Literature Thesis
Mirror-touch “synesthesia”
The Tactile Mirror System account
Date: July 19, 2015
Name: Chantal Keijlard
Student ID: 5960428 Supervisor: Dr. O. Colizoli Co-assessor: Dr. R. Rouw
MSc in Brain and Cognitive Sciences, University of Amsterdam Cognitive Neuroscience
Introduction
The underlying neural mechanisms of synesthesia, also referred to as a “crossing of the senses” (Simner et al., 2006), have been studied extensively over the past decade (Ward, 2013). Synesthesia is customarily defined as a singular experience in which one modality of perception (i.e. the concurrent) is triggered by stimulation in the same or another perceptual modality (i.e. the inducer; Grossenbacher & Lovelace, 2001). Since the first reports from over two centuries ago, many different types of synesthesia have been identified (see Ward, 2013). Although it is not the most common form of synesthesia (see Ward, 2013), the most studied form to date is grapheme-color synesthesia (GCS; Simner et al., 2006; Rothen & Meier, 2013). This form is characterized by a color experience when looking at a grapheme (i.e. a letter, a number, a symbol; see Colizoli et al., 2014). Examples of other types of synesthesia include lexical-gustatory synesthesia (taste sensations induced by certain sounds; e.g., Colizoli et al., 2013), visual-pain synesthesia (experience of photisms induced by pain; Dudycha, G.J. & Dudycha, M.M., 1935), and spatial-sequence synesthesia (sequences which are perceived in visual space, SSS; e.g., Eagleman, 2009). To date, no consensus exists on the definition of synesthesia (e.g. Simner, 2012). Most diagnostic tests are focused on determining the degree of consistency regarding the synesthetic experience (Eagleman et al., 2007; Novich et al., 2011). However, the consideration of consistency as a defining characteristic of synesthesia is a topic of debate (Simner, 2012; Eagleman, 2012; Cohen Kadosh & Terhune, 2012).
A ‘new’ type of synesthesia has been suggested by Blakemore et al. (2005); namely mirror-touch synesthesia (MTS). This type of synesthesia is characterized by a vicarious tactile sensation upon observing touch to another person. An extensive investigation of the prevalence of MTS revealed a prevalence of 1.6% of the general population, deeming it the most common form of synesthesia discovered to date (Banissy et al., 2009). Within MTS, two subtypes can be distinguished. In the first subtype, the tactile sensations are experienced with an ‘anatomical frame of reference’, which means that the tactile sensation in the observer’s body is experienced at the anatomically corresponding site of the observed touch. In the second subtype, the observer holds a ‘specular frame of reference’, meaning that the experienced touch is on the mirrored site of the observed touch, as if the observer were look into a mirror (Banissy & Ward, 2007). Mirror touch synesthesia is a much-debated type of synesthesia (e.g. Banissy & Ward, 2007; Rothen & Meier, 2013). In fact, it has been argued that it is not a form of synesthesia, but rather a activation in a common neural network (Rothen & Meier, 2013). Others have argued that hyper-activation in such a system may be mediated by structural as well as by functional mechanisms that were previously proposed to underlie validated forms of synesthesia (for a review see Bargary & Mitchell, 2008), thus deeming MTS a legitimate form of synesthesia (Fitzgibbon et al., 2012).
Most reported cases of MTS are developmental, although acquired forms have also been described (e.g. after limb amputation; Goller et al., 2013). In amputees, MTS may be related to phantom limb sensations, which seem to be caused by a cortical reorganization following the loss of input from the affected limb(s) (Flor et al., 1998). Since MTS appears to be more prevalent among non-amputees, it was suggested that it might rely on common brain networks rather than abnormal crosstalk between regions (Blakemore et al., 2005). If this is the case, the question remains why MTS does not occur in the majority of the population. Several researchers have suggested that MTS may be a cause of a disinhibition within the sensorimotor equivalent of the
mirror neuron system (the tactile mirror system; Keysers et al., 2004; Blakemore et al., 2005; Bolognini et al., 2013), leading to hyperactivity in sensorimotor areas.
The aim of this paper is to examine whether MTS should really be considered a form of synesthesia. In order to successfully resolve this issue, it is important to come to a general agreement upon a solid definition of synesthesia itself. In this paper it will be evaluated whether MTS corresponds to the current defining characteristics of synesthesia. Furthermore, it will be argued that MTS is a product of a disinhibited tactile mirror system driving hyperactivity in somatosensory areas, which are implicated in both the observation and experience of tactile stimulation. Synesthesia, on the other hand, seems to be the cause of abnormal crosstalk between the involved brain regions. Thus, it is argued that MTS is not a form of synesthesia and should be redefined. Since it is unclear whether it is the sight of touch or the concept of touch that elicits the tactile experience, it may be more fitting to refer to mirror-touch experiences as vicarious experiences. A vicarious experience is an experience caused by empathy for, or simulation of, an observed experience of another person (e.g. Keysers & Gazzola, 2009). Hence, MTS will be referred to as VTE (Vicarious Tactile Experience) hereafter. The importance of separating VTE from synesthesia lies in our understanding of vicarious experiences, synesthesia, and by extension perception itself. Furthermore, by providing a broader understanding of the neural mechanisms of VTE, we may shed additional light on the neural underpinnings of similar phenomena such as phantom limb sensations. Building on evidence from amputees and non-amputees with and without VTE, an attempt will be made to provide more insight into the neural mechanism(s) underlying this intriguing condition. Finally, it is argued that higher empathic ability may mediate hyperactivity in the tactile mirror system, giving rise to VTE.
1. Does VTE conform to the defining characteristics of synesthesia?
Although its definition is debatable, there is a general consensus in the literature on the following criteria that a condition must meet in order to bear the name ‘synesthesia’ (Colizoli et al. 2014; Rothen & Meier, 2014; Deroy & Spence, 2014; Bor et al., 2014). It has been established that synesthesia is minimally defined as a collection of arbitrary, consciously perceived perceptual or percept-like experiences, which are triggered automatically upon being presented with an inducer (Ward & Mattingley, 2006; Banissy & Ward, 2007). Synesthesia has a substantial bandwidth within individuals, allowing for idiosyncracy across individuals (Asher et al., 2006; Rothen & Meier, 2013). In the case of GCS for example, no two synesthetes’ inducer-concurrent pairings are exactly the same (Day, 2005; Meier & Rothen, 2009). Bandwidth and idiosynchracy allow for the use of consistency tests as ‘measures of genuineness’, due to the fact that they complicate faking behaviors in non-synesthetes (e.g. Asher et al., 2006). Another characteristic of synesthesia is that, unlike experiences in the real world, synesthetic experiences are not perceptually present, allowing synesthetes to easily distinguish between the synesthetic percept and the real world (Banissy et al., 2009; Seth, 2014). For instance, grapheme color synesthetes show no difficulty distinguishing between a synesthetic color of a grapheme and the physical color of that grapheme (e.g. Ward, 2013). In this section, it is argued that VTE does not conform to several defining characteristics of synesthesia. Firstly, VTE has insufficient bandwidth, rendering the criteria idiosyncracy and consistency trivial. Second, the concurrent in VTE is perceptually present. Finally, synesthetes report better mental imagery than non-synesthetes do (Barnett & Newell, 2008), whereas VTE individuals report higher empathy than non-VTE individuals do (e.g.
Banissy & Ward, 2007). Thus, it is argued that there are major qualitative and quantitative differences between VTE and synesthesia. For a summary of the comparison between the discussed characteristics of GCS and VTE, see table 1.
1. 1. Consciously perceived perception or percept-like experience
When thinking of synesthetic experiences as ‘percept-like experiences not perceived by others’, a question that arises immediately is: How does a synesthetic experience deviate from a hallucination? Much like the synesthetic concurrent, a hallucination is defined as an anomalous percept-like experience with the same impact as actual perception (perceptual presence), which is not in the observer’s direct control (Cytowic, 2002). Another shared feature between the two perceptions is that both hallucinations and synesthetic concurrents alike can involve any combination of the senses (Ohayon, 2000; Ward & Mattingley, 2006). While it has been argued that the synesthetic inducers and concurrents may also involve higher order constructs (e.g. a notion of a personality type for ordered linguistic sequences; Simner & holenstein, 2007), the authenticity of such types of synesthesia is a topic of debate (see Hochel & Milán, 2008). An essential difference between hallucinations and synesthetic experiences is, that unlike a synesthetic concurrent, a hallucination occurs in the absence of an eliciting stimulus (see Marks & Mulvenna, 2013). According to Cytowic (2002), a hallucination is more complex, irregular, and more unpredictable than a synesthetic experience. Ward (2013) described hallucinations as concurrents without an inducer, while Sollberger (2013) described synesthetic experiences as “reliable and useful hallucinations.” The reliability of such experiences is evidenced by the consistency of inducer-concurrent pairings. These pairings are speculated to be biologically useful through their multisensory nature, recruiting multiple sensory modalities where others would only recruit one. It has often been suggested that synesthesia may bring about cognitive enhancements in an individual. However, findings regarding memory and visual search task performance in synesthetes compared to non-synesthetes are inconsistent and thus are a topic of debate (for a review on the topic, see Rothen et al., 2012; also see Meier & Rothen, 2013). Although the sensory modalities involved in synesthesia do not give rise to any abnormal experiences (see Sagiv & Robbertsen, 2005), the relationship between these modalities in synesthesia is arbitrary. In the case of GCS, for instance, a grapheme consistently elicits a specific color (e.g. Colizoli et al., 2012). Although neither the perception of a grapheme nor the perception of a color is strange, a grapheme does not typically evoke a random color experience when there is no actual color present. Moreover, a hallucination is perceptually present; meaning it cannot (generally) be differentiated from perceptions existing in the external world, while synesthetic experiences co-exist with and show a clear distinction from such perceptually present experiences (Cytowic, 2002; Colizoli et al., 2014; Seth, 2014). Hence, in consideration of the aforementioned delineations it may be inferred that a synesthetic experience, while similar to a hallucination (Grossenbacher & Lovelace, 2001), is a separate phenomenon. In the case of VTE, however, the concurrent is perceptually present (see section 1.2.), meaning that it is just as vivid as a hallucination. Another issue regarding VTE is that it is unclear whether in the inducer is the sight of touch, or merely the concept of touch. If the latter is the case, this means that VTE is not a vision to touch inducer-concurrent pair, but rather an empathic association between a concept and the tactile modality. While it is clear that VTE is anomalous and consciously perceived, it is less clear whether it gives rise to perceptions or percept-like experiences. However, unlike the synesthetic concurrent of GC-synesthetes, during the Stroop task, VTE
individual have difficulty differentiating between the vicarious touch and the actual touch for the reason that the vicarious touch is perceptually present (as was mentioned in section 1.1) and thus is experientially identical to actual touch in the case of VTE. This difficulty is denoted by the occurrence of the so-called mirror touch errors found in the VTE individuals in the experiment of Banissy & Ward (2007). The perceptual absence of the synesthetic concurrent in GCS is demonstrated by the low error rates of GC-synesthetes on the synesthetic Stroop test. However, the longer RTs indicate that there is some interference caused by the synesthetic Stroop test. Following this line of reasoning, the above findings indicate that there is automaticity regarding perceptual information processing in both VTE and GCS.
1.2. Automaticity of the synesthetic/vicarious concurrent
The automatic nature of synesthesia refers to the fact that the experiences cannot be controlled at will, to a point that they interfere with task demands. Processes showing strong automaticity are not positively or negatively influenced by selective attention directed respectively toward or away from a stimulus. However, the degree of automaticity determines the extent to which selective attention can influence perceptual information processing (see Yantis & Jonides, 1990). A paradigm that is often used as a measure of partially automatic processes is the Stroop task. This task aims to examine the degree to which an individual can inhibit a prepotent response through the recruitment of selective attention. In synesthesia research, Stroop-like tasks are often used to quantify the degree of automaticity in this condition as well (e.g., Ward & Mattingley, 2006; Banissy & Ward, 2007; Banissy et al., 2009; Colizoli et al., 2012). For instance, the GCS version of the Stroop test (e.g. Mills et al., 1999) is comprised of visually presented graphemes, which are either presented in colors congruent or incongruent with the synesthetic colors of the participant. The instructions given to the participants are to ignore the synesthetic color of the grapheme and to report the color in which the grapheme is presented as fast and as accurately as possible. The Stroop effect denotes the differences in reaction times (RTs) and error rates between congruent and incongruent trials. Typically, grapheme color (GC) synesthetes show low error rates on this task, while still exhibiting increased RTs during incongruent compared to congruent trials (see Rouw et al., 2013). This difference inplies that the synesthetic concurrent is perceptually absent, while still interfering with task demands. Despite its wide use in determining the automaticity of synesthetic concurrents, multiple researchers have noted that such Stroop effects are insufficient in determining the authenticity of synesthesia types seeing as how they are found in both synesthetes and non-synesthetes (Colizoli et al., 2012; Meier & Rothen, 2009; Rothen & Meier, 2013). Moreover, one could argue that overlearned associations may be responsible for the occurrence of this Stroop effect in combination with the subjective reports of actual sensations in synesthetes (Robertson & Sagiv, 2005). It is unclear whether the interference occurs at the semantic level, the perceptual level or both. Multiple researchers have suggested that automaticity should not be a defining characteristic of synesthesia altogether (Price & Mattingley, 2013), based on the premise that the synesthetic concurrents seem to be under the strong influence of selective attention (Mattingley, 2009). If synesthetic experiences truly occurred automatically one would expect that they would also occur when the synesthete is not consciously aware of the inducer (see Mattingley, 2009). Rich and Mattingley (2003) examined this, using synesthetic variants of Navon-like stimuli (Navon, 1977). Participants were to identify the displayed color of the letters, ignoring the synesthetic color that the letters induced. The congruence effect denoted the differences in RTs between congruent (displayed colors and synesthetic colors were the same)
and incongruent (displayed colors and synesthetic colors were different) trials. If the synesthetic color is elicited automatically, there should be a large congruency effect for both global and local features regardless of where attention was allocated. However, the researchers proposed that if the congruency effect was reduced for the unattended feature, synesthetic experiences might rely on selective attention to the inducer rather than being a fully automatic response. Participants were slower to identify the displayed color of the letters in the incongruent condition versus the congruent condition. This congruency effect, while still significant, was reduced in the mixed congruent/incongruent conditions. Thus, the researchers concluded that while the synesthetic colors were elicited automatically upon presentation of an inducer, they were indeed partly modulated by selective attention.
When considering VTE, the “concurrent” is perceived at a specific location on the body, prompted by observed touch to a corresponding location on someone else’s body. Since two distinct sensory modalities (i.e. vision and touch) may be involved in VTE, a cross-modal congruency test may be of use when assessing the automaticity of this form of synesthesia. Such a paradigm aims to elucidate underlying mechanisms of failures in cross-modal selective attention. During this test participants attempt to focus attention on one sensory modality while ignoring input from another sensory modality (Driver & Spence, 1998; Martino & Marks, 2000; see Spence et al., 2004). A somewhat similar test that may be able to test for automaticity specifically in individuals with VTE is the visual-tactile spatial congruity task developed by Banissy & Ward (2007). This test was originally intended for the authentication of VTE. For the purpose of examining whether mirror touch synesthetes showed impairments in the ability to discriminate between simultaneously occurring synesthetic touch and actual tactile stimulation, the researchers designed the visual-tactile spatial congruity task. Two experiments were carried out. In the first experiment participants watched videos consisting of a person being touched on either the left or right cheek and in the second experiment participants watched videos consisting of either a left or right hand being touched. Congruency was determined on the basis of the frame of reference of the participant (i.e. anatomical versus specular). Participants were instructed to ignore the synesthetic experience elicited by the touch video and to report the site of the actual touch (i.e. on the left side, on the right side, or both). Higher error rates and longer response times were expected for the incongruent condition compared to the congruent condition. Because actual touch was administered to one side of the face/hand only, reports of touch on both sides of the face/hand were considered the error of interest (i.e. the mirror-touch error). Mirror-touch error rates were higher in individuals with VTE (VTE individuals) compared to non-experiencing controls (non-VTE individuals) for both the face and the hand experiment. Furthermore, VTE individuals were significantly slower to identify the site of touch when it was incongruent to the observed touch. These results demonstrate that the vicarious tactile experiences in VTE individuals are partially automatic.
1.3. Bandwidth
Synesthesia requires a substantial bandwidth to allow for idiosynchratic inducer-concurrent pairings across individuals. For instance, in GCS, inducer-concurrent pairings are arbitrary, meaning that no two individuals’ pairings are exactly the same (e.g. Day, 2005). However, while individual synesthetic mappings are difficult to predict, they are often fairly straightforward (Sagiv & Ward, 2006). For instance, lower pitch sounds are frequently associated with darker colors while higher pitch sounds are repeatedly associated with lighter colors in both synesthetes and
non-synesthetes (e.g. Marks, 1974; Melara, 1989). However, in sound-color synesthetes, the specific inducer-concurrent pairings show a much greater degree of consistency compared to the pairings of non-synesthetes (Ward et al., 2006). While several researchers have shown effects related to training grapheme-color associations in non-synesthetes (Meier & Rothen, 2009; Colizoli et al., 2012), the validity of trained forms of synesthesia has been questioned (Deroy & Spence; 2013) since the trainees generally lack the subjective color experience (Rothen & Meier, 2014; Colizoli et al., 2014; but see Bor et al., 2014). This is also the case for VTE paradigms; where somatosensory brain activity in non-synesthetes is evoked by the observation of touch to another, while the conscious VTE remains absent in these individuals (e.g. Blakemore et al., 2005). However, Bolognini et al. (2014) successfully elicited VTE in non-VTE individuals by lowering the threshold of perceptual awareness during the observation of touch to another person (see section 2.2). This finding points to a functional, rather than a structural, brain difference between individuals with VTE and non-VTE individuals and is thus in line with previous notions that VTE may rely on a common neural network; a topic which will be discussed in section 3.
The bandwidth in VTE is very narrow. There is only one “inducer” (the observation of touch to another person) and one “concurrent” (the experience of touch at a corresponding site on one’s own body). The inducer-concurrent relationship in VTE is quite straightforward seeing as how empathic processes may already prompt one to imagine the experience of touch to others when presented with the visual input of others being touched (Sagiv & Ward, 2006; Rothen & Meier, 2013). Furthermore, if VTE were to be considered a form of synesthesia, it would be the only form that is restricted to only one inducer (Rothen & Meier, 2013), meaning that idiosyncracy, which is a defining characteristic of synesthesia, is trivial in the case of VTE.
1.4. Consistency of inducer-concurrent pairings
The consistency over time of the reported inducer-concurrent experiences has been used widely as the ‘test of genuineness’ for synesthesia (see Baron Cohen et al., 1987; Asher et al., 2006). This is because it is difficult to simulate the remarkable consistency that synesthetes show long after the first measurement (e.g. Baron-Cohen et al., 1987). As was mentioned before, the arbitrary nature as well as the bandwidth of synesthetic associations allows for the validity of consistency tests. If all synesthetes had the exact same synesthetic associations, there would be no reason to test for consistency. Moreover, if the associations were idiosyncratic across individuals but there were very little possible inducer-concurrent pairings it would be easy for non-synesthetes to simulate synesthetic consistency, rendering it trivial as a measure of authenticity (see Colizoli et al., 2014). However, Simner (2012) argued that, while synesthetes seem to show consistent inducer-concurrent pairings, there might be subgroups that lack such consistency while still conforming to all other defining characteristics of synesthesia. Furthermore, since synesthetes were found to be 80-100% consistent in the reports of their inducer-concurrent pairings as opposed to 30-50% consistency in non-synesthetes for learned associations (e.g. Asher et al., 2006), it is not unthinkable that there is a spectrum of consistency. In such a spectrum, lower scores might denote weaker forms of synesthesia, meaning that an individual with a 65% consistency could still be considered synesthetic. These individuals would, under the current conditions, not be chosen to participate in most synesthesia studies, even though they clearly show a higher degree of consistency than a non-synesthetic control would. For instance, it was suggested that the consistency shown in most synesthetes might very well be biased simply due to
the fact that most researchers tend to select their synesthetic participants on the basis of a consistency score above 80% (Simner et al., 2012).
Perhaps the consistency does not apply to the inducer-concurrent pairings per se, but rather refers to the consistency with which any inducer can elicit any concurrent. As was mentioned before, when considering VTE, the question of whether or not the inducer-concurrent pairings are consistent is trivial due to the narrow bandwidth of this condition (Rothen & Meier, 2013). While other more established forms, such as GCS, have multiple inducers (e.g. the letters of the alphabet) eliciting multiple concurrents (e.g. a specific color/hue for each grapheme), VTE only has one possible inducer (i.e. observation of touch to another) and one possible concurrent (i.e. tactile experience on the corresponding site of the observed touch). This means that consistency in VTE is uninformative, seeing as how VTE is not arbitrary.
Table 1: Comparison between the characteristics of vicarious tactile experiences and grapheme color synesthesia.
VTE GCS
Dimension Amputee Non-amputee
Consciousa √ √ √ Perceptualb ? ? √ Bandwidth/a × × √ Idiosyncratic Automatica ? ? √ Consistenta ? ? √ Perceptuallyb ? √ × present Acquirable/b √ ? ? Trainable Relies onb ? √ × common neural networks Relies onb ? × √ different neural networks Increasedb √ √ ? reported empathy Increasedb ? ? √ reported imagery Concurrentb × × √ sense absent
a Defining characteristic of synesthesia
2. Neural Mechanisms of VTE
Earlier research indicated that the mere observation of touch to another person elicits overlapping neural activation with the actual experience of touch to the self (Keysers et al., 2004; Blakemore et al., 2005). Interestingly, this functional overlap is found in healthy synesthetes and non-synesthetes alike. Evidently, the conscious tactile experience during the observation of touch to another person, occurring in the absence of any deafferentation caused by either brain injury or amputation, remains unique to the individuals with VTE (Halligan et al., 1996; Goller et al., 2013; Holle et al., 2013). It has previously been suggested that VTE relies on hyperactivity within common neural networks (Blakemore et al., 2005). However, with an estimated prevalence as low as 1.6% of the general population (Banissy et al., 2009), one might wonder which additional factors may play a role in the development of this condition. It has been suggested on multiple occasions that VTE may be caused by a disinhibition in the somatosensory mirror neuron system (also the tactile mirror system; Keysers et al., 2004; Blakemore et al., 2005; Keysers & Gazzola, 2009). This disinhibition may be responsible for activity in somatosensory regions to exceed a supposed threshold of conscious tactile experience (Blakemore et al., 2005).
2. 1. The Mirror Neuron System
Mirror neurons were first documented in monkey area F5 (i.e. premotor cortex). These neurons specifically respond during the execution of an object-directed action or during the observation of another biological agent carrying out such an action (Rizzolatti et al., 1996). Neurons with similar properties, which respond to actions performed by others, were also found in monkey superior temporal sulcus (STS). It was previously hypothesized that the functional role of the MNS is to mediate action understanding (Rizzolatti et al., 2001; Rizzolatti & Craighero, 2004). Umiltà et al. (2001) found evidence for this hypothesis by presenting monkeys with partially hidden object-directed actions while measuring mirror neuron activation. They positioned the monkey opposite an experimenter with a frame placed in their midst. The monkey watched the experimenter through the frame. Attached to the frame was a plane on which objects could be placed. The experiment consisted of three conditions: 1) a full vision condition during which the experimenter touched, grasped, manipulated, held, released, or placed an object on the plane; 2) a full vision miming condition where the experimenter executed the same actions as in condition one with the exception that there was no object involved; 3) a hidden miming condition in which the experimenter carried out the same actions as in condition 2, only this time the hand-object interaction (the crucial moment) was hidden behind an opaque screen. A subset of mirror neurons was activated upon the observation of object-directed actions, even during the hidden miming condition. These neurons were not activated during the full vision miming condition, suggesting that the activity was not triggered by the visible initial part of the action during the hidden miming condition. Moreover, meaningless actions did not seem to activate the MNS altogether. These findings indicated that the monkey was able to infer the crucial, hidden information of meaningful actions in the absence of visual input. The researchers concluded that these results provided further evidence for the MNS hypothesis of action understanding.
To date, no single neuron recordings in supposed mirror-neuron areas of the human brain have been conducted. Therefore, current evidence of a MNS in humans is indirect (Rizzolatti & Craighero, 2004). Nevertheless, the neurophysiological and neuroimaging data collected thus far provide a strong argument for a MNS in humans (see Rizzolatti 2005). For
instance, in a quest to provide evidence for the human MNS, Fadiga et al. (1995) stimulated the motor cortex of participants using transcranial magnetic stimulation (TMS) during four different observation conditions. Motor evoked potentials were recorded from hand muscles. During the first observation condition, participants watched an experimenter grasp an object (transitive movement). The second observation task required participants to observe the object that was used in the first condition (object without movement). During the third observation condition, participants were to look at the experimenter who traced a geometric shape in the air with an extended arm (intransitive movement). Finally, the fourth observation condition required participants to detect and report as quickly as possible the dimming of a light stimulus (dimming detection). Excitability of the motor cortex was increased during the action observations compared to the object observation and the dimming observation. The pattern of muscle activity was similar to muscle activity during action execution. These results indeed provided support for the notion that an MNS exists in humans. Interestingly, ‘human mirror neurons’ seem activated both for transitive (object-based) and intransitive (meaningless) movements, whereas monkey mirror neurons only respond to movements with a clear objective, suggesting that the supposed human MNS might be more encompassing compared to its non-human primate counterpart (Rizzolatti et al., 2004). Further support for this notion came from Rizzolatti et al. (2005) who found structural differences between the human and monkey frontal lobe, with the human frontal lobe showing a large expansion comparatively. Since the prefrontal lobe has previously been implicated in imitation learning, the researchers suggested that this region might also be involved in the construction of complex motor patterns from basic motor actions in humans specifically.
Further research into the ‘human MNS’ led to the suggestion that mirror neurons do not exclusively reside in premotor cortex. Keysers & Gazzola (2009) theorized that networks similar to the mirror neuron network for actions might be involved during the observation of emotional faces and sensations of others as well. In line with their theory, the existence of a sensorimotor equivalent of the MNS was proposed (the tactile mirror system), which activates when an individual is being touched as well as during the mere observation of touch to another. Using functional magnetic resonance imaging (fMRI), Keysers et al. (2004) investigated whether movies depicting touch to the leg would automatically activate the sensorimotor cortex of an observer. There were two tactile conditions and two visual conditions. During the tactile conditions, the experimenter stroked one leg of the participants with a washing glove while they lay in the scanner with their eyes closed. During the visual conditions, the participants were instructed to watch short clips in which a person was shown laying on a medical examination bed while their leg was being touched by one of three objects (a wooden rod, a metal rod, or a brush). Additionally there were visual control blocks in which participants watched similar clips with the exception that the object only approached the leg and touch did not actually occur. The researchers found that the SI (primary somatosensory cortex) representation of the legs was located in the contralateral dorsal aspect of the postcentral gyrus, whereas the SII (secondary somatosensory cortex) representation of the leg was located in bilateral frontoparietal operculum, with a lateral extension onto the convexity of the inferior parietal lobule. Although there was an increase of activation in SI during presentation of the visual touch clips compared to the visual control clips, this difference did not reach significance. However, activation of SII was increased during presentation of the visual touch clips compared to the visual control clips. The activation shown during the observation of touch to the leg was similar to the activation during the tactile
conditions. This was also the case when the participants observed two objects touching. Thus, the researchers suggested that personal experience of touch and observed touch to others or to objects is mediated by a shared circuitry in the brain, which allows an implicit understanding of touch. Although these results are in line with an MNS account of tactile experience, it remains unclear why this circuit leads to overt tactile experiences during the observation of touch in some, but not all, individuals.
In summary, the monkey MNS is thought to comprise of area F5 (premotor cortex) and the superior temporal sulcus. The human MNS, however, is thought to be more expansive, recruiting not only premotor cortex, but also potentially the prefrontal lobe. Finally, the tactile mirror system is thought to involve SI and SII.
2.2. VTE and the tactile mirror system
In order to investigate whether VTE is simply caused by a hyper-activation in the common tactile mirror system, Blakemore et al. (2005) compared neuroimaging data between one individual with VTE and non-experiencing controls during tactile stimulation and observation of touch. First, participants lay on the MRI bed with their eyes closed while an experimenter applied touch to their neck and their cheek with a piece of felt attached to a wooden rod. In addition there was a baseline condition during which no touch occurred. Stimuli consisted of videos showing touch to another person’s face or neck. After scanning, the participants indicated whether they had felt the observed touch or not. The VTE individual reported experiencing tactile sensations during presentation of the videos while, as expected, none of the non-VTE individuals reported having such sensations. In line with the expectations, all participants showed activation in both SI and in SII during the observation of touch to another person versus an inanimate object as well as during actual touch to the self. Furthermore, all participants showed activation in premotor cortex during the observation of touch to another person. The VTE individual, however, showed increased somatosensory activation compared to non-VTE individual during the observation of touch. Activation in the left premotor cortex was also increased in the VTE individual compared to the non-VTE individual. Thus, the researchers concluded that these results provided evidence not only for a mirror system specific to the observation of touch, but also for the notion that this system seems to be overactive in VTE. Additionally, the VTE individual showed increased activation in insular cortex during the observation of touch compared to non-VTE individual. Considering that the anterior insula contains tactile receptive fields (e.g. Olaussen et al., 2002), the researchers suggested that this region possibly mediates the conscious perception of touch during the observation of touch to another. This is consistent with the suggestion that the insular cortex is involved in the somatosensory awareness of one’s own body (e.g. Craig, 2002). Therefore, it was concluded that a combination of anterior insula activity with a hyperactive tactile mirror system in VTE might explain why vicarious tactile experiences do not occur in healthy non-synesthetes. The increased anterior insula activity that was found in the VTE individual implied additional functional differences between VTE individuals and non-VTE individuals. Furthermore, while it has been established that the tactile mirror system is overactive in individuals with VTE compared to non-VTE individuals, the underlying mechanisms of this hyper-activation remain unclear.
To further investigate the role of SI in VTE and social touch Bolognini et al (2014) investigated whether synesthetic touch could be induced in non-synesthetes through direct or indirect stimulation of SI using paired pulse TMS (ppTMS). A test pulse was administered to the
right SI, directly after a conditioning pulse (prime) was administered to the right SI, right premotor cortex or the posterior parietal cortex. Participants carried out a visual-spatial congruity task, which was designed to authenticate VTE (Banissy & Ward, 2007). This task required participants to watch videos of another person being touched on the left or the right hand. During presentation of the videos, participants were touched on their own hand. The actual touch location was either congruent (on the corresponding hand of the observed touch) or incongruent (on the hand not corresponding to the observed touch) to the observed touch location. Participants were instructed to ignore the synesthetic experience elicited by the touch video and to report the site of the actual touch (i.e. on the left hand, on the right hand, or both). Higher error rates and longer response times were expected for the incongruent condition compared to the congruent condition. Since actual touch was administered to one hand only, reports of touch on both hands were considered the error of interest (i.e. the mirror-touch error). The task was carried out once with SI ppTMS (ppTMS condition) and once without SI ppTMS (Control condition). The results from the control condition corroborated earlier findings (Blakemore et al., 2005) indicating SI activation below the threshold of conscious perception during the observation of touch in non-VTE individuals. However, during incongruent trials, where observed touch was contralateral to the ppTMS side and actual touch was ipsilateral to the ppTMS side, error rates were higher when SI ppTMS was administered within 150 ms to the observed touch. These mirror-touch errors were specific to the observation of touch as opposed to the observation of movement. These results indicated that SI ppTMS was successful in inducing VTE in non-VTE individuals and provided evidence for the hypothesis that VTE is caused by a somatosensory overactivation during the observation of social touch. Furthermore, VTE was also induced when SI activity was primed through the ipsilateral posterior parietal cortex (PPC). According to the authors, these results provided evidence for a potentially visually driven effective connectivity between the PPC and SI, which is in line with previous reports of a tight functional connectivity between the SI, PPC, and premotor cortex.
With the intention of illuminating the underlying neural mechanisms of VTE, Holle et al. (2013) mapped the functional and structural brain differences between individuals with VTE and non-VTE individuals. They found bilateral activity in both SI and SII during touch observation in both groups, which was more extensive and stronger in the left hemisphere. Note that this finding was in conflict with a previous report of Keysers et al. (2004) who found activity in SII, but not in SI during the observation of touch to an object in non-VTE individuals. The data of Holle et al. (2013) implied that the VTE individuals held the more common ‘specular frame of reference’ (Banissy et al., 2009), as they had indicated at the start of the study. Another interesting finding was that non-experiencing controls showed activation of the left premotor cortex, while VTE individuals showed reduced activity in this region during the observation of touch to a dummy’s face. Since the premotor cortex receives tactile input from SII (Disbrow et al., 2003) and is implicated in passive touch (e.g. Whang et al., 2005), the researchers proposed that these results might have been caused by core functional differences in SII function for VTE. The structural analyses indicated decreased gray matter density in the right temporo-parietal junction (TPJ) and a more dorsal part of the medial prefrontal cortex (i.e., the superior medial gyrus) in the VTE individuals compared to non-VTE individuals. These regions have reportedly been implicated in empathy, self-other distinctions (e.g. Decety & Lamm, 2007; Brass et al., 2009; Banissy et al., 2011) and self-representations (e.g. see Lombardo et al., 2010; Wang et al., 2011). The researchers noted that these structural findings in addition to the functional findings in the
tactile mirror system were in line with previous propositions that faulty self-other monitoring is responsible for the disinhibition of normal somatosensory mirror mechanisms, leading to the conscious overt tactile experiences in VTE (Banissy et al., 2009).
In summary, the same neural regions (i.e. SI, SII, and premotor cortex) are recruited for the observation of touch and the experience of touch in VTE individuals and non-VTE individuals. However, this activity does not typically exceed the threshold of conscious perception (Blakemore et al., 2005; Bolognini et al., 2014). It seems that decreases in gray matter density in regions related to social cognition (i.e. the right TPJ, and the superior medial gyrus) lead to a disinhibition in the tactile mirror system, giving rise to VTE. Considering the apparent similarity between the underlying mechanisms for actions, emotions and sensations through mirror systems, it is possible that disinhibition is a cause for anomalous experiences in each of these modalities. For instance, it has been established that during the observation of an action, the same regions are activated as would be during the execution of the same action (Gallese et al., 2004). However, the actual motor movements are inhibited by the frontal lobe and thus the observation of an action does not typically automatically lead to the execution of that action. Nevertheless, reports have been made of individuals who automatically copy observed action; a condition dubbed echopraxia. This condition is known to occur after frontal lobe damage or basal ganglia dysfunction. It was previously proposed that a disinhibition occurs in the MNS, as a result of a lack of inhibition from the frontal lobe leading to the overtly expressed action that is performed when observing that action in someone else (see Rizzolatti et al., 2009). Considering the aforementioned research, it is not unimaginable that the underlying mechanisms through which actions, emotions, and sensations are disinhibited during the observation of stimuli presented in each of these modalities are related to mechanisms modulating empathy through feelings of likeness with others (see Serino et al., 2009).
2.3. Simulation Theory of Empathy
It has previously been suggested that empathy for emotional states of others is established through a process called simulation. Simulation is thought to recruit mirroring mechanisms through which an observer can understand (and potentially replicate) the state (physical or mental) of another person (Gallese & Goldman, 1998; Gallese, 2001; Gallese et al., 2004). Gallese et al. (2004) defined simulation as the neural ability to interconnect one’s own behavioral and emotional experiences to those of others. According to the simulation theory of action, the visual information of the observed action is translated into neural activation that normally occurs during the execution of that action (see Gallese et al 2004). Research on the simulation mechanism for emotions mainly involved the basic emotion disgust (see Gallese et al., 2004). For instance, the left anterior insula has been found to activate during the presentation of disgusting stimuli (Small et al., 2003), during the observation of disgusted faces (Philips et al., 1997; Carr et al., 2003), and during imitation of disgusted faces (Carr et al., 2003).
More recently, it was proposed that empathy for tactile sensations relies on a simulation mechanism as well (e.g. Schaefer et al., 2012). Research conducted by Banissy & Ward (2007) provided evidence for the simulation account of tactile empathy in VTE. Firstly, the researchers proposed that VTE is linked with empathy and thus suggested that VTE individuals should have higher empathic ability compared to non-VTE individuals. Empathic ability was measured using the Empathy Quotient (EQ; Baron-Cohen & Wheelwright, 2003), which consists of three subscales; 1) cognitive empathy, 2) emotional reactivity, 3) social skills. Secondly, the researchers
suggested that if VTE indeed relies on a tactile mirror system, VTE individuals should have impaired discriminative ability when vicarious touch and actual tactile stimulation occur simultaneously. The VTE individuals scored significantly higher on the emotional reactivity subscale than the non-VTE individuals. This finding, combined with the previously described results on the visual-tactile spatial congruity task (see section 1.3), was in line with the notion that the simulation process indeed mediates our understanding of and empathy for social touch to others (Gallese et al., 2004). The higher reported empathic ability in VTE individuals may thus be a reflection of an enhanced simulation ability, which may have ultimately led to a selective perceptual sensitivity for touch above the threshold of conscious experience.
Since the observation of touch activates the tactile mirror system in VTE and in non-VTE individuals, Serino et al. (2008) suggested that it should be possible to manipulate the threshold of conscious tactile experience in non-VTE individuals so as to give rise to VTE. They conducted a task similar to the visual-tactile spatial congruity paradigm (Banissy & Ward, 2007) with a few changes. Visual stimuli consisted of the participants’ own face, someone else’s face, or a house. There were several observation conditions in which the faces/houses were touched on the left side, on the right side, or on both sides simultaneously. In the control condition, faces and houses were merely approached with either one or two fingers, but touch did not occur. During the observation of touch, participants received tactile stimulation on the left side, the right side, or both sides of their face by means of electrodes attached to both cheeks. Tactile stimulation was set below the threshold of conscious tactile perception. This threshold was quantified per participant using a staircase method. Participants were instructed to ignore observed touch from the video clips and to report consciously perceived touch on their own cheeks. Results showed an enhanced tactile perception during the observation of touch to faces compared to the observation of a face being approached by hands, or the observation of a house being touched. These findings indicated that the mere observation of touch to a face might indeed lower the threshold for conscious tactile perception in non-VTE individuals. This effect, however, was even more pronounced during the observation of touch to the participant’s own face compared to the face of another person. The researchers suggested that the enhanced tactile processing, induced by the observation of touch, was mediated by the degree of similarity between the observer and the observed subject from the video clip. In order to provide additional evidence for this hypothesis, Serino et al. (2009) conducted two experiments using the same paradigm with different visual stimuli. The visual stimuli still consisted of video clips of faces. In the first experiment, the faces were either ethnically similar or ethnically dissimilar (i.e. Caucasian or Maghrebian) to the participant. In the second experiment, faces of politicians were used. The politicians were representatives of a political party that was either compatible or incompatible with the political preference (i.e. democratic or conservative) of the participant. The results showed an enhanced tactile perception during the observation of touch to faces, replicating the results of Serino et al. (2008). This effect was more pronounced during the observation of touch to people with a similar ethnicity/political preference as the participants. These results indicated that similarity between the observer and the receiver of touch indeed lowered the threshold for conscious tactile perception, providing evidence for the hypothesis that reported empathic ability is modulated by self-other similarity. It is not clear, however, to what extent this effect is influenced by representations of one’s own body.
In order to investigate the malleability of this own body representation (or bodily self; Longo et al., 2008) in healthy individuals, paradigms eliciting illusions of body ownership can be
useful. One such paradigm is known as the rubber hand illusion (RHI), which can induce a sense of ownership over a rubber hand in healthy, normal-bodied individuals (Botvinick & Cohen, 1998). During this illusion, a participant observes touch to a rubber hand while touch is also administered to the participant’s own hand, which is hidden from view. In order to elicit the RHI, it is essential that the observed touch occurs simultaneously to and is synchronous with touch to the self (Botvinick & Cohen, 1998). However, Shimada et al. (2009) found that the illusion still occurs when the observed and actual touch are administered asynchronously, as long as the lag between the two does not exceed 300 msec. When carried out appropriately, the rubber hand becomes integrated with the representation of one’s own body, leading to a sense of ownership over the rubber hand (Costantini & Haggard, 2007).
Interestingly, Davies and White (2013) found that it was possible to induce the RHI in individuals with VTE without applying touch to the real hand. Furthermore, the VTE individuals also felt vicarious tactile sensations on their own hidden hand during the observation of touch to the rubber hand. It appears as though the observed touch to the prosthetic hand initially induced a vicarious touch experience in these individuals, after which the touch sensation was transferred to the prosthesis. Moreover, unlike in the non-synesthetes, the prosthesis did not need to be in the synesthetes perceived body space in order to elicit the illusion. These results suggested that the own-body representation of VTE individuals might be even more malleable than that of non-synesthetes. The researchers proposed the possibility that the perceived body space of mirror touch synesthetes is more extended compared to that of non-synesthetes. This extension might reach beyond the synesthete’s own body and may, in specific circumstances, incorporate the bodies of others as well. This, in turn, might lead to activity in the parietal network, which is involved in the processing of multisensory information and peripersonal space (Davies & White, 2013).
Maister et al. (2013) noted that while the RHI elicits changes in one’s own-body representation, individuals with VTE might experience changes in self-identity representation as well. In order to investigate this hypothesis, they presented VTE individuals with the enfacement illusion (a facial equivalent of the RHI; Sforza et al., 2010). The researchers hypothesized that self-other boundaries may be blurred in VTE, which potentially caused these individuals to transfer observed touch to another’s face onto their own face through feelings of self-other resemblance. Seeing as how the face is a key component of self-identity (Sforza et al., 2010), it is not unthinkable that self-other boundaries would not typically be blurred as easily when observing touch to a face compared to observing touch to a hand (Maister et al., 2013). This difference, however, might be less pronounced for individuals with VTE, due to their seemingly atypical self-representation, possibly incorporating the whole body rather than just the face. Therefore, Maister et al. (2013) aimed to elicit the enfacement illusion in addition to a change of body representation in VTE individuals without administering tactile stimulation. The procedure of the enfacement illusion consisted of presenting the participants with photographs of two faces (self and other) morphed into one. These morphed faces either contained more features of the self or of the other and participants carried out a self-face recognition test in which they were to indicate whether the observed face was their own or whether it belonged to the other. After the self-face recognition test, participants were presented with videos of different levels of morphed faces. In the no touch videos, the morphed faces were presented for 2 minutes. In the enfacement videos, the same faces were presented only this time a hand holding a cotton bud approached each face and touched their cheek. Participants carried out the self-face recognition
task both before and after the enfacement illusion was elicited. Afterwards, participants completed a questionnaire about their subjective tactile experiences during the observation of the faces. During the enfacement videos, VTE individuals reported tactile sensations on their own face, implying that the VTE was genuine. Interestingly, they also reported a growing resemblance between their own face and the face of the other. In fact, they indicated that the other’s face felt like it was their own face. However, they did not indicate a sense of ownership over the other’s face. Self-face recognition scores were higher during enfacement videos compared to no-touch videos in VTE individuals. As expected, this difference was not found in the non-experiencing controls. These results suggested that self-other similarity was strengthened in individuals with VTE through the incorporation of other’s facial features into one’s own facial representation. The effect shown in the VTE individuals was similar to the effect of the enfacement illusion in non-synesthetes with the exception that it was elicited in the absence of tactile stimulation. The researchers concluded that these results were in line with the notion that self-representations are modulated by the integration of multiple sensory experiences, which either help preserve or update one’s self-representation. These changes in self-face representation and self-other boundaries seem to occur more readily in individuals with VTE compared to non-synesthetes. It is still not clear whether higher reported empathic ability underlies these changes or whether the changes are the cause of a higher reported empathic ability in VTE individuals.
3. Prevalence of VTE among Amputees
It has previously been proposed that the incorporation of phantom limbs in amputees’ body-representation is sustained by disinhibition of mirror systems involved in empathy for touch and empathy for pain (Giummarra et al., 2007). It is possible that the body-representation of an amputee is more malleable than the body representation of a normal-bodied individual. Hence, it would be interesting to investigate whether the sense of body ownership can be changed in amputees who might already have anomalous body-representations. To this end, Ehrsson et al. (2008) examined the possibility of inducing the RHI in upper-limb amputees. They found that there was a shift in location of felt touch towards a prosthetic hand during synchronous tactile stimulation of the index finger on the prosthetic hand and to the location on the stump corresponding to a phantom index finger. Furthermore, they found that some, but not all, of the amputees started experiencing a sense of ownership over the rubber hand. Thus, although the RHI was weaker in the amputees than is typically found in normal-bodied individuals, the results indicated that body-representations of amputees retain their capability to update themselves using multisensory input even in the physical absence of the to-be-manipulated limb. The researchers noted that the RHI in these individuals was unlikely to be VTE-related due to the fact that it did not occur in the absence of actual tactile stimulation to the stump.
It has been argued that phantom limbs, experienced by amputees, are caused by the memories of the affected limbs (see Katz & Malzack, 1990). Due to these limb-memories, the representation of amputees may no longer be congruent with their actual body (i.e. their self-representation consists of an intact body, whereas their own body self-representation is now incomplete). It is possible that through retaining their original self-representation, amputees can still experience empathy for actions and sensations of the limb corresponding to the memory of their own amputated limb. If the self-representations of amputees are still intact, it should be possible to induce the vicarious experiences, which were previously induced in normal-bodied
individuals (Bolognini et al., 2013; Bolognini et al., 2014), in amputees as well. These vicarious touch sensations should be inducible in both the intact limbs of these individuals as well as in their phantom limbs. Thus, Goller et al. (2013) investigated whether VTE can be acquired following loss of limb by presenting upper-limb amputees and normal-bodied MT-synesthetes with videos of people, dummies, or objects being touched. The videos depicted either non-painful touch, or slightly non-painful touch. Out of 32 amputees, 9 reported vicarious touch sensations. However, these VTE-amputees mainly reported vicarious touch sensations during the observation of slightly painful touch to another person, while the normal-bodied VTE individuals reported tactile sensations during both non-painful and slightly painful trials. Furthermore, the tactile experiences of the VTE individuals were felt on the corresponding site of the observed touch, whereas the tactile experiences of the amputees were mostly mapped to the phantom limb or to the remaining stump, regardless of the site of the observed touch. Thus, the bandwidth of VTE is even more restricted in amputees than in normal-bodied VTE individuals. Finally, the VTE-amputees scored higher on the emotional reactivity scale of the EQ compared to the non-synesthetic amputees. The researchers proposed that loss of limb could cause disinhibition in the somatosensory network, leading to VTE-like symptoms in amputees. Furthermore, they suggested that the connections might be strengthened even further due to the formation of new synaptic connections caused by the sensory loss of the amputated limb, which in turn might cause the activity for observed touch to exceed the threshold of conscious perception. These results are in line with the hyperactive tactile mirror system account of vicarious tactile experience and the account of enhanced empathy for touch. Additional evidence comes from earlier research conducted by Ramachandran & Brang (2009) who proposed that the lack of sensory input from an amputated limb causes the disinhibition in the tactile mirror system, leading to the conscious tactile awareness during the observation of touch to another person. They aimed to find evidence for this claim by administering a touch-observation task to 4 amputees. The amputees were seated in front of an assistant whose hand was placed on a table close to the phantom hand of the amputee. Amputees were instructed to watch, as the hand of the assistant was being stroked and rubbed. They were to report the occurrence and exact location of any sensation that might be felt on their own body, including their phantom limb. All four amputees indeed reported experiencing tactile sensations on their phantom limb grossly corresponding to the site of the observed touch. These findings suggested that VTE could indeed be acquired after the loss of a limb. Furthermore, it is possible that highly empathic amputees may be more susceptible than amputees with lower empathic ability to experiencing vicarious tactile sensations elicited by the observation of touch to others. On the other hand, it is also possible that the amputation itself leads to higher empathic ability, perhaps through the traumatic experience of losing a limb.
Giummarra et al. (2007) proposed that the MNS system helps humans form a template body representation through the observed actions of others. In amputees, this template remains the same as it is formed through the observation of actions of others by the MNS. In turn, the mirror neuron activity evoked by observed action in a limb of someone else may evoke sensations of a moving corresponding phantom limb in the amputee. Considering the evidence of the involvement of mirror neuron systems in brain areas other than premotor cortex (outlined in section 2.2), it is possible that these body representations are not formed solely by the visual input of actions, but also by observing other types of interactions with the corresponding limb(s). Evidence of this has been found for scratching (an itch sensation disappears when the amputee
watches a prosthesis being scratched, see Giummarra et al., 2007), touching (tactile sensations in a phantom arm emerge when observing touch to someone else’s corresponding arm, e.g. Ramachandran & Brang, 2009), movement (a sensation of movement in the phantom limb is elicited by observation of movement of the intact limb through a mirror, Ramachandran & Rogers-Ramachandran, 1996), and pain (pain sensations in the phantom limb elicited by observations of pain in others, Fitzgibbon et al., 2010). However, in the examples of itching and actions, an illusion of ownership of the observed corresponding limb was induced, which may have led to the vicarious effects in these cases. It is unclear why the tactile mirror system activity exceeded the threshold of conscious perception during observation of someone else being touched on the limb specularly corresponding to the phantom limb of the amputee, while the observation of actions only exceeded this threshold when the observed limb occupied the body space of the phantom limb, which created the illusion of ownership in the amputee. Perhaps this difference indicates that actions do not rely on empathy as heavily as bodily sensations and emotions do. The MNS for actions is said to promote action understanding, which may be necessary for imitation learning (Rizzolatti & Craighero, 2004). The utility of action simulation could be to learn how to copy actions of others, which have been proven useful in the past (Pellegrino et al., 1992). In the case of emotion -or sensation understanding, the goal is to empathize with others so we can act according to their needs (see Decety & Jackson, 2004). Additionally, the roles of actors and receivers of actions may play a role. The observation of touch elicits a vicarious tactile sensation in combination with heightened empathy in VTE. Specifically, this effect seems to be related to the anticipated sensations of the receiver rather than to the actor of the touch. This is indicated by the vicarious touch being experienced on the corresponding site to the observed touch rather than to the tip of the finger of the person executing the touch (e.g. Serino et al., 2008).
If the MNS indeed mediates the formation of a body representation template, phantom limbs should also occur in individuals with congenitally missing limbs (aplasics). Multiple cases of such aplasic phantoms have been documented (see Price, 2006). For instance, Brugger et al. (2000) described a case of a woman who was born without forearms and legs. Ever since she could remember, she reported vivid mental images of these missing limbs. Her phantom limbs would disappear whenever others moved through her phantoms’ body space, as well as when she looked at herself through a mirror. Thus, the visual information of the absent limbs interfered with the illusion of their presence. Through the use of TMS in the sensorimotor cortex, sensations of a phantom hand and phantom fingers were elicited on the contralateral side of stimulation. These sensations were accompanied by motor evoked potentials (MEPs) in the stump on the side where the phantom was felt. During stimulation of the parietal cortex as well as the premotor cortex, similar sensations were elicited with the exception that there were no MEPs recorded in the stump. Additionally, an implicit reaching task was conducted during which the aplasic was presented with drawings of hands and feet on a computer screen. These hands and feet were either presented at a common angle or an uncommon angle. In the uncommon angle condition, mental rotation was required in order to determine whether the drawing depicted a right or a left hand or foot. The aplasic was instructed to press a button with her left arm stump for drawings depicting a left foot or hand and with her right arm stump for drawings depicting a right foot or hand. Latencies for correct trials were recorded. The difference in latency between the rotation and the non-rotation trials denoted the extra effort required for mental rotation. Performance on this task was similar for the aplasic as it was for healthy control
subjects, indicating that the body representation may be intact in aplasics. Finally, an fMRI study was carried out during which the aplasic was instructed to “move” her phantoms. There were 4 phantom movement tasks: 1) Sequential movements of the right index finger to the right thumb; 2) Sequential movements of the left index finger to the left thumb; 3) pronation (turning the palm of the hand downward) and supination (turning the palm of the hand upward) of the right hand; 4) flexion and extension of the right foot. During these phantom movements the researchers found bilateral premotor activity, indicating that the phantoms were indeed represented in the body template of the aplasic. However, during a finger-to-thumb movement of either phantom hand, there was no activity found in the hand-representation of the primary motor cortex. This was also the case during the pro -and supination of the right phantom hand. Furthermore, unlike previous findings regarding traumatic amputees (Ersland et al., 1996), there was no activation of SI during phantom finger movements. The researchers concluded that the primary motor cortex might mediate phantom movements in amputees through memories of the lost limb(s). However, in the case of congenitally absent limbs, this region does not seem to be involved.
Other factors may play a role in the process of updating the body template in aplasics. One system that may mediate this process is the MNS. Through observations of other’s interactions with their limbs, an aplasic may experience a “vicarious sense of presence” for these limbs (e.g. Brugger et al., 2000). Perhaps the body template is innate at first, but is updated and reformed through integration with multimodal perceptual information about the human body. The MNS account would mainly operate through the visual modality, seeing as how the observation of another’s interaction with their own limbs may form the body representations of amputees and aplasics. In congenitally blind individuals, the visual modality is absent, forcing these individuals to use a different sensory modality to form the same representations as individuals with intact sight (Pascual-Leone et al, 2005). It has been shown, for instance, that congenitally blind individuals show social face, hand, and arm movements similar to those made by non-blind individuals (Brugger et al., 2000). These findings indicate that congenitally blind individuals form representations of such social actions through an innate template of social behavior that is formed by multisensory processes (Iverson & Goldin-Meadow, 1998). Thus, individuals with the ability to use the combined force of all perceptual modes may retain their body –and self-representation when one of these modalities is suddenly inactive. For individuals with congenitally absent perceptual modalities, these representations may be formed in a slightly different way while still retaining the same core innate features as those present in normal-bodied individuals. Thus, it seems as though human body representations, while partly innate, are highly plastic (Brugger et al., 2000).
4. Discussion
In summary, it is argued that VTE is not a form of synesthesia. Several of the defining characteristics of synesthesia are not met by VTE. Firstly, the bandwidth of the inducer-concurrent pairings in VTE is very restricted, rendering both the criteria ‘idiosyncracy’ and ‘consistency’ trivial. Secondly, the VTE concurrents are indistinguishable from real-world perceptions, whereas the synesthetic concurrents are not. In this paper it was argued that the reason VTE does not meet these defining characteristics is that it is rooted in a common neural network that is hyperactive in certain individuals. This is unlike synesthesia where brain areas
