Subjective reports and experimental evidence show that there is no imagination involved in the experiences of synaesthetes; they literally see letters or whole words as colours, or hear a symphony when someone familiar walks into the room. Moreover, the synaesthetic associations between the different sensory modalities involved are persistent, not random. As a result, any given stimulus will reliably induce the same effect in the `dependent' sense in an individual. This characteristic has formed the basis of a `gold standard' test for synaesthetes, discussed later.
Synaesthesia is unique in that it is perhaps the only psychological trait that `routinely inspires envy in those who study it experimentally' ; the majority of synaesthetes interviewed have said that they would not want to lose their rare form of experience. Whether or not synaesthesia confers more traditional advantages over non-synaesthetes, such as memory or intelligence, has provoked much interest - for example, could synaesthesia have been responsible for some part of Feynman's flair with physics? Small studies have demonstrated that some grapheme-colour synaesthetes can recall a number array with significantly more accuracy than non-synaesthetes, but this performance advantage was not shown in all synaesthetes tested. While anecdotal reports of synaesthetes possessing exceptional memory in facts or dates abound , there have been no large scale trials comparing the memory or intelligence of synaesthetes to non-synaesthetes, so currently we have no answer to synaesthesia's possible cognitive benefits.
Synaesthesia has been studied in one way or another since the nineteenth century, although only recently has it reappeared in the limelight following a number of pioneering experiments demonstrating that it was a genuine phenomenon. Currently there are many different aspects of synaesthesia being examined, from its possible genetic basis to how it develops in the brain. Synaesthesia is also proving to be a useful tool in investigating other psychological phenomena. This essay will first discuss the epidemiology of synaesthesia and then move onto a historical account of its study by science. Current theories on the developmental basis of synaesthesia in terms of neurology will be summarised, as well as more recent research.
Estimates vary for the prevalence of synaesthesia, partly due to a difference in opinion on what exactly constitutes synaesthesia, and partly due to different diagnostic methods. Figures ranging from 1 in 20,000 to 1 in 20 have been stated, while more recent studies suggest the prevalence is between 1 in 2000 to 1 in 200. Following a survey of synaesthetes conducted by placing advertisements in Cambridge Evening News and Varsity, a minimum prevalence of 1 in 2000 was estimated, over six times more females responded than males and a quarter of the respondents reported a first degree relative with synaesthesia . Taken as a whole, this survey provided strong evidence for synaesthesia having a genetic basis.
A leading hypothesis based on this survey suggests that a single dominant gene linked to the X chromosome could be responsible for synaesthesia . This would explain the sex bias observed in synaesthetes towards females, but would only account for a 3:1 ratio, not the 6:1 shown in data. However, if the `synaesthesia gene' was lethal in males, killing half of all male foetuses, the discrepancy between theory and data is handily removed. This theory consequently predicts that female synaesthetes should encounter significantly higher miscarriage rates than non-synaesthetes; as yet, this prediction has not yet been tested.
It was long believed that synaesthesia could not be due to a gene residing on the X-chromosome because Vladimir Nabokov, a noted synaesthete, had a son who was also synaesthetic. Hence, if Nabokov had managed to pass the trait to his son, it could not possibly reside on the X-chromosome as sons only inherit the Y-chromosome (of the sex chromosomes) from their fathers. It was later discovered that Nabokov's wife possessed coloured hearing and so was a synaesthete, solving the apparent mystery.
A real problem in determining the prevalence of synaesthesia is that many people either do not realise that they are synaesthetes, thinking that everyone experiences the world in the same way that they do, or they do not report their experiences at all. It has been suggested that there are more female synaesthetes reported simply because they are more willing to respond to surveys; while this factor may have affected the survey described above, the sex bias for synaesthesia is so dramatic that it could not account for it completely. Even so, further epidemiological studies of synaesthesia would be of great use.
The history of synaesthesia
Research into synaesthesia began in the 19th century with a classic report by Sir Francis Galton, in which he outlined the experiences of several synaesthetes he had studied with colour associations . Galton remarked that while the colour associations of individual synaesthetes were very stable over time, they were not shared between individuals, leading to rather heated arguments when more than one synaesthete was in the same room. That synaesthetic associations are highly idiosyncratic was an important finding that has continued to this day, and other major findings of that age included that certain forms of synaesthesia are more common than others. Most research conducted in the late 19th and early 20th centuries was based on single case studies and often the subjects were also being seen by psychologists for other issues such as migraines. Combined with the fact that practically all research was based on simple subjective reports with no proper testing, only a few broad facts were discovered about synaesthesia.
In the 1920s and 1930s, synaesthesia fell out of favour with psychologists as a consequence of the demise of cognitivism and rise of behaviourism. Behaviourism originated from a push by psychologists to establish the field as a rigorous, empirical science that could be studied by inferential statistics. To them, this dictated that only the observable, objective behaviour of a subject should be studied; subjective reports of conscious experiences were not to be trusted. The loss of interest in synaesthesia is readily made apparent by looking at the number of papers published on the subject; papers declined precipitously from the 1920s and only forty years ago did they begin to increase.
After fifty years of relative obscurity, cognitivism re-emerged in the 1960s. Exactly when behaviour's influence began to diminish cannot be pinpointed, but a significant event was Noam Chomsky's critical review of leading behaviourist B. F. Skinner's book on language acquisition, in which he revealed major problems in the behaviourist approach. Following the rise of cognitivism, psychologists and neuroscientists were free to discuss the inner workings of the mind without fear. Even so, synaesthesia had been forgotten by many psychologists.
Tests for `genuineness'
The turning point in the revival in synaesthesia research at the end of the 20th century can arguably be attributed to a number of experiments carried out by Prof. Simon Baron-Cohen, in which he determined that synaesthesia was a genuine phenomenon, primarily through a test for the persistency of synaesthetic grapheme-colour associations.
Today, there are several tests used to detect synaesthesia in subjects. Since there are in theory as many different types of synaesthesia as there are combinations between the five senses, a battery of tests must be used to cover all possibilities. In practice, two forms of synaesthesia predominate over the others - grapheme-colour association, and hearing-colour association. The former has been the subject of the most attention due to the ease of devising and controlling experiments.
For some time it was believed that synaesthesia was not a `real' psychological condition and that subjects were merely confabulating - making up - the entire experience, or that they were speaking metaphorically (e.g. 'She has a very sharp voice.'). Prof. Simon Baron-Cohen's `gold standard' test changed this markedly. Baron-Cohen's test was simple: he recorded subjects' grapheme-colour associations, and then tested them using the same lists several months or years later . Synaesthetes performed significantly better than control subjects, showing that they could not be confabulating.
A later test by Ramachandran further established the legitimacy of synaesthesia . While Baron-Cohen's test illustrated the stability of grapheme-colour associations, it did not show that they were necessarily perceptual. In other words, Baron-Cohen's synaesthetes could have had a photographic memory and simply memorised the associations between the graphemes and colours without actually experiencing them perceptually. Ramachandran's test employed `pop out', a phenomenon well known to psychologists. This phenomenon can be easily demonstrated by looking at a field of identical characters and asking subjects to pick out the characters that are anomalous, for example, by dint of having a different colour or shape to the rest. The anomalous characters appear to `pop out' to the subject, allowing for near instantaneous identification.
Figure 1: Displays used in Ramachandran's test. The box on the left shows a matrix of 5s and 2s. The similarity of the shapes of the numbers made it difficult for control subjects to find the embedded shape. The box on the right is a simulation of a synaesthete's experience, who will see the numbers as coloured. The embedded triangle, composed of 2s, pops out immediately.
Subjects taking Ramachandran's test were asked to state which shape they saw in the box on the left in figure 1. Normal subjects failed to see any shape, which was reflected in the test results. Synaesthetes, whose grapheme-colour associations had already been recorded and used to tailor the test, immediately saw the shape formed by the coloured photisms of the characters and successfully reported it significantly faster than normal subjects. Thus, this experiment showed that synaesthesia was in fact a perceptual phenomenon and synaesthetes really did see letters as coloured.
Such tests have vastly improved the ability of researchers to identify synaesthetes and characterise the trait. However, the causes and mechanisms of synaesthesia are far from understood. In the past twenty years, there have been several distinct theories advanced, the most significant of which are discussed below.
The first theory states that synaesthesia is nothing more than a set of associations learned early in life. This would suggest that grapheme-colour synaesthesia is caused by the synaesthete having played with coloured alphabet blocks as an infant, or perhaps having read books with coloured letters or words. While this appears to make sense intuitively, learned association cannot reconcile a number of findings, most notably that most synaesthetes cannot recall having learned their associations from coloured objects as infants, and that even if they are mistaken, their synaesthetic associations are not the same as those in any books or toys they had as infants. Another problem for this theory is that books and toys for infants with coloured alphabets tend to make letters coloured significantly differently from adjacent letters to aid discrimination; however, nearly all grapheme-colour synaesthetes report that their letters are not coloured in such a way. Instead, they are closer together in hue and less bright.
Furthermore, learned association cannot account for the sex bias observed in synaesthetes or the results of experiments conducted on its perceptual features. For these reasons, learned association is not seen as a useful explanation among synaesthesia researchers, although it may apply to the use of metaphor in language.
Dr. Richard Cytowic, a researcher who was at the forefront of the revival of synaesthesia in the 1980s, has proposed that synaesthesia is caused by different parts of the brain becoming disconnected from each other, allowing the `normal processes of the limbic system [a set of networks in the brain involved in instinct and mood] to be released, bared to consciousness, and experienced as synaesthesia.'  Cytowic's theory is not particularly appealing because it is inconsistent with current models of brain cognition, and its evidence relies on a type of brain scan (SPECT xenon-133 measuring regional blood flow) that cannot probe deep enough into the brain to actually show limbic system activity; hence, this theory too has not been pursued by the majority of synaesthesia researchers.
A more conventional theory recently put forward by Ramachandran and Hubbard states that anomalous cross-activation of different brain areas is responsible for synaesthesia, with the cross-activation being a product of a single gene mutation . During the development of the brain, many connections between neurones are made initially, and then subsequently these connections are pruned so that only some remain. Ramachandran and Hubbard's mutation would act early in the development of the brain by creating an excess of cross-connections or insufficient pruning of connections between brain areas; indirect evidence for this theory comes from imaging studies of prenatal monkeys . In the foetal macaque, there are significantly more connections between higher areas and visual areas than in adult macaques; thus, it is suggested, a mutation causing defective pruning could preserve those connections and cause synaesthesia. This fits in with the observation that children appear to experience synaesthesia at a higher prevalence than adults.
This cross-activation theory is not based solely on physical cross-wiring of neurones between different areas; it is acknowledged that various forms of disinhibition of existing feedback connections could also cause the same effect. This would account for the existence of `acquired' synaesthesia , which is believed to operate in the same way as `phantom limbs', in that existing connections between areas that do not otherwise directly exchange information are disinhibited when their associated input areas are disrupted (e.g. visual deprivation causing tactile input to activate visual areas).
Another somewhat related theory is based on the observation that there are vast numbers of reciprocal connections between neurones in different areas of the brain. Most neuroscience textbooks show a unidirectional flow of information from sensory processing areas (e.g. the visual cortex) to higher processing and multimodal areas (e.g. areas of sensory integration) without exploring the purpose of backward connections to sensory areas; this is unsurprising since they are almost always inhibited. Peter Grossenbacher has suggested that when the backward connections are disinhibited - when they convey information back to the sensory areas - synaesthesia will occur . This is not as similar as it appears to Ramachandran's theory, as Grossenbacher requires feedback connections all the way from higher areas to sensory areas, while in Ramachandran's theory cross-activation is only required at a more local level.
Grossenbacher believes that his theory can account for the observation that LSD elicits synaesthesia more satisfyingly than Ramachandran's theory (that presumes preserved neonatal pathways), claiming that on consumption of LSD, people `obviously aren't growing new connections in their brains. They're using connections we all have, but in a novel way.' This assumes that LSD induced synaesthesia is identical to traditional synaesthesia when in fact it is not; among the many differences, there are no persistent cross-modal associations and LSD users only experience hearing-colour synaesthesia.
Unsurprisingly, there has been insufficient evidence for any single theory to persuade all neuroscience and psychology researchers, although most tend towards some variation of the latter two theories discussed. Current research into synaesthesia is continuing to seek a resolution to this uncertainty by way of further imaging studies and psychophysical experiments. However, a growing minority of researchers are beginning to use synaesthesia as a tool to better understand other psychological phenomena. This last section of the essay will discuss the techniques being used in synaesthesia research today in addition to selected research topics.
Imaging studies using functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) have already been used in synaesthesia studies, mainly to see which areas of the brain are activated during synaesthetic experiences. These techniques measure the regional blood flow within the brain, which is thought to indicate the activity of brain areas. As the spatial and temporal resolution of these techniques improves, particularly in fMRI, they will prove increasingly useful in refining theories of synaesthesia.
In recent years, a new technique called transcranial magnetic stimulation (TMS) has allowed researchers to safely and non-invasively stimulate specific brain areas using what is essentially a sophisticated electromagnet. Previously, stimulation of the brain was only possible during brain surgery, which drastically limited its use. At least one group in the US is planning to use TMS to stimulate areas postulated to be responsible for synaesthesia to examine whether their synaesthesia is temporary altered or removed.
Psychophysical experiments into synaesthesia track the differences in performance of synaesthetes in specific cognitive and perceptual tasks to non-synaesthetes; the perceptual pop-out test displayed earlier is a prime example. By using existing experiments on synaesthetes to examine perceptual phenomena that are already reasonably well documented, for example, the various forms of visual masking (when a `signal' object - perhaps a grapheme or colour - is masked by a `distracter' object and the subject asked to identify the signal), it is possible to pinpoint exactly where synaesthetes' perception diverges from normal perception.
One study has shown that the colours experienced by grapheme-colour synaesthetes prevents a particular type of visual masking called object substitution, where a grapheme is displayed for a split-second and then replaced (`substituted') by another object in the same location . Control subjects are not able to identify the grapheme, but synaesthetes are, because the grapheme elicits an associated colour that they can identify. This study provides a good example of how research in synaesthesia is shedding light on other psychological questions; the mechanism underlying why object substitution prevents normal subjects from identifying the masked object (in this case, the grapheme) is not yet known, but the discovery that synaesthetes do not respond in the same way raises several interesting possibilities.
Despite the reports of Galton and numerous others stating that synaesthetes do not agree on their cross-modal associations, there have been a number of studies searching for trends among synaesthetes, particularly for word or grapheme-colour associations. Proof of even a weak trend has so far remained elusive, and some researchers believe that new theories of synaesthesia must be taken into account in order to properly look for trends. Once such theory suggests that there are higher and lower synaesthetes . Lower synaesthetes would possess sensory associations with graphemes or discrete experiences, whereas higher synaesthetes would have associations with concepts. For example, while a higher synaesthete might associate the colour red with all instances of the number 1 (e.g. the Arabic numeral `1', the roman numeral `I', a single dot on a die, etc.), a lower synaesthete would only associate red with one instance of the number (e.g. the Arabic numeral only). Under this theory, a lower synaesthete might have cross-activation between colour and grapheme areas, while a higher synaesthete would have a hypothetical `sequence' or `cardinality' area (possibly the angular gyrus) cross-activated with the colour area. The existence of a cardinality region has yet to be proven, although research is ongoing.
If there really are higher and lower synaesthetes, then attempts to find trends within a mixed group may be futile. Consequently, to produce valid results the groups must be examined separately. Other factors confound `trend studies' such as the fact that for some, word-colour associations are actually `first letter of the word'-colour associations. Some of the latest research in this area addresses whether there are any semantically related trends across synaesthetes .
An interesting psychological experiment carried out at the University of California, San Diego, used synaesthesia as a technique in itself to explore how we recognise graphemes. The idea for the experiment was inspired by a synaesthete who commented that the same grapheme would elicit different shades of the same colour when it was set in different typefaces. By displaying and iteratively modifying a number of typefaces, it was possible to find the typeface for which the synaesthete experienced the most vivid colour; this, it was thought, corresponded to the `prototypical' representation of the grapheme recognised by the synaesthete's brain. This simple experiment provided useful insight not only into how we recognise graphemes, but also objects and scenes.
In general, the more bizarre manifestations of synaesthesia have not been explored by researchers, partly because they are more difficult to verify and draw conclusions from. However, for the purposes of this essay one intriguing example is provided by a synaesthete (`EP') at San Diego. EP reported seeing a visual calendar every time she thought about dates. The calendar took the form of a linear coiled coil, where one horizontal rotation of the coil represented a year, and one vertical rotation represented a week. EP appeared to have an exceptional memory for the day on which events occurred, but not for the exact year and month. As yet, no research has been conducted on EP although, needless to say, exactly how her synaesthesia operates is baffling.
Synaesthesia somehow seems to short-circuit the normal perceptual and cognitive processes that occur in human brains to cause the strange experiences that have fascinated so many. In doing so, it gives us unprecedented access to the nature of those processes that underpin our thoughts and consciousness. As we increase our knowledge of the genetic, neural and cognitive aspects of synaesthesia, we will find that we are beginning to understand the brain more completely. Researchers may wish that they possessed synaesthesia, but being able to explore a new and strange trait that may hold the answers to many fundamental questions is reward enough.
1 Feynman, R. (1988), `What do you care what other people think?' pp. 59. New York, Norton.
2 Downey, J.E. (1911), `A case of colored gustation'. American Journal of Psychology, 22, pp. 528-539.
3 Harrison, J. (2001), `Synaesthesia: the strangest thing', pp. 169-74. New York, Oxford University Press.
4 Farber, I. (2001), `...but what does `blue' smell like?', Nature, 410, pp. 744-5.
5 Luria, A. (1968), `The mind of a mnemonist.' New York, Basic Books.
6 Baron-Cohen, S., Burt, L., Smith-Laittan, F., Harrison, J., Bolton, P. (1996), `Synaesthesia: Prevalence and familiarity', Perception, 25 (9), pp. 1073-80.
7 Bailey, M. E. S., Johnson, K. S. (1997), `Synaesthesia: is a genetic analysis feasible?', in Synaesthesia: classic and contemporary readings (ed. S. Baron-Cohen & J. E. Harrison). Oxford, Blackwell Publishers, Inc.
8 Galton, F. (1883), `Inquiries into Human Faculty and its Development.' London, Dent.
9 Baron-Cohen, S., Wyke, M., Binnie, C. (1987), `Hearing words and seeing colours: an experimental investigation of synaesthesia', Perception, 16 (6), pp. 761-7.
10 Ramachandran, V. S., Hubbard, E. M. (2001), `Psychophysical investigations into the neural basis of synaesthesia', Proceedings of the Royal Society of London, B, 268, pp. 979-83.
11 Cytowic, R. E. (1993), `The man who tasted shapes', pp. 163. London, Abacus Books.
12 Ramachandran, V. S., Hubbard, E. M. (2001), `Synaesthesia: A window into perception, thought and language'. Journal of Consciousness Studies, 8 (12), pp. 3-34.
13 Kennedy, H., Batardiere, A., Dehay, C., Barone, P. (1997), `Synaesthesia: Implications for developmental neurobiology', in Synaesthesia: classic and contemporary readings (ed. S. Baron-Cohen & J. E. Harrison). Oxford, UK: Blackwell Publishers, Inc.
14 Armel, K. C., Ramachandran, V. S. (1999), `Acquired synaesthesia in retinitis pigmentosa', Neurocase, 5 (4), pp. 293-6.
15 Grossenbacher, P. G. (1997), `Perception and sensory information in synaesthetic experience', in Synaesthesia: classic and contemporary readings (ed. S. Baron-Cohen & J. E. Harrison). Oxford, Blackwell Publishers, Inc.
16 Wagar, B.M., Dixon, M.J., Smilek, D. & Cudahy, C. (2002), `Coloured photisms prevent object-substitution masking in digit colour synaesthesia.' Proceedings of the 12th annual meeting of Theoretical and Experimental Neuropsychology (TENNET 12), Brain and Cognition, 48, 606-611.
17 Ward, J., Simner, J. (2002), `Phoneme-taste synaesthesia: Linguistic and Conceptual Factors', (submitted).
This work is licensed under the Creative Commons Attribution-NoDerivs-NonCommercial License.