The human brain is a complex structure that controls sophisticated cognitive behaviour. Anatomically, the cerebral cortex is divided into frontal, temporal, parietal and occipital lobes, and these regions control thinking, language, movement, sensation, vision and other functions. The formation of these distinct functional regions during cortical development is called regionalization (or arealization)1, 2, 3, 4. Two models for the formation of cortical functional regions have been proposed5, 6, 7. The protomap model suggests that intrinsic signals from the 'proliferative units' in the ventricular zone regulate functional regionalization, whereas the protocortex model argues the importance of extrinsic influences, such as the thalamocortical inputs5, 6, 7. Accumulating evidence indicates that both models are applicable to the regulation of cortical patterning and the establishment of cortical regionalization2, 3, 4.
The cerebral cortex is also divided into left and right hemispheres. The left hemisphere is normally dominant for language and logical processing, whereas the right hemisphere is specified for spatial recognition8, 9. Additionally, the segregation of human brain functions between the left and right hemispheres is associated with asymmetries of anatomical structures, such as the Sylvian fissures and the planum temporale10, 11. One of the striking features of motor control in humans is that more than 90% of the population is more skilful with the right hand, which is controlled by the left hemisphere12. Similar to the left-hemisphere dominance of handedness, language ability is dominant in the left hemisphere in more than 95% of the right-handed population but in only 70% of the left-handed population12.
Is it coincidental that both language ability and hand use are dominant in the left hemisphere in most of humans? Is there genetic control of both brain asymmetry and handedness? Using molecular and neurological approaches, we are beginning to tackle these questions and discover the neurological circuitries that regulate brain asymmetry. Here, we describe brain asymmetries that have been measured using modern imaging techniques and discuss the genetic correlation between brain asymmetry and preferential hand use. Furthermore, we propose evolutionary and molecular mechanisms that might regulate brain asymmetry and handedness.
Functional and anatomical brain asymmetries
The first detailed description of functional asymmetry in the human brain was made in the 1860s by a French doctor named Paul Broca. He found that there was a lesion in the left hemisphere of the post-mortem brain of a patient with a one-word vocabulary. Broca claimed that language ability in the human brain is lateralized and supplied perhaps the first strong evidence of functional asymmetry in the brain13. This brain region, which controls speech, is called Broca's area in honour of his discovery. In 1874, a German neurologist named Carl Wernicke discovered that damage to a region of the left hemisphere could cause a type of aphasia that resulted in an impairment of language comprehension14. This area is called Wernicke's area.
Brain functional asymmetry is not limited to language ability. Whereas the right cerebral cortex regulates movement of the left side of the body (and the left cerebral cortex regulates movement of the right side), more than 90% of the human population is naturally more skilled with the right hand than with the left12. Cognitive studies on patients with unilateral lesions and on patients with split-brain surgery have revealed many other differences between the left and right cerebral cortex15. For example, the left hemisphere is dominant for mathematical and logical reasoning, whereas the right hemisphere excels at shape recognition, spatial attention, emotion processing and musical and artistic functions15, 16, 17.
Using modern imaging techniques, particularly MRI, scientists can map the asymmetries of anatomical structures in the human brain. Among the most studied regions are the Sylvian fissures, which separate the frontal and temporal lobes. For example, the posterior end of the Sylvian fissure in the right hemisphere is higher than in the left, whereas the left fissure has a more gentle slope10, 11, 18 (Fig. 1a,b). The planum temporale, a region in the posterior portion of the superior temporal sulcus, is larger in the left hemisphere than in the right in more than 65% of adult brains and 56–79% of fetal or infant brains examined19, 20, 21, 22. More recently, digital brain maps have generated three-dimensional images of human brains and further revealed cortical asymmetries10. Moreover, a new population average, landmark- and surface-based (PALS) atlas approach has shown the most consistent asymmetries to be in and near the Sylvian fissures (Fig. 1a,b) and the superior temporal sulcus11 (Fig. 1c).
Figure 1 | Anatomical asymmetries in the human cerebral cortex.
Coronal (a) and horizontal (b) MRI of the Sylvian fissures (red arrows) in the human brain. The Sylvian fissures separate the frontal lobes and temporal lobes in both hemispheres. The left (L) fissure is more ventral (V) and extends further towards the posterior (P) than does the right (R). Illustration of differences of sulcal depth between the left and right hemispheres (c). Cortical regions that are deeper in the right hemisphere are shown in red and yellow, whereas regions that are deeper in the left are shown in blue and green. A, anterior; D, dorsal. Modified, with permission, from Ref. 11 © (2005) Elsevier Science.
The differences in neuronal cell type or cell organization that might underlie these gross anatomical differences are unclear. Studies have shown that language-related areas of the left cortex might contain more and larger layer 3 pyramidal cells than corresponding areas in the right hemisphere23. Rosen24 and Galaburda25 used histological studies to suggest that the asymmetrical regions in the cortex might be the results of differences in neuron numbers but not packing density. However, the tremendous size of the human cortex and its extensive and variable folding pattern make corresponding areas difficult to compare with certainty.
In addition to the asymmetries related to language abilities, such as those of the Sylvian fissures and the planum temporale, anatomical asymmetries associated with hand use have also been detected in other regions in the human cerebral cortex. In the primary somatosensory cortex (S1), studies using magnetic source imaging have shown that the cortical representation of the right hand is larger than the one of the left hand in right-handers, and vice versa in left-handers26. Moreover, the left central sulcus, a large inward fold marking the division between the frontal and parietal lobes, is deeper than the right central sulcus in right-handers27. Inter-hemispheric comparison has further revealed a significant increase of the hand and finger movement representation in the primary motor cortex opposite to the preferred hand28.
In contrast to these findings, other reports have shown no obvious correlation of handedness and brain asymmetries. For example, using voxel-based morphometry, Good et al.29 did not detect effects of hand use on asymmetrical morphology in sensorimotor regions of more than 465 normal adult brains. Although different methodologies used in these studies could lead to opposite conclusions, the analyses of anatomical asymmetries associated with handedness in the primary sensory and motor cortices are compelling.
Handedness and language ability are two of the most obvious lateralized behaviours in humans. Taking note of the convergence of functional and anatomical studies, the asymmetrical cortical controls that regulate handedness are tightly correlated with those for language ability. But how are these controls established in humans?
Correlation of hand use and language ability
That most humans (more than 90%) prefer to use their right hand has been observed in almost all cultures and ethnicities throughout history12, 30. Statistical studies suggest that handedness might be under genetic control. There are at least two well-known genetic models of handedness31, 32 (Box 1), and although these models seem to reflect genetic mechanisms of cortical asymmetry and handedness, genes that regulate these asymmetries have not been identified. Furthermore, the question of whether a single gene can control such complex processes in the CNS is still unanswered31. Nevertheless, the single-gene models proposed by Annett33 and McManus34 fit statistical data of cerebral dominance for handedness in humans. Identifying the gene(s) that regulates brain asymmetry and handedness remains an appealing but challenging task.
Why is there a left-hemisphere bias for handedness and language ability? Preferred hand use has been observed even at embryonic and fetal stages in humans, long before language ability is developed. For example, in most human embryos, the right hand is more developed than the left at 7 weeks35. Using ultrasound, it has been observed that at 15 weeks most fetuses prefer to suck their right thumb, hinting that handedness is present prior to birth36. Interestingly, Hepper et al.37 followed up this study of 75 individuals. They found that the 60 fetuses that preferred to suck their right thumb were indeed right-handed as teenagers, and of the 15 fetuses that preferred to suck their left thumb, 5 were right-handed and 10 were left-handed. Moreover, several early studies have shown that some cortical sulci and gyri, such as the temporal gyri, are asymmetrical in human fetal brains from 10–44 weeks22, 38, 39. Using the measurement of cerebral blood flow, Chiron et al.40 found that the maturation of the right hemisphere precedes the left in the brains of human infants between 1 and 3 years of age. This asymmetrical pattern shifts towards the left hemisphere during the process of development of language abilities at about the age of 3 (Ref.40). Additionally, Trevarthen41 observed that expressive gestures, such as communicative hand movement, were asymmetrical in infants.
These results imply that anatomical and functional brain asymmetry precedes uptake of information from the environment and cognitive development. This in turn suggests the existence of intrinsic controls that regulate brain asymmetries at early stages. Although these kinds of study are interesting, one has to keep in mind whether this early right-hand preference is controlled by high-level regulation in the left hemisphere of the cortex, or by spontaneous movement regulation in the spinal cord. Furthermore, whether early brain asymmetries contribute more to handedness or to language ability still remains an intriguing and challenging question12.
Because anatomical asymmetries of certain areas in the human brain are associated with language ability, several researchers have made efforts to map asymmetries of the planum temporale and Broca's area in the brains of chimpanzees and great apes42, 43, 44. They found that simian brains also have asymmetries, which resemble those of humans. These studies suggest that brain structures associated with language ability might have existed before humans evolved. However, it is not clear whether vocal communication is asymmetrical in non-human primate brains or how these asymmetrical structures are involved in vocal processing45. Furthermore, because of the complex structure of the cerebral cortex, the mapping of areas that correspond in human and primate brains is difficult46.
Which hemispheric asymmetry (for handedness or for language ability) appeared first in evolution still remains a puzzle. Further comparative studies of brain asymmetry and handedness in non-human primates will help us to understand the relationship between handedness and language ability in humans47.
Evolutionary mechanisms of biased hand use
More than 90% of the human population is right-handed, and biased hand use is also observed in non-human primates and other mammals. But whether there is a dominant preference for one hand at a population level is still debatable. What has made most humans right-handed during evolution is still unknown.
Handedness in non-human primates. There are many contradictory reports about hand use in non-human primates, such as chimpanzees. A broad range of manual tasks have been observed in chimpanzees, including simple reaching, bimanual feeding, coordinated bimanual actions, throwing, manual gestures and so on48, 49. Although these observations have led to the argument that, for some measures, chimpanzees are right-handed, most of these findings are from captive great apes; evidence of population-level handedness in wild apes is extremely sparse48, 49. Therefore, these studies do not conclude that there is a dominant preference for hand use in non-human primates at the population level.
A recent report of hand preferences during termite-fishing/probing actions of chimpanzees is interesting. First, Lonsdorf and Hopkins50 studied wild chimpanzees living in the Gombe National Park, Tanzania, but not captive chimpanzees. Second, they observed termite-fishing actions, which require fine motor skill. They claimed that directional biases in hand use vary depending on the type of tool use. Therefore, the question of whether there is strong handedness in non-human primates might be confounded by biases in the types of motor skill required. Tests that better discriminate behavioural biases in wild primates are needed before any definitive conclusions can be drawn.
Paw preference in other mammals. A well-studied lateralized manual behaviour of many mammals is the food-reaching task, defined by paw preference. Although paw preference has been observed and studied in mice, rats, cats and dogs, it does not seem biased to either the left or the right front paw at a population level51, 52, 53, 54, 55, 56. For example, paw preference was observed among domestic cats, but no significant bias in preference was found at the level of the group53. In mice, although there is paw preference in each individual mouse, approximately half of the mice studied preferred to use the left paw and half preferred to use the right57, 58. There are also differences in the strength and direction of paw preference between mouse strains, indicating that genetic background is an important influence on this behaviour55.
Paw preference in mice has encouraged scientists to find the genetic causes of this manual lateralization. Collins59 attempted to breed left- or right-handed mice, and although he was unable, by inbreeding, to create a mouse strain that prefers to use only the left or the right front paw, he did succeed in generating mice that show a strong lateralization. For example, the HI strain was bred using mice that showed consistent right or left paw use in a food-reaching task, and the LO strain was bred using mice with little overall paw preference59.
Is paw preference associated with lateralized brain anatomy and/or function in mice? The direction of paw preference seems to correlate with the dominance of dopamine expression levels in the brain: a mouse that prefers to use the left front paw has a higher dopamine level in the left hemisphere than in the right, although the physiological implications of this correlation are unknown60, 61. Selectively bred mice (the O/AP strain) with supernumerary whiskers on the right side of the face and corresponding supernumerary barrels in the left barrel field showed a higher preference for using the left front paw. Likewise, mice with supernumerary whiskers on the left side of the face preferred to use the right front paw62. This biased front-paw use might be the result of competition for cortical representation between the size of the motor cortex and the somatosensory barrel field in which the whiskers are represented. A larger S1 (the barrel field) in the left hemisphere, for example, might make the size of the left motor cortex smaller, and lead to biased left front-paw use controlled by the right hemisphere.
Moreover, the areas of whisker-pad representation in the S1 between the left and right hemispheres of adult rats have shown striking variations (that is, asymmetries) in individuals. However, these asymmetries are not biased to either the left or the right hemispheres63, 64. Does the asymmetry of the S1 provide a hint that paw preference has corresponding and asymmetrical cortical structures that control it? It will be interesting to map the sizes of the S1 regions and the direction of paw preference in mice.
In terms of the distribution of hand use, there is a consistent 9:1 ratio of right/left hand preference reported in humans, higher than has been reported for any other mammal12, 50. The directional manual task in mice, like paw preference, might be regulated by the formation of stable neural circuits, but it has a random distribution at the population level. Given these examples of low or no bias in chimpanzees and mice, it is an intriguing puzzle as to how consistent right-hand dominance evolved in humans. Corballis31 has proposed that there was a genetic mutation in hominid evolution that promoted preferential use of the right hand and is now seen in modern humans. This evolutionary bias might be advantageous as it could increase brain capacity and social cohesion65. Applying genomic approaches, particularly the complete sequencing of the human and chimpanzee genomes, will provide a considerable insight into the evolutionary mechanisms of lateralized human behaviours and human brain development and asymmetry47, 66, 67, 68, 69.
Body and brain asymmetries
Little is known about the genetic causes of brain anatomical and functional asymmetries70. By contrast, studies of molecular regulation of asymmetries in the visceral organs, such as the heart and lungs, have made encouraging progress (Box 2). Inspired by the identification of molecules that have essential roles in visceral organ asymmetry, researchers have succeeded in identifying molecules that regulate brain asymmetry in zebrafish, perhaps the only species that has been well studied with respect to brain asymmetry (Box 3). Conserved molecules that regulate body asymmetry, such as Nodal and ion channel related gene products, are also essential for regulating asymmetry of the epithalamus, a small structure of the diencephalons71, 72.
Do the same molecular mechanisms that regulate body asymmetry also cause human brain asymmetry? The complete reversal of normal organ position, such as heart and lungs, is called situs inversus. With the exception of the reversed frontal and occipital petalia observed using anatomical and functional MRI techniques, the left-hemisphere dominance for language was still found to be similar in individuals with situs inversus and in normal subjects73. Nor did lateralization of auditory processing show any differences between individuals with situs inversus and normal subjects74. Moreover, 50% of individuals with Kartagener's syndrome, a disorder caused by cilia with a decreased or total absence of motility, have been found to have situs inversus75. This disorder might confirm the function of cilia in regulating visceral organ asymmetry, as defects in cilia mobility might result in the random distribution of NODAL molecules (Box 2). However, the situs inversus patients with Kartagener's syndrome developed normal handedness76. Therefore, it seems reasonable that the molecules and mechanisms that regulate visceral organ asymmetries might be distinct from those that regulate brain asymmetries and handedness70, 77.
Our recent studies, using a genomic screening approach, further support this idea. Using a serial analysis of gene expression (SAGE) technique, we measured gene expression levels in the left and right hemispheres of human fetal brains78. We verified 27 genes that are differentially expressed in the hemispheres of 12-week-old human fetal brains by using either real-time reverse transcription (RT)-PCR or in situ hybridization (Fig. 2). Most genes identified using SAGE analyses function in signal transduction and gene expression regulation47. Among them, the transcription factor Lim domain only 4 (LMO4) showed consistent asymmetry of expression in human fetal brains at 12 weeks and 14 weeks, and less so at 16 and 17 weeks78. However, we did not detect genes that have essential roles in visceral organ asymmetry, such as genes involved in the sonic hedgehog (SHH) or NODAL pathways, which are also differentially expressed in human fetal brains78. Because the earliest stage analysed was in the human fetus at 12 weeks, it cannot be ruled out that molecules regulating body asymmetry might also be differentially expressed in human embryonic brains (for example, at 8–10 weeks). It will be interesting to measure gene expression levels in the left and right hemispheres in human embryonic brains (8–10 weeks) using SAGE or cDNA microarray approaches.
Figure 2 | Asymmetrically expressed genes in 12-week-old human fetal brains, detected by serial analysis of gene expression and real-time reverse transcription (RT)-PCR.
The cDNA made from the perisylvian regions of the left and right hemispheres of two 12-week-old human fetal cortices were used as templates for real-time RT-PCR. The relative gene expression levels are average ratios of gene expression detected by RT-PCR between the left and right hemispheres of two brains. These differential gene expression levels also match those measured by serial analysis of gene expression (SAGE)78. 27 genes showing consistent differential expression are listed. Among them, 17 genes were highly expressed in the left perisylvian regions, whereas 10 genes were highly expressed in the right.
Molecular regulation of brain asymmetry
An essential step leading towards asymmetry is to break symmetry79. Although the initiation mechanisms of breaking symmetry are still unknown, an uneven distribution of molecules that are essential for left–right body axis patterning could be important for this biological event.
How, then, is symmetry broken in the CNS? Neuroepithelial cells divide vigorously, fold dorsally and form a neural groove during early embryonic development80. The neural groove continues to grow, and the dorsal parts meet at the midline and fuse to form a neural tube80. Neural tube development is accompanied by the formation of the notochord and the induction of the floor plate81. Numerous studies have shown that the notochord is a patterning centre for the ventral neural tube82. In the forebrain, a structure anterior to the notochord is called the prechordal plate83. Molecules secreted from the notochord, such as SHH, function as morphogens to induce and maintain the ventral property and neural cell types in the spinal cord82. Similar morphogens also induce and pattern the forebrain, and are probably secreted from the notochord or the prechordal plate84. Moreover, the patterning centres are not limited to the notochord and the ventral neural tube. Morphogens, such as bone morphogenetic proteins and WNTs, are secreted from the roof plate in the neural tube85.
One possible mechanism for breaking symmetry in the brain is that the morphogens secreted from the ventral (floor plate or prechordal plate) and/or dorsal (roof plate) midlines are distributed differently between the left and right (Fig. 3a,b). The different expression levels of morphogens induce differential expression of downstream transcription factors, such as LMO4 (Ref. 78), and eventually lead to brain asymmetry.
Figure 3 | Three models of molecular induction of brain asymmetry.
a | Whereas the neural tube (red) is derived from the ectoderm, the notochord (blue) is derived from the mesoderm and accompanies the neural tube formation. During neural tube development, the most dorsal part of the neural tube becomes the roof plate (green), whereas the most ventral part of the neural tube becomes the floor plate (blue). The forebrain is developed from the most rostral region in the neural tube. In the forebrain, the morphogens secreted from the notochord or the prechordal plate (not shown) might be differentially distributed between the left and right neural tube. b | Similarly, uneven secretion of molecules might also occur in the roof plate. Different morphogen expression levels in the left and right neural tube might break the symmetry of brain patterning and induce asymmetrical expression of downstream genes. c | Anterior signals might also induce cortical asymmetry. The patterning centre in the most rostral neural tube — for example, the anterior neural ridge (green) — could be a source for cortical asymmetrical patterning. The distribution of molecules secreted from this area might be different in the left and right hemispheres.
Recent studies of the molecular regulation of cortical regionalization have identified a patterning centre in the anterior cortex2, 3, 4. An important molecule that is secreted from the anterior cortical region is fibroblast growth factor 8 (FGF8)86. The ectopic expression of FGF8 can expand the motor cortex and shift the functional regions of the cortex caudally86. It is possible that the expression levels of morphogens secreted from the anterior cortical region might be different in the left and right hemispheres (Fig. 3c). The asymmetrical expression of regional markers, such as LMO4, could reflect asymmetrical topographic mapping of functional regions along the anterior–posterior axis in the cortex78.
Future work will include the identification of more of the morphogens that pattern the early cortex and their downstream targets. Consistent with regionalization in the cortex, which involves complex gene expression and regulation4, brain asymmetry and handedness are a conjugated result of molecular regulation, neural connections and plasticity.
Conclusions and future perspectives
The challenge of studying brain asymmetry is that because the obvious anatomical and functional asymmetries have been identified largely in humans, we cannot carry out direct experiments. The recent behavioural studies in non-human primates, such as the investigation of handedness in chimpanzees, might help us to better understand how human handedness has evolved50. In particular, the comparison of human and chimpanzee genomes has enriched our knowledge of the evolutionary mechanisms of human brain development47, 66, 67, 68, 69. Similar studies might help us to understand the evolutionary regulation of human brain asymmetry.
Several human neurological disorders show disrupted normal brain asymmetry. For example, reduced and reversed anatomical brain asymmetry has been reported in individuals with schizophrenia, autism or dyslexia87, 88, 89, 90, suggesting a potential indirect relationship between the causes of these disorders and the asymmetrical development of the human cerebral cortex. Recently, several studies have reported clinical cases of polymicrogyria — a malformation of cortical development that is characterized by many small gyri in the cortex — that occurs only on one side of the cortex; this is known as unilateral polymicrogyria91, 92 (Fig. 4). Patients have seizures, motor dysfunction and mental retardation. A genetic cause of unilateral right-sided polymicrogyria is suggested by the existence of several pedigrees in which the disorder is present in more than one individual of an affected family92. These studies indicate that unilateral polymicrogyria can be inherited as a Mendelian trait, suggesting that there might be a gene that is required for the development of the right perisylvian region92. Using forward genetic approaches to map genes that cause disrupted brain asymmetry might reveal their normal function in asymmetrical development of the brain.
Figure 4 | Unilateral polymicrogyria detected using MRI.
Polymicrogyria (indicated by arrows) is detectable in the right hemispheres in both brains shown. An apparent increase in cortical thickness is observed in the right (R) hemispheres, whereas the cortices of the left (L) hemispheres appear entirely normal. Modified, with permission, from Ref. 92 © (2006) Lippincott Williams & Wilkins.
Faster development and improvement of large-scale screening approaches at the genomic level could make the identification of asymmetrically expressed genes in human and mouse brains easier and quicker. Generating genetically engineered mice can help us to understand the functions of these genes in brain development. Using these mouse models can also help to reveal the neural circuitries that regulate brain asymmetry and lateralized behaviours. However, unlike visceral organ asymmetries, which are easy to detect, brain asymmetry relies largely on fine brain mapping and reliable behavioural tests. Therefore, the development of molecular imaging techniques and an improved understanding of lateralized behaviours in rodents will be extremely useful for studies of brain asymmetry.
