Comparing synesthetic experience with other additional sensations: possible parallels between underlying neural mechanisms

(1)

Comparing synesthetic experience with other

additional sensations: possible parallels between

underlying neural mechanisms

ABSTRACT

Synesthesia is characterized by the experience of extra sensations that do not relate to the external world in a straightforward manner. In order to investigate the neural mechanisms of synesthesia, a comparison is made with other phenomena that involve additional sensations. Similarities between synesthesia and the other phenomena would suggest a common mechanism underlying all types of additional sensations. The other phenomena with additional sensations discussed here include tinnitus, phantom limb sensations, and various types of hallucinations.

In order to investigate if a more general mechanism could underlie all additional sensations, a comparison is made first on a phenomenological level: it is discussed which characteristics of synesthetic experience are also present in other additional sensations. Although several parallels were found, there are important phenomenological differences between synesthetic experience and other additional sensations.

These parallels in phenomenology give reason to compare synesthesia with other conditions involving additional sensations on a neurobiological level. If there are neural similarities, they could underlie all

additional sensations. This was done on the basis of three theories on synesthetic neural mechanisms that are not mutually exclusive: cross-activation through structural abnormalities in sensory cortex, cross-activation through functional abnormalities in sensory cortex, and parietal hyperbinding.

Of the three investigated theories on synesthesia neural mechanisms, functional cross-activation seemed to be most related to additional sensations in other conditions. However, the evidence is tentative and should be interpreted with caution. In addition, only tinnitus and phantom limb sensations could be related to all three suggested neural mechanisms of synesthesia. Yet, these neural similarities are inconclusive and there are major limitations to be kept in mind. Furthermore, there was very little evidence for a neurobiological parallel between synesthesia and various types of hallucinations. A lack of structural neuroimaging research in this field might be the cause for this finding.

In sum, there was some indication of neural similarities between all conditions where additional sensations play a role. Nevertheless, due to several neural differences and methodological limitations, these parallels are not strong enough to suggest a common neural mechanism for the experience of additional sensations in general. Further limitations and possible implications of these findings are discussed. Finally, directions for future research are suggested.

R. San Giorgi 5929105 23-01-2015 University of Amsterdam Literature Thesis supervisor: Dr. R. Rouw

(2)

Co-assessor: Dr. H.S. Scholte

1. Introduction

“Thus the matter of belief is, in all cases, different in kind from the matter of sensation or presentation, and error is in no way analogous to hallucination. A hallucination is a fact, not an error; what is erroneous is a judgment based upon it.”

(Russell, 1914, p. 173)

In this quote, Lord Russell emphasizes that additional sensations, like hallucinations, are seemingly incomprehensible but nevertheless a part of our lives. Some people experience extra sensations, additional to those that most of us experience when perceiving the external world. In developmental synesthesia, a particular stimulus in one modality (‘inducer’) located in the external world, can elicit an internal perception (‘concurrent’) in the same or a different modality. Further synesthesia characteristics, as listed by Colizoli, Murre, and Rouw (2014a) include inducer-concurrent pairings that are consistent over time. This means that one inducer always evokes one concurrent. For example, in a grapheme-colour synesthete, the letter ‘A’ is always paired with ‘red’ and never with another colour. Another synesthesia criterion is automaticity of the concurrent sensation after being exposed to the inducer. That is, the synesthetic reaction is non-voluntary and cannot be controlled. Furthermore, synesthetes have insight in the additional quality of their perception: they do not experience the concurrent to be ‘in the real world’ in the way that inducers are. Finally, synesthesia is idiosyncratic: the pairings of inducers and concurrents are specific and different per individual

(Bargary & Mitchell, 2008).

Synesthesia is not the only phenomenon where additional sensations play a key role. In several other conditions people experience extra sensations which cannot be perceived by others and do not seem to be a common reaction to stimuli in the external world. If there are phenomenological similarities between synesthesia and other conditions, the question arises if there are also parallels between their underlying neural mechanisms. If this is the case, these conditions could originate from a broader neural mechanism related to additional perception in general. This would be an interesting finding affecting all research directed at perception, since it implies that genuine and additional sensation are related to separate brain mechanisms.

The current thesis will tempt to find an answer to this question by comparing synesthesia with tinnitus (an auditory phantom phenomenon), phantom limb sensations, and four types of hallucinations. First, synesthetic experiences will be compared with the additional sensations in the other conditions on a phenomenological level. After a concise description of the conditions discussed in this thesis, a comparison will be made with respect to five characteristics of synesthesia. If the additional sensation is elicited by stimuli in the external world, then inducer-concurrent pairing is present. If an inducer always elicits the same concurrent sensation, pairs are considered consistent over time. Furthermore, automaticity is present if the additional sensation is non-voluntary and if it cannot be controlled. Insight is determined by if the individual is aware of the additional aspect of the sensation. Finally, there is idiosyncrasy if the additional sensations have unique properties in each individual. Then, three existing theories regarding the underlying neural mechanisms of synesthesia will be explained and compared with neuroimaging findings in the other conditions. These

(5)

hyperbinding in the parietal cortex, and are not mutually exclusive. Next, the outcome of these comparisons will be discussed. Lastly, a conclusion will be drawn and the potential consequences of this outcome will be reported. Possible directions for future investigations will be suggested.

(6)

2. Comparison of phenomenology

2.1 Tinnitus

There are two types of chronic tinnitus: subjective tinnitus, where a tone with a specific frequency and loudness is perceived in the absence of an external auditory source or other

stimulation of the auditory nerve, and objective tinnitus, a rarer form where sounds originating from the body are perceived (Meyer, Langguth, Kleinjung, & Møller, 2014). As opposed to its chronic form, acute tinnitus is temporary, often caused by exposure to loud noises. In this paper, tinnitus will refer only to chronic subjective tinnitus; both objective and acute tinnitus are a common reaction to external stimuli. Tinnitus often arises after a hearing impairment, where hair cells in the inner ear are damaged (Mühlau et al., 2006), and is often accompanied by a feeling of distress. However, there is no clear correlation between the level of tinnitus-related distress and tinnitus loudness (Wallhäusser-Franke et al., 2012).

Unlike synesthetic experience, tinnitus does not involve inducer-concurrent pairs, although tinnitus is often more prominent under silence. Tinnitus is often consistent over time. The auditory additional sensation is automatic and can usually not be controlled. In a rare form, tinnitus loudness can be modulated with specific face movements (Lockwood et al., 1998). Similar to synesthesia, tinnitus patients can differentiate between externally generated auditory stimuli and tinnitus-related sounds. Tinnitus is perceived to be located in either one of the ears or both (Seydell-Greenwald, Raven, Leaver, Turesky, & Rauschecker, 2014). Thus, patients have insight in the additional aspect of their sensation. Individual differences in tinnitus are limited to tone frequency and loudness. Therefore, there is no high idiosyncrasy as in synesthesia. An overview comparing tinnitus and all other conditions with synesthesia can be found in table 1.

2.2 Phantom limb sensations

Phantom limb sensations involve painful or non-painful tactile sensations in a deafferented or amputated limb. The majority of research on phantom limb sensations focuses on painful sensations, possibly because there is need for improved effective treatment and because it is less rare than non-painful sensations. Yet, experiencing a non-non-painful phantom sensation is more similar to synesthetic experience: there is often referral of tactile sensations on the face to a phantom limb (Vilayanur S. Ramachandran & Rogers-Ramachandran, 2000). This type of inducer and concurrent is similar to mirror-touch synesthesia. Mirror-touch synesthetes experience a tactile sensation on their own body after observing touch to another person (Banissy, Cohen Kadosh, Maus, Walsh, & Ward, 2009). Thus, the referred sensation mapping in non-painful phantom limb sensations can be interpreted as inducer-concurrent pairs. Although the similarity with mirror-touch synesthesia makes non-painful phantom limb sensations more relevant for this thesis than painful phantom sensations, the amount of neuroimaging research on non-painful phantom limb sensations is limited. Therefore, this thesis will include research on painful phantom limb sensations, and, where possible, focus on the underlying neural mechanisms of non-painful phantom limb sensations specifically.

In non-painful phantom sensations, a one-to-one correspondence between the stimulated face areas and the location of phantom limb sensations is present, similar to synesthesia

(7)

inducer-concurrent pairs. Besides the face, other common inducing mapping places are located near the

amputated site, or at the contralateral limb (Hebb, Ramachandran, & Hirstein, 1998). This pairing is fairly consistent over time, although there are sometimes slight shifts in location of specific reference fields (Hebb et al., 1998). The phantom limb sensations occur automatically, and the patient cannot influence them. Patients are aware of the fact that there is no actual stimulation on their amputated limb, although it feels that way, thus giving them insight into the additional aspect of their phantom sensation. However, the phenomenon is not entirely idiosyncratic; inducing mapping locations are usually at the ipsilateral lower face, at the contralateral limb, or at the site near the amputated limb (Vilayanur S. Ramachandran & Rogers-Ramachandran, 2000). This restricts the amount of individual difference in non-painful phantom limb sensations.

2.3 Charles Bonnet syndrome

The Charles Bonnet syndrome (CBS) is a condition where people with visual impairment as a result from ocularpathology (Bergman & Barak, 2013) experience complex visual hallucinations, but are psychologically healthy. The hallucinations often occur in older adults and can consist of detailed images of people, animals, landscapes, but also more simple forms like patterns and textures, for example a brickwork wall (Ffytche, 2005; Howard, Brammer, & David, 1998). Not much is known about the mechanisms underlying CBS, and it has even been suggested that multiple

pathophysiologies can result in the CBS phenotype (Ffytche, 2005). Currently there is a respectable amount of case reports on CBS, but systematic neuroimaging studies are absent.

Hallucinations can be triggered by visual stimulation in CBS patients with reduced vision (Howard, Brammer, et al., 1998), but the hallucinations are not predictable as in synesthetic

inducer-concurrent pairs. In CBS patients with acute loss of vision, the onset of hallucinations can be hours or

days after loss of vision, and the hallucination duration can vary from seconds to hours (Howard, Brammer, et al., 1998). Thus, the additional experiences in CBS are not consistent over time like synesthetic experiences. Similar to synesthetic experiences, CBS hallucinations occur automatically and are recognized as unreal (Braun & Dumont, 2003), indicating that patients have insight. Furthermore, there is individual difference in types of hallucinations: they can consist of simple repetitive patterns, but also highly detailed complex scenes involving a variety of objects, animals, and people (Ffytche, 2005). Therefore, idiosyncrasy in CBS is high.

2.4 Brain damage related hallucinations

Patients sometimes experience hallucinations following lesions in various brain areas. Spontaneous visual or auditory hallucinations are a very common form of hallucination after brain damage (Braun & Dumont, 2003). There are reports of hallucinations with characteristics similar to synesthetic experience after brain damage (Ro et al., 2007), or damage to the anterior optic pathways (Afra, Funke, & Matsuo, 2009). These synesthesia-like experiences involve different modalities, where inducers are for example auditory, and concurrents visual. This inducer-concurrent pairing can last for months or years, meaning that this phenomenon is consistent over time, like developmental

synesthesia. However, these synesthesia-like hallucinations are very rare; most brain damage related hallucinations are unimodal and not elicited by an inducer (Braun & Dumont, 2003). These

(8)

spontaneous hallucinations are automatic and cannot be controlled. Patients are not always aware of the hallucinatory nature of their experience (Braun & Dumont, 2003), which means that the amount of insight varies among individuals. Although some hallucinations are very simple, others can be complex and unique per individual, as images are sometimes recognized from memory (Braun & Dumont, 2003). Therefore, brain damage related hallucinations are considered idiosyncratic.

2.5 Auditory verbal hallucinations

An estimated 70% of schizophrenia patients have auditory verbal hallucinations (AVHs) (Daalman et al., 2011), more commonly described as “hearing voices”. These voices differ from each other with respect to factors as gender (male or female voices), familiarity (voices from people they do or do not know), or structure (systemized or repetitive voices). Interestingly, Sommer and

colleagues (2010) reported that 10%-15% of the general population experience AVHs as well, and are otherwise psychologically healthy. The authors suggest that these non-clinical subjects with AVH are more vulnerable for schizophrenia. This indicates a similar underlying neural mechanism for both clinical and non-clinical AVH. Since these mechanisms are the main interest of this thesis, both conditions will be discussed under the sections that refer to AVH.

There is no inducer-concurrent pairing in both clinical and non-clinical AVH. Frequency of AVHs is higher in patients than in non-clinical individuals, and AVH content can be repetitive or variable (Stephane, 2013), making AVHs inconsistent over time. Patients can exert less control on AVHs than non-clinical individuals with AVH (Daalman et al., 2011), indicating higher automaticity in patients than in control subjects. Insight also varies individually, but is, perhaps surprisingly, not significantly higher in patients (Daalman et al., 2011). Idiosyncrasy is high, as the number of different voices varies, and some voices belong to people they know.

2.6 Drug-induced hallucinations

Synesthesia-like hallucinations can be elicited with hallucinogens like psilocybin, lysergic acid diethylamide (LSD), or mescaline. Although there is no lack of behavioural research in this area, most experiments possess severe methodological limitations (see Luke and Terhune (2013) for an

overview). Combining hallucinogen intake with a neuroimaging paradigm has rarely been done. However, drug-induced hallucinations will still be included in this thesis in order to give an overview of the limitations of existing literature.

Cross-modal inducer-concurrent associations can be present in drug-induced hallucinations, where the inducer is most commonly auditory, and the concurrent visual (Luke & Terhune, 2013). Other pairings have also been reported, with inducers consisting of touch, smell, or even emotions, but the majority of studies report a visual concurrent sensation (Luke & Terhune, 2013). These associations are however acute and temporary (Luke & Terhune, 2013), and their coupling is not

consistent (Sinke et al., 2012). Subjects can control the strength of their hallucination, for example by

opening or closing their eyes, although this controllability is dose-dependent (Luke & Terhune, 2013). However, they cannot influence the pairing of inducers and concurrents, as this seems to be an

(9)

aware of the fact that their experiences are drug-induced (Sinke et al., 2012). Drug-induced hallucinations are idiosyncratic, with high inter- and intrapersonal variance.

Table 1 : Comparing phenomenology: synesthesia versus other phenomena involving additional sensations. Similarities with synesthesia phenomenology are shown in bold.

Is inducer-concurrent pairing present? How consistent is the experience over time? Is it an automatic experience? Are the additional sensations recognized as unreal? What is the amount of individual difference?

Tinnitus No, but silence _{can be a trigger} High Yes Yes Low

Phantom limb

sensations Yes High Yes Yes Low

Charles Bonnet syndrome

No, although visual stimuli can be a trigger

Low Yes Yes High

Brain damage related hallucinations

Yes, but rarely Varies Yes Varies Varies

Auditory verbal

hallucinations No Low Varies Varies High

Drug-induced hallucinations

Yes, but pairs are not consistent

(10)

3. Comparison of structural hyperconnectivity in sensory cortex

3.1 Is there structural hyperconnectivity within or between sensory brain

areas in synesthesia?

One theory proposed by Hubbard (2007) suggests that synesthetic experience can be explained by cross-connections between cortical areas where the sensory modalities of both

synesthetic inducer and concurrent are processed. For example, in grapheme-colour synesthesia, this theory predicts higher structural connectivity between the area dedicated to the inducer and the concurrent. These areas are respectively the inferior temporal cortex, associated with grapheme processing (Cohen et al., 2000), and the ventral occipitotemporal cortex, the location of visual area four (V4), which is associated with colour processing (Mckeefry & Zeki, 1997).

Using a variety of neuroimaging techniques, many studies have investigated the structural neural characteristics related to synesthesia, and a multitude report supporting evidence for this ‘cross-activation’ theory (Hänggi, Wotruba, & Jäncke, 2011; Jäncke, Beeli, Eulig, & Hänggi, 2009; Rouw & Scholte, 2007, 2010). A diffusion tensor imaging (DTI) study of Rouw and Scholte (2007) found that grapheme-colour synesthetes showed, as compared with non-synesthetes, increased structural connectivity in inferior temporal cortex, adjacent to V4. This suggests a structural hyperconnection between the two unimodal brain areas related to the synesthetic experience of grapheme-colour synesthesia. Furthermore, Jäncke and colleagues (2009) investigated grapheme-colour synesthesia with T1-weighted magnetic resonance imaging (MRI). They found altered neuroanatomical

differences in the fusiform gyrus (where V4 is located) and adjacent regions. These differences included increases in cortical thickness, volume, and surface area. This again supports the theory that corticocortical connections between areas related to the modalities of inducer and concurrent are the neurobiological base of grapheme-colour synesthesia. Similarly, Hänggi and colleagues (2011) found fusiform gyrus hyperconnectivity in grapheme-colour synesthetes when using T1-weighted MRI on grapheme-colour synesthetes.

Rouw and Scholte (2010) investigated the difference between projector and associator synesthetes. Projector synesthetes describe their synesthetic experience as if they were located “in the outside world”, while associator synesthetes describe them as being “in the mind” only. By using voxel-based morphometry (VBM), the authors found that projector synesthetic experiences (located in the external world) were more related to grey matter increases in sensory cortical areas, such as the visual and auditory cortex, as compared to associator synesthetic experiences (located inside the head). Therefore, projector synesthesia seems to be more related to sensory cortical involvement than associator synesthesia. If the additional sensation related to projector synesthesia can be considered stronger than the additional sensation in associator synesthesia, the results of this VBM study could imply a role for sensory cortical involvement in other additional sensations as well.

Thus, there is plenty of literature that argues in favour of structural cross-activation between the sensory cortical areas involved in synesthesia. The following section will review if the cross-activation theory could also explain other phenomena with additional sensations. In order to do so, these phenomena are evaluated for possible increased structural connectivity within or between brain areas that are associated with the modality of their additional sensation.

(11)

3.2 Is there structural hyperconnectivity within or between sensory brain

areas in other phenomena involving additional sensations?

3.2.1 Tinnitus

Not many studies have investigated structural neural connectivity in tinnitus properly, possibly because it is difficult to control for confounding factors of tinnitus, such as hearing loss, depression and anxiety. Despite these challenges, there are reports on tinnitus-related increases in structural connectivity located in the auditory pathway (Crippa, Lanting, van Dijk, & Roerdink, 2010; Seydell-Greenwald et al., 2014). Seydell-Greenwald and colleagues (2014) compared tinnitus patients with a control group matched for age, depression and anxiety. They used DTI and found that tinnitus patients showed increased tract density in the auditory pathway, including the inferior colliculus (IC), and the auditory cortex. Similar results were found by Crippa and colleagues (2010), who compared tinnitus patients with control subjects. They also found increased structural connectivity in the auditory pathway: from the IC, through the medial geniculate nucleus of the thalamus (MGN) to the auditory cortex.

However, these results are attenuated by reports of no increases in structural connectivity located in the auditory pathway in tinnitus. For example, a VBM and DTI study by Husain and colleagues (2011) found no tinnitus-related brain anomalies in the auditory pathway. In order to control for confounding factors, the authors compared three groups: hearing loss patients with tinnitus, hearing loss patients without tinnitus, and control subjects. By comparing both groups with hearing loss, they limited the chance of mistakenly attributing the neural correlates of hearing loss, which is a confounding factor, to tinnitus.

Interestingly, Husain and colleagues (2011) report that hearing loss, as opposed to tinnitus, has a far greater influence on changes in grey and white matter in the vicinity of the auditory cortex. This casts doubt on the value of previous reported studies on tinnitus. While (Seydell-Greenwald et al., 2014) successfully controlled for age, tinnitus patients still had significantly more hearing loss than control subjects. They do investigate whether their effects might be due to hearing loss in post hoc correlation analysis, and find reduced, but similar, effects. Crippa and colleagues (2010) do not report any control measures for hearing loss, stating that it is not crucial since hearing loss is present in only 20% of tinnitus patients.

Thus, there is only tentative indication for higher structural connectivity in the sensory brain areas related to tinnitus – the auditory pathway. Since hearing is the only sensory modality involved in tinnitus, cross-activation between modalities has not been investigated, and is not expected. Where the ‘cross-activation’ theory of synesthesia involves cortical sensory brain areas, tinnitus seems to be related to subcortical neuroanatomical changes. An overview comparing tinnitus and the other phenomena with synesthesia can be found in table 2.

3.2.2 Phantom limb sensations

A well-known model for phantom limb sensations suggests that it can be attributed to maladaptive reorganization of the primary somatosensory (SI) cortex (V S Ramachandran, Brang, &

(12)

McGeoch, 2009). Many hand amputees have a topographically organized map on the lower face (Vilayanur S. Ramachandran & Rogers-Ramachandran, 2000). The hand and face are represented adjacently in SI, which supports the maladaptive reorganization model. Similar to the synesthesia cross-activation theory (Hubbard, 2007), this model suggests cross-connections between parts of SI dedicated to the body part with the map (inducer) and the amputated body part (concurrent).

However, no reports on such increased structural connectivity between SI areas have been related to phantom limb sensations. Examples of acute phantom limb sensations in current literature even argue against increased structural connectivity. Borsook and colleagues (1998) reported a case study where a hand amputee had non-painful phantom limb sensations from stimulation of the face within 24 hours after amputation, and phantom limb pain was felt after two days. Additionally, mirror feedback (V S Ramachandran et al., 2009) can relieve phantom pain immediately. These interesting findings question the role of cortical reorganization in phantom limb sensations.

Furthermore, several neuroimaging studies argue against a structural cross-activation theory in phantom limb sensations (Makin et al., 2013; Simões et al., 2012). A DTI study of Simões and colleagues (2012) found a decrease in microstructural connectivity in the somatosensory region of the corpus callosum, when comparing patients with lower limb non-painful phantom sensations with control subjects. No cortical structural hyperconnectivity is reported. Rather, their results support a hypothesis where less organized connectivity leads to decreased interhemispheric inhibition, which is the cause for phantom sensations. This decrease in structural connectivity is the opposite of what is implied by the synesthesia cross-activation theory, and interhemispheric connections are not known to be related to synesthetic experiences. Additionally, Makin and colleagues (2013) showed reduced grey matter volume in the phantom hand area in SI, compared to contralateral (intact) hand area in amputees. This reduced grey matter volume in the phantom limb cortex was also found when compared with the hand area in congenital one-handed control subjects, and intact control subjects. The authors suggest that this reduction in grey matter implies structural degeneration caused by the sensory deprivation after amputation. Therefore, the presence of structural hyperconnectivity between SI areas associated with “inducer” body part and “concurrent” phantom limb is unlikely.

Makin and colleagues (2013) do however find grey matter volume in the phantom cortex of amputees to be correlated positively with an increase in phantom limb pain. The authors suggest that phantom limb pain is related to preserved local structure in the deprived phantom cortex. Although cross-activation theory in phantom limb sensations would imply activation between cortical areas, Makin an colleagues (2013) found greater grey matter volume, implicating increased local structural connectivity, only within the phantom cortical area to be related to phantom limb pain. Therefore, applying the synesthesia cross-activation theory to phantom limb sensations is not fully supported by their results.

In conclusion, although some indication of increased or preserved structural connectivity within the phantom cortex is found, structural connectivity between the phantom limb cortex and the cortex related to the inducing body part, similar to the synesthesia cross-activation theory, has not been found.

(13)

3.2.3 Charles Bonnet syndrome

The structural neurobiological properties of Charles Bonnet syndrome have not been investigated systematically, although several case reports mention cortical anomalies in the visual cortex (Gubernick et al., 2014; Kishi, Uegaki, & Kitani, 2000). One case study (Kishi et al., 2000) investigated a 73-year old woman with CBS and found mild atrophy in the medial occipital cortex, where primary and association visual areas are located. Another case study (Gubernick et al., 2014) investigated a 74-year old woman who experienced olfactory sensations in addition to visual hallucinations, after suffering from an infarction affecting inferior temporal and occipital lobes. Although there is no inducer in CBS, the cortical brain area related to the “concurrent” additional sensation in CBS would be the visual cortex, located in the occipital cortex. Both these case studies describe not hyperconnectivity, but impairment of the visual cortex.

Thus, although it has not been researched systematically, there is currently no indication for hyperconnectivity in the visual cortex in CBS. Rather, the CBS visual hallucinations seem to be related to lesions in or dysfunction of the visual cortex.

3.2.4 Brain damage related hallucinations

When unimodal hallucinations occur after brain damage, the lesion is often located in the pathway of the sensory modality of the hallucinations (see Braun and Dumont (2003) for an extensive review of case studies). It is therefore highly unlikely that structural hyperconnectivity, as proposed by the synesthesia cross-activation theory, plays a role in brain damage related hallucinations. However, no structural investigations of white matter properties in brain damage related hallucinations have been found.

3.2.5 Auditory verbal hallucinations

Applying the structural cross-activation theory on AVH, an increase in local structural connectivity in the auditory cortex would be expected. However, existing literature suggests that a decrease in local connectivity in the auditory cortex is more likely related to AVH (Allen, Larøi, McGuire, & Aleman, 2008; Modinos et al., 2013; Nenadic, Smesny, Schlösser, Sauer, & Gaser, 2010; Palaniyappan, Balain, Radua, & Liddle, 2012). Allen and colleagues (2008) review structural

neuroimaging studies on AVH in schizophrenia and show eight studies reporting reduced grey matter in the superior temporal gyrus (STG), which is where the primary auditory cortex (PAC) is located. This is supported by more recent studies (Modinos et al., 2013; Nenadic et al., 2010; Palaniyappan et al., 2012), but further evaluation of these findings is beyond the scope of this thesis.

There is no indication for increases in white matter in the auditory cortex related to AVH. Rather, Plaze and colleagues (2011) used VBM and found differences in white matter in subtypes of AVH, but these were not located in the auditory cortex. Since the white matter anomalies were located in the STG, and near the temporo-parietal junction (TPJ), their findings will be discussed in further detail in chapter 5.

Thus, reduced grey matter volumes in the auditory cortex indicates that increased structural connectivity is not likely, and there is no evidence for a similar ‘cross-activation’ theory AVHs in schizophrenia.

(14)

3.2.6 Drug-induced hallucinations

In drug-induced hallucinations, the additional sensation is only present shortly after drug intake, and lasts only for a small period of time. With regard to the temporary aspect of drug-induced hallucinations, structural differences in grey or white matter related to those hallucinations are improbable. This is probably the reason that there is currently no investigation of structural neural differences related to drug-induced hallucinations.

3.3 Conclusion

Both tinnitus and phantom limb syndrome seem to be associated with higher structural integrity in the auditory and somatosensory pathway respectively. Although there is no cross-connection between different modalities in tinnitus, higher structural integrity in the auditory pathway has been found. However, where synesthesia is often related to changes in cortical structures, tinnitus is more related to anomalies in subcortical structures, such as the IC and MGN. Painful phantom limb sensations seem to be related to increased, or preserved, structural

connectivity within the phantom cortex, but no report of structural connectivity between the cortex of inducing body parts and the phantom limb cortex has been found. CBS, brain damage related hallucinations, AVH, and drug-induced hallucinations show no indication of being related to increased structural connectivity in sensory brain areas.

(15)

Table 2: Is increased structural connectivity within or between sensory brain areas, as found in synesthesia, also present in other phenomena involving additional sensations? Similarities with synesthesia findings are shown in bold.

Tinnitus

- Some experiments found evidence in favour of increased structural connectivity in the auditory pathway

- These areas are more subcortical (IC, MGN) than in synesthesia

- Some experiments found no evidence of increased structural connectivity in the auditory pathway

Phantom limb sensations

- Some experiments found evidence in favour of increased structural connectivity in SI related to phantom limb pain

- No evidence for structural cross-connections between phantom limb cortex and cortex of inducing body part have been reported

- No evidence for increased structural connectivity in SI related to non-painful phantom sensations have been reported

Charles Bonnet syndrome

- Currently no systematic neuroimaging research present

- Case studies suggest lesions in visual cortex, making structural hyperconnectivity unlikely

Brain damage related hallucinations

- Currently no systematic neuroimaging research present

- Case studies suggest lesions in sensory brain areas, making structural hyperconnectivity unlikely

Auditory verbal hallucinations

- Large amount of neuroimaging research suggest reduced grey matter in PAC, making structural hyperconnectivity unlikely

Drug-induced

hallucinations - Not investigated, but highly unlikely due to its acute and temporary characteristics IC = Inferior Colliculus, MGN = Medial Geniculate Nucleus SI = Primary Somatosensory Cortex, PAC = Primary Auditory Cortex

(16)

4. Comparison of functional hyperconnectivity in sensory cortex

4.1 Is there functional hyperconnectivity within or between sensory brain

areas in synesthesia?

There is debate regarding the nature of the neural mechanisms underlying cross-activation. These mechanisms could be based on structural or functional mechanisms, and one mechanism does not exclude the other. Structural evidence has been evaluated in previous section, but many studies suggest that synesthetes show differences in sensory brain area activation. Here, some studies that give indication for hyperactivation in or between sensory brain areas in synesthesia are discussed.

The visual cortex can be hyperactivated during a synesthetic experience with a visual concurrent, even in the absence of external visual stimulation. For example, Nunn and colleagues (2002) compared a group of coloured-hearing synesthetes with control subjects using fMRI and found that V4, but not primary (V1) or secondary visual cortex (V2) of synesthetes is activated by auditorily presented words. When control subjects were imagining colours during auditory word presentation, V4 showed no activity, even after intensive training for word-colour associations. This way, the authors controlled for the possibility of hyperactivation caused by visual imagery. They conclude that synesthetic perception is more similar to genuine colour perception than to imaginary colour

perception. In a similar experiment by Aleman and colleagues (2001), words were presented auditorily to one coloured-hearing synesthete in an fMRI task. V1 activation was found, again indicating that visual cortex can be activated by the synesthetic experience (seeing the colour after hearing the word) in absence of visual stimulation. Furthermore, Hubbard and colleagues (2005), let grapheme-colour synesthetes execute several behavioural tests where additional colour perception could benefit performance. The results were compared with those of control subjects and an fMRI experiment was performed where subjects were visually presented with black graphemes. The researchers found increased performance to be correlated with hyperactivation of V1, V2, V3, and V4. This leads to the conclusion that stronger synesthetic experience, which increased performance on behavioural tests, is correlated with hyperactivation of the visual cortex.

These studies show arguments in favour of functional cross-activation in synesthesia. They report hyperactivation of the sensory cortex related to the modality of the concurrent additional sensation in synesthesia. The following section will discuss reports of hyperactivation in the sensory cortex related to the modality of the concurrent additional sensation in other phenomena.

4.2 Is there functional hyperconnectivity within or between sensory brain

areas in other phenomena involving additional sensations?

4.2.1 Tinnitus

The modality of the additional sensation in tinnitus is auditory. If the functional

cross-activation theory of synesthesia is also a theory of tinnitus, hyperactivity in the auditory cortex would be expected. Several investigations link tinnitus with hyperactivation in the auditory cortex (Leaver et al., 2011; Lockwood et al., 1998). Some patients can control the loudness of their tinnitus by

(17)

tomography (PET) to investigate this rare form of tinnitus. After controlling for changes in cerebral blood flow (CBF) due to the movements that patients made, they found that a reduction of tinnitus loudness was associated with a decrease of CBF in the temporal lobe. This implies that stronger additional sensation (louder tinnitus) is related to hyperactivity in the auditory cortex (located in the temporal lobe). In addition, Leaver and colleagues (2011) found hyperactivity in the primary auditory cortex (PAC) in tinnitus patients compared to control subjects, during an fMRI experiment where subjects were presented with white noise bursts. Unlike previously mentioned experiments with synesthetes, this experiment does not investigate the additional sensation involved in this

phenomenon. Although the PAC hyperactivation is not directly related to the additional sensation in tinnitus, it does show that functional differences in the auditory cortex are present in tinnitus patients.

However, the role of the PAC in tinnitus is not indisputable. Plewnia and colleagues (2007) suggest a limited role for the PAC in tinnitus. They used PET to determine that the inferior parietal cortex, and not the PAC, was most related to tinnitus-related increases of CBF. Based on this finding, they proceeded with repetitive transcranial magnetic stimulation (rTMS) on the inferior parietal cortex. Results indicated that tinnitus loudness was reduced after rTMS. This suggests a limited role for the auditory cortex in tinnitus, and that higher-order processing is more essential. An overview comparing tinnitus and all other phenomena with synesthesia can be found in table 3.

If the functional cross-activation theory of synesthesia is also a theory of phantom limb sensations, hyperactivation would be expected in areas of the primary somatosensory cortex (SI) related to the body part with the map (inducer), and phantom limb (concurrent). The previously mentioned case study where phantom limb sensations are present within 24 hours after amputation (Borsook et al., 1998) emphasizes that functional, instead of structural, changes in SI can lead to phantom limb sensations.

Indeed, several studies have found associations between functional anomalies in SI and phantom limb sensations. Amputees were investigated using fMRI years after their amputation by Makin and colleagues (2013), and were instructed to move their phantom limb. Interestingly, significant activation in the SI was found, located at the cortex associated with the amputated limb (identified as contralateral to the intact limb). Contrary to the maladaptive model of phantom limb sensations, they found that greater phantom pain was associated with greater local activity within the phantom cortex. Similar to the synesthesia cross-activation theory, hyperactivation in a sensory area (SI) is associated with an additional sensation (phantom limb pain).

However, painful phantom sensations are likely to be related to other neural correlates than non-painful phantom sensations, as shown by Bolognini and colleagues (2013). They make the distinction between phantom sensations with and without the presence of pain. When using anodal (positive) transcranial direct current stimulation (tDCS) on the primary motor cortex (M1), phantom pain was temporarily reduced. However, non-painful phantom sensations were found to be reduced by cathodal (negative) tDCS of the posterior parietal cortex (PPC). This indicates that painful and non-painful phantom limb sensations are dissociable.

(18)

Similar to what these studies indicate, Flor and colleagues (2006) review a large amount of literature and draw the conclusion that functional SI reorganization seems to be related to painful phantom sensations only. Thus, hyperactivation in SI, as predicted by applying the synesthesia cross-activation theory to phantom limb sensations, was only found to be associated with phantom limb pain, and not with painful phantom limb sensations. If this hyperactivity is not found in non-painful phantom limb sensations, the cortical hyperactivity in phantom pain is probably more related to correlates of pain than correlates of additional sensations. If it were related to additional

sensations, this SI hyperactivity would have also been expected in non-painful phantom limb

sensations. If SI hyperactivity is more related to pain processing than additional sensations, a parallel with the synesthetic cross-activation theory is unlikely.

The modality of the additional sensation in CBS is visual; applying the functional cross-activation theory would predict hyperactivity in visual cortex, in the absence of external visual stimulation. As previously mentioned, systematic investigation of CBS is lacking. However, several case studies mention hyperactivation in the visual cortex (Kazui et al., 2009; Meppelink, de Jong, van der Hoeven, & van Laar, 2010). Kazui and colleagues (2009) examined two CBS patients who

experienced complex hallucinations, including images of humans, animals, and landscapes. Using single photon emission computed tomography (SPECT), increases of CBF were found in V1 and V2 in both patients. Furthermore, Meppelink and colleagues (2010) described the case of a 50-year old CBS patient who experienced visual sensations of changing colours flowing at different speeds and directions. FMRI revealed cortical activation in cortical areas related to colour and motion (V4 and V5, respectively) when she attended either motion or colour features of her hallucination. A previous study on colour imagery did not show V4 activation (Howard, Barnes, & McKeefry, 1998), so the neural correlates of this hallucination are likely different from those of visual imagery.

Thus, even though CBS research is very limited, there seems to be some evidence for activation of the visual cortex related to the additional visual sensations, and in absence of visual stimulation. This is in favour of a possible parallel with the functional cross-activation theory of synesthesia.

A 40-year old woman experienced sound-touch synesthesia after suffering from a

ventrolateral thalamic lesion, and was investigated with fMRI by (Beauchamp & Ro, 2008). Results showed hyperactivation of the secondary somatosensory cortex (SII) in response to an auditory stimulus. Similar to developmental synesthesia, the cortical brain area related to the concurrent is hyperactivated during synesthetic experience.

However, hallucinations after brain damage with a very similar phenomenology as developmental synesthesia seem to be a rarity. No other studies were found to investigate the functional brain properties of these multimodal hallucination patients. As mentioned in the previous section, unimodal hallucinations are often related to lesions in the brain areas associated with corresponding sensory modality. In sum, there is some tentative indication of similarities between

(19)

hyperactivation in brain damage related hallucinations and developmental synesthesia, but there is currently not enough literature in order to draw a proper conclusion.

(Kompus, Westerhausen, & Hugdahl, 2011) review existing literature on AVH neuroimaging studies and found consistent reports of reduced activity (hypoactivation) in PAC related to

schizophrenia patients with AVHs during external auditory stimulus presentation, but hyperactivation when they experienced AVHs during absence of external auditory stimuli.

Furthermore, (Kompus et al., 2013) present an fMRI investigation comparing non-clinical subjects with AVH, schizophrenia patients with AVH, and control subjects. The non-clinical AVH group showed PAC hypoactivity for speech sounds compared to controls.

Thus, similar to the hyperactivity of sensory brain areas during synesthetic experience, the experience of AVH seems to be associated with hyperactivity of the PAC. However, PAC hypoactivity is found when AVH patients are presented with external stimuli. In conclusion, there is support for PAC hyperactivation related to the additional auditory sensations during AVH, and in absence of auditory stimulation. There seems to be some similarities with the hyperactivation of sensory cortex during synesthetic experience.

Safety concerns restrict (neuroimaging) experiments with hallucinogens on humans. For example, (Johnson, Richards, & Griffiths, 2008) state that intoxicated subjects should not be exposed to potentially distressing environments. Using proper safety measures, (Carhart-Harris et al., 2012) investigated the effects of psilocybin, the hallucinogenic component of magic mushrooms, in an fMRI paradigm for the first time. A decrease in anterior cingulate cortex activation correlated positively with subjective reports on intensity of hallucinations. However, there is no report on abnormal activity in sensory cortical regions, making parallels with the synesthesia cross-activation theory unlikely. No other neuroimaging studies on the effects of hallucinogens have been found.

4.3 Conclusion

Although hyperactivation of the PAC in tinnitus has been found in several studies, other reports question the role of the PAC in tinnitus. Systematic research on both CBS and brain damage related hallucinations is absent, but case studies give tentative evidence for hyperactivity in the cortical areas related to the sensory modality of the hallucination. AVHs can evoke PAC hyperactivity in absence of auditory stimulation, similar to sensory cortical activation during synesthetic

experience. Thus, even though some phenomena with additional sensations are not properly

investigated neurobiologically, there is tentative evidence for hyperactivation in sensory brain areas in tinnitus, CBS, brain damage related hallucinations, and AVHs. Non-painful phantom limb sensations and drug-induced hallucinations are highly unlikely to have neural mechanisms similar to those implied by the synesthesia cross-activation theory.

(20)

Table 3: Is hyperactivation within or between sensory brain areas, as found in synesthesia, also present in other phenomena involving additional sensations? Similarities with synesthesia findings are shown in bold.

Tinnitus

- Some experiments found evidence in favour of PAC hyperactivity

- Some experiments found no evidence of PAC hyperactivity

Phantom limb sensations

- Some experiments found evidence in favour of SI hyperactivity related to phantom pain, but not to non-painful phantom sensations

- Therefore, SI hyperactivity is more likely related to pain processing than synesthesia

Charles Bonnet syndrome -_- Currently no systematic neuroimaging research present_{Case studies suggest hyperactivation of the visual cortex}

Brain damage related hallucinations

- Currently no systematic neuroimaging research present - One case study found evidence in favour of hyperactivation

in SII during synesthetic-like tactile hallucination

Auditory verbal hallucinations

- Some experiments found evidence in favour of PAC hyperactivity

- Some experiments found evidence against PAC hyperactivity during external auditory stimulus presentation

Drug-induced hallucinations

- Very limited neuroimaging research

- One experiment found no evidence of hyperactivation in sensory cortex

(21)

5. Comparison of structural and functional connectivity of the parietal

cortex

5.1 Is there structural or functional hyperconnectivity in the parietal cortex in

synesthesia?

The parietal cortex plays an important role in cross-modal binding in non-synesthetes. For example, Bien and colleagues (2012) found that integration of auditory and spatial information was impaired on a behavioural task after disrupting the right intraparietal sulcus (IPS) with TMS. In this chapter another neurobiological model of synesthesia will be addressed, one which highlights the role of the parietal cortex. This model is first described by Esterman and colleagues (2006). They investigated two colour-grapheme synesthetes who were performing a behavioural test, where their synesthetic experience interfered with task performance. After rTMS of the right parietal cortex, the synesthetic interference decreased and performance increased. The authors suggested a

‘hyperbinding’ theory of synesthesia, where synesthetic experiences originate from increased activation of parietal binding mechanisms.

Numerous studies have highlighted the important role of the parietal cortex in synesthetic experience. The review of Rouw and colleagues (2011) list several whole-brain studies that report activation of the posterior (superior and inferior) parietal cortex during synesthetic experience. Furthermore, Rouw and Scholte (2007) found greater connectivity in the parietal cortex of

synesthetes as compared with control subjects. In accordance, the resting state EEG study of Jäncke and Langer (2011) showed that coloured-hearing synesthetes had stronger interconnections within the parietal cortex than control subjects.

In the following section, the role of the parietal cortex in other phenomena with additional sensations will be evaluated. More specifically, the posterior parietal cortex (PPC) will be reviewed. If similar increases in structural and functional connectivity are found, the hyperbinding theory could play a role in creating additional sensations in general.

5.2 Is there structural or functional hyperconnectivity in the parietal cortex in

other phenomena involving additional sensations?

5.2.1 Tinnitus

Results from several tinnitus neuroimaging studies highlight the importance of the parietal cortex (Giraud et al., 1999; Plewnia et al., 2007; Schlee et al., 2009). As previously mentioned, Plewnia and colleagues (2007) used PET to find tinnitus-related increases in CBF in the inferior parietal cortex. Additionally, tinnitus loudness was reduced after rTMS on inferior parietal cortex. This suggests that higher-order processing is crucial in tinnitus. Furthermore, Giraud and colleagues (1999) investigated tinnitus patients who could modify tinnitus loudness with eye movements using PET. They found increased CBF in the temporoparietal cortex to be related to tinnitus sensations. This is again in parallel with the synesthesia parietal hyperbinding theory. In addition, Schlee and colleagues (2009) used magneto-encephalography (MEG) to investigate long-range projections in tinnitus patients and control subjects. Results indicated a flow from higher cortical areas, including the

(22)

parieto-occipital cortex, towards the temporal cortex, with a positive correlation between the flow and tinnitus loudness.

These studies suggest tinnitus-related parietal modulation located in posterior parts (close to the temporal or occipital cortex), similar to the synesthesia hyperbinding theory. However, the majority of tinnitus neuroimaging studies do not report any abnormalities in parietal functioning. This leads to the conclusion that a similar hyperbinding mechanism might be possible in tinnitus, but that existing literature is inconclusive. An overview comparing tinnitus and all other phenomena with synesthesia can be found in table 4.

Several studies have implicated that parietal hyperbinding could also play in phantom limb sensations (Bolognini et al., 2013; Töpper, Foltys, Meister, Sparing, & Boroojerdi, 2003). As previously mentioned, Bolognini and colleagues (2013) found a reduction of non-painful phantom limb

sensations during cathodal tDCS of the posterior parietal cortex (PPC). They conclude that PPC hyperactivation is associated with non-painful phantom limb sensations. This is very similar to the experiment by Esterman and colleagues (2006), where synesthetic experiences were abolished by parietal disruption. In addition, Töpper and colleagues (2003) investigated the phantom pain of two patients with cervical root avulsion with TMS. They found temporary reduction of phantom limb pain after rTMS of the PPC. No such effect was found in the four control subjects, where pain was

experimentally induced. This suggests a role for PPC unique in phantom pain, and not pain in general. These studies argue in favour of parietal hyperbinding in phantom limb sensations. They also successfully exclude the possibility of PPC involvement to be related to pain processing in general, and not to the additional sensation. However, results from most studies highlight the role of the somatosensory cortex (as discussed in previous chapters) in phantom sensations, and do not mention PPC abnormalities. Those results would argue against a parietal hyperbinding mechanism in phantom limb sensations. It is however important to notice that the studies that did describe PPC involvement specifically excluded the possibility of it being related to more general pain mechanisms. Thus, it can be concluded that there is indication for a parietal hyperbinding mechanism in phantom sensations, but further research is necessary.

The CBS patient of Meppelink and colleagues (2010) showed decreased glucose metabolism located in the superior parietal cortex, measured with PET. However, this hypometabolism was not limited to the parietal cortex, and also included occipital and thalamic regions.

As mentioned before, neuroimaging research on CBS is lacking. There are no known theories on CBS that involve parietal involvement in CBS. This case study alone does not give sufficient evidence in order to propose a role of the parietal cortex in CBS, similar to the synesthetic hyperbinding mechanism.

(23)

As mentioned before, brain damage related hallucinations are often paired with lesions in the pathway of the sensory modality of the hallucinations (Braun & Dumont, 2003). There is currently no theory present that predicts parietal involvement in brain damage related hallucinations. Although systematic research is lacking, case studies do not mention any indication of parietal hyperbinding playing a role in brain damage related hallucinations.

Vercammen and colleagues (2010) performed a resting state fMRI experiment on

schizophrenia patients with AVH and control subjects, and found decreased functional connectivity between the TPJ and frontal and temporal brain areas. This finding of decreased connectivity makes parietal hyperbinding in AVH unlikely.

As previously mentioned, synesthetes can be subdivided in ‘associator’ and ‘projector’ synesthetes. Interestingly, Plaze and colleagues (2011) divided schizophrenia patients with AVHs into similar groups: the first group experienced AVHs ‘inside the head’ (similar to associator synesthetes), while the second group experienced the AVHs as if they were in external space (similar to projector synesthetes). Using VBM, they found decreased white matter volume located in the superior

temporal gyrus (STG), near the TPJ, in the ‘external AVH’ group compared to a control group, while an increase in STG white matter was found in the ‘internal AVH’ group. Although Rouw and Scholte (2010) found greater connectivity in the superior PPC for synesthetes versus control subjects, differences in mechanisms for projector and associator synesthetes were not located in the parietal cortex.

Thus, there is some indication of parietal involvement in AVH. However, opposite to the increased connectivity found in synesthesia, AVH is more likely related to a decrease in parietal connectivity. Additionally, while there seems to be a parallel in phenomenological distinction in both AVH and schizophrenia between ‘internal’ and ‘external’ sensations, differences between groups can only be related to white matter changes in the STG (near the TPJ) in AVH.

As mentioned before, Carhart-Harris and colleagues (2012) were the first to investigate the effects of psilocybin using fMRI. They do not report abnormalities in parietal functioning. No other neuroimaging studies on the effects of hallucinogens were found. This leads to the conclusion that there is currently no indication of functional hyperbinding during drug-induced hallucinations.

5.3 Conclusion

Both tinnitus and non-painful phantom limb sensations have been related to hyperactivation in the PPC, although these findings are far from conclusive. In contrast, AVH seems to be associated with a decrease in structural parietal connectivity. CBS, brain damage related hallucinations and drug-induced hallucinations have not been thoroughly investigated yet, and there is currently no reason to

(24)

suspect parietal hyperbinding in these phenomena. In conclusion, parietal hyperbinding as found in synesthesia is only potentially present in tinnitus and phantom limb sensations.

(25)

Table 4: Is increased structural or functional connectivity of the parietal cortex, as found in synesthesia, also present in other phenomena involving additional sensations? Similarities with synesthesia findings are shown in bold.

Tinnitus - Some evidence in favour of parietal hyperconnectivity

- Some evidence against parietal hyperconnectivity

Phantom limb sensations - Some evidence in favour of parietal hyperconnectivity, specifically related to non-painful phantom sensations

Charles Bonnet syndrome

- Currently no systematic neuroimaging research present - One case study found evidence against parietal

hyperactivation

Brain damage related hallucinations

- Currently no systematic neuroimaging research present - Case studies found no evidence of hyperconnectivity in

parietal cortex

Auditory verbal hallucinations - Some evidence against parietal hyperconnectivity

Drug-induced hallucinations

- Very limited neuroimaging research

- One experiment found no evidence of hyperactivation in parietal cortex

(26)

6. Discussion

6.1 Relating the findings to the research question

The current comparison between neural correlates of synesthesia and other phenomena involving additional sensations gives no strong arguments in favour of a more general neural

mechanism underlying all additional sensations. The neural correlates of the six phenomena included in this thesis show little resemblance with the three suggested neural mechanisms of synesthesia. An overview of all comparisons is shown in table 5.

6.1.1 Evidence for structural cross-activation in other conditions

The synesthesia cross-activation theory suggests increased structural connectivity within or between sensory cortical areas. Tentative evidence for hyperconnectivity was only found for tinnitus and phantom pain. However, in tinnitus this hyperconnectivity seemed to be located more in subcortical parts of the auditory pathway, such as the IC and MGN. Additionally, other experiments found no indication of increased structural connectivity in the auditory pathway. Therefore, there is not enough evidence for a similar cross-activation mechanism in tinnitus. In phantom limb

sensations, SI hyperconnectivity was related to painful sensations only. It is therefore more likely to be related to processing of pain than processing of additional sensations. Also, no evidence of cross-connections between phantom limb cortex and cortex of inducing body part has been reported. Thus, the evidence supporting the structural cross-activation theory is limited, and therefore not sufficient to assume parallels in underlying neural mechanisms between synesthesia and other phenomena involving additional sensations.

6.1.2 Evidence for functional cross-activation in other conditions

The synesthesia cross-activation theory can also be interpreted as a functional mechanism where hyperactivity within or between sensory cortical areas is related to the synesthetic experience. Hyperactivation in sensory brain areas has been reported in all discussed phenomena, except in drug-induced hallucinations. However, experimental results in tinnitus research are contradictory; while some investigations show primary auditory cortex (PAC) hyperactivity, others do not. In phantom pain, hyperactivity in SI seems to be related to painful phantom limb sensations only, similar to SI structural hyperconnectivity. However, no co-activation between phantom limb cortex and cortex of inducing body part has been reported. Case studies of Charles Bonnet syndrome show hyperactivity of visual cortex during hallucinations, when external stimuli are absent. Unfortunately, systematic research is lacking. Similarly, functional properties of brain damage related hallucinations are also not properly investigated. Although one case study reported SII activation during hallucinatory touch experience, this evidence is insufficient for suggesting a cross-activation mechanism in brain damage related hallucinations. AVH is however often related to PAC hyperactivation. At first glance, it may seem as if the functional cross-activation theory is a potential candidate for explaining several forms of additional sensations. But closer examination reveals that, with the exception of AVH, the evidence in favour of functional cross-activation within or between sensory brain areas holds methodological limitations and is inconclusive in all conditions.

(27)

Table 5: A summary of the findings: Comparing neurobiology between synesthesia and other phenomena involving additional sensations. Similarities with synesthesia neurobiology are shown in bold.

Increased structural connectivity within or between sensory brain areas

Hyperactivation within or between sensory brain areas

Increased structural or functional connectivity of the parietal cortex

Ti n n it u s

- Investigated and parallels found - Sub-cortical instead of cortical

- Investigated and parallels found

- Investigated and parallels found P h an to m lim b se n sa ti o n s

- Investigated but no parallels found with non-painful phantom limb sensations

- Investigated but no parallels found with non-painful phantom limb sensations

- Investigated and parallels found with non- painful phantom limb sensations

C h ar le s B o n n et Sy n d ro m e

- Not sufficiently investigated

- Highly unlikely

- Parallels are likely present

- Highly unlikely B ra in d am ag e re la te d h al lu ci n ati o n s

- Highly unlikely

- Highly unlikely - Not sufficiently investigated

A u d it o ry ve rb al h al lu ci n ati o n s

- Investigated but no parallels found

- Investigated and parallels found

- Investigated but no parallels found D ru g-in d u ce d h al lu ci n ati o n s - Not investigated

- Highly unlikely - Not sufficiently investigated - Not sufficiently investigated

IC = Inferior Colliculus, MGN = Medial Geniculate Nucleus, PAC = Primary Auditory Cortex, SI = Primary Somatosensory Cortex, SII = Secondary Somatosensory Cortex

Comparing synesthetic experience with other additional sensations: possible parallels between underlying neural mechanisms