Introduction
In recent years, there has been increasing interest in the experience of the own body. In various disciplines such as cognitive science, philosophy, and robotics, as well as in neuroscience, researchers started to investigate how we experience our own body, and which role this body has for our understanding of the conscious experience of the world and ourselves (see, for example, Churchland, 2002; Clark, 1999; Thompson & Varela, 2001; Wilson, 2002). Although body-related disorders have been documented in neurological and psychiatrical patients for more than a century (Corradi-Dell’Acqua & Rumiati, 2009; de Vignemont, 2010), approaching these questions from an experimental point of view remained a challenge.
A milestone in this regard was the so-called rubber hand illusion (RHI), introduced in the late 1990s (Botvinick & Cohen, 1998). In this illusion, participants perceive a model hand as part of their own body, as if it is their own hand. Surprisingly, this illusion is rather simple to induce: the participant has one hand hidden behind a screen and sees a life-like model hand in front of him/her. The experimenter then touches both hands (the model hand and the participant’s occluded hand) at the same time and at the same location. Often only after seconds, participants start to experience something astonishing: it feels as if the touch they feel on their hand originates from the model hand, and as if the model hand is part of the own body (see Figure 1.1).
Participants report this illusion typically when the hands are stimulated simultaneously and at the same location. The experience can be measured with questionnaires that contain statements reflecting the experience of perceiving the model hand as part of the own body (e.g. “It felt as if the rubber hand was my own hand”). Interestingly, when the model hand is threatened with a knife, a heightened skin conductance response (SCR) can be measured (Armel & Ramachandran, 2003; Petkova & Ehrsson, 2008). Another way to measure the illusion is the proprioceptive drift: here, participants estimate the felt position of the (real) hand, and during the illusion, a slight shift in position sense towards the model hand can be observed (Botvinick & Cohen, 1998; Tsakiris & Haggard, 2005). The shift is usually not more than 20%, but is often correlated with the actual strength of the illusion.
Figure 1.1 Illustration of the classical variant of rubber hand illusion.
This astonishing experiment sparked a whole new line of research investigating the cognitive and perceptual principles of the experience of our own body (Blanke, 2012; Ehrsson, 2012; Tsakiris, 2010). Since then, a number of variations have been introduced: for example, participants can experience three arms (Ehrsson, 2009; Guterstam, Petkova, & Ehrsson, 2011) or even an invisible arm (Guterstam, Gentile, & Ehrsson, 2013). The principles found in these experiments have been extended to the whole body (Petkova & Ehrsson, 2008; Petkova et al., 2011), to be inside a Barbie doll (Van Der Hoort, Guterstam, & Ehrsson, 2011), to induce out-of-body experiences (Ehrsson, 2007; Lenggenhager, Tadi, Metzinger, & Blanke, 2007), and many more.
What most of these studies have in common is typically one aspect: in the majority of these experiments, the participant is passive and receives the visual and tactile stimuli. So, the participant actually never moves, and is often explicitly instructed to remain completely still. Moving the hand while the rubber hand is immobile would actually interfere with the experiment, as the illusion would break in the very moment. In a way, this is a narrow and restricted view of the way we experience our body. As pointed out by Tsakiris and colleagues, it “lacks ecological validity, mainly because it does not involve bodily movement” (Tsakiris, Schütz-Bosbach, & Gallagher, 2007, p. 650). The experience of the body in a static way can be considered to be somewhat unnatural because, most of the time, we experience our body in motion. We constantly move, and especially our hands carry out an endless number of actions. The situation that we just process information of our body as a static and passive recipient is often not the case. Rather it is the case that our perceptions are used to guide our actions (Johansson & Flanagan, 2009). And when moving, we create new input: we get signals from our own body, which changes its positions and states, and also from the world around us, which flows by as we move around (Franklin & Wolpert, 2011). This will inevitably create some challenges for our perceptual apparatus. Following this line of thought, we can hardly separate our perceptual system from the motor system, and the motor system is in many ways implied in our perceptual system (Hommel, Müsseler, Aschersleben, & Prinz, 2001; O’Regan & Noë, 2001).
If we accept that our body is a moving body, then we need to understand also how we come to the experience of our body while moving. We need to understand how we come to the Sensation of Movement itself and how this sensation ties into the experience of the body. Investigating a static body will give us only parts of the story we need to understand.
How we experience the body: the sense of ownership and agency
Acting and interacting with the environment is a challenge in many ways. When we move, we face an overwhelming amount of sensory information from all our senses: some originate from the outside world (e.g. the object I am looking at) and some come from my own body (e.g. the touch I feel when I grasp the object). But how can our perceptual apparatus distinguish a sensation that refers to itself (my body), as opposed to something that originates in the external world (the object)?
Gallagher (2000) suggested that two distinct aspects contribute to the experience of the (bodily) self. The “sense of ownership” refers to the experience that “my body is moving regardless of whether the movement is voluntary or involuntary”. The “sense of agency” instead refers to the experience “that I am the one who is causing something to move” (Gallagher, 2000, p. 15). Both ownership and agency are present in every voluntary movement, and therefore it is hard to distinguish between them. However, a simple way to disentangle them is passive movements: when someone takes my arm and moves it, I do not experience a sense of agency over this movement, as I do not actively cause it. Still, I perceive this arm to be my own arm. Thus, we can subtract the experience of agency, and we are left with an experience of ownership of my arm. Both these mechanisms may be used independently in self-recognition, but feed into the processes enabling us to identify a perceptual stimulus as self-produced or self-related, as opposed to something from the external world.
Ownership and agency are certainly not the only contributors to the experience of the body. Particularly, interoceptive cues play an important role in the way we perceive our body (Craig, 2002; Tsakiris, Tajadura-Jiménez, & Costantini, 2011). However, the role of interoception is beyond the scope of this text.
Probing the sense of ownership: the rubber hand illusion
The rubber hand illusion (RHI) is an elegant experiment to manipulate the sense of ownership (SoO) that is the experience that it is my body. An object (the model hand) that is clearly not part of the own body becomes part of the own body. The rubber hand illusion created a wealth of insights into the principles of this experience of ownership (Ehrsson, 2012; Tsakiris, 2010). A core principle of this illusion is temporal congruency, thus stimulation of the two hands must occur simultaneously. If stimulation occurs asynchronously, i.e. the stimulation to the hand is delayed, then the illusion typically does not occur (Botvinick & Cohen, 1998; Tsakiris & Haggard, 2005). Also, a spatial mismatch by positioning the model hand either too far away from the real hand (Lloyd, 2007; Preston, 2013) or aligning it in an anatomically incongruent position will substantially diminish the illusion (Ehrsson, Spence, & Passingham, 2004; Tsakiris & Haggard, 2005). The illusion will also not occur when the object in question is not shaped like a human hand (or body, in the case of full-body illusions) (Petkova & Ehrsson, 2008; Tsakiris, Carpenter, James, & Fotopoulou, 2010). The latter shows that not only bottom-up processes determine the illusion, but also top-down factors, such as the knowledge of the shape of the human body. Interestingly, certain features do not seem to play a significant role, such as skin color of the rubber hand or even the gender of the body (Petkova & Ehrsson, 2008). Relatively early, however, it has been questioned whether the illusion is strictly dependent on visual and tactile stimulation alone. Ehrsson, Holmes, and Passingham (2005) introduced a variation of the RHI, in which participants had their eyes closed and instead touched the model hand with their own finger while they felt a touch at the same time on their own hand. Thus, a similar experience can be induced by using sensory channels other than vision and touch.
From these observations, we can conclude, however, that the integration of visual and somatosensory information (but probably other modalities too, and constrained by certain top-down factors) underlies the SoO over our body (Ehrsson, 2012). This mechanism allows us to differentiate between my hand or another person’s hand, or any other object in the world.
Probing the sense of agency: mismatch detection paradigms
The sense of agency (SoA) is unfortunately less straightforward to explain, not only because theories of agency increased in recent years and led to new models, but also because the SoA is inherently multi-facetted (David, Newen, & Vogeley, 2008; Gallagher, 2007; Haggard, 2008; Moore & Obhi, 2012; Synofzik, Vosgerau, & Newen, 2008a), For example, Balconi lists the “awareness of a goal, of an intention to act, and initiation of action, as well as awareness of movements, sense of activity, sense of mental efforts, sense of control, and the concept of authorship” as putative elements of the SoA (Balconi, 2010, p. V). Given this heterogeneity, it is not surprising that experiments investigating the experience of agency are diverse. Typically, the SoA is tested with paradigms, in which participants execute an action and receive feedback (e.g. a button press followed by a tone). This feedback can be manipulated by introducing a temporal delay or spatial distortion, or any other feature of the action outcome. Some studies use a button press followed by a beep tone (Sato & Yasuda, 2005), control of a cursor (Farrer, Bouchereau, Jeannerod, & Franck, 2008), recorded hand movements (Farrer, Frey et al., 2008), computer-animated hand movements (Nahab et al., 2011), or control of an avatar with a joystick (Yomogida et al., 2010). In sum, these studies have shown that a temporal delay of approximately more than 150ms or a spatial distortion of around 15° results in an altered experience of the action (Jeannerod, 2003; see also the chapter by Christensen and Grünbaum, Present volume). Typically, these observations have been described in terms of a motor control framework (Frith, 2005; Frith, Blakemore, & Wolpert, 2000b). According to this account, the motor system detects discrepancies between planned actions and sensory feedback by means of an efference copy (Bays & Wolpert, 2006; Crapse & Sommer, 2008; Holst & Mittelstaedt, 1950). In recent years, this account has been criticised, as it may not be sufficient to explain the variety of observations in agency studies (Synofzik et al., 2008a). For example, the “helping hands” experiments by Wegner and colleagues have shown that we can experience agency for events clearly not caused by us (Wegner, 2002). In these experiments, participants were cued with an instruction such as “clap your hands”, and without performing the action, they looked at a pair of arms that performed this action. When there was a match between the cue and the action they saw, participants reported an increased experience of agency (Wegner, Sparrow, & Winerman, 2004). Therefore, the SoA could be best understood as a multifactorial process, drawing on a variety of cues in which the sensorimotor part (i.e. movement) represents only one likely important cue.
Probing ownership and agency: moving a rubber hand
When looking at the literature of both SoO and SoA, it is apparent that they show only little overlap. In fact, most studies do not directly address the issue of the other. Thus, RHI studies only rarely discuss the role of agency, and agency experiments often do not address the dimension of ownership. Although we understand a great deal about both these aspects in isolation, we understand only little about how they are experienced together (Gallagher, 2000; Synofzik, Vos...
