The authors made a good revision, and improved the manuscript. I just added few minor comments that will help to keep the review updated and balanced.
1) Please add the study by Ro et al. (2007, Annals of Neurology) to the part that mentions sensory deprivation.
2) The electrophysiological findings by Schiltz et al. (1999, The Journal of Neuropsychiatry and Clinical Neurosciences) and Cohen Kadosh et al. (2007, Journal of Cognitive Neuroscience) should be cited as counter evidence for the argument for the perceptual nature of synesthesia.
3) "The importance of number forms in studying cognitive processes in normal people was largely overlooked until recently (Ramachandran and Hubbard, 2001b)." It should be noted that Seron et al. (1992, Cognition) studied this issue before Ramachandran and Hubbard (2001b).
4) The studies by Smilek et al (2007, Consciousness and Cognition) and Price & Mentzoni (2008, Cortex) should be cited when you discuss the representation of months.
5) The behavioral and neuroimaging findings by Cohen Kadosh and colleagues should be cited in the section of bi-directionality.
6) Please update the reference by Beeli et al. (in press, Cerebral Cortex).
Response to Review 2
1) We have included the Ro reference
2) We thank reviewer 2 for this suggestion, but argue that these two studies are not relevant to the argument on the perceptual nature of synesthetia
Schiltz et al. 1999 utilyzes an oddball paradigm to compare group differences in event-related waveforms between synesthetes and controls. On showing group differences relating to the positivity of the wave over anterior sites, Schiltz et al. offer two possible explanations: inhibition or multisensory neurons. First, spatial information from EEG must be considered with the utmost care as the electrophysiological response projected by any grouping of cortical neurons can be detected at any EEG scalp site. Second, comparisons of waveform morphology and amplitude between groups can produce spurious, uninterpretable results. Third, the time window in which these effects are reported occurs after N1 modulations reported by Brang et al., 2008 and Beeli et al, 2008, which suggests these latent anterior effects (if these effects are in fact related to the synesthetic percept) are resultant of downstream modulation.
Cohen Kadosh et al. studied a unique case in which the synesthetic percepts are subjectively reported to be bidirectional in numerical and size comparison tasks with the aim of understanding the origin of this rare form. This study does not relate to the perceptual nature of synesthesia.
3) We have included the Seron reference
4) We have included the Smilek reference as this relates to the perceptual reality of timespace-synesthesia, but not the Price reference as it does not directly relate to any of the topics in this article.
5) We have included the Cohen Kadosh reference
6) The Beeli reference has been corrected
The authors made some changes, and I have a few more suggestions.
As there is growing evidence for bi-directionality in grapheme-color synesthesia, the authors should include this in the review.
The coordinates in Rouw and Scholte's study indicate that the area in the parietal lobe is the superior parietal lobule, rather than the angular gyrus. They also did not examine higher synesthetes but projectors vs. associators (and the parietal lobe did not show any distinction). In addition, numerical processing is believed to be implemented in the intraparietal sulcus rather than the angular gyrus.
The study by Cohen Kadosh et al. (2007) (cited here as Kadosh) indicates that it is the magnitude of the numbers rather than the linguistic frequency.
Response to Reviewer 2
We again appreciate the reviewers helpful comments on this article. We have added a short review on the area of "bi-directionality in synesthesia" as we agree with the reviewer that it is a pertinent topic. Discussion of Rouw and Scholte's work has been clarified to better represent the distinction between anatomical structures and forms of synesthesia. We have corrected the reference to Cohen Kadosh's paper to reflect the author's interpretation of numerical magnitude over grapheme frequency.
Overall I am very pleased to see that the authors made some of the changes that were suggested by myself and Reviewer 2. There ar estill some issues that I feel need to be tackled.
1) As I understand it, the purpose of the section on “Martian colours” is to illustrate the point that synesthesia is more than memory associations. Fine. But I still don’t think it merits its own section early on in the article. I think it would be better placed in one of the sections that deals with more general issues (nature/nurture, physiology) in which it could potentially make a more direct contribution. 2) I am still uncomfortable with the word ‘exotic’. Although I do understand the gist of it, I’d appreciate something more meaningful such as “Rarer types of synesthesia” or “Less studied varieties of synesthesia”. 3) Related to this, it isn’t obvious what the criterion is for ‘exotic’. For example, the touch-emotion synesthesia gets its own section (ditto for martian colours) but other types of synesthesia are bundled together as ‘exotic’. I don’t see the reason why touch-emotion synesthesia should be given its own unique section. Is it less exotic? It should be placed in the section presently labelled as exotic (that I sincerely hope will be re-labelled). Unlike other types of synesthesia, it has yet to be peer reviewed and we, as a community, do place importance on this (rightly or wrongly). I have noted the authors’ objections to peer review (and I have some sympathy with it), but at the end of the day we all have to work within these constraints – and that includes you. When I give authors specific advice (e.g. please downplay this section because it has not been peer reviewed and please add it to the ‘exotic’ section), I do not expect to be given a lecture on the History of Scientific Discovery (from Galileo to Mendel). 4) Please check again the sentence “How does one account for the observation that it is 7 times more common among poets and artists? (Rich, Bradshaw, Mattingley, 2005; Ward, Yaro, Thompson-Lake, Sagiv, 2007; Domino, 1989).” The first two of these studies investigated the occupations and hobbies of people with synesthesia, and did find that certain occupations and hobbies were indeed more common in synesthetes but they don’t cite the 7:1 ratio anywhere that I can see. Also note that the studies didn’t show that “synesthesia is more common in artists and poets” it showed that “art and poetry are more likely to be pursued by synesthetes” (a subtle but potentially important difference). The study of Domino (1989) did attempt to assess the prevalence of synesthesia in fine arts students and found a prevalence of 25%. This is almost 7 times the latest population prevalence rate (of 4%, Simner et al 2006), but the problem is that Domino relied on self-report of synesthesia. Simner et al. (2006) also found a self-report of about 25% giving a ratio of 1:1 when methods are equated. Some clarification about how you got this number and the nature of the studies you cite would be helpful.
Response to reviewer 1
We have restructured much of the article in accordance with reviewer 1's comments. The key change is a new category titled "Less studied variants and aspects of synesthesia" under which many of the sub-categories have been placed. We have also created a subheading for Touch-emotion synesthesia and removed the references to 'exotic forms.' This area also now includes the reference to our in-press paper on the topic.
In terms of art and synesthesia, we have changed this bullet to reflect the reviewers comments, and now reads "How does one account for the observation that poetry and art are more likely to be pursued by synesthetes?"
This is deeply disappointing, and a very poor counterpart to the wikipedia entry. If I were to mention every omission, then I would have to rewrite it myself. The authors need to read and cite the wider literature of their subject. My main comments are listed below and I give a few examples to illustrate my point (I could give more). This seems like a misuse of scholarpedia to promote one particular viewpoint and data from one particular research group, rather than present a fair and balanced overview of the field.
1) The authors rely on conference abstracts (often from their own group) when in fact there are many other published papers that could be discussed. For example, they mention their own conference abstract on synesthesia and memory but miss the three recent papers to have appeared in peer-reviewed journals (Smilek et al 2002, Mills et al 2006, Yaro & Ward, 2007) and some classic descriptions such as Luria (1968).
2) The article focuses on very rare and under-explored aspects of synesthesia, and completely neglects common types of synesthesia that are well documented (by other research groups). For example, an anecdotal report of ‘martian colors’ in a colorblind synesthete appears as a whole section early on. There is another complete section devoted to a type of synesthesia that has never been reported in a peer-reviewed journal (touch-emotion synesthesia). This comes at the price of a complete omission of types of synesthesia that have been well explored in peer-reviewed journals with thorough empirical studies: sound-vision, mirror-touch (or vision-touch), taste-color, lexical-gustatory… The list goes on.
3) The is no sense of which ideas are commonly agreed upon by the research community and which are more speculative. Some of the most commonly agreed on ideas are given short treatment (consistency, automaticity) whereas longer treatments are given to ideas that I have never heard of before (e.g. a serotonin theory of synesthesia) and that are by no means universally accepted.
4) A discussion of some of the most commonly asked questions about synesthesia is completely absent. How common is it? Is it more common in women, children, etc.? Which are the most common types? What do we know about its genetic basis? Can synesthesia be acquired? Can synesthesia be lost? Are the colors distracting? I think focussing on some of the most commonly asked questions and providing succinct answers about what we know (AND being honest about what we don’t know) is the way forward. Again, one must distinguish between speculation/ideas and empirical research.
5) Some things are plain wrong. There is no proper prevalence study that compares the prevalence of synesthesia in artists versus others. There is a self-report claim that synesthesia is reported in 25% of artists (by Domino). However, it is also self-reported by 25% of non-artists but that figure drops dramatically if given an objective measure (Simner et al 2006). These statements (“7 times more common in artists”) just can’t be included as facts. There is evidence that synesthetes are more likely to have art as an occupation/hobby (i.e. art is more common amongst synesthetes; Rich et al. 2005) but there is no reliable data about how common synesthesia is in artsists.
6) Some things are contradictory. They go to great lengths to show that synesthesia is a sensory phenomenon and then include sections that say that experiencing emotions from touch is synesthesia, or thinking of letters/numbers in terms of gender and personality is synesthesia. Either synesthesia is sensory and this is not synesthesia, or synesthesia can incorporate non-sensory associations which contradicts their original definition.
Response to Reviewer 1
We thank the reviewer for his/ her suggestions for revision. Some of his ideas are useful; others, gratuitous or even just downright silly. Scholarpedia performs an extremely valuable service to the academic community, but its practice of allowing referees to remain anonymous while having the review posted publicly is unfairly skewed in the referees favor since it means he/she can publish any nonsense that he chooses without exposing himself. There is no accountability of the kind you see in (say) BBS (which posts signed RESPONSES to an article, and replies, but not the initial anonymous reviews). His remark at the end of paragraph one “If I were to mention every omission, then I would have to rewrite it myself,” etc.) is very revealing for he seems to be hinting that he is peeved not to have been invited to write the piece himself.
Here are more detailed responses.
1) He says we should include a section dealing with more commonly asked questions such as incidence, sex ratio, genetics etc. This is a perfectly legitimate criticism and we now include a paragraph dealing with it.
2) He argues that the article is “speculative”- a word which has pejorative connotations among scientists who lack imagination. If he had any knowledge of the history of science he would have realized that speculation is the driving force of all science so long as testable conjectures are made and so long as the author makes it clear WHEN he is speculating and when he isn’t (we make this distinction clear throughout our entry). The referee should read Karl Popper or Peter Medawar (recall, especially, Medawar’s memorable phrase that scientists who don’t speculate are like “cows grazing on the pasture of knowledge”). For example, in 1999 we speculated that “cross wiring” in the fusiform gyrus causes “lower synesthesia” and cross wiring in the vicinity of the angular gyrus leads to “higher” synesthesia - a prediction that has now been directly and elegantly confirmed by the DT imaging results of Rouw and Scholte (2007). The referee is probably unaware that this degree of concordance between prediction and subsequent confirmation is rare in cognitive neuroscience.
3) His second criticism is that our review is not a comprehensive. Our reason for this is the space limitation imposed by SCHOLARPEDIA. Given this limitation we decided to focus on just one or two types of synesthesia in some detail and simply MENTION some of the others. We chose to focus on grapheme color synesthesia because a) It is the most common b) It is one of the types on which most experiments have been done. c) It is the one for which we have a clear-cut physiological explanation and experimentally verified anatomical basis (partly based on our own psychophysical results and brain imaging studies conducted many years ago). Perhaps the referee regards this tendency to explain psychological phenomena in neural terms as unhealthy.
We do talk about number forms in some detail and also briefly mention many other types of synesthesia but our goal is not to provide an exhaustive catalog of published literature. This can be found in Google or in Wikipedia.
We do apologize, though, if we have not dealt in detail with the referee’s particular favorite form of synesthesia.
4) He says we don’t cite recent literature on how synesthesia facilitates memory. This is incorrect; we DO cite Smilek, Dixon, Cudahy, & Merikle (2002) and others but the referee seems to have missed this. We agree with him we should have cited Luria (and now do so).
5) He says we cite (published) abstracts presented at meetings. This is common practice in science; like it or not it establishes priority. The objection is usually raised by people (the referee?) who are peeved that they “missed the boat”. If abstract references are not allowed, nothing prevents someone who lacks originality from scanning recent meetings for new abstracts and then conducting and publishing longer studies on the same experiments. But we agree that there were too many of these in our review and have now deleted all but three.
6) The referee says we discuss only “exotic forms” of synesthesia and not common ones. This is patently absurd. Much of our review focuses on two of the most common types of synesthesia Grapheme- color and number line (both originally described by Galton). What’s “rare and exotic” about an incidence of 1 in 20?
7) Pharmacology; the referee says this is an idea he has never heard of before. If he had done his homework he would have realized that it was presented at the Cognitive Neuroscience meeting in New York this year AND published as a paper in a journal. His second comment that “by no means universally accepted” is irrelevant; what he means is HE doesn’t accept it. In any event, Science isn’t decided on the basis of democratic vote – on the basis of how many people know about something. There was a time when it was “not widely known” that the earth was round.
8) Higher prevalence of synesthesia in artists, poets and novelists (and our theory to account for this). The referee contests the empirical data. Again he should do his homework. There are at least three studies since Galton which describe a much higher incidence. Rich AN, Bradshaw JL, Mattingley JB.(2005). A systematic, large-scale study of synaesthesia: implications for the role of early experience in lexical-colour associations. Cognition. 2005 Nov; 98(1):53-84. Domino, G. (1989), ‘Synaesthesia and creativity in fine arts students: An empirical look’, Creativity Research Journal, 2 (1–2), pp. 17–29. Ward J, Yaro C, Thompson-Lake D, Sagiv N (2007). Is synaesthesia associated with particular strengths and weaknesses? UK Synaesthesia association meeting.
The referee is correct to point out that we should back up our claim (about higher incidence) with citations. We appreciate his pointing this out and have added references.
9) The referee says in so many words that synesthesia is either sensory or it isn’t; you can’t have it both ways. This is nonsense – such binary divisions are the exception rather than rule in biology. Even the “central dogma” of Biology – that DNA dictates RNA structure which in turn dictates proteins – is full of exceptions but no one doubts that the basic genetic code exists.
The evidence for at least some (and perhaps many) forms of synesthesia being sensory is overwhelming. People who have this are an important category since they allow us the opportunity to explore the underlying physiological mechanisms in some detail. It doesn’t follow that ALL phenomena that we choose to call “synesthesia” are sensory; in any event this is a semantic, not scientific issue. Nor do higher “exotic forms” negate the evidence (both psychophysical and physiological) that some of the common forms are indeed sensory in nature. We don’t see why the referee thinks that the new form of synesthesia (which we describe) in which sensations evoke emotions is a stumbling block for our “sensory cross wiring” hypothesis. Again, he seems to be getting muddled in terminology. If the “cross wiring” gene(s) is over-expressed not just across senses but in the connections between sensory systems and amygdala (See Ramachandran and Hubbard, 2001b) it could lead to spontaneous evocation of highly specific emotions in response to specific sensations. Whether you wish to call this “synesthesia” is simply a matter of definition (preoccupation with semantic hygiene is unhealthy in science). The point is surely that it is ANALOGOUS to synesthesia and can help illuminate neural mechanisms underlying such phenomena.
10) He says that the essay should provide a balanced overview and not be skewed ( toward physiological explanations?) What he really means is that it isn’t skewed in his particular direction. Anyhow, we have already been over this under (2); we chose to emphasize the one form of synesthesia that’s most common and for which the physiology is (at least partly) known. We also discuss the broader implications of the phenomenon and, yes, this is a bit slanted toward our ideas about the possibility that the over-connectivity in synesthesia might predispose to metaphor. But what use is a list of empirical findings on a quirky phenomenon like synesthesia unless one speculates on its potential theoretical implications for phenomena OTHER THAN the one that one is studying? (This is called Science, in case your referee is wondering). SCHOLARPEDIA provides an especially valuable forum in this regard, making it different from WIKEPEDIA which serves a different but equally important purpose.
11) The referee is correct to note that we slant the review toward the particular form that we have studied – grapheme-color. We have now added a paragraph making it clear why we chose to do this. We have also slightly altered the tone of the article to make it more suitable for an encyclopedia.
In summary we thank the referee for his suggestions. Even though the majority of his comments are irrelevant, we found some of them useful and they will be incorporated in the revision.
VS Ramachandran, David Brang Center for Brain and Cognition, UCSD
I read the review, as well as the comments made by reviewer 1 and the authors' reply. I must say that although the review is interesting it remains very speculative. I do not think that this approach is very suitable for this type of review and for this audience. I believe that the review should be more solid, basic and backed-up with some scientific evidence (for example, what is the evidence that 15% of the population are lower synesthetes?). In addition, there is a lot of research on grapheme-color synesthesia, but again, this is a review and should be integrative (or should be classified as a review on grapheme-color synesthesia). You cannot draw conclusions on all the synesthetes based on a specific subgroup, and while ignoring evidence from other subgroups. There are many other points of selectivity in this review such as the case of cross-activation and gene mutation. For example, what about the disinhibition hypothesis? The findings by Rouw and Scholte are correlational, and therefore cannot lead to the conclusion that hyper-connecitivity leads to synesthesia. How does the pharmacological basis of synesthesia fit in? This is a new result, but was published in a negligible journal and without any control group. Again, you might want to soften your straightforward conclusions a bit.
Also the use of conference abstracts should be kept to a minimum (especially when the posters were presented 3+ years ago). It is very easy to publish abstracts, which are not subjected to a rigorous peer review. Again, as speculation, it has some value, but putting them in a review can give them credibility, at least for the lay reader, and this should be avoided.
The review of the literature is also very selective. I can understand that the authors try to emphasize their contribution to the field, and at the moment more than 33% (!) of the cited work is by their group, however, this amount of citations does not reflect the situation in the field). This bias for self citation affects the quality of the review. For example, it is inconceivable why the reader needs to be informed in detail about some abstract (e.g., Azoulai, Hubbard, Ramachandran, 2005), or a book chapter (e.g., Ramachandran & Hubbard, 2005a), instead of important, major experiments that were published in very good peer review journals (Sagiv, Simner, Collins, Butterworth, and Ward, 2006; or Piazza et al., 2006, Cogn Neuropsychol). The same is also true for the description of synesthesia for months and their spatial form. There is a description of a finding in their lab (without any reference), rather than an experimental paper by another group (Smilek et al., 2007, Conscious Cogn). The section on mirror neurons and synesthesia is again very biased, and instead of describing the work on this type of synesthesia led by Bannisy and Ward (2007, Nature Neurosci), it deals mainly with phantom limbs…
To sum up, the instruction for review specifies that "Scholarpedia enforces the same rigorous anonymous peer review process as most printed journals. This is done primarily to insure the accuracy and quality of information". I do not think that this review in its current form lives up to this requirement. I hope that the authors will not take my evaluation in a negative way, but I am sure that for a review paper, there is a need for more balanced writing, and while some speculations and abstracts can be included, it should not be done on the account of leaving out important experimental works by other labs.
Response to reviewer 2
We thank the referee for his comments which, unlike those of referee 1, we found useful. As Semir Zeki once said “Referees are swine. But swine can sometimes lead you to truffle.”
We are pleased he/she read our original review, referee 1’s comments and our response.
First, the points of disagreement. He says the article is “speculative” and that it should represent the “consensus view” of the field. We have already dealt with the issue of speculation in our response to referee 1; without speculation science becomes indistinguishable from a catalog of facts and scientists become mere technicians or (to quote Medawar) “Cows grazing on the pasture of knowledge”. Our suggestion made in 1998, when the field was barely getting started, that number/color synesthesia was based on cross activation in the fusiform gyrus was regarded to be pure speculation when we first presented it but now the evidence for it, from our group and from many other groups, is overwhelming. The notion of synesthesia being caused by hyperactive “back-projections” or being influenced by top-down phonemic representations was also discussed by us in several early papers (Armel and Ramachandran, 1999; Ramachandran and Hubbard, 2001a; Ramachandran and Hubbard, 2001b). The referee says that the findings of Rouw and Scholte (increased white matter in the fusiform gyrus of lower synesthetes) is “just correlational” but this criticism could be made of most imaging studies. It isn’t conclusive by itself but provides strong support for the hyper-connectivity theory. (His criticism could just as easily be applied to ANY science; the claim that smoking causes lung cancer is “just correlational” as the Tobacco companies have long argued). In particular the degree of cross-connectivity in the fusiform correlates with the degree of “lower” synesthesia – so there is a clear quantitative relationship.
Referee 2 also suggests that we should present a “consensus view”. How can there be a consensus view “on solid foundation” ( to quote him ) when the whole field is only 7 years old – there were essentially no experiments before then (in any event adhering to the consensus view isn’t always a good idea; it was once the consensus view that continents cant possibly drift apart or that the earth was flat).
Some aspects of the phenomenon (e.g. grapheme-color) can now be explained in terms of the known physiology and we deliberately decided to place emphasis on these aspects (and we now say so explicitly at the very outset). To label all kinds of sensory associations “synesthesia” and use their complexity to detract from the ones which can be explained anatomically is not a wise strategy in science (e.g. the fact that viruses don’t have DNA doesn’t undermine the whole enterprise of exploring DNA as the basis of heredity). The term synesthesia is loosely applied to a large number of phenomena that may or may not be related but our inability to explain the more exotic variants (e.g. “F is strong and silent and reminds me of blood”) doesn’t detract from our ability to explain – and experimentally explore- the more common versions. We have now added a paragraph at the end stating this.
However, we agree that more coverage should be given to other forms of synesthesia and we have now added an extra paragraph to address this. Even in the original version of the review we do make it clear that not all kinds of synesthesia – nor even all aspects of grapheme /color synesthesia - can be explained in terms of the hyper-connectivity model as it presently stands. In fact our 2001 Journal of Consciousness paper points out top down effects and linguistic/phonemic categorization (e.g. an ambiguous H in the middle of CAT or THE) as strongly modulating the cross activation. These are now frequently ‘rediscovered’ without proper attribution.
Referee 2’s criticism that our pharmacological model was published in a “negligible journal” (which, by the way, has an impact factor of 1.299) is irrelevant. Perhaps he is unaware that Mendel’s paper - which marks the beginning of biology as a science- was published in an obscure – “marginal”- Austrian botanical journal. Conversely, a great deal of rubbish is published in ‘’nature’’ and ‘’science’’ even though ‘’on average’’ the quality is high. In the last decade the papers that have appeared in these journals have been heavily skewed in the direction of brain imaging even when the experiments don’t tell us a great deal. Both referees suggest that we remove the section on pharmacology. We don’t understand their reasons but in response to their suggestion we have decided to delete the whole section.
We have already discussed the question of citing abstracts in our response to referee one; they establish priority. “Peer review” is useful but a double – edged sword since it is usually done by one’s competitors in the field who have their own axe to grind (“peer review” isn’t useful in a newborn field in which one has only 10 “peers”!). In any event, most conference abstracts are vetted by a committee of experts and the data are “peer reviewed” by all those who attended the meeting rather than two people (on the other hand if abstract citations are not allowed nothing stops someone making a career out of taking last years conference abstracts and conducting “proper studies” or publishing longer papers without proper citation). Surely it’s obvious that those who recognize the importance of a problem and get a whole enterprise started (whether synesthesia, number lines, or anything else) deserve just as much credit as those who specialize in conducting follow up studies.
A case in point is our abstract on the number line/ memory experiment (Azoulai, Hubbard, Ramachandran, 2005). Which was presented in 2004 long before the Sagiv et al, 2006 paper which the referee suggests we cite. We agree with him but don’t see why the Sagiv reference should ‘’replace’’ the reference to our abstract which anticipates the basic conclusions of the Sagiv et al paper by at least two years. The solution to this is that ‘’both’’ should be cited and we now do so (we credit Galileo for his “N of 1” observation leading him to discover the moons of Jupiter; we don’t just credit the voyager space probes that recently photographed the moons in great detail).
We agree with referee 2 that there is no direct evidence that lower synesthetes constitute only 1 out of 6; we have corrected this by adding qualifying remarks.
Finally both the editor and Referee 2 (last paragraph) point out that the ‘’style’’ of the review should be changed and that it should cite more of the other recent studies in the field. We agree with this. We have removed first person pronouns (“We did”, “We showed,” etc.) as much as possible and have added two new paragraphs on other important work that has recently been done on synesthesia.
In summary we have found some of the referees’ comments to be useful and have incorporated them into the revised version. Noblesse oblige.