The Role of Visual Active Self-Touch in Body Ownership

Table of Contents

1 Introduction

2 Theoretical background
2.1 Body ownership
2.2 The Rubber Hand Illusion
2.2.1 Psychometrics of body ownership
2.2.2 The neuropsychological mechanisms of the RHI
2.3 Self-touch
2.3.2 Self-touch and body-ownership illusions
2.3.3 State of the art
2.3.4 Does active visual self-touch induce a greater ownership illusion than external touch?
2.4 Conceptual Hypotheses

3 Method
3.1 Experimental Design and operationalization of the independent variables 26
3.2 Measures of illusory ownership
3.2.1 RHI Questionnaire
3.2.2 Proprioceptive drift
3.3 Experimental setup
3.4 Participants
3.4.1 COVID-19
3.5 Procedure
3.5.1 Preparation
3.5.2 Illusion induction
3.6 Empirical hypotheses

4 Results
4.1 Data preparation
4.2 Descriptive statistics
4.3 Hypothesis testing
4.3.1 Testing of hypothesis 1
4.3.2 Testing of hypothesis 2
4.3.3 Testing of hypothesis 3
4.3.4 Explorative analysis

5 Discussion
5.1 Summary and discussion of results
5.2 Limitations
5.3 Future research and implications

5 Conclusion

Abstract

The goal of the present study was to investigate how active visual self-touch modulates ownership illusions during the rubber hand illusion (RHI). 23 healthy test subjects participated. A 2x2x2 repeated measures ANOVA was performed, in which self-touch vs. external touch during RHI induction was examined. Additionally, the effect of touching the rubber hand during the RHI vs. not touching it was measured, as well as the location of the rubber hand within or outside of peripersonal space (PPS). The measures used were the behavioral measure of the proprioceptive drift (PD) and a 10-item subjective embodiment questionnaire, including the components “ownerhip”, “location” and “agency”. Based on previous studies and theories regarding self-touch, it was hypothesized that self-touch would positively modulate body ownership. This could not be confirmed, as no difference in degree of embodiment was found between “self” and “external” conditions. A further postulation was that touching the rubber hand during the RHI, instead of just performing the stroking motion without the touch, would induce a higher sense of embodiment, which could be confirmed. Lastly, it was hypothesized that embodiment would be positively modulated by the rubber hand being inside the PPS of the subject's own hand. The results however showed that a greater sense of embodiment of the rubber hand came from the rubber hand being outside of PPS. In an explorative analysis a significant three-way interaction between all three factors could be shown.

Key words: body ownership, embodiment, rubber hand illusion, peripersonal space, proprioceptive drift, body plasticity, self-touch

Table of Figures

Figure 1. Schematic diagram of the experimental set-up in the rubber hand illusion task.

Figure 2. RHI questionnaire results from the original study of Botvinick and Cohen (1998). 11 Figure 3. Component loadings from the PCA on the embodiment of the rubber hand components found by Longo et al. (2008).

Figure 4. The four constraints determining hand ownership

Figure 5. A schematic representation of the different regions of space surrounding the human body

Figure 6. Body ownership ratings on a scale between 0 and 1 for “other” (external) conditions, “self” conditions and “self no-touch” conditions.

Figure 7. A predictive forward model of self-generated actions

Figure 8. A schematic experimental set-up of the present study

Figure 9. The rubber hands used for the RHI

Figure 10. The subject's view during RHI induction

Figure 11. The experimental setup during RHI induction in the “inside” PPS conditions.

Figure 12. The available measures of body ownership

Figure 13. A histogram of the total (referenced) PD value across all experimental conditions.

Figure 14. A histogram of the total subjective embodiment score across all experimental conditions.

Figure 15. A profile plot of type of touch and location of rubber hand.

Figure 16. A summary of the statistically significant main effects found in the present study.

Figure 17. A profile plot of type of touch and location of rubber hand during external touch.46 Figure 18. A profile plot of type of touch and location of rubber hand during self-touch

Table of Tables

Table 1. The factors, which were and the resulting eight conditions of the present study

Table 2. Results from the single sample t-test, testing which experimental PD values differ from the baseline PD mean

Table 3. Descriptive statistics regarding the PD for all experimental conditions and the baseline measurement

Table 4. The mean questionnaire rating scores of the various experimental conditions

1 Introduction

When touching or looking at our own body parts, we instantly have the feeling that they belong to us. We have the ability to perceive our own body as distinct from other entities and objects in the environment. Typically, we also have the feeling of being in control of our own bodily actions. These experiences are all essential parts of self-consciousness (Ehrsson, Holmes & Passingham, 2005) and fall under the term of “body ownership” (Gallagher, 2000).

While for most of the population, the feeling of body ownership seems to be only natural, this is not the case for everyone. A number of neurological disorders are characterized by a dysfunction in bodily experience, which may include feelings of disembodiment towards body parts, “filling in” of amputated limbs, or even a reduplication of certain body parts (Giummarra, Gibson, Georgiou-Karistianis & Bradshaw, 2008). An example is the case of body integrity dysphoria, a condition in which patients feel a strong sense of alienation from one or several body parts, which is often accompanied by a strong desire for amputation (Brugger & Lenggenhager, 2014). The “counterpart” of this clinical group would be that of amputees, who may suffer from phantom limb pain, or feelings of disownership towards their prosthesis (Bekrater-Bodmann, 2020).

Body ownership is assumed to be formed both through top-down and bottom-up experiences: it relies on sensory input from the environment (Tsakiris, Hesse, Boy, Haggard & Fink, 2007), but also on the coherence of that sensory input with mental representation of our bodily structure (Blanke, 2012). Despite the crucial role these bodily representations and experiences of body ownership have for our sense of self, neuropsychology and experimental psychology have proven that mental representation and conscious awareness of our own body are not cognitively impenetrable but can be manipulated (Botvinick & Cohen, 1998). Several experimental manipulations of body ownership -like the Rubber Hand Illusion- have shown that the formation of bodily self-representation occurs through the integration and correspondence of multisensory signals (Hara, Rognini, Higuchi & Pozeg, 2015). By systematically manipulating sensorimotor signals, promoting their integration, or generating conflicts and illusions, it is possible to study how the bodily self is constructed and maintained in the mind (Samad, Chung & Shams, 2015), which may help understand how body ownership, but also dysfunctions of body ownership operate.

A particular growing area of interest in processes of body ownership is self-touch. Self-touch is crucial in the development of the sense of self in infants (Rochat & Morgan, 1998) and seems to play an important role in body ownership, since it represents a unique situation: Two simultaneous tactile inputs are combined, due to the hand that is touching also being the hand that is touched. This added dimension of sensory input is the reason previous studies have hypothesized a positive modulation of body ownership through self-touch (e.g., Hara et al., 2015). However, the number of studies about this concrete topic is relatively scarce, and more experimental evidence is required in order to make definitive statements about how self-touch modulates embodiment and why.

To further investigate the role of self-touch in body ownership, the present study focuses on how self-versus externally induced touch modulates body ownership in healthy participants. The results of this study could give a deeper insight into the relevance of self­touch in terms of maintaining an intact sense of body and reveal mechanisms in which this occurs, in order to gain an overall larger understanding of body ownership.

2 Theoretical background

In the following chapter the concept of body ownership and its theoretical and practical relevance will be highlighted, before addressing the rubber hand illusions and the neuropsychological mechanisms behind it. Following that, the essence of this study - self touch - will be discussed, by investigating its role in the formation of a body representation and body ownership. Previous research will be reviewed before deriving the hypotheses regarding visual active self-touch's influence of body ownership processes, which are to be tested in this study.

2.1 Body ownership

The feeling that the body we inhabit is our own (body ownership) is a fundamental aspect of self-consciousness (Gallagher, 2000; Jeannerod, 2003): The relationship between ourselves and our bodies is unique and differs both from the relationship between our body and other people's bodies and the relation between our bodies and external objects (Tsakiris, 2011).

When we feel ownership towards our body, our bodily sensations seem unique to ourselves and we feel as though we are in control of our body (body agency). This special perceptual state of one's body is, under normal circumstances, ever present in one's mental life (Gallagher, 2000). Under circumstances, in which one's bodily self-consciousness is clinically altered, this is however not the case: Some patients for example experience an “absence” of a body part. This feeling can range from the inattention towards a certain body part (e.g. personal neglect) to the delusional belief of a body part not belonging to them (e.g. body integrity identity disorder). The common denominator of these disorders seems to be that the body (or a part of it) no longer belongs to one's body representation. Another relevant clinical field involving body ownership is that of amputees, prosthesis, and phantom limb (pain). Amputated patients, who chose to wear a prosthesis often do not embody their prosthesis, meaning they don't have the feeling of their prosthesis belonging to their body. This in turn can cause less prosthesis-satisfaction and possibly an increase in phantom limb pain (Bekrater-Bodmann, 2020). Therefore, the understanding of bodily self-consciousness and its mechanisms is highly relevant. Empirical research regarding self-consciousness has only recently started to investigate how the connection between a body and the experience of that body belonging to oneself is maintained, disrupted and what the mechanisms and modulating factors behind this phenomenon could be.

In neuropsychology, the focus of this research has been on the integration of multisensory body-related signals: Body ownership is thought to be based on and formed by integrating corresponding sensory signals from different modalities (visual, tactile, motor, proprioceptive etc.). When experimenting with the feeling of body ownership, one can use multisensory stimulation to alter the experience of the body by eliciting the feeling of body ownership in one condition, while diminishing it in another. The Rubber Hand Illusion (RHI) has been created as an experimental paradigm, allowing the controlled manipulation of the body ownership experience. In the RHI, observing a rubber hand being stroked at the same time as one's own unseen hand results in an observer attributing the rubber hand to their own body (Tsakiris, 2011).

2.2 The Rubber Hand Illusion

Botvinick and Cohen (1998) first discovered an illusion, in which participants felt ownership over a rubber hand. Participants were seated with their left arm resting on a table, while a standing screen was positioned next to the arm to hide it from the subjects' view. Meanwhile, a realistic looking, life-sized (left) rubber hand was placed in front of the participant. During the experiment, participants viewed the rubber hand (in an anatomically plausible condition) being stroked by a small paintbrush, while at the same time their real hidden hand was being stroked in precise synchrony (by a hidden paintbrush), resulting in a false attribution of the tactile sensations to the rubber hand. This meant that participants had the feeling the rubber hand in front of them could “sense” the touch, since they experience one united multisensory event (seeing and feeling the brush stroking the rubber hand), instead of two distinct unimodal events (feeling the one brush and seeing the other). Additionally, there was a displacement of the real (hidden) hand's felt location towards the visible dummy hand, meaning the subject felt their hand was closer to the rubber hand and in some cases even “inside” of it. Participants even reported feeling as though the rubber hand was their own hand. All these phenomena are known as the rubber hand illusion (RHI) (Botvinick & Cohen, 1998). Figure 1 depicts a schematic diagram of the classical RHI experimental set-up.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1. Schematic diagram of the experimental set-up in the rubber hand illusion task. (LH, left hand; part, partition; RH, right hand; Ru, rubber hand (Downey et al., 2014))

Post-experimentally, the induced illusion was measured implicitly, via a behavioral test and explicitly, using a subjective questionnaire. The latter included questions about the distinct perceptual effects caused by the manipulation, which were to be confirmed or denied (e.g. “I felt as if the rubber hand was my own hand) and are shown in figure 2. The resulting scores indicated that participants had mistakenly regarded the rubber hand as their own body part. They seemed to feel the touch of the viewed brush (which wasn't actually touching them), instead of the hidden brush (which was actually touching them). Essentially, participants experienced the tactile stimulation of their own hand from the proprioceptive location of the rubber hand.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2. RHI questionnaire results from the original study of Botvinick and Cohen (1998). The original questionnaire by Botvinick and Cohen (1998) included nine statements. Statements describing the predicted embodiment phenomena are underlined. Subjects indicated their response on a seven-step visual-analogue scale ranging from ‘agree strongly' (+++) to ‘disagree strongly' ( ). Points indicate mean responses and bars indicate response range. The questions underlined showed a statistically significant tendency to evoke affirmative responses (P<0.001) for underlined questions).

The implicit behavioral measure of the RHI, which is known as the proprioceptive drift (PD) and describes the felt displacement of the real hand towards the rubber hand, was measured by asking participants to blindly indicate (with the non-touch-receiving hand) where their real (touch receiving) hand is located. If the illusion has successfully been induced, the participants' indicated position will be displaced towards the rubber hand. In the original study of Botvinick and Cohen (1998) the mean displacement amounted to 23mm ( p < .04) (Botvinick & Cohen, 1998).

The PD and the subjective illusory rating scale have generally been used to measure the strength of an illusion during the RHI. It has been indicated that the PD distance and the agreement rating positively correlate (e.g., Ehrsson et al., 2005; Kalckert & Ehrsson, 2012). However, it is important to mention that the effect of subjective ownership and the effect of the PD may be caused my separate multisensory mechanisms. While subjective illusory ownership seems to rely on visuo-tactile integration, the PD seems to rely on visuo- proprioceptive integration. This dissociation implies that conclusions about ownership feelings cannot be drawn from one measure alone and that it is worth looking at both measures when performing the RHI (Rohde, Di Luca & Ernst, 2011).

Studies extending and replicating the RHI have found that not only the pattern of (matching) stimulation determine the induction of an embodiment illusion, but also the match between pre-exististing body representations and the rubber hand (Constantini & Haggard, 2007). Tsakiris and Haggard (2005) were able to demonstrate that rotating the rubber hand by 180° and 90° to the participant's hand diminishes the illusion. This proves that for illusory ownership to arise, the rubber hand must match the subject's body schema. Furthermore, the illusion was reduced when subjects looked at a wooden block, or the “wrong-sided” hand, as shown by Tsakiris and Haggard (2005), demonstrating that the to-be-embodied object must be visually similar to the actual body part being touched.

Over the years, further studies have confirmed and even extended Botvinick and Cohen's (1988) results of the RHI. Additional and alternative illusory measures were used, including verbal reporting of the felt hand position (Constantini & Haggard, 2007), somatosensory evoked potentials and EEG (Kanayama, Sato & Ohira, 2007), skin conductance responses that occurred when there was a “threat” to the rubber hand (Armel & Ramachandran, 2003) and fMRI (Ehrsson, Spence & Passingham., 2004).

2.2.1 Psychometrics of body ownership

Longo, Schüür, Kammers, Tsakiris and Haggard m(2008) took a psychometric approach to try and answer the question of what it is like to have a body. In their study they performed the classical RHI and asked subjects to fill out a 27-item questionnaire. A Principle Component Analysis (PCA) was used to extract various components involved in the feeling of body ownership. The results showed the subjective ownership experience consists of several distinct components: ownership (e.g., the rubber hand belonging to one's body); location (i.e. the rubber hand and the subject's own hand seemed to be in the same location) and agency (i.e. having control over the rubber hand). The items and their component loadings from the PCA are visualized in figure 3.

Abbildung in dieser Leseprobe nicht enthalten

Figure 3. Component loadings from the PCA on the embodiment of the rubber hand components found by Longo et al. (2008).

2.2.2 The neuropsychological mechanisms of the RHI

Their discovery of the RHI led Botvinick and Cohen (1998) to believe that the identification of the bodily self and the feeling of body ownership is formed through the spatiotemporal integration of visual, tactile and proprioceptive signals. They hypothesized that the illusion “involves a constraint-satisfaction process operating between vision, touch and proprioception - a process structured by the correlations normally holding among these modalities” (Botvinick & Cohen, 1998). In other words, the participants received tactile, visual, and proprioceptive inputs. These, along with their internal mental body representation served as the constraints and through an (unconscious) problem solving process, a solution was found, in order to integrate these inputs in a way that would make sense: i.e., embodying the rubber hand. In line with this the assumption was made that an embodiment of the rubber hand and the consequent linking of multisensory inputs would inevitably mean a distortion of the sense of position, which resulted in the behavioral illusory phenomenon of the proprioceptive drift (Botvinick & Cohen, 1998).

It has become clear that this process of multisensory integration relies on both bottom- up and top-down processes. As Botvinick and Cohen (1998) already demonstrated in their original RHI Version, the embodiment of the rubber hand depends on a matching tactile stimulation of the rubber hand and the subject's real hand. This was termed the “embodiment constraint” by Blanke, Slater & Serino (2015) and can be defined as a bottom-up process of intermodal matching (Constantini & Haggard, 2007). It describes that certain stimulation patterns that are implemented during ownership illusions can change multisensory integration of body signals. More concretely, the embodiment constraint is the way synchronous and prolonged stimulation impacts illusory ownership of a rubber hand, as visualized in figure 4d (Blanke et al., 2015). This constraint was already discovered in the original version of the RHI by Botvinick & Cohen. In their experiment they implemented a control condition, in which the rubber hand and the participant's hand were stroked asynchronously. It could be shown that this temporal asynchronicity prevented an illusion induction as synchronicity is necessary for the brain to successfully bind different multisensory signals onto the dummy hand (Botvinick & Cohen, 1998).

The incorporation of the rubber hand into the body however also depends on the pre­existing mental body representation, meaning the proprioceptive body configuration and the visual body image matching that of the rubber hand. Blanke et al. (2015) termed this the proprioceptive constraint and the body-related visual information constraint, respectively. Both can be defined as top-down matching processes since they rely on internal representations rather than sensory input (Constantini & Haggard, 2007). Regarding the former, studies have shown that proprioceptive signals have a large impact on the perception and integration of bodily signals. It was discovered that visual and tactile cues can be integrated more easily when they are presented on the same side of the body (e.g., visual cues are presented on the left side of the body's midline, while the left hand is being stimulated). However, when participants are asked to cross their arms, the opposite effect can be observed: A stronger cross-modal congruency is induced when visual cues are presented on the opposite side of the hand being stimulated (e.g., the visual cues are presented on the left side, while the left hand, which is on the right side, is stimulated). This shows that during multisensory signal integration a bodily re-referencing takes place, called the spatial remapping of touch, which is based on proprioceptive cues (Heed, Buchholz, Engel & Röder, 2015). In summary, the physical position of the hand as determined by proprioception, seems to play a big part in multisensory signal integration (Blanke et al., 2015). Multiple studies have shown that no (or a weaker) illusion is induced if the rubber hand is presented in a spatial position that is incompatible to the participant's real hand (Ehrsson et al., 2004; Constantini & Haggard, 2007). Figure 4a shows a case in which no illusory ownership is induced, due to the non­matching position of the fake hand to the real hand.

The constraint of body-related visual information describes the phenomenon that multisensory integration of bodily signals is reliant on visual information about the body and its parts in space. Studies have found that simultaneously occurring visual and tactile cues induced a larger cross-modal congruency when the visual cues were presented on dummy hands vs. on an object that does not resemble a hand (Pavani, Spence & Driver, 2000), which is visualized in figure 4b.

When inducing an ownership illusion, neither the correlation of multisensory signals, nor the matching to an internal body representation (proprioceptive or body-image related) are sufficient on their own. The multisensory stimulation and the body representation must be matched between the participant's hand and the rubber hand, resulting in a supra-additive interaction and a feeling of ownership (Tsakiris & Haggard, 2005).

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. The four constraints determining hand ownership. (A) proprioceptive constraint; (B) body-related visual information constraint; (C) PPS constraint; (D) embodiment constraint. Red and green body of test subject indicates an unsuccessful or successful induction of ownership over the rubber hand, respectively (Blanke et al., 2015).

2.2.2.1 Embodiment mechanisms within peripersonal Space

Apart from the three constraints mentioned in the previous chapter, there is one more constraint that seems to play a large role in the induction of embodiment, which Blanke et al. (2015) termed the peripersonal space (PPS) constraint. The PPS is defined as the immediate space surrounding a certain body part and is particulary relevant for ownership processes, since it is this space in which the multisensory integration, as well as the various top-down and bottom-up processes that lead to body ownership take place in (Blanke et al., 2015). A schematic representation of the different regions of space surrounding the human body is visualized in figure 5.

During the RHI, if a rubber hand is presented outside of the PPS of the subject's real hand, no illusion or a weaker illusion is induced (Lloyd, 2007), as also shown in Fig. 4c. However, when the various presented stimuli (e.g., the visual stimuli of watching the RH being stroked and the tactile stimuli of feeling one's own hand being stroked) are both presented within the PPS, a positive modulation of perception and integration occurs (Blanke et al., 2015), therefore enhancing illusion strength (Lloyd, 2007).

In order to understand why this is the case, the PPS and embodiment mechanisms within it will be looked at more closely:

The PPS is a specialized brain system, which acts as a boundary zone between one's body and the environment. It seems to contribute to the self-attribution of sensory signals (Makin, Holmes & Ehrsson, 2008), as RHI relies on the processing of sensory information from the PPS, making it an important spatial component for body ownership experience. In monkeys, the PPS is represented by multisensory neurons in the premotor area F4 and the ventral intraparietal area (VIP). These neurons respond to visual, tactile and auditory events, as long as they occur within the limited space around the body, which for the hands can extend from 5 to 35 cm (Fogassi et al., 1996).

The multisensory neurons representing the PPS seem to have tactile, visual and auditive receptive fields (RFs), that have a spatial correspondence. In other words, the RFs VIP-F4 neurons are anchored to various parts of the body and coded in somatic coordinates (Ferri, Chiarelli, Merla, Gallese & Constantini, 2013). These neurons show a selective response to visual, auditory and tactile stimuli only when they are presented near the relevant body part (i.e. in this body part' RF). When visual and tactile stimuli are presented simultaneously and within the same receptive field, neurons representing the PPS showed evidence of multisensory integration (i.e., responding in a non-linear way to the inputs). This suggests a mechanism of linking separate visual and tactile signals occurring within the PPS to a single multisensory experience, given the signals are presented synchronously and within the same receptive field (Makin et al., 2008).

This mechanism of multisensory integration within the PPS is what the RHI and body ownership in general rely on. When the various presented stimuli (e.g., the visual stimuli of watching the RH being stroked and the tactile stimuli of feeling one's own hand being stroked) are both presented within the PPS, as opposed to one or both of those stimuli being in the extra-personal space (the space beyond the PPS), a positive modulation of perception and integration occurs (Blanke et al., 2015), therefore enhancing illusion strength (Lloyd, 2007).

Ferri et al. (2013) made the discovery that a RHI can be induced without touching the rubber hand. In their experiment participants were asked to look at a RH, while an experimenter moved their hand towards it without touching it. It was found that the anticipation of touch is sufficient to elicit ownership over a fake hand. This seems to be a proactive process of rather than a reactive process and could be explained through predictive processing. This is a psychological or neural process, in which information from the past or present is used to form expectations about future states or events. The information used to form the expectations can be past experiences or innate priors, which both act as top-down modulators of bottom-up sensory input (Ferri et al., 2013). However, an illusion of ownership is only induced when the experimenter's hand entered the PPS of the rubber hand. This is due to human's ability to dynamically form a body representation in space (Makin et al., 2008). The advantage of this is that it allows an efficient mapping of what is really near, which either gives us the possibility of avoiding a threat or taking advantage of an opportunity. For this reason, the PPS should not be seen as a fixed region around the body, but rather as a barrier between the body and the environment (Ferri et al., 2013).

Abbildung in dieser Leseprobe nicht enthalten

2.3 Self-touch

One interesting factor, which has been suggested to contribute to body representations is self-touch (van Stralen, Zandvoort & Dijkerman, 2011), since it involves interaction of body ownership and body agency (Hara et al., 2015). Self-touch allows a basic form of self­awareness (Gallagher & Metzloff, 1996), helps to form a structural and conscious representation of the body (Dieguez, Mercier, Newby & Blanke, 2009), contributes to development (Rochat & Hespos, 1997) and self-other distinction (Boehme, Hauser, Gerling, Heilig, & Olausson, 2019; Hara et al., 2015). During self-touch, one's own body is simultaneously on the receiving and on the giving end of tactile stimulation. In other words, the body acts as a perceiver, while it is also the object of perception. Therefore, the two fundamentals of the experience of self - body ownership and body agency - are combined. Active visual self-touch (AVST), which is defined as a person observing themselves stimulating oneself, includes not only the integration of visual and motor signals, but also of two different simultaneous sources of tactile inputs (e.g., the hand that is touching and the hand being touched). The integration of these intermodal redundancies and temporospatial contingencies enables one to finetune and update the sense of body (Schütz-Bosbach, Musil, & Haggard, 2009) and is a unique way to distinguish one's own body from other objects in the environment (Hara et al., 2015).

[...]