Tuesday, March 30, 2010

Synesthesia: crossed wires or free association?

For millennia, philosophers have mused over the nature of perception, how closely it mirrors “reality” and whether different people might, quite without knowing it, subjectively perceive the world in very different ways.  We might agree that an apple is “red”, but is the quality of your experience of its redness the same as mine?  This has seemed an almost impossible nut to crack, but the fascinating condition of synesthesia provides a stark example where the quality of subjective experience is very definitively and demonstrably different.  This may be due to genetic variants which affect the functional segregation of specialized circuits in the brain which makes synesthesia even more interesting.   

The following is a book review I recently wrote for the American Journal of Human Biology, of: Wednesday is Indigo Blue: Discovering the Brain of Synesthesia. By Richard E. Cytowic and David Eagleman. X + 309 pp. Cambridge, MA: MIT Press. 2009. 

For people with synesthesia, particular stimuli automatically and involuntarily elicit a characteristic secondary percept or feeling.  Thus, hearing particular sounds may stimulate the very real perception of colored shapes in the visual field, tasting flavors may induce the tactile sensation of specific objects, and different textures may induce specific emotions.  Curiously, in addition to basic sensory stimuli, many learned categories such as numbers, letters, words, calendar units, musical notes or chords can act as inducers.  Indeed, colored letters and numbers (grapheme-to-color synesthesia) and the association of numbers with specific positions in space (numbers forms) are by far the most common types of synesthesia.  The particular pairings, between inducing stimuli and concurrent, secondary percepts are quite idiosyncratic, though very consistent and stable over time in each individual.  As such, they form an enduring and essential component of the qualitative experience of perception in synesthetes. 

Cytowic can rightly claim some of the credit for bringing synesthesia – a once popular subject that languished while the strict doctrine of behaviorism held sway in psychology – back to the attention of the scientific community and the general public.  Here he parlays his extensive personal research and experience with synesthetes into a wonderfully detailed description of the phenomenology of this condition. 

The personal accounts of synesthetes, teased out by probing questioning, richly convey a sense of what the experience of synesthesia feels like, in as much as that is possible to convey to people who are not synesthetes.  One synesthete who perceives shapes in response to sounds says: “The shapes are not distinct from hearing them – they are part of what hearing is”.  These accounts, including descriptions of various synesthetes’ efforts to convey the nature of the experience through art, graphically illustrate the diversity and idiosyncrasy of synesthetic experiences.  They are counterbalanced by data derived from the extensive and systematic characterization of large numbers of synesthetes by Eagleman and colleagues, and other researchers, which provide an informative and detailed overview of the condition across the population.  One of the amazing findings of recent years is just how common this condition is – exact figures vary but are on the order of one to four percent of the population.

The condition often runs in families and most synesthetes report that it “has always been there”.  Synesthesia has therefore attracted a great deal of recent interest as a possible case where mutation of a gene or genes results in altered organization or function of circuits mediating perception.  The use of clever psychophysical experiments has verified objectively the existence of an extra percept as part of the experience of people with synesthesia.  More recently, neuroimaging studies of synesthetes with sound-to-color or grapheme-to-color synesthesia have directly demonstrated activation of visual areas subserving color perception in response to aurally presented words or visually presented, achromatic graphemes.  The hypothesis has thus emerged that the condition results from cross-activation of one specialized cortical area by another, normally separate one. 

There are two major mechanistic explanations: the first, that there are extra axonal connections between the relevant cortical areas in synesthetes; the second is that such connections exist in all people, but are disinhibited in synesthetes.  The authors consider both models as plausible explanations but advance several arguments in favor of the latter.  These are based mainly on the premise that synesthetic experiences represent an extension of normal processes of multisensory integration.  This resemblance seems quite superficial.  In particular, while multisensory effects are ubiquitous in normal perception, they tend to be modulatory rather than driving; that is, stimulation in one sensory modality can alter the perceived timing or spatial location of sensory information in another modality (as in the ventriloquist effect) but does not generally trigger a totally new percept.  In addition, recent studies have found that synesthetes do not generally show greater multisensory integration of this sort. 

More pertinent evidence in favor of the disinhibition model is the fact that even non-synesthetes have a tendency to map stimuli from one sense to another (e.g., matching higher pitched sounds with brighter colors) and also that some hallucinogens can induce a pseudo-synesthetic state.  The authors suggest that such mechanisms obviate the need for any extra connections in synesthetes, though why they think this would be a difficult situation to engender is not clear.  There is substantial phenotypic variation in brain circuitry in the normal population, which is highly heritable, and now many examples of genetic lesions in humans that alter nervous system circuitry with behavioral consequences.  A model of innate structural differences thus seems equally plausible, and even parsimonious (see review by Bargary and Mitchell).

The subtitle of the book promises to “discover the brain of synesthesia”.  Exploration of this topic is largely postponed to chapter nine, but is foreshadowed so frequently in the rest of the book that I admit I skipped ahead to read it.  I confess to being disappointed that the discussion remains at a somewhat metaphorical level, with abstract models of connections between cortical areas driving activity above a threshold of conscious perception.  Much more realistic and detailed schemes of cortical connectivity could have been considered that would place important constraints on the models proposed to explain synesthesia.

Nevertheless, if this section leaves one somewhat unsatisfied this only serves to illustrate how much we have to learn about how the brain works and how much synesthesia, as an exception to the normal rules, will likely tell us.  Whether the genes involved primarily influence the structure or function of cortical circuits, synesthesia will provide an important model to reveal the principles and processes controlling the dynamic balance between segregation and integration of cortical areas. 

In particular, the authors highlight the fact that synesthesia presents a tractable paradigm of the important interaction between nature and nurture.  Innate genetic effects on circuitry must interact with the experience-dependent processes that lead to the specialization of cortical areas for specific types of stimuli (such as letters and numbers, for example).  Such interactions are likely to be crucial in many other psychological or psychiatric conditions such as autism and dyslexia, where cascading effects of a primary mutation over cognitive development determine the eventual phenotype.  This book thus provides both an engaging window into the subjective world of synesthesia and a thoughtful introduction to the broader scientific issues it raises. 

 Bargary G, & Mitchell KJ (2008). Synaesthesia and cortical connectivity. Trends in neurosciences, 31 (7), 335-42 PMID: 18550184

Mitchell, K. (2010). Book review: Wednesday is Indigo Blue: Discovering the Brain of Synesthesia American Journal of Human Biology DOI: 10.1002/ajhb.21039

Friday, March 26, 2010

Intelligence a matter of the right connections

What makes some people smarter than others?  Is intelligence innate?  Is it under genetic control?  Is there something different about the brains of people with high versus low intelligence, and if so, what is the nature of the difference?  Some answers to these important questions on this often touchy subject are emerging.  

Many would bristle at the very notion that some people are “smarter” than others, if that is meant to imply an innate difference in ability.  There is however a wealth of evidence that that is precisely the case, though it is important to define exactly what is meant by intelligence.  When people are examined on a variety of tests, spanning different cognitive abilities – verbal ability, spatial reasoning, abstract logic, memory – it is found that people who do well on one of these tests tend to also do well on the others.  Psychologists use the term “g”, for general intelligence, to denote a statistical construct which captures this correlation and which is thought to reflect some underlying characteristic that contributes to success on all these measures.

Results from a large number of twin, family and adoption studies agree that the heritability of g is very high – at least 50% and perhaps as high as 80%.  (This means that 50-80% of the variance in g across the population is due to differences in genes).  Some effects of a shared family environment are seen at early ages but these tend to disappear when examined in older individuals.  Whatever the effect of the family environment on IQ measures when an individual is within that environment, these effects seem to be temporary and diminish in later life.

(An important aside: Note that the heritability within populations does not tell us anything about what might cause a difference in the mean of a trait between two populations.  This is a common misinterpretation – differences in mean between populations may be entirely due to differences in environment, even if the trait is very highly heritable within each population, where environmental variance is low). 

Presumably, some parameter of brain structure or function that correlates with intelligence is being affected by these genetic differences.  While a correlation with overall brain size has been repeatedly noted (“Check out the big brain on Brett!”), this leaves a lot of the variance in intelligence unexplained (and is also not particularly informative).  Is intelligence localized to a certain brain region or is it a distributed property of the entire network? 

A number of recent studies, taking very different approaches, arrive at the same conclusion – it is the connectivity between areas of the brain that best correlates with intelligence.  Paul Thompson and colleagues analysed brain connectivity and intelligence in a large twin study (with 92 pairs of twins).  Using diffusion tensor imaging, they were able to assess the size, organization and “integrity” of axonal tracts connecting all areas of the brain.  They found, when comparing these measures across pairs of either monozygotic or dizygotic twins, that these parameters were more heritable for some tracts than for others.  Most importantly, they also found that intelligence was correlated with connectivity measures across numerous tracts in the brain.  They could also show a substantial shared genetic effect – the genes affecting structural connectivity were also affecting intelligence.

Ralph Adolphs and colleagues used a very different approach – they analysed a large collection of patients with lesions in different parts of the brain.  By looking across this collection for sites where lesions were consistently correlated with reduced intelligence they were able to map a network of important regions.  The most striking finding is that many of the “regions” thus defined were actually within the white matter – they were not restricted to specific cortical areas but rather reflected the connections between areas of the brain.  Again, the implication is that it is the efficiency and effectiveness of brain connectivity which are the major parameters affecting intelligence.

While the overall genetic effects are very robust, only a few specific genes have been identified that seem to influence intelligence.  For now, the neurodevelopmental processes which mediate the effects of these and other genes on the parameters of brain connectivity remain a mystery.


Chiang, M., Barysheva, M., Shattuck, D., Lee, A., Madsen, S., Avedissian, C., Klunder, A., Toga, A., McMahon, K., de Zubicaray, G., Wright, M., Srivastava, A., Balov, N., & Thompson, P. (2009). Genetics of Brain Fiber Architecture and Intellectual Performance Journal of Neuroscience, 29 (7), 2212-2224 DOI: 10.1523/JNEUROSCI.4184-08.2009


 Glascher, J., Rudrauf, D., Colom, R., Paul, L., Tranel, D., Damasio, H., & Adolphs, R. (2010). Distributed neural system for general intelligence revealed by lesion mapping Proceedings of the National Academy of Sciences, 107 (10), 4705-4709 DOI: 10.1073/pnas.0910397107

Thursday, March 18, 2010

LRR proteins help neurons find a partner

Matching 100 billion neurons with their appropriate partners is a daunting task, especially when each neuron can make synaptic contact with about 1,000 other cells. Nevertheless, the developing brain accomplishes this feat with remarkable specificity – neurons from each area of the brain send out axons which follow a stereotyped pathway to find their appropriate targets, guided by signpost proteins along the way. Once in the right general area they have to select specific cell types with which to form a synapse, often limited to a certain layer or sub-region. Many cells will even form synapses specifically with distinct subcellular compartments of their target cells – on distal or proximal parts of the dendrites or directly on to the cell body, for example.

Specifying this level of connectivity, with the numerical complexity of the mammalian brain, obviously requires a large number of labels that can be used to match synaptic partners (even allowing for some level of combinatorial logic that increases the coding capacity). One class of molecules that has emerged recently as important in this process are members of the leucine-rich repeat (LRR) superfamily. These are secreted or transmembrane proteins which are characterised by a domain called the leucine-rich repeat in their extracellular region. LRR domain proteins are well known in the immune system, where they recognize a diversity of pathogenic factors. In the nervous system, these proteins are also involved in recognizing a diversity of factors, but in this case, the factors are protein labels on the surface of other neurons.

A role for these proteins in the process of neuronal target selection first came to light in studies in fruitflies, where different LRR proteins act as labels to match specific motor neurons with their appropriate target muscles. In parallel, studies in mammals have identified a number of subfamilies of LRR proteins which are capable of inducing one neuron to form a synapse with another one in a cellular assay in culture. A couple of recent studies have highlighted the important roles of one such subfamily, called the LRRTM proteins in this process. These studies found that LRRTM2 can induce the formation of a synapse and that it accomplishes this by binding to the protein Neurexin-1 on the opposing cell.

Neurexins were already known to be involved in this kind of process, through binding to a different family of proteins, the Neuroligins. What makes these studies particularly interesting is not just the identification of the additional role of LRRTM proteins in thise process, but that Neurexins, Neuroligins and the LRRTM genes have all previously been implicated in autism and/or schizophrenia (and in the case of LRRTM1 in the genetics of left-handedness). Failing to form the right kinds of connections, especially the correct balance of excitatory and inhibitory connections, can, it seems, lead to the kind of dysfunction of neural circuits and networks that ultimately results in psychopathology.

The complement of LRR proteins has expanded dramatically over mammalian evolution, and individual members have diverged rapidly in protein sequence (see Dolan et al., below). These findings, in the light of the functions of various members of this superfamily, suggest that LRR proteins collectively contribute to the complexity of connectivity of the mammalian brain and may have been important in its evolution.

Ko J, Fuccillo MV, Malenka RC, & S├╝dhof TC (2009). LRRTM2 functions as a neurexin ligand in promoting excitatory synapse formation. Neuron, 64 (6), 791-8 PMID: 20064387

de Wit J, Sylwestrak E, O'Sullivan ML, Otto S, Tiglio K, Savas JN, Yates JR 3rd, Comoletti D, Taylor P, & Ghosh A (2009). LRRTM2 interacts with Neurexin1 and regulates excitatory synapse formation. Neuron, 64 (6), 799-806 PMID: 20064388

Dolan J, Walshe K, Alsbury S, Hokamp K, O'Keeffe S, Okafuji T, Miller SF, Tear G, & Mitchell KJ (2007). The extracellular leucine-rich repeat superfamily; a comparative survey and analysis of evolutionary relationships and expression patterns. BMC genomics, 8 PMID: 17868438

Friday, March 12, 2010

Wired for Music

Music has a bizarre power to engage and affect us – to move us emotionally or literally, whether it’s foot-tapping, finger-drumming or booty-shaking.  It seems to have properties that make it automatically and powerfully salient for human beings.  An obvious question is whether this reflects some innate properties of the human brain or whether it emerges over time due to experience with types of music.  Put another way, does the brain shape the music or the other way around?  Does music show particular structures because those are inherently salient and pleasant to humans or is this reaction caused by the brain’s tendency to specialise in processing stimuli that occur with some statistical regularity in its environment? 

A new study by Perani and colleagues demonstrates very convincingly that the human newborn brain already shows strong functional specialisation for music processing.  By performing functional magnetic resonance imaging on newborns, all under 3 days old, they found a strongly lateralized pattern of activations in response to music.  These responses were much stronger in the right hemisphere, as is observed in adult humans. 

More interestingly, they found that modifications to the music, which introduced infrequent key shifts or dissonance through shifting one component a half-tone higher, resulted in a very different response pattern.  This type of altered music did not engage the right hemisphere network as efficiently as the original music, and did engage regions in the left hemisphere that were not responsive to the original music.  Thus, the right hemisphere auditory network is not just specialized for music, it is specialized for music with the appropriate structure (consonance) which most listeners agree is most pleasant.

Previous studies have shown that the left hemisphere is preferentially activated by language stimuli even in newborn infants.  The authors speculate that there may be a general division of labour between the two hemispheres, with the left more specialized for processing temporal characteristics of stimuli and the right for processing spectral characteristics (including frequency or pitch).  Interestingly, the right hemisphere is also involved in processing prosody – the melodic components of natural speech – modulations in emphasis and inflection that communicate emotional content and tone.  It seems likely that the apparent specialisation “for” music reflects the fact that these circuits are pre-tuned to be most responsive to stimuli with specific acoustic characteristics.  We did not evolve to enjoy music – music evolved (or was actively designed) to best stimulate our natural preferences.

Newborns thus arrive in the world pre-wired to process different types of acoustic stimuli in specialized circuits, localized to either hemisphere, one for detecting and distinguishing sequences of sounds in time and the other for decoding oscillatory components of the stimuli, including tone, pitch, timbre, rhythm, etc.  How this specialisation arises during development is a fascinating topic, and one that is poorly understood.  The mechanisms underlying the initial establishment of left-right differences in early embryos are fairly well-established but how these affect the developmental programmes in the brain is far less clear, though a number of genes have been found that are differentially expressed in the two hemispheres of developing human brains (see Sun et al, below).   

The fact that functional lateralisation depends upon a genetic programme also suggests that variation in the responsible genes might lead to differences in the degree or direction of lateralisation in different people.  This is known to occur for language lateralisation (which can vary with handedness, itself under genetic influence).  Lateralisation is also known to be affected in a range of psychatric disorders, most notably schizophrenia.  How the kinds of mutations that result in these disorders affect lateralisation and how this contributes to psychiatric symptoms are important questions for the future. 

Perani, D., Saccuman, M., Scifo, P., Spada, D., Andreolli, G., Rovelli, R., Baldoli, C., & Koelsch, S. (2010). Functional specializations for music processing in the human newborn brain Proceedings of the National Academy of Sciences, 107 (10), 4758-4763 DOI: 10.1073/pnas.0909074107

Sun, T. (2005). Early Asymmetry of Gene Transcription in Embryonic Human Left and Right Cerebral Cortex Science, 308 (5729), 1794-1798 DOI: 10.1126/science.1110324 

Friday, March 5, 2010

Is Mental Illness Good For You?

Mental illness is surprisingly common.  About 10% of the population is affected by it at any one time and up to 25% suffer some kind of mental illness over their lifetime.  This has led some people (many people in fact) to surmise that it must exist for a reason – in particular that it must be associated with some kind of evolutionary advantage.  Indeed, this is a popular and persistent idea both in scientific circles and in the general public.  (See the recent article “Depression’s Upside” from the New York Times Magazine, for example).

Such theories come in two main varieties – the first, that mental illness confers some specific advantage to those afflicted; and second, that the mutations which cause mental illness in one person’s genetic background may confer an advantage when they are in a different genetic background (balancing selection).  Both of these suffer from some misconceptions about how evolution by natural selection works.  The intuitive appeal of the “survival of the fittest” metaphor may have something to do with this – the actual mechanisms of natural selection are more nuanced.

Natural selection works by changing the frequencies of genetic variants in the population.  If a particular gene, gene X, comes in two varieties, X and X’, and one of these (say X’) tends to increase the evolutionary fitness of the people who carry it, this means they will have more offspring than people who do not carry that variant and the frequency of the X’ variant will increase in the next generation (at the expense of the X variant).  Variants that increase fitness a lot rapidly out-compete the alternative version and soon all copies of the gene will be of that type (it becomes “fixed” in the population).  In contrast, new variants that arise which seriously decrease fitness will tend to be rapidly weeded out of the population. 

We know that psychiatric disorders can be caused by such genetic variants (mutations). What are the likely effects of such mutations on fitness?  Is there any reason to think they may confer some kind of advantage on carriers?  Some examples of the types of advantages that have been postulated include that schizophrenics may have been seen as shamans in ancient societies, that people with bipolar disorder may be more creative (it is especially common among poets, for example), or that depressed people are actually more realistic and better able to concentrate on a problem.  The trouble with these theories is that natural selection doesn’t care whether you are good at poetry or solving problems through prolonged rumination.  Natural selection only cares how many children you have – more accurately, how many children you have who survive to breed themselves.

If it were true that people with mental illness live longer and have more surviving offspring than people without, then this kind of theory might be viable.  In fact, the evidence is overwhelmingly the opposite.  First, mortality rates before or during the reproductive period for depression and schizophrenia are two to three times higher than the general population.  Second, surviving patients with schizophrenia have only 1/3 the average number of children – this general trend seems to also hold for other psychiatric disorders.  Variants that increase risk of mental illness thus demonstrably and significantly decrease fitness and should be rapidly selected against; i.e., they would never rise to a high frequency.  (This is one of the major arguments against the so-called common disease/common variants hypothesis – see post of July 7th for more on this). 

But wait – maybe these disorders are deleterious in our current environment but conferred an advantage of some kind in primitive environments.  (Much as variants that predispose to obesity or diabetes in our current environment might have been adaptive in an environment where high-fat food was very scarce).  This is certainly conceivable, although there is no evidence or even good reason to think that this was the case.  In fact, through some simple modeling it can be shown that the current rates of mental disorders do not fit such a scenario (see Keller and Miller).

An alternative suggestion is that the variants that predispose to mental illness do so only in some genetic contexts – only in the presence of additional variants.  In other genetic backgrounds, perhaps they confer an increase in fitness which counterbalances the decreased fitness in mental illness sufferers.  This scenario of balancing selection is modeled on very rare cases like sickle-cell anemia, where a particular mutation causes a deleterious condition when present in two copies, but confers an advantage (increased resistance to malaria, in this case) when present in only one copy.  While this kind of model is difficult to disprove, there are strong arguments against it and no evidence that it applies – relatives of schizophrenics do not show increased fertility rates, for example.

So why is mental illness so common?  If it’s so bad and is caused by genetic variants, why hasn’t natural selection gotten rid of them all by now?  Well, the answer is it does – in fact, it’s very good at getting rid of them.  Unfortunately, new mutations keep arising all the time.  Rates of mental illness are higher than those of other genetic disorders because it takes the combined actions of thousands of different gene products to wire the staggeringly complex human brain.  If any one of a large number of these genes gets mutated, then development of the brain may be compromised and this may ultimately result in a psychiatric disorder.  There is thus no paradox to explain – common disorders like schizophrenia are really umbrella terms for lots of distinct genetic disorders, each of which is extremely rare, due to the efficient action of natural selection.  All of these theories are thus offering solutions to a problem that doesn’t exist. 

For more on this topic see the extremely insightful article by Keller and Miller. 


Keller, M., & Miller, G. (2006). Resolving the paradox of common, harmful, heritable mental disorders: Which evolutionary genetic models work best? Behavioral and Brain Sciences, 29 (04) DOI: 10.1017/S0140525X06009095