What is synaesthesia?
Synaesthesia is a cross-sensory phenomenon where one stimulus elicits a perceptual experience of another sensory area. To a synaesthete, listening to a particular tone may activate a sour taste on the tongue, or the letter A may consistently take on the colour red. The stimulus that causes the experience is known as the inducer, while the actual experience is called the concurrent (Hubbard, 2007). Agreement on the prevalence of synaesthesia seems uncertain, although most recent estimates indicate it is reasonably common, occurring in roughly 4.4% of the population (Hochel & Milan, 2008; Mattingley, 2009). There is also some debate about whether it occurs more often in women than in men, with some studies indicating this is true, and others showing no difference between the sexes (Hochel & Milan, 2008; Mattingley, 2009).
Types of synaesthesia
Among synaesthetes, perceptions involving colour are the most common (Eagleman & Goodale, 2009). Of these, cases of grapheme-colour synaesthesia are predominant, with estimates ranging from around 70% to 90% (Mattingley, 2009; Ward, Tsakanikos, & Bray, 2006). Coloured- hearing synaesthesia is reasonably well-reported, with one study estimating occurrence in about 20% of coloured synaesthesia cases (Ward et al., 2006). Synaesthesia involving concurrents other than colour are less frequently reported in the literature, although there have been several cases of music to taste (Beeli, Esslen, & Jancke, 2005), words to taste (Ward, Simner, & Auyeung, 2005), and tones to textures (Pearce, 2007). It is also possible for one person to experience more than one type of synaesthesia, with Ward et al. (2006) recording two synaesthetes who, according to self-reports, each experienced grapheme-colour, sound-colour, and taste-colour synaesthesia, among others. Regardless of how many types of synaesthesia exist, the experience for each individual is unique - while one person will always experience 'A' as red, another will consistently perceive it as gold.
Research Findings
Synaesthesia examples have been reported since the 1800s, however until recently there was controversy over whether it was a real perceptual phenomenon or a learned memory association (Hochel & Milan, 2008). Part of the problem was the lack of empirical research, with most studies focusing on cataloguing case-studies of self-reported subjective experiences (Mattingley, 2009). The first empirical studies looked at the consistency of the phenomenon, testing whether a particular synaesthete's experience was constant over time (Hochel & Milan, 2008). While most studies found a consistency of close to 100%, this was not considered evidence of a real perceptual experience, as it provided no indication of the neural processes underlying the process (Hochel & Milan, 2008).
In the 1980s researchers set out to resolve this issue by giving grapheme-colour synaesthetes a modified version of the Stroop task, where participants were asked to name the print colour of various letters, some coloured congruently with their reported synaesthesia colour for that letter, others incongruently. Synaesthetes were significantly slower to name incongruent colours than controls, indicating the synaesthesia experience was involuntary (Hochel & Milan, 2008; Mattingley, 2009). Many similar experiments have since been performed with the same results, including stroop-like tasks with coloured-hearing synaesthesia (Ward et al., 2006) and even tone-gustatory synaesthesia (Beeli et al., 2005). However, these results are still insufficient to distinguish between real perception and learned association, as it has been shown that interference effects can also occur in normal participants from a well-learned memory association (Hochel & Milan, 2008; Mattingley, 2009).
To overcome this limitation, Ramachandran and Hubbard (2000) used two clever experiments that strongly imply synaesthesia, at least in the grapheme-colour form, is a genuine perceptual experience. In the first experiment they showed matrixes of number pairs to synaesthetes and normal controls. The matrixes were designed to take advantage of perceptual grouping, or the tendency to group items together according to similar features. It was found that synaesthetes tended to group the matrix of numbers together according to their induced colours, while controls tended to group the matrix of numbers on the basis of shape (Ramachandran & Hubbard, 2000). In the second experiment, they showed participants an apparently random monochrome placement of similarly-shaped numbers or letters, with the placement of one of the letters or numbers forming a hidden shape. They found synaesthetes were significantly faster than controls at finding the hidden shape, indicating their induced colours made the shape pop out from the rest of the image (Ramachandran & Hubbard, 2000).
Brain Imaging Studies
Brain imaging studies have also revealed differences between synaesthetes and normal controls, particularly for synaesthetes who experience induced colours. Studies using PET scans have shown unusual activity in several areas of the synaesthetic brain associated with vision, including the posterior inferior temporal cortex and the parieto-occipital junctions, the former being associated with colour/shape integration (Pearce, 2007). To support these findings, two recent TMS studies have found that stimulation of the right parieto-occipital junction produces a reduction of interference in stroop-related tasks in synaesthetes, as would be expected if this area of the brain was involved in synaesthetic perceptions (Esterman, Verstynen, Ivry, & Robertson, 2006; Muggleton, Tsakanikos, Walsh, & Ward, 2007).
Finally, several fMRI experiments have implicated the left-hemisphere colour processing areas of V4/V8 in colour-induced synaesthesia (Hochel & Milan, 2008; Mattingley, 2009; Pearce, 2007). Activity in these areas was present in synaesthetes but not controls when they were presented with tones or spoken words (Hochel & Milan, 2008). Interestingly, the majority of experiments so far show no activity in visual areas V1 or V2, implying that the perceptual processing occurs higher up in the visual system (Hochel & Milan, 2008).
Theories of Synaesthesia
There are two main theories of synaesthesia. The first is known as 'cross-activation', and was initially postulated by Ramachandran and Hubbard (2001). The basis of this theory is that different cortical regions in the synaesthete brain have extra connections that do not normally exist. These extra connections could potentially have been present from birth, as newborns are born with many more neural connections than they will eventually require. During normal maturation, redundant or unused neural connections are eliminated in a process called neural pruning (Hochel & Milan, 2008). It is possible that a genetic mutation caused this normal pruning process to be impaired in synaesthetes, leading to cross-wiring in particular areas of the cortex (Hochel & Milan, 2008). A common criticism of this theory is that affected areas must therefore be next to each other in the brain in order to be cross-wired, and not all sensory experiences in synaesthesia fulfil this requirement. However, Hubbard (2007) argues this suggestion may not necessarily be true. According to Hubbard (2007), the critical feature is the presence or absence of early connections, not where in the brain the connecting regions exist.
The second major theory of synaesthesia concerns disinhibition of top-down sensory connections (Hochel & Milan, 2008; Hubbard, 2007). This theory suggests that the synaesthete brain may contain faults in the feedback system of multimodal sensory areas. Sensory perceptions initially travel along separate pathways until they converge at a multisensory nexus (Hubbard, 2007). When feedback signals from these multisensory areas are sent back to the lower sensory processing regions, a system of inhibition normally prevents the signals from going to the wrong sensory area (Hochel & Milan, 2008). However, it is possible that control of this feedback system is disinhibited in the synaesthete brain, allowing different signals from different sensory regions to become muddled together (Hochel & Milan, 2008; Hubbard, 2007). Support for this theory stems from evidence that synaesthesia-like experiences have been reported by non-synaesthetes under the influence of hallucinogens, something that would assumedly be impossible if the cross-activation theory were correct (Hochel & Milan, 2008). However, Hubbard (2007) points out that visions stemming from hallucinogens often contain detailed and complex scenes, while synaesthete visions contain only colour and movement. This could imply that the two experiences stem from separate neural mechanisms.
While current research into the genetic origins of synaesthesia is limited, there is some evidence that synaesthesia runs in families, although the specific type of synaesthesia can vary between family members (Hubbard, 2007). This has implications for any final theory of synaesthesia, as it implies that the multiple types of synaesthesia all have a common underlying infrastructure, even though the actual subjective experience may be very different.