the experiments began all subjects were asked to read written instructions describing the specifics of the procedure of the experiments. Then right handedness was tested. Before every condition subjects were asked to move their fingers freely and see how the set up felt for 20 seconds. After this short practice sessions a baseline drift was measured and baseline ownership question was asked. Possible adjustments were made before starting the reaction time task.

Picture 1 and 2. left: The right hand was placed in a box and attached to the model hand. On top of the box were 2 red LED lights indication either an index or a middle finger response. right: The left hand wore a blue glove and was resting on a black fabric covered platform. The right hand was attached with the model hand with rods. *movement sensors are not shown in this picture.

Conditions

The experiment consisted of three conditions. In order: A non-switched condition (NS1), a switched condition; where the participant’s index finger was attached to the model’s middle finger and his or her middle finger was attached to the index finger of the model hand, followed by another non-switched condition (NS2). Each condition consisted of 3 blocks of 50 trials; making a total of 9 blocks and 450 trials. After each block the participants were asked to point with their eyes closed to where they thought their right hand was located (proprioceptive drift measure). Furthermore they were asked to answer the questionnaire. One trial is a 2 sec frame wherein a light in front of either the index or the middle finger of the rubber hand is lit. The timing of this light was randomized between 300, 500, 700, 900 and 1100 milliseconds after the start of a trial. The subject was asked to react to the correct light corresponding to the finger in front of it by making a swift finger movement as fast and as correct as possible. By measuring

(11)

11 the movements with two joint angle movement sensors the onset of the response of the finger to the light could be calculated. The time between the onset of the light and the onset of the movement is considered the reaction time in these tasks.

Questionnaires

After each block we used a 9-statement questionnaire loosely based on the questionnaire designed by Botvinick et al. (1998). Two other statement questions were asked at the end of each condition. Subjects were asked to answer to the statements on a 7-point Likert scale, ranging van -3(totally disagree) to +3 (totally agree). The questionnaire can be divided into questions considering ownership, agency and control questions. Full Questionnaire is attached. Scores of ownership questions are controlled with pilot data.

Results

Statistical analysis of questionnaire, drift and reaction time task

All statistical tests were based on a priori hypothesis unless explicitly stated otherwise. The proprioceptive drift measures and the questionnaire ratings of ownership, agency and task dependent issues served as dependent variables. Ownership, agency and task dependent questions were compared over time, i.e. blocks, and between conditions. We carried out correlation analyses (spearman correlations) to explore the relationship between drift and ownership and agency and ownership in the three different conditions. All data sets were checked on normal distribution using Shapiro-Wilk test (p >.05). But since it is highly disputed that 7-point Likert scale questionnaires are distributed in a normal fashion, we decided to use non-parametric analyses. Drift measure data was assessed for normality using a Shapiro-Wilk test (p > 0.05). Since all drift data was normally distributed we used a two-way paired sample t-test between the different conditions and blocks and ANOVA to compare between responders and non-responders (see next section for explanation of differentiation of responders).When testing for normality of the reaction time data, the switched condition showed not to be normally distributed. Therefore the appropriate non-parametric tests were used to analyze the switched finger response reaction time data.

(12)

12

Proportion of responders

We reasoned that the answers to the two ownership questions;” During the experiment there were times when it felt as if the rubber hand were my hand” and “During the experiment there were times when I felt my fingers moving at the location of the rubber hand” should have a median rating equal to or higher than 1 in order to classify a participant as a responder of the rubber hand illusion. The non-responders were described as participants that had an ownership score of 0 or lower. In the first, non-switched, condition 9 of the 15 participants (60%) showed to be clear responders, whereas 4 of 15 (26.7%) showed to be clear non-responders and 2 subjects showed to be indecisive. Surprisingly in switched condition more than half of the participants, namely 53.3%, 8 of the 15 participants, were reported as clear responders. One was indecisive and the other 6 were marked as non-responders (40%). In the last, non-switched, condition 66.7% was classified as an ownership responder.

The mean rating of the ownership score (n=15) in the first non-switched condition is (0.72, SD ±1.56), in the switched condition is (-0.39, SD ±1.34) and in the third, non-switched, condition is (0.95, SD ±1.69). When the groups are divided in non-responders and non-responders, based on the scores that differ per

condition as previously described, then the mean ratings for illusion responders showed: (1.68, SD ±0.72) for the first, non-switched condition, (1.48, SD ±0.50) for the switched condition, and (1.96, SD ±0.55) for the third, non-switched, condition. Whereas the non-responders showed respectively in the same order (-1.17, SD ±1.46), (-0.94, SD ±0.71) and (-1.07, SD ± 1.32). The overview of ownership score over time is depicted in figure 1.

Figure1. Mean ownership ratings based on the two ownership questions per block divided in Responders and non-responders. Illusion C1.1 C1.2 C1.3 C2.1 C2.2 C2.3 C3.1 C3.2 C3.3 Responders Mean 1.42 1.78 1.83 0.94 1.88 1.63 1.75 2.03 2.10 SD 1.47 .84 .83 0.56 .58 .74 .68 .56 .57 Non-Responders Mean -1.50 -1.13 -0.88 -1.83 -0.33 -0.67 -1.10 -1.30 -0.80 SD 1.47 1.19 1.75 .91 .93 .82 1.25 1.20 1.64 Total (n=15) Mean 0.38 0.80 0.97 -0.40 0.90 0.67 0.80 0.92 1.13 SD 1.95683 1.60 1.59 1.69 1.30 1.35 1.63 1.80 1.73

(13)

13

Questionnaire: Ownership questions

Since we chose a design that includes the subject to answer the questionnaire 9 times, once after each block, we decided to exclude some of the question asked in the original study by Botvinick and Cohen (1998) for time and boredom related reasons. Our pilot data (n=9) showed that on several control questions like “It felt as if my hand were turning rubbery” and “It felt as if I had more than one right hand or arm” people rated negative and significantly different from the ownership ratings when compared to this experiment in all three conditions, Wilcoxon signed rank test shows, conditions in order of testing (Z = -2.395, p = 0.017; Z = -2.703, p = 0.007; Z = -2.296, p = 0.022).

Ownership ratings differences over blocks

The combined ownership data from the questionnaire indicate that there is no significant difference within Condition 1, the first non-switched condition. However, as expected there is a difference between the last block of the non-switched condition and the first block of the switched condition (Z = - 3.048, p = 0.002, Wilcoxon signed ranked test). After the finger responses are switched, the participants experience less ownership sensation over the wooden model hand. However, surprisingly this experience does go up over time within the switched condition. In the first block within this condition is lower than the 2nd block (Z = 3.083, p = 0.002) and also differs from the last block (Z = 2.631, p = 0.009). The second and the third block do not differ from each other. The third condition, non-switched, showed a difference between the beginning and the end (Z = 2.153, p = 0.031). Interpretation of these results can be found in the discussion.

Ownership rating differences between conditions

When looking at the differences between conditions it is surprising to see that there is no difference between the first non-switched condition and the switched condition. You can see in figure X that there seems to be a time effect appearing in the switched condition. Illusion responders tend to rate the sense of ownership higher after getting used to the switched fingers. This point will be elaborated on further in the results and discussion section. What we did expect is that when subject, in this case responders, go back to the ‘normal’ non-switched condition, the third condition, then in general ratings of ownership go up (Z = 1.995, p = 0.046) and also the amount of responders goes slightly up (see above).

(14)

14

Agency ratings

The agency ratings are controlled with the agency control question, asked after each condition. The agency ratings consisted of the mean rating of the two following questions: “It felt as if the rubber hand moved in a way I expected it to move” and “It felt as if I was moving the rubber hand”. This rating was compared with the following agency control question: “During the experiment there were times when it felt as if the rubber hand was moving my fingers”. As expected, participants experienced a significantly higher sense of agency in both non–switched conditions 1 and 3 compared to their condition related agency control questions (respectively Z = -3.408, p = 0.001, and Z = -3.240, p = 0.001). The sense of being in control was also rated significantly higher from the control question when seen over the entire condition (Z = -3.353, p = 0.001). However, agency ratings in the switched condition did not reach the levels of the non-switched conditions (condition 1 vs. condition 2 Z = -3.328, p = 0.001; condition 2 vs. condition 3 Z = 3.409, p = 0.001). The agency ratings in both non-switched conditions did not differ from each other. As hypothesized, agency ratings did go up over time when finger responses were switched. In block 3 of the switched condition the participants rate the control over their movement higher than in block 1 (Z = 2.454, p = 0.014). That agency ratings in the switched condition did not reach the levels of the non-switched conditions could be due to the fact that two subjects reported that they were completely unable to move when looking at the hand when finger responses were switched. These two participants will also be excluded from the reaction time task analysis. As a control to see whether or not participants would experience more agency over the index finger than the middle finger we asked in every block if they felt in control of their index finger and of their middle finger, in separate questions. In none of the blocks were these questions rated differently, meaning they experienced the same extent of agency over the middle finger as over the index finger.

Correlation between agency and ownership

To investigate if there is a connection between agency and ownership over an artificial limb we looked at their correlations in all different condition. In condition 1 the related ownership and agency ratings were correlated (Spearman: r = 0.541, p = 0.037). Interestingly in condition 2 an even stronger correlation was found (Spearman: r = 0.669, p = 0.006). Contrary to our expectations the third condition did not show a correlation between ownership and agency.

(15)

15

Proprioceptive drift

As described in the methods section, proprioceptive drift measures were taken before the participants saw the setup, before every condition and after every block (a total of 13 times) to be able to generalize the data and have a clear overview and a good baseline measure for every individual subject. One participant was excluded because he mentioned after a couple of blocks that he was pointing towards a horizontal position of his hand and not so much to the vertical (height) of his hand, resulting in a lack of baseline measure for the rest of his proprioceptive drift pointing data. The baseline drift was subtracted from all other proprioceptive drift measures to get the relative drift.

Figure 2. This graph shows the mean proprioceptive drifts in cm in every block of all participants (n=14). Condition 1 is blue, condition 2 (switched) is red and condition 3 is depicted in green.

Using a paired-samples t-test we found that there is a tendency, but no significant drift between the start baseline and after 20 seconds of practice before the unset of first condition (t(13) = - 2.132, p = 0.053). The baseline drifts before condition 1,2 and 3 do not differ from each other. Only block 2 in the first condition seems to be significantly different from the baseline measure (t(13) = -2.759, p = 0.016). See figure 2 for mean proprioceptive drift in all blocks.

Proprioceptive drift compared between responders and non-responders

When differentiating between responders and non-responders we investigate drift in every condition using independent t-tests. The baseline drift before the first condition did not differ when comparing between ownership responders and non-responders. Although the baseline measures of condition 2 seems to be

(16)

16 slightly higher for participants that experience an illusion compared to the non-responders, there is no significant difference. The baseline drift in condition 3 is higher in responders t(15) = 3.266, p = 0.006. It seems that in responders ownership builds up over time and does not get completely abolished by switching finger responses (or switching back). See figure 2 and figure 3. Because it does not abolish the illusion completely baseline drift measures are relatively high. Therefore responders only showed

significantly higher proprioceptive drift than baseline in block 2 and 3 of the first condition (block2 t(14) = 2.436, p= 0.040; block 2 t(14) = 2.821, p = 0.021). In all other blocks there seems to be a trend, look at figure 3, but no significantly higher ratings. In line with our hypotheses, non-responders never showed a difference from the baseline drift pointing data.

Figure 3. Means of proprioceptive drift measured in cm in all conditions divided over responders and non-responders.

Figure 4. Graphs of average proprioceptive drift in cm divided in condition and over responders and non-responders. Condition 1 is blue, condition 2 (switched) is red and condition 3 is depicted in green.

Proprioceptive Drift

Baseline-C1 C1.1 C1.2 C1.3 Baseline-C2 C2.1 C2.2 C2.3 Baseline-C3 C3.1 C3.2 C3.3 responders Mean 2.30 3.54 5.30 5.93 3.24 4.20 4.35 4.39 4.09 5.25 4.80 5.64 N 8 8 8 8 8 8 8 8 10 10 10 10 SD 3.2 3.4 3.2 4.0 3.1 3.2 2.9 2.9 3.2 2.8 3.2 3.0 non-responders Mean 0.10 0.15 0.05 -0.78 -0.42 -0.24 -0.66 -0.70 -1.45 -1.45 -1.30 -1.25 N 4 4 4 4 5 5 5 5 4 4 4 4 SD 1.0 0.7 2.7 1.7 2.2 1.4 1.8 2.0 1.1 1.2 1.2 1.3 Total Mean 1.55 2.22 3.41 3.18 2.25 2.84 2.84 2.97 2.51 3.34 3.06 3.67 N 14 14 14 14 14 14 14 14 14 14 14 14 SD 2.7 3.1 3.9 4.5 3.5 3.5 3.7 4.0 3.8 4.0 3.9 4.2

(17)

17

Correlations between proprioceptive drift and ownership

As in the original study by Botvinick and Cohen (1998) we looked if proprioceptive drift and ownership ratings correlate. Using a Spearman’s rho correlation it showed that in every block and even in every condition there is a correlation between proprioceptive drift and the ownership ratings of the participants. Again, surprisingly even in the switched condition this correlation appeared and even went up over time. Figure 5 shows all correlations between drift and ownership ratings.

Correlations between ownership and proprioceptive drift

MeanC 1 C1.1 C1.2 C1.3 Mean C2 C2.1 C2.2 C2.3 MeanC 3 C3.1 C3.2 C3.3 Spearman's rho Cor. Coefficient ,703** ,795** ,588* ,686** ,775** ,593* ,636* ,760** ,793** ,753** ,776** ,767** Sig. (2-tailed) .005 .001 .027 .007 .001 .026 .014 .002 .001 .002 .001 .001 Figure 5: Correlations between ownership rating and proprioceptive drift in all blocks.

Reaction time task

As mentioned earlier, the reaction times were measured with two small sensors. With the help of these sensors the onset of the movement and therefore the reaction time could be determined per trial. Before looking at the reaction time data, we excluded outliers, described as reaction that deviated more than 2 times a standard deviation from the mean. In total 4.4% of all trials were excluded in condition1, 3.7% in condition 2, and 4.1% in condition 3. To get a general idea of the data we performed a repeated measures ANOVA. This showed that, as expected, there was a main effect for condition. The switched condition has a larger reaction time than the two non-switched conditions (F = 15.053, p = 0.001). There showed to be no time-effect, against our hypothesis, for illusion responders nor for non-responders. We hypothesized that especially in the beginning of the switched condition illusion responders will have a slower reaction, but at the end will be acquainted with the situation and reaction times will decrease and will be of the level of the non-responders. Since the reaction time data in the second condition turned out to be not normally

distributed we used the suited non-parametric test to investigate our hypotheses. No such effect was found (Z = -0.594, p = 0.552). Furthermore we hypothesized that illusion responders will experience after effects from the switched condition when starting with the non-switched condition. Again a Wilcon signed rank test showed no after effect (first block of third condition vs. second block of third condition; Z = 2.84, p = 0.776). See figure 6 below for mean reaction times of all participants over all blocks.

(18)

18 Figure 6: Mean reaction times in milliseconds in all blocks.

Correct responses in reaction time task

Parallel to the reaction times, the amount of correct trials also showed to be worse in the switched condition (F=18.530, p<0.001). Also an interaction effect between condition and block was found, leading to a significantly lower amount of correct trials in the first block of the switched condition (F=2.994, p=0.028). The amount of correct responses in the switched conditions does no reach the level of the non-switched conditions. In addition no after effect from the non-switched on the beginning of the non-non-switched condition was found. When looking closer at the data a sex effect appeared. Also In the first block of the first conditions females responded correctly to less trials than males (Z = -2.716, p = 0.007).

(19)

19

Closing questionnaire

After all three conditions participants were asked to answer several closing questions. From this closing questionnaire we could obtain that not everyone experienced the different conditions in the same way. Illusion responders experienced a difference between the first and the third condition whereas the non-responders did not (Z = -2.471, p = 0.019). Responders also experienced a difference in the beginning and end of the third condition (Z = -2.255, p = 0.024). The answers to these questions could suggest that participants that are susceptible for the illusion experienced an after-effect from the switched condition over the normal condition. Subjective reports of some of the participants also mentioned that they felt confused in the beginning of the third condition. This could be a clue that participants really perceived the model hand with the switched finger as part of their body; however more experiments are needed to draw conclusions.

Discussion

This study brings forward a couple of main findings. First we successfully used the setup of the moving rubber hand illusion based on the set up used by Kalckert and Ehrsson (2012). We further developed the set up by adding an extra movable finger and introduced the switched finger condition. This switched finger condition showed not to be capable of breaking the illusion, as what was expected, but surprisingly when index and middle finger responses are switched the movement of the model hand was still able to elicit an illusion of ownership. The experience of ownership was supported by subjective reports, the RHI

questionnaire and proprioceptive drift data. Closing questionnaire data suggests that Illusion responders experienced an after-effect caused by the switched condition, although nothing could be found in reaction data or in the amount of correct trials. Even though we put forward a set up that challenges the boundaries of ownership over an artificial limb, future studies are needed to strengthen the body of knowledge. This discussion will give possible explanations about the outcomes of the experiment and why some deviated from the expected results. Furthermore it will suggest improvements and further research.

As described in the introduction Lloyd (2007) explored the spatial limits of the rubber hand illusion and showed that the illusion significantly decayed when the hand was placed more than 30 cm away from the real hand in a horizontal direction. The closer the fake limb was placed to the real hand (the closest being 17.5 cm) the stronger the illusion was rated. Interestingly the time taken to elicit the illusion followed the same trend. This could explain why the switched finger response condition in our study had lower ownership ratings in block 1 compared to block 2 and 3, it simply takes a longer time to get used to

(20)

20 the situation and elicit an ownership illusion. It is possible that the brain learns to deal with the

multisensory mismatch between proprioceptive information and visual feedback and is therefore able to incorporate a switched body part as small as a finger into its bodily image. Dealing with a mismatch like this is not unique (van Beers, Wolpert and Haggard, 2002) but it is not yet shown in body ownership associated studies. In contrast to the other, previously described manipulations, the switched condition is apparently easier to adopt within a person’s body image. The movement in itself is a natural movement, and the position of the hand and even of the individual fingers is normal opposed to a very distant or an

incongruent hand position used in earlier studies. A replication of this study, possibly combined with fMRI to look for body-ownership related area activation, could tell us more about the possible incorporation of switched fingers within our body image. Ideally this replication study will be combined with the study of Kalckert and Ehrsson (2012) in order to look for a double dissociation of agency and ownership within the switched moving rubber hand illusions paradigm.

Several difficulties should be addressed when discussing these outcomes and designing future studies. An important issue concerning the rubber hand illusion is that it’s an illusion that is not continuously as strong over trials. The experience has been described as one that comes and goes. Small malfunctions of the setup, like scratching of the rods against the side or loosening of the finger caps, among movements of the non-participating fingers of the real hand are able to (partly) destroy the illusion. In accordance with questionnaires of Botvinick & Cohen (1998) and Kalckert and Ehrsson (2012) among others, we tried to design the ownership questions to control for these ‘bouts’ of illusion. By instructing the

participants that the illusion does not have to be continuous and by introducing the questions with “During the experiments there were times when..”’ we tried to catch the strength of the illusion independent of frequency or constancy of it. Nevertheless one should keep in mind that it is likely that participants can take the amount of illusion bouts into account when rating the ownership questions. This could explain why the ownership ratings in the non-switched condition are somewhat lower than the ownership ratings of Kalckert and Ehrsson (2012). It is also noticeable that the ratings of responders go up over time, but the ratings of non-responders also show a trend of becoming less negative. We address this phenomenon to habituation, the participants seems to be getting more and more familiar with the situation and are

therefore also likely to rate ownership related questions higher. This is one of the dangers when addressing the rubber hand illusion with the help of questionnaires.

Due to time and tiredness reasons we decided to use the ownership control questions of the pilot data to control for the ownership rating questions. We know that from a methodology point of view it is not the best way to exclude these ownership control questions in the real experiment, but we believe the

(21)

21 pilot data in combination with the ownership questions correlation with the drift is a strong enough claim to put forward the statements about rubber hand illusion. Furthermore it does not take away that this a point of discussion in every study using subjective report and specifically studies trying to grasps the experience of ownership over a body part with the use of questionnaires. A large group of subjects might be able to even out individual differences like these. Unfortunately this study held a relatively small group of participants due to time consuming analysis of the individual reaction time response trial for trial. The individual differences might also be reflected in the answers of the questionnaires. Some questions might be multi-interpretable. For example: “During the experiment there were times when I could not feel my own right hand”. This could be interpreted literally or as moments of awareness loss of the real hand. Therefore we did not include this question in the final analysis. Additionally Kalckert and Ehrsson (2012) following the illustration of Gallagher et al. (2007) and David et al. (2008) make a distinction between external agency and body agency. Body agency can be described as the experience of being a causative agent over a movement of one’s own body. Whereas external agency does not have to be confined to a part of one’s body and is more closely related to the sense of being a causative agent during interactions with surrounding objects. This is a good point, but unfortunately no study using movement in a rubber hand illusion paradigm, including this one, makes a distinction in their questionnaire trying to address these different kinds of agency. Although it is likely that when people experience a model hand as their own they also experience body agency, it would be interesting to see if the non-responders than also claim external agency over a model hand.

The link between agency and ownership revealed an interesting point of discussion. A stronger correlation between agency and ownership was found in the, second, switched condition compared to the first, non-switched condition. This could be interpreted in multiple ways. It could be that the participants that experienced more agency, therefore experienced more ownership over the model hand. Another explanation could be: people that are more susceptible for the rubber hand illusion will more easily embrace the model hand and therefore also feel more in control of its movements. Against our hypotheses the second condition showed that asynchronous visual feedback as displayed with the switched fingers is not enough to break the illusion.

Next to individual differences, another concern is that the questionnaire answers about actions that are already in the past. Although in this study it is sufficient to ask about the generalizable experiences over an entire block, both ownership and agency over bodily movements are experiences that might be positively addressed to oneself easier when asked afterwards. It could be that agency acts within some kind of control system that notifies us when we make a mistake and lose control over certain actions.

(22)

22 An online measure method would be preferable in future studies. Furthermore it is an outstanding

question whether or not we need agency at all in a small reaction time task like the one that is used in this study. An automated motor controlled movement might play a larger role than the sense of being in control of that movement.

The drift measures were another compelling result in this study, especially when looking at responders vs. non-responders. Rohde, de Luca & Ernst (2011) claim that the mere sight of a rubber hand in the classical set up results in proprioceptive drifts, and that these drift are never larger than the drifts seen when the hand is being stroked synchronously. Our data showed that ownership sensation and drift do go hand in hand. Both the responders and the non-responders had a small drift after only seeing the hand, but after movement only the proprioceptive drift of the RHI responders went up. However in a next design it would be advisable to destroy the illusion entirely between conditions. The data now shows that

proprioceptive drift remains high and therefore not significantly differs from the baseline drift in the second and the third condition. Keeping these possible improvements in mind, it does not take away the fact that the drift and ownership score correlated in all 9 blocks.

Even though there might be some possible improvements showing helpful for replication studies or future directions using the moving rubber hand illusion paradigm, this study clearly sheds new light on the boundaries of ownership over body parts. By unraveling these boundaries, body ownership and agency studies like this will benefit potential clinical applications. The mirror box designed to relief

phantom limb pain described by Ramachandran et al. (1995; 2009) is a well-known clinical application that aids patients with a distorted body representation. But especially the creation of artificial limbs for

amputees will benefit when the precise relations between bodily representation, ownership and agency over specific body parts becomes more and more clear.

References

Armel, K. C., & Ramachandran, V. S. (2003). Projecting sensations to external objects: evidence from skin conductance response. Proceedings. Biological sciences / The Royal Society, 270(1523), 1499-506. doi:10.1098/rspb.2003.2364

Barnsley, N., McAuley, J. H., Mohan, R., Dey, a., Thomas, P., & Moseley, G. L. (2011). The rubber hand illusion increases histamine reactivity in the real arm. Current Biology, 21(23), R945-R946. Elsevier. doi:10.1016/j.cub.2011.10.039

van Beers, R. J., Wolpert, D. M. & Haggard, P.(2002). When feeling is more important thanseeing in sensorimotoradaptation. Current Bioogy,12, 834−837

(23)

23 Blakemore, S.-J., & Frith, C. (2003). Self-awareness and action. Current Opinion in Neurobiology, 13(2),

219-224. doi:10.1016/S0959-4388(03)00043-6

Botvinick, M., & Cohen, J. (1998). Rubber hands “feel” touch that eyes see. Nature, 391(6669), 756.doi:10.1038/35784

Cadieux, M. L., Whitworth, K., & Shore, D. I. (2011). Rubber hands do not cross the midline. Neuroscience letters, 504(3), 191-194. Elsevier Ireland Ltd. doi:10.1016/j.neulet.2011.09.010

Costantini, M., & Haggard, P. (2007). The rubber hand illusion: Sensitivity and refer- ence frame for body ownership. Consciousness & Cognition, 16, 229–240.

David, N., Newen, A., and Vogeley, K. (2008). The “sense of agency” and its underlying cognitive and neural mechanisms. Conscious. Cogn. 17, 523–534.

Dummer, T., Picot-Annand, A., Neal, T., & Moore, C. (2009). Movement and the rubber hand illusion. Perception, 38(2), 271-280.doi:10.1068/p5921

Ehrsson, H. H., Spence, C., &Passingham, R. E. (2004). That’s my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science (New York, N.Y.), 305(5685),

875-7.doi:10.1126/science.1097011

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

Ehrsson, H. H., Rosén, B., Stockselius, A., Ragnö, C., Köhler, P., &Lundborg, G. (2008). Upper limb amputees can be induced to experience a rubber hand as their own. Brain, 131, 3443–3452.

Ehrsson, H. H., Wiech, K., Weiskopf, N., Dolan, R. J., &Passingham, R. E. (2007). Threatening a rubber hand that you feel is yours elicits a cortical anxiety response. Proceedings of the National Academy of Sciences of the United States of America, 104(23), 9828-33.doi:10.1073/pnas.0610011104 Gallagher, I. (2000). Philosophical conceptions of the self: implications for cognitive science. Trends in

cognitive sciences, 4(1), 14-21. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10637618 Gallagher S. (2007). The natural philosophy of agency.Philos. Comp.2, 347–357. doi:

10.1111/j.1747-9991.2007.00067

Haggard, P. (2005). Conscious intention and motor cognition. Trends in cognitive sciences, 9(6), 290-5. doi:10.1016/j.tics.2005.04.012

Ijsselsteijn, W. A., Kort, Y. A. W. D., & Haans, A. (n.d.). Is This My Hand I See Before Me ? The Rubber Hand Illusion in Reality , Virtual Reality , and Mixed Reality. Hand, The.

Kammers, M. P. M., de Vignemont, F., Verhagen, L., & Dijkerman, H. C. (2009). The rubber hand illusion in action. Neuropsychologia, 47(1), 204-11. doi:10.1016/j.neuropsychologia.2008.07.028

Kalckert, A., & Ehrsson, H. H. (2012b). Moving a Rubber Hand that Feels Like Your Own: A Dissociation of Ownership and Agency. Frontiers in Human Neuroscience, 6(March),

(24)

24 Lloyd, D. M. (2007). Spatial limits on referred touch to an alien limb may reflect boundaries of visuo-tactile

peripersonal space surrounding the hand. Brain and cognition, 64(1), 104-9.doi:10.1016/j.bandc.2006.09.013

Longo, M. R., Schüür, F., Kammers, M. P., Tsakiris, M., & Haggard, P. (2009).Self-awareness and the body image.ActaPsychol (Amst), 132(2), 166–172.

Makin, T. R., Holmes, N. P., & Ehrsson, H. H. (2008). On the other hand: dummy hands and peripersonal space. Behavioural brain research, 191(1), 1-10.doi:10.1016/j.bbr.2008.02.041

Maravita, A., & Iriki, A. (2004). Tools for the body (schema).Trends in Cognitive Sciences, 8, 79–86.

Maravita, A., Spence, C., & Driver, J. (2003). Multisensory integration and the body schema: Close to hand and within reach. Current Biology, 13, R531–R539.

Moseley, G. L., Olthof, N., Venema, A., Don, S., Wijers, M., Gallace, A., & Spence, C. (2008). Psychologically induced cooling of a specific body part caused by the illusory ownership of an artificial counterpart. Proceedings of the National Academy of Sciences of the United States of America, 105(35), 13169-73. doi:10.1073/pnas.0803768105

Newport, R., Pearce, R., & Preston, C. (2010). Fake hands in action: embodiment and control of supernumerary limbs. Experimental brain research.ExperimentelleHirnforschung.

Expérimentationcérébrale, 204(3), 385-95. doi:10.1007/s00221-009-2104-y

Oldfield, R. C. (1971). The assessment and analysis of handedness: the Edinburgh inventory.

Neuropsychologia, 9(1), 97-113. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/5146491 Petkova, V. I., & Ehrsson, H. H. (2008). If I were you: Perceptual illusion of body swapping. PLoS ONE, 3(12),

e3832.

Ramachandran, V.S., Rogers, D., & Cobb, S. (1995). Touching the phantom limb. Nature, 377, 489-490 Ramachandran, V.S., Brang, D., McGreoch, P.D. (2009). Size reduction using Mirror Visual Feedback (MVF)

reduces phantom pain. Nature, 15, 357-360

Rohde M., Di Luca M., Ernst M.O.(2011) The Rubber Hand Illusion: feeling of ownership and proprioceptive drift do not go hand in hand. PloS ONE (6), e21659.

Sanchez-Vives, M. V., Spanlang, B., Frisoli, A., Bergamasco, M., and Slater,M. (2010). Virtual hand illusion induced by visuomotor correlations. PLoSONE(5), e10381.

Tastevin, J. L’Encephale 1, 57–84 (1937).

Tsakiris, M. (2010). My body in the brain: a neurocognitive model of body-ownership. Neuropsychologia, 48(3), 703-12.doi:10.1016/j.neuropsychologia.2009.09.034

(25)

25 Tsakiris, M., & Haggard, P. (2005). The rubber hand illusion revisited: visuotactile integration and

self-attribution. Journal of experimental psychology. Human perception and performance, 31(1), 80-91. doi:10.1037/0096-1523.31.1.80

Tsakiris, M., Prabhu, G., & Haggard, P. (2006). Having a body versus moving your body: How agency structures body-ownership. Consciousness and cognition, 15(2), 423-32.

doi:10.1016/j.concog.2005.09.004

deVignemont, F. (2011). Embodiment, ownership and disownership.Consciousness and cognition, 20(1), 82-93.doi:10.1016/j.concog.2010.09.004

Walsh, L. D., Moseley, G. L., Taylor, J. L., & Gandevia, S. C. (2011). Proprioceptive signals contribute to the sense of body ownership. The Journal of physiology, 589(Pt 12), 3009-21.

(26)

26

Pilot: Transcranial Magnetic stimulation during the moving rubber hand

illusion

The rubber hand illusion is a well-known multisensory paradigm trying to explore the relationship between the body and the sense of ’self’. As previously described, the existing body of knowledge has mainly investigated this sort of awareness by researching the boundaries of body ownership and agency. Thus far the anatomy of these traits has been underexposed.

With the help of functional magnetic resonance imaging Ehrsson, Spence and Passingham (2004) found neural activity in the premotor cortex that reflects the feeling of ownership over a fake rubber hand. A control of the sight and touch of the rubber hand alone was included. They even found a

relationship between the subjective rated strength of the illusion and the level of neural activity in the ventral premotor cortex. To exclude that activity does not reflect a visual representation of the brush, or any other object near the hand, Ehrsson, Holmes and Passingham (2005) introduced a somatic rubber hand illusion follow up study. The experimenter moved the left index finger of the blindfolded participant so that it stroked the fake rubber hand. Simultaneously the experimenter touched the participant’s real right hand. After touching the hands in synchrony for around 10 seconds the participants experience an illusion of touching one’s own hand. This illusionary experience was scanned using fMRI and compared with two control conditions. They suggest that the activity found in the ventral premotor cortices, intraparietal cortices and the cerebellum reflects the detection of multisensory bodily signals and therefore could be the underlying mechanism for the feeling of body ownership.

With the use of positron emission tomography Tsakiris et al. (2007) identified a neural correlate of the shift in perceived position of the hand during the rubber hand illusion. When the fake rubber hand was attributed to the self, activity in the right posterior insula and the right frontal operculum was found. Kammers et al. (2008) applied offline low frequency repetitive transcranial magnetic stimulation (rTMS) over the inferior posterior parietal lobule (IPL), an area that shows to be active during a working rubber hand illusion (Ehrsson et al. 2005). Their results suggest that IPL is involved in immediate perceptual (re) location of the limbs when the RHI is elicited, but the subjective experience of ownership is not

associated with this region. Using the rubber hand illusion paradigm, Tsakiris et al (2008) applied single pulse TMS to disrupt the right temporo parietal junction (rTPJ). They suggest that this area is involved in maintenance of the bodily sense of one’s own body, while differentiating it from external objects.

(27)

27 So far transcranial magnetic stimulation has only been used as a disruptive tool within the rubber hand paradigm. As shown by Pascual-Leone (2000) among others, the technique can also be used to show functional connectivity between different brain regions. Here we propose to use paired pulse TMS as a tool to map functional connectivity between brain regions that has shown to play a role in attributing a body part to one’s own body. Since we used a moving rubber hand, we stimulated over M1. After Pilot work of single pulse TMS we plan to use paired pulse TMs over M1 and over the intraparietal sulcus (IPS) as Ehrsson et al. (2005) showed it to be a correlate of ownership. We measured the electromyographic (EMG) signal of the first dorsal interosseous (FDI) muscle during movement in the moving rubber hand illusion pilot-experiment. Yahagi and Kasai (1998) showed with the help of MEP’s of the FDI muscles that the primary motor cortex plays a role in the mental representation motor movements. We hypothesize that the fronto-parietal connectivity is increased during a working illusion. Unfortunately this hypothesis could not yet be tested, due to problems in the single pulse stimulation over M1 in the pilot phase. Therefore the most recent of single pulse stimulation over M1 pilot data is discussed in this report.

Methods

Participants

In this pilot study we used 6 naïve participants, only five of them participated through the entire pilot since one was not susceptible for the rubber hand illusion. All participants had normal or corrected to normal eye sight. The study protocol was approved by the local ethics committee of the Capital Region Copenhagen.

Procedures

The pilot consisted of 3 different conditions. In each condition the participant has his left hand in a box similar as described earlier in Knip (2012), based on a set up described by Kalckert and Ehrsson (2012). Participants were seated in front of a box placed on the right edge of a table. The left hand was placed inside the box and the index finger of the left hand was attached to the index finger of the wooden hand model resembling a left human hand. The participant’s right hand was hanging freely along the

participant’s body. The top and the sides of the box were covered with fabric and at the right side of the box a sheet covered plate was placed to use as a pointing tool to measure proprioceptive drift.

On the left index finger of the participants two electrodes were placed to measure MEP’s of the FDI muscle. Before the experiment started the participants were instructed to move their finger in their own pace from side to side. They were allowed to change pace or pause whenever they liked, as long as they were aware of the set up, focused their attention on it and only moved their left index finger.

(28)

28 After all three conditions the participants were asked to answer a questionnaire used by Kalckert and Ehrsson (20012), originally based on the RHI questionnaire by Cohen and Botvinick (1998). For a more elaborate explanation see methods section of the previous report.

Conditions

The entire pilot consisted out of three conditions. One condition with the index finger attached to the model hand that was placed in a congruent position to the body of the participant. This condition is later referred to as the Agency and Ownership condition (AO). In another condition; the index finger was attached and the wooden model hand was placed in an incongruent position to the body (turned around 180 degrees). This condition is later referred to as the Agency, no Ownership condition (ANO). In the last condition, referred as the No Agency and No Ownership condition (NANO), the index finger was detached from the wooden model hand which was placed in an incongruent position. The three different conditions were randomized. Each condition consisted out of 3 states. With the help of a Signal script, one state triggers a transcranial magnetic stimulation over M1. The other two states were without TMS stimulation. One condition lasted as long as 20 trials. This was usually around 5 minutes.

Apparatus

Two electrodes were placed on the left index finger. The MEP’s were measured and collected with the help of the Software program Signal. To find the motor threshold of the FDI muscle we used a technique

described by Rossi et al. (2009). We defined the motor threshold to be at the level when the lowest intensity still to reach 50 microvolt at least 5 out of 10 stimulations over M1. During this procedure the subjects were asked to keep their eyes open and their hands at rest.

The transcranial magnetic stimulation was performed with the use of a Biphasic Magstim Rapid 200. The direction of the current was from posterior to anterior. During this pilot stimulation was only over motor area M1.

The TMS pulse was triggered with the help of an amplifier measuring the amplitude of the MEP’s of the FDI muscle. Therefore movement of the finger would directly result in receiving a magnetic pulse depending on what pseudo randomized state was active during this movement.

(29)

29 Results

As previously described this pilot study used a questionnaire, proprioceptive drift measure and motor evoked potentials (MEP’s) to study the moving rubber hand illusion.

Questionnaires

In the Agency and Ownership (AO) condition the agency scores were rated higher than the agency control questions (Wilcoxon signed rank test; Z = -2.041, p = 0.041). Also the ownership score was rated

significantly higher than ownership control question (Wilcoxon signed rank test; Z = -2.023, p = 0.043). When looking at the Agency and No Ownership (ANO) condition the agency scores showed the same trend and were higher than the control questions in this condition (Z = -2.032, p = 0.043). As expected, since the rubber hand was placed in a incongruent position, the ownership scores were rated negatively and did not differ from the ownership control question (p = 0.66). As anticipated the No Agency and No Ownership (NANO) condition, where the model hand was not only in an incongruent position but also not attached to the subjects finger, showed no difference in both agency and ownership scores compared to scores of the control questions (respectively p = 0.317, p = 0.317).

Figure 1: Average Agency and Ownershipscores and their control ratings in all three conditions (AO, ANO and NANO).

Proprioceptive drift

Even before seeing the set up and with their eyes closed, the subjects were asked to point 3 times with their right hand towards were they think their left hand is located to determine the baseline. After each condition the subjects were asked to close their eyes and point 3 times to the location of their left hand.

(30)

30 The proprioceptive drift data showed to be different in Agency No Ownership compared to the No Agency and No Ownership condition t(5) = 3.426, p = 0.027. Due to a larger deviation the drift measure in the Agency and Ownership condition did only show a trend to be larger than the condition where agency and ownership were both absent t(5) = 2.522, p = 0.065.

Figure 2: Average, baseline corrected proprioceptive drift measures of all three conditions in centimeters.

Motor evoked potentials

The motor evoked potentials were measured from the FDI muscle during the 3 conditions. As previously defined each condition is divided in three states; one with (state 1) and two without (state 2 and 3)

transcranial magnetic stimulation. The potentials (in mV) shown in figure 3 depict the average amplitude of one state. The different colors represent the different conditions. The states without TMS did not showed to differ from each other in any condition (paired sample t-test; p = 0.650, p = 0.367,p = 0.074). However when looking at the states where TMS was applied differences come to light. MEP amplitudes were smaller in the Agency and Ownership (AO) condition compared to the Agency without Ownership condition (ANO) ( t(4) = -2,789, p = 0.049). The condition without Ownership and without Agency (NANO) also showed smaller MEP amplitudes than the condition with Agency and without Ownership (ANO) ( t(4) = 3.516, p = 0.025). When comparing the conditions AO and NANO during the TMS state no differences in MEP amplitudes were found ( t(4) = -1.623, p = 0.180).

Figure 3: Graph of average MEP’s in all three conditions in all states. Blue line represents Agency and Ownership condition, red line represents Agency and No Ownership condition and blue line

represents No agency and No Ownership condition. On the horizontal axes the different states are haracterized ; 1 is the state were TMS was applied,2 and 3 were without TMS.

(31)

31 Discussion and concluding remarks

The outcomes of this pilot showed multiple points of interest for further research. Firstly it showed that TMS can be used in a non-disruptive way within the rubber hand paradigm. Even though all subjects received magnetic stimulation over M1, 4 out of 5 subjects rated the ownership questions positively. Secondly it successfully replicated part of the study by Kalckert and Ehrsson (2012) by showing dissociation between ownership and agency. Furthermore this pilot opens the way to the previously proposed research of evaluating fronto-parietal connectivity during an ownership illusion.

Since a great deal of the pilot is similar to the earlier described moving rubber hand illusion study, only specific points of improvement for the pilot will be discussed. It is argumentative to note that parts of the pilot may benefit from possible improvements. The amount of participants is a probable cause of weak statistical power. A larger group of subjects might be necessary to make stronger claims about the differences between conditions. The use of subjective reports with the help of Likert-scale questionnaires is also debatable. Since a condition consists of only one state with applied TMS and two states without, one could question if the ownership ratings of one condition do not only reflect the bouts of illusion during the states without TMS. Verbal reports after the pilot claimed, however, that the single pulse did not always disrupt the illusion. Surprisingly the drift measures showed no difference between the Agency and

Ownership (AO) condition and the Agency and No Ownership (ANO) condition. The model hand was placed in an incongruent position during the ANO condition; consequently it was unexpected to see a

proprioceptive drift in this condition. A possible explanation could be that the sight of the movement is causing the proprioceptive drift, this however is not in line with Kalckert and Ehrsson’s (2012) findings and needs more data in order to make a non speculative claim.

A possible outcome of this pilot, that only showed a minor trend, is the difference in MEP amplitudes between the Agency and Ownership and the No Agency and No Ownership condition. When this trend, shown in Figure 3 on the previous page, becomes a significant difference it could mean that the subjects are attributing the model hand to their own body while excluding one’s own hand from the image of the bodily self in a similar way as shown by Moseley et al. (2008) with the help of temperature drop and in Barnsley et al. (2011) with increased levels of histamine in the real hand during a working ownership illusion.

All in all, this pilot is very promising and needs to be carried out in full in the near future. The single pulse TMS set up needs to be expanded and modified in a paired pulse TMS set up stimulating over M1 and the intraparietal sulcus (IPS) to address the fronto-parietal connectivity during a working rubber

Moving rubber hand illusion: a switched perspective