By: K Chang

Synaesthesia is a neurological condition that does not interfere with normal day-to-day functioning1. Synaesthetic experiences occur when activation of one sensory modality, induces a second, involuntary sensory experience, different from the first1. Various models have been proposed, both neurophysiological and genetic, to explain this condition. The neural models are either anatomical, which focuses on structural differences in the brain between synaesthetes and non-synaesthetes and/or functional, which involve differences in how circuits in the brain work 2,3. Furthermore, the brain areas implicated in this condition vary depending on the form of synaesthesia. For instance, grapheme-colour synaesthetes have shown specific activation in the fusiform gyrus, which processes visual stimuli, and an adjacent area that is involved in the perception of colours, whereas in auditory-colour synaesthetes, associated brain areas are auditory areas and the fusiform gyrus1.

Currently, research in genetics involve linking specific forms of synaesthesia to certain genes and observing differences in expression levels or forms within families4,5. Popular research methods include BOLD fMRI for neural studies and genome mapping for genetic studies.

Researchers have also looked towards acquired forms of synaesthesia to help elucidate which neural models are more likely to be the underlying cause of synaesthesia2. Furthermore, several case studies have shown that synaesthetes have greater memory capacities than non-synaesthetes6. Therefore, if the underlying mechanisms can be discovered, then it can potentially be used to help patients with learning or memory deficits and/or expand the current abilities that people have now.

The video(s) below describe the life of Daniel Tammet, who has Asperger's syndrome and synaesthesia. Julian Asher, a scientist who studied Daniel Tammet and a synaesthete himself believes that Daniel's complex imagery (i.e. synaesthetic experiences) is the "key to his gigantic memory." Other synaesthetes with greater-than-normal memory are also presented in this documentary.

Part 2/5 begin at 7:15

Part 3/5

The Boy with the Incredible Brain: [Part 1] [Part 2] [Part 3] [Part 4] [Part 5]Entire documentary (for interest)

In addition, researchers in other fields have also used neural models in synaesthesia to help discern possible mechanisms for other neurological or psychiatric disorders and the basis of symptoms that arise from these conditions7. For instance, the knowledge gained from studying synaesthesia (including techniques) can also be used to better understand hallucinations or aberrant perceptions in schizophrenics.

Possible Neural Models

Increased Structural Connectivity

A popular anatomical model proposed to mediate synaesthetic experiences is the idea of increased projections existing between two (or more) areas involved in the sensory experience perceived by synaesthetes in comparison to non-synaesthetes. Increased projections then allow cross-activation of two or more brain areas following an appropriate stimulus.

Rouw and Scholte were the first scientists to validate that increased structural connectivity or “hyperconnectivity” contributed to synaesthetic experiences. With the use of diffusion tensor imaging (DTI) and fMRI, they were able to demonstrate that there is a significant increase in white matter tracts in grapheme-colour synaesthetes than non-synaesthetes1. The calculated fractional anisotropy (FA) was higher in four brain areas of synaesthetes, indicating increased fibre density and a stronger connectivity in these regions. It was found that the left and right frontal cortex, the right inferior temporal cortex, and the left superior parietal cortex, showed a greater fibre density in synaesthetes. The differences in white matter tracts are displayed in Figure 1.
Figure 1. Increased anisotropy in synaesthetic brains. Green indicates the normal white matter tracts on an MNI brain. Yellow shows where increased anisotropy was found. Higher FA values in the A) right inferior temporal cortex and B) left parietal cortex and both right (not shown) and left frontal cortices. (Image adapted from Rouwe and Scholte, 2007).

Synaesthetes and non-synaesthetes were asked to perform a grapheme task that would cause either a strong, weak or no synaesthetic experience while being monitored by BOLD fMRI1. Rouw and Scholte observed increased brain activation (or increased oxygenated blood flow) in several areas in close proximity to the hyperconnections found in the DTI study, but the area of most interest to the researchers was the high BOLD signal from the right inferior middle temporal lobe region. This region corresponded with the fusiform gyrus, which is known to process visual stimuli and the adjacent region mediates colour perception. Since increased brain activity was observed in the inferior temporal cortex during the synaesthetic experience and increased structural connectivity was also found in that region, it suggests that synaesthetic experiences are generated by hyperconnectivity (leading to cross-activation) between the areas involved with the sensory experiences. It is possible that other forms of synaesthesia are mediated in the same manner.

With increased projections between brain regions, scientists also predicted that an increase in grey matter accompanying the increase in white matter tracts7. Using voxel-based morphometry, this hypothesis was supported. Increased grey matter was found at the intraparietal sulcus (IPS) and the right fusiform gyrus, but not in the left fusiform gyrus for grapheme-colour synaesthetes. Structural differences in synaesthetes are further supported by a study conducted on a patient who has both taste-tone interval and tone-colour synaesthesia, E.S. He showed greater connectivity and increased white and grey matter in the areas associated with audition, taste and colour perception8.

In a more recent study, a group of scientists took a different innovative approach to investigating the neural mechanism of synaesthesia. The combined use of both MRI surfaced-based morphometry techniques and graph-theoretical network analysis of 154 anatomical structures were used to quantify cortical thickness of the whole brain3. Not only did the results show an altered connectivity in the fusiform gyrus and intraparietal sulcus in grapheme-colour synaesthetes (compared to non-synaesthetes), but a global hyperconnectivity in the brain demonstrated by reduced small-worldness mediated by increased clustering of local anatomical connections.

In contrast, some studies did not find a significant difference in the amount of structural connections between the areas that are thought to mediate certain synaesthetic experiences3. For example, the connectivity in the inferior temporal cortex, which was found to be altered in synaesthetes in Rouw and Scholte's study, were not significantly different in a different study.

Although some scientists have been able to show the existence of increased white matter tracts in synaesthetes, which support the anatomical model and have implicated specific areas that mediate synaesthetic experiences, they do not discredit functional models like the disinhibited feedback theory.

Possible Molecular Pathways?

Bargary and Mitchell in a review article suggested that defects in one of the three specific processes related to neurodevelopment could result in increased structural connectivity between brain areas responsible for the perception or processing of the sensory information involved in the synaesthetic experiences. The three proposed pathways are: neural pruning, axon guidance, and border formation9. Although, these pathways are proposed to produce hyperconnectivity in the brain by either failing to eliminate synaptic connections or make appropriate synapses, these claims have yet to be validated experimentally. In addition, scientists have not been able to narrow down which molecular mechanisms within these neurodevelopmental pathways are altered to generate synaesthesia. For instance, is the abnormality at the protein level or at the gene level? If there is a defect in the protein, is it a result of improper protein folding, creating reduced or non-functioning proteins or is there excessive degradation? Or is there a gene mutation in the promoter, impacting downstream events? Many of these questions have yet to be answered, but there is on-going research to find them.

Neural Pruning
Figure 2. Golgi stain of developing pyramidal neurons of a newborn to 2 years old. Complexity of synaptic connections increase overtime. (Image adpated from Courchesne et al., 2007).

Synaptic pruning is a natural regulatory process critical for proper neurodevelopment10. After birth, infants develop more neurons and form more synaptic connections (Figure 2)10. However as the brain matures (reaching adolescence), synaptic connections are eliminated or refined, decreasing the density of neural connections and neurons10. Bargary and Mitchell proposed that insufficient synaptic pruning allows for these synapses to remain, leading to greater connectivity between brain areas. Defective synaptic pruning during neurodevelopment is a plausible mechanism for the structural differences observed when comparing synaesthetes and non-synaesthetes9.

Axon Guidance

Many components are required for proper axonal pathfinding. The expression of chemoattractants or chemorepellants and its receptors as well as the timing of expression on the growth cone are critical for target-selection of a growing axon or growth cone9. Failure in any of these processes can result in improper axonal guidance and as a result, synapses can be formed with incorrect targets9. This is a possible explanation for the increased connectivity between brain areas observed in synaesthetes, not normally found in normal people.

Border Formation

The expression of certain ligands in distinct parts of the brain tissue such as Ephrins helps direct different cell types to specific locations in the cortex9. A border forms when neighbouring regions express different ligands. In axon guidance, growth cones rely on the expression of these ligands in order to reach their target area. If a border (or axonal barrier) is not formed, then the region can be inappropriately innervated by cells as shown in Figure 39. The innervation of 2 (or more) areas in the brain by neurons from one region of the cortex is suggested to mediate the cross-activation of sensory areas involved in the synaesthetic experience.

Figure 3. The importance of border formation in neurodevelopment. Red and green colours represent different ligands or regions of the brain. In the absence of an axonal barrier, neurons born in a separate region makes synapses in both red and green areas, while in normal development, these cells should only be making synapses in the green region. (Image adapted from Bargary and Mitchell, 2008).

Disinhibited Feedback

The disinhibited-feedback model differs from the increased structural connectivity model in that it focuses on functional differences between neural connections rather than abnormalities in the connections itself. It suggests that the anatomical connections in the brain are the same between synaesthetes and non-synaesthetes, but the activity (i.e. function) in these connections are different, resulting in synaesthetic experiences11.

Multimodal convergence areas receive sensory signals from various pathways (bottom-up signalling). Sensory information for one sense is projected to multimodal convergence areas, interacting with other types of stimuli. In the disinhibted feedback theory, this occurs, but information also propagates back down (top-down signalling) to activate pathways for a different sense, leading to a synaesthetic experience11. For example, in tone-colour synaesthesia, the auditory pathway is activated from the perception of sound and it feeds into these multimodal convergence areas. Once in these areas, activation of a second, visual pathway occurs by reciprocal feedback connections2. Normally, in non-synaesthetes, reciprocal feedback is inhibited, preventing top-down signalling and the perception of 2 (or more) concurrent senses (induced by just one sense) 2. This neural model suggests that synaesthetic experiences are mediated through the disinhibition of reciprocal feedback connections, allowing a concurrent sensory pathway to be activated.

Kadosh et al. tested the idea that synaesthetic experiences can be induced if brain function was altered in non-synaesthetes via posthypnotic suggestion (PHS). PHS was used because previous research on this technique showed that participants had no recollection of being asked to follow specific instructions11. Therefore, if synaesthesia was induced, it was not was because the participants remembered what they learned beforehand or what the researchers instructed them to do. In this study, the PHS group was told to associate the numbers 1-7 with specific colours under hypnosis. For instance, the number 4 was to be associated with the colour green. The same participants then had their PHS eliminated under hypnosis (i.e. subjects were told to ignore the digit-colour associations).

When a black digit was presented, participants that received PHS saw a coloured digit as indicated in Figure 2A. In addition, the perception of the colour to number was not random, it was the correct digit-colour association learned prior to the digit detection task. However, when the same experimental group was told to ignore the digit-colour associations under hypnosis (the non-PHS group), only a black digit was seen (Figure 2B).

Figure 4. The posthypnotic suggestion group experienced synaesthesia similar to those with actual grapheme-colour synaesthesia. A) Participants that received posthypnotic suggestion perceived the assigned colour to the digit, even when the digit was black. The image represents the induced synaesthetic experience (the digit presented to the participant was black). B) Without posthypnotic suggestion, they did not perceive any additional colour associated with the number. The colours associated with each number is indicated at the top of Figure A,B. (Image adapted from Kadosh et al., 2009).

Implications of this study:
If synaesthesia is mediated by increased structural connectivity, then PHS would not have been able to induce synaesthetic-like experiences in non-synaesthetes because they are not supposed to have hyperconnectivity in the brain. However, Kadosh et al. demonstrated that synaesthesia was inducible in non-synaesthetes. This suggests that communication between brain areas, normally inhibited in non-synaesthetes, are disinhibited after PHS. In addition, the researchers proposed that the formation of new connections between brain areas after PHS is unlikely due to the short time frame of the experiment. Although the induction of synaesthetic-like experiences in non-synaesthetes occurred, Kadosh et al did not directly show that disinhibition of reciprocal feedback connections occurred. A possible way to improve this study would be to use PHS in conjunction with other techniques such as BOLD fMRI and/or graph-theoretical analysis. That way, if hyperconnectivity was not found, but synaesthesia was induced, it would better disprove the increased structural connectivity model.

Genetic Basis

X-linked versus Somatic Chromosome-linked

Francis Galton was one of the first to scientifically recognize that there was a familial component to synaesthesia, suggesting a that there was a genetic basis for the occurrence of synaesthesia12. Family studies and pedigree analyses have also revealed a disparity in the ratio between female synaesthetes and male synaesthetes13. Further investigation on the modes of inheritance showed that transmission from father to son was significantly lower than mother-daughter, mother-son, and father-daughter transmission. These findings have been consistently shown across several studies (Table 1)13.

Table 1. Father-son transmission was found to be the lowest in three separate studies. A meta-analysis was also conducted, combining all of the information from previous studies. Transmission of synaesthesia from affected mothers or fathers to offspring was observed. The percentage of each mode of transmission is indicated. The bracketed values indicate the number of offspring over the total affected. (Image adapted from Ward and Simner, 2005).

Since father-son transmission was the lowest amongst the other types, combined with less males having synaesthesia, researchers have suggested that synaesthesia is an X-linked dominant trait13. This hypothesis was supported in a case study on female monozygotic twins, one with synaesthesia and one without. The results suggested that the observed discordance is due to differences in X-inactivation5. However, a study on male monozygotic twins contradicted this finding and the hypothesis that synaesthesia is X-linked. One twin had grapheme-colour synaesthesia, while the other did not12. Upon genotypic analysis, the two boys were confirmed to be monozygotic twins. Therefore, these findings reject the idea of synaesthesia being X-linked or that it could be X-linked, but there are other factors at work such as incomplete penetrance or epigenetics to create the discordant male monozygotic twins.

Furthermore, it has also been proposed that synaesthesia occurs less in males because lethality is associated with the X-linked dominant trait13. However, Ward and Simner failed to show a link between synaesthesia and lethality in males.

In contrast to the X-linked hypothesis for synaesthesia, other studies have shown specific types of synaethesia to be linked to somatic chromosomes. Performing a whole-genome scan and fine-mapping linkage study, Asher et al. were able to link auditory-visual synaesthesia to chromosomes 2q24, 5q33, 6p12, and 12p12. Auditory-visual synaesthesia was significantly linked to chromosome 2q24 (across all subjects), while the linkage to the 3 other regions were not as strong4. In addition, they were not able to find evidence of auditory-colour synaesthesia to be X-linked.

Ultimately, this study changes the perspective that current scientists have about this condition as well as the future approaches to the study of synaesthesia. Evidence for multiple sites of linkage for one type of synaesthesia reveals that the genetic basis of synaesthesia is more complex than previously thought4. Their findings demonstrate that the genetic basis of synaesthesia is not the result of a single gene, rather the interaction of various genes (with the possibility of epigenetics at work as well). It also makes discerning the modes of inheritance more difficult to answer as researchers now have to consider multiple loci. In addition, coloured-sequence synaesthesia has also been linked to chromosome 16q14. Therefore, researchers would also have to consider different forms of synaesthesia while genome-mapping, performing linkage studies, and investigating modes of inheritance.

Gene Mutations

As of right now, specific genes involved in synaesthesia have yet to be identified. Genes related to neurodevelopmental processes have been proposed to be altered in synaesthetes, but evidence to support these hypotheses remain unknown9. However, the outlook in this field is bright. With linkage-mapping revealing specific loci being strongly associated with certain forms of synaesthesia, it helps narrow the search for possible gene mutations. By comparing the genetic sequences of synaesthetes and non-synaesthetes in the regions shown to be linked to synaesthesia, researchers can locate variants within these loci and possibly find the gene(s) responsible for synaesthesia. If the variants are found in all subjects that have one form of synaesthesia and are not present in non-synaesthetes, a candidate gene can finally be proposed. However, more research still needs to be done in order to reach this stage.

Acquired Forms


Patient S.R. suffered from a thalamic stroke (ventrolateral nucleus lesion, where thalamocorticol tracts were disrupted), losing sensation from the entire contralateral side of her body15. 18 months after the initial injury, patient S.R. reported auditory-tactile synaesthetic-like experiences. Using BOLD fMRI, researchers were able to observe increased brain activity in the parietal operculum in response to sound, but no activity in response to somatosensory input (shown in Figure 3). Researchers propose that during recovery and the regeneration of the damaged white matter tracts, abnormal connections were formed between sensory areas (e.g. in the parietal operculum area). In this case, synaesthesia was caused by stroke-induced neural plasticity, where the formation of new connections in patient S.R. was incorrect. Consequently, aberrant connectivity supports the anatomical model of synaesthesia rather than the functional models like disinhibited feedback.

Figure 5. Brain activity in response to sound. A,C) Non-synaesthetes and B,D) patient S.R. Patient S.R. shows increased activation of the parietal operculum as indicated by the white arrows in D). Orange indicates brain activity and the black box indicates the thalamic lesion. (Image adapted from Beauchamp and Ro, 2008).


Ingestion of hallucinogenic drugs such as lysergic acid diethylamide (LSD) and/or mescaline as well as methamphetamines have been reported to induce synaesthesia in non-synaesthetes2,16. It is hypothesized that these drugs make use of existing neural connections, networks, and pathways to generate synaesthetic-like experiences2. Therefore, forms of drug-induced synaesthesia support the concept of altered function in synaesthetic brains rather than anatomical models. If synaesthetes have abnormal connectivity, synaesthetic induction via drugs would not occur in non-synaesthetes because they lack hyperconnectivity between brain areas.


Some scientists have been able to show participants (non-synaesthetes) having a form of learned synaesthesia through implicit associative learning17. Since normal participants had synaesthetic experiences, this supports the idea that synaesthetic brains are anatomically similar to non-synaesthetic brains, but the how these connections operate differ between synaesthetes and non-synaesthetes. It is proposed by these scientists that associative learning somehow alters function to result in synasthesia. However, other scientists have contested the occurrence of learned synaesthesia. In the video below, he explains one can learn or train themselves to associate colours to graphemes, but key components that characterize synaesthesia are lacking. The example that Ward gives is that non-synaesthetes who learn to associate red to the letter A, would not actually see the letter A in red. In addition, to think of the associations take time, which loses a key feature of synaesthesia, in that it is an automatic and involuntary experience. However, Ward does explain cases of acquired synaesthesia, but as a result of injury or loss of a sense.


  1. Rouw, R.K. & Scholte, S. Increased structural connectivity in grapheme-colour synesthesia. Nature Neuroscience. 10, 792-797 (2007).
  2. Grossenbacher, P.G. & Lovelace, C.T. Mechanisms of synesthesia: cognitive and physiological constraints. TRENDS in Cognitive Sciences. 5, 36-41 (2001).
  3. Hanggi, J., Wotruba, D. & Jancke, L. Globally altered structural brain network topology in grapheme-colour synesthesia. The Jounal of Neuroscience. 31, 5816-5828 (2011).
  4. Asher, J.E. et al. A Whole-genomes scan and fine-mapping linkage study of auditory-visual synesthesia reveals evidence of linkage to chromosome 2q24, 5q33, 6p12, and 12p12. The American Journal of Human Genetics. 84, 279-285 (2009).
  5. Smilek, D. et al. Synaesthesia: a case study of discordant monozygotic twins. Neurocase. 8, 338-342 (2001).
  6. Bor, D., Billington, J. & Baron-Cohen, S. Savant memory for digits in a case of synaesthesia and Asperger Syndrome is related to hyperactivity in the lateral prefrontal cortex. Neurocase. 13, 311-319 (2007).
  7. Weiss, P.H. & Fink, G.R. Grapheme-colour synaesthetes show increased grey matter volumes of parietal and fusiform cortex. Brain: A Journal of Neurology. 132, 65-70 (2009).
  8. Hanggi, J., Beeli, G., Oechslin, M.S., & Jancke, L. The multiple synaesthete E.S. - neuroanatomical basis of tone-colour synaesthesia. NeuroImage. 43, 192-203 (2008).
  9. Bargary, G. & Mitchell, K.J. Synaesthesia and cortical connectivity. TRENDS in Neuroscience. 31, 335-342 (2008).
  10. Courchesne et al. Mapping early brain development in autism. Neuron. 56, 399-413 (2007).
  11. Kadosh et al. Induced cross-modal synaesthetic experience without abnormal neuronal connections. Psychological Science. 20, 258-265 (2009).
  12. Smilek, D., Dixon, M.J. & Merikle, P.M. Synaesthesia: discordant male monozygotic twins. Neurocase. 11, 363-370 (2005).
  13. Ward, J. & Simner, J. Is synaesthesia an X-linked dominant trait with lethality in males? Perception. 34, 611-623 (2005).
  14. Nelson, et al. The genetics of colored sequence synesthesia: evidence of linkage to chromosome 16q and genetic heterogeneity for the condition. Nature Preceedings. doi:10.1038/npre.2009.3987.1 (2009).
  15. Beauchamp, M.S. & Ro, T. Neural substrates of sound-touch synaesthesia after thalamic lesion. The Journal of Neuroscience. 28, 13696-13702 (2008).
  16. Ahmadi, J., Keshtkar, M., & Pridmore, S. Methamphetamine induced synesthesia: a case report. The American Journal on Addition. 20, 306 (2011).
  17. Stevenson, R.J. & Boakes, R.A. Changes in odor sweetness resulting from implicit learning of a simultaneous odor-sweetness association: an example of learned synesthesia. Learning and Motivation. 29, 113-132 (1998).