Asymmetry of brain and behavior in animals: Its development, function, and human relevance

Developmental Changes in Hemispheric Control

Age-related changes in lateralization of the human brain have long been recognized but usually as increasing lateralization directed toward the adult-typical condition (e.g., Giedd et al., 1996; Toga and Thompson, 2003). Since myelination of the corpus callosum is not complete until a child is at least 10 years old or even until mid-adolescence (Gbedd et al., 1999; Luders et al., 2010; and see Table 5.1 in Sarnat, 2008), it is assumed that full expression of brain lateralization is delayed until this age when interhemispheric communication matures. Behavioral studies do indeed support the stabilization of, for example, left hemifield (right hemisphere) advantage in recognizing faces at around 10 years of age (Chiang et al., 2000; Reynolds and Jeeves, 1978). However, it is possible that at least some of the changes in behavior that occur during development depend not merely on maturation of weak to stronger hemispheric asymmetry but also on shifts in control of behavior by one hemisphere to the other. Results from studying chicks show that shifts in the hemisphere in charge of behavior occur quite precisely during early development post-hatching and that they coincide with marked changes in behavior (Table 2; and see Rogers, 1995).

Asymmetry in the brain in domestic chicks was discovered by injecting the protein synthesis inhibitor, cycloheximide, into either the left or right hemisphere on day 2 after hatching (Rogers and Anson, 1979). Injection of the left hemisphere impaired the chick's ability to peck at food grains and avoid pecking at pebbles in a task subsequently called the pebble-floor task (grains of mash scattered on a background of small pebbles stuck to the floor). In contrast, treatment of the right hemisphere with cycloheximide had no effect on performance of this task. Thus the ability of the left hemisphere to categorize stimuli (grain versus pebbles) had been revealed (Rogers and Anson, 1979) and later was shown to be localized in the hyperpallial (Wulst) region of the left hemisphere (Deng and Rogers, 1997). This laterality can also be demonstrated by testing chicks monocularly since each eye sends its inputs mainly to the opposite hemisphere: chicks can perform the pebble-floor task when using their right but not left eye (Mench and Andrew, 1986).

In a subsequent experiment, injection of cycloheximide into the left or right hemisphere was applied to groups of male chicks of different ages after hatching and then the chicks were all tested on the pebble-floor task on day 14. The results (summarized in Rogers, 2010, and see Table 2) showed that cycloheximide impaired object categorization (grains versus pebbles) when it was injected into the left hemisphere once only at any time between days 2 and days 5 or 6 or on day 8 but not at other ages. Something crucially dependent on protein synthesis in the left hemisphere was happening at these stages of development. Cycloheximide treatment of the right hemisphere had no effect until days 10 and 11, when it then impaired the chick's ability to make the distinction between grain and pebbles. Interestingly, these transitions in susceptibility to cycloheximide coincide with obvious and precise shifts in behavior of untreated chicks. The left-hemisphere-sensitive period from day 2 to day 6 coincides with the onset and offset of the chick's ability to imprint on a visual stimulus. Under natural conditions chicks imprint on day 2 but not before (Workman and Andrew, 1989) and the imprinting period can be extended by raising chicks in the dark but only until day 6 (Parsons and Rogers, 1997). The day-8 susceptibility of the left hemisphere coincides with the first age at which a chick searches for food on its own; prior to that age chicks peck mostly with their beak closed and under instruction by the hen (Workman and Andrew, 1989). Also on day 8, chicks show a peak in looking at each other using their right eye and left hemisphere (Workman and Andrew, 1989) and their magnetic compass becomes functional (Denzau et al., 2013): the latter is known to be a functional specialization of the right eye and left hemisphere (Rogers et al., 2008). Presumably, development of all of these specializations of the left hemisphere would be disrupted by cycloheximide treatment between days 2 and 6 or on day 8.

The right-hemisphere-sensitive period on days 10 and 11 coincides with the time when chicks begin to go out of the hen's sight and there is a sudden increase in sparring and frolicking (play involving low level aggression) and perching (Workman and Andrew, 1989). Days 10 and 11 are also days when structural changes are seen in the hippocampus (Freire et al., 2004) and chicks learn to navigate spatially using distal spatial cues (Freire and Rogers, 2005, 2007). During this stage of development chicks seek transitory separation from the hen (or the imprinting stimulus), followed by restored contact (Gottlieb and Lickliter, 2004), and do so under right hemispheric control. Later research showing that the right hemisphere is specialized to use geometric spatial information based on distal cues, as compared with the specialization of the left hemisphere to use proximal (landmark) cues (Tommasi and Vallortigara, 2001), was consistent with these specializations of the right hemisphere. However, why would treating the right hemisphere with cycloheximide at a time when the right hemisphere is in charge of behavior impair the chick's ability to categorize stimuli? The parsimonious explanation is that interhemispheric connections become functional at this stage of development and this may be the reason for the transient period on days 10 and 11 when disruption of development of the right hemisphere affects pathways in the left hemisphere concerned with object categorization. It might be said that the chick enters a sort of “adolescence” at this age, when it begins to adopt some independence from the hen and the right hemisphere interferes with decisions made on object categories. Other behavior characteristic of the right hemisphere, such as attention to novelty and expression of more intense emotions (e.g., aggression), also predominate during this stage of control by the right hemisphere.

The results obtained using chicks suggest that transitions in behavior might reflect which hemisphere is in charge at a given age. We should consider whether human development follows any such pattern of left and right hemisphere transitions. Indeed, some evidence from fMRI studies of three-month-old human infants indicates that the left hemisphere is dominant in early life, as it is in chicks. The evidence is based on the fact that, in human infants of this age, the left hemisphere responds to a range of sounds, not solely speech sounds (Dehaene-Lambertz et al., 2006). This finding is at odds with the previously held view that the right hemisphere leads the left during ontogenetic development (e.g., Taylor, 1969, and reviewed in Geschwind and Galaburda, 1987). However, Mahmoudzadeh et al. (2013) have recorded asymmetry of activity in the premature human brain equivalent to about one or two months before birth: left-right differences were found in response to hearing syllables (larger on the right) and changes in phonemes (only in the left frontal region). These results show very early lateralization in linguistic processing and point to complementary roles of both hemispheres. This appears to be another area of interest in which research on animal models could be informative for understanding the sequential development of brain lateralization in humans.

In humans, adolescence stands out as a particularly clear phase of development. It is characterized not only by changes in neural structure and connectivity (Brenhouse and Andersen, 2011) but also by changes in emotional reactivity and increased risk-taking behavior (Arnett, 1992; Steinberg, 2008). Casey et al. (2008) argued that delayed functional connectivity of structures in the prefrontal cortex is associated with impulsivity in adolescents. It is known that the right anterior insular cortex and other regions of the right hemisphere are involved in the feelings experienced in response to events happening in the immediate time-frame (Craig, 2009) and in risk taking, as in gambling (Seeley et al., 2006). Other evidence shows that lower basal levels of neural activity in the right prefrontal cortex are associated with increased risk-taking behavior (Gianotti et al., 2009), possibly because there is lesser inhibition of functions of the right hemisphere. This evidence tends to point to adolescence as a stage of development during which the right hemisphere has a special role. In fact, greater involvement of the right hemisphere during adolescence would also be consistent with the expression of intense emotions, including hostility and aggression.

Incidentally, this does not imply that during adolescence, or any other stage of development, the right hemisphere is used exclusively while the left hemisphere is silent. Biases in activity, and control, would be relative and most likely vary from region to region in the brain. The popular concept of a person being “left-brained” or “right-brained” is absurdly simplified and erroneous. As recent neuroimaging procedures have revealed, all people use both sides of the brain to perform the same task (Nielsen et al., 2013) although there are lateral biases of activity to the left or right depending on brain region and task. It is known that left- and right-handed primates differ in behavior (Cameron and Rogers, 1999; Gordon and Rogers, 2010), and something akin to this may occur in humans (Faurie and Raymond, 2004), but it is not a matter of a left-handed person being right-brained and a right-handed person being left-brained. This qualification is also applicable at the structural level: right-handed common marmosets have a larger SII in the right hemisphere than in the left hemisphere but left-handed ones are not the mirror image of this (Gorrie et al., 2008).

Experience and Development of Lateralization

Although early concepts of lateralization in the human brain recognized to some extent age-dependent changes in the presence and strength of laterality, the general picture was static in the sense that development involved change toward the final, preprogrammed specialization of the hemispheres. Furthermore, any consideration of influences outside the genetic program were not taken into consideration, apart from a potential role of sex hormones and that as a direct consequence of the individual's genetic sex (e.g., Geschwind and Galaburda, 1987).

Studies of animal models have shown how incorrect that view was. A crucial effect of exposure to light on the development of asymmetry of visual processing was first found in the chick model (Rogers, 1982). During the last three days of incubation, the chick embryo is turned in the egg so that its left eye is occluded by its body and only the right eye can be stimulated by light passing through the shell and membranes (Rogers, 1990, 1997). This postural asymmetry is encoded in the genes (embryos that turn in the opposite direction are extremely rare) and the consequent asymmetry of light stimulation leads to asymmetrical development of the visual pathways projecting to the pallium (forebrain). Hence, there are long lasting effects on lateralized behavior expressed in response to visual stimuli. Chicks hatched from eggs exposed to light for a brief period at any time during the final days before hatching use their left hemisphere to distinguish grain from pebbles and their right hemisphere to respond to predators, attack other chicks and copulate. Chicks incubated in the dark during the final days of incubation lack these asymmetries and chicks that have been manipulated so that the left eye of the embryo is exposed to light while the right eye is occluded have reversed asymmetry in that they use their right and not their left hemisphere to distinguish grain from pebbles (Rogers, 1990; summarized in Rogers, 2008).

Exposure to light during the final stages of incubation also establishes the ability of chicks to discriminate objects based on whether they are placed on their left or right side (Chiandetti and Vallortigara, 2009). Chicks hatched from eggs incubated in the dark are unable to use left-right cues to distinguish between objects. It is possible that the light-induced ability to use cues indicating the difference between left and right explains the results of Casey and Sleigh (2013), who recently reported that, in quail chicks, unilateral visual experience before hatching generates a population bias in turning behavior measured posthatching.

Similar effects of light on the development of visual lateralization have been shown in the pigeon: light exposure of the pigeon embryo induces specialization of the left hemisphere to categorize stimuli and decreases the right hemisphere's visuomotor speed (Skiba et al., 2002). In response to light exposure, a structural asymmetry develops in the tectofugal visual system, whereas in the precocial chick it does so mainly in the thalamofugal visual system. This species difference was thought to be determined by the stage of development of each visual system at the time of light exposure (Deng and Rogers, 2002), but it now seems to be a clear species difference since no asymmetry has been found in the thalamofugal pathway of the pigeon regardless of light exposure, dark rearing or exposure of either the left or right eye to light (Stöckens et al., 2013). Exposure to light induces asymmetry at a cellular level also: for example, in the visual regions of the pigeon's midbrain, light exposure leads to larger GABAergic cells on the left side and parvalbumin-immunoreactive cells on the right side (Manns and Güntürkün, 2003).

Apart from the particular importance of these findings in understanding the development and behavior of birds, they raise the possibility that light might influence the development of laterality in other species. Indeed, the zebrafish model has provided another means of investigating the influence of light exposure on development of lateralization (Andrew et al., 2009). Budaev and Andrew (2009a, b) exposed embryos of zebrafish, in their transparent eggs (Fetcho and Liu, 1998), to light or raised them in the dark. They found that the fry hatched from eggs exposed to a normal light/dark cycle were lateralized in that they avoided a model predator when it was presented to them on their left side and approached it when it was on their right side. This asymmetry was not present in fry raised in the dark. Prenatal exposure to light can also generate lateralization in live-bearing fish, as shown by exposing gravid goldbelly topminnow females, which are largely transparent, to high and low levels of light (Dadda and Bisazza, 2012): lateralization was present only in those exposed to high levels of light.

Budaev and Andrew (2009a) found sensitive periods for the effect of light exposure on the development of asymmetry in zebrafish as early as days 1 and 3 of incubation and pointed out that this was too early for it to be due to the development of asymmetry in the visual pathways, as in the example of light effects on lateralization of birds discussed above, but instead it may be an effect on the habenular nuclei. There are no retinal photoreceptor cells at these early stages of development but there are several melanopsins and other opsins expressed in extraocular tissues. Budaev and Andrew (2009a, b) have suggested that light exposure may affect the activation of genes important in left-right differentiation (see Gamse et al., 2005; Liang et al., 2000; and a summary by Roussingé et al., 2012). Stem cells of the habenulae are of two kinds (Concha et al., 2009), those on the left side appearing as early as 24 h after the beginning of development and those on the right side not appearing until 48 h after this time (Aizawa et al., 2007): that is, there is a temporal asymmetry of first left then right neurogenesis. Light exposure on day 1 could perhaps alter the sequential activation of neurones on the left and right sides and so change the pattern of lateralization. An alternative route by which very early exposure to light might influence the development of behavioral lateralization is via melatonin containing cells: De Borsetti et al. (2011) have shown that light exposure and melatonin schedule the differentiation of neurons in the habenular nuclei and start the circadian oscillator in the pineal cells.

The hypothesized role of the habenular nuclei in the lateralization generated by light exposure has yet to be tested by conducting anatomical studies but it would be well worth doing so since the habenulae have extensive connections to and from other brain regions (Aizawa, 2013; Aizawa et al., 2005; Villalón et al., 2012) some of which are asymmetrical, such as the asymmetrical inputs from the olfactory bulbs predominantly to the right habenula (Miyasaka et al., 2009). Moreover, asymmetries of the habenular nuclei and connected structures, such as the parapineal (see Gamse et al., 2005), have clear effects on behavior (Barth et al., 2005; Facchin et al., 2009). Zebrafish with spontaneous reversal of asymmetry of the epithalamus, specifically of the parapineal, show longer latencies to initiate swimming compared to controls and swim shorter distances than controls (Facchin et al., 2009). However, Facchin et al. (2009) found no differences in motor responses to dark flashes or looming shadows presented on the left or right in wild-type zebrafish than in zebrafish with reversed asymmetry.

Much research is needed to test whether light exposure at an early age has lateralized effects on gene activation and neurogenesis and to see whether this is in any way linked to asymmetry of behavior. Available methodologies should now make it possible to integrate genetic and experiential influences on asymmetry and do so in terms of not only molecular and structural asymmetries but also behavioral lateralities.

As a sequel to the discovery of early light effects on laterality in zebrafish, similar evidence for a very early effect of light on laterality in chicks has now been reported. Chiandetti et al. (2013) have found that exposure of chick embryos to light early during incubation, well before the visual connections to the forebrain develop, affects the development of a particular behavioral lateralization. Embryos were exposed to light from the first to the third day of incubation inclusive and on day 4 after hatching they were tested on a task in which they stood with their head though a hole and pecked an array of grains placed in a grid pattern. They showed a preference (not an exclusive one) to peck at the grain on their left side. The same bias was seen in chicks hatched from eggs exposed to light only during the last 3 or 4 days of incubation but chicks hatched from eggs incubated in the dark were not lateralized in this task. It is interesting that both early and late exposure to light had the same effect on this measure of laterality, especially since the two times of exposure must depend on different routes of action (i.e., provided that the late effect on asymmetry in this task depends on development of asymmetry in the thalamofugal visual pathway, which has yet to be determined). Certainly, early exposure to light has no influence on the development of lateralization measured in the pebble-floor task (explained above) or on lateralization of attack and copulation responses, as demonstrated in the original experiment showing the effects of light on development of lateralization (Rogers, 1982; Zappia and Rogers, 1983). In this experiment, chick embryos received exposure to light up until the final 5 days of incubation and were then incubated in the dark or exposed to light. Only the latter group showed lateralization on the pebble-floor task and laterality of attack and copulation. Therefore, late but not early exposure to light influences the development of lateralized behavior clearly known to depend on asymmetry of the thalamofugal visual pathway. Hence, the mode of action and outcome in lateralized behavior of the early light exposure shown by Chiandetti et al. (2013) is distinctly different from the previously studied effects of light during the late stages of incubation. It is an area important for future research comparing chicks and zebrafish. Perhaps, it depends on gene activation caused by exposure to light, as the researchers suggest, but let us consider what stages of development the chick embryo undergoes in the first three days of incubation.

The neural developments that take place from day 1 to day 3 of incubation include the initiation of brain differentiation starting on day 2 and including the neurohypophysis, most other basal brain structures and the midbrain, differentiation of the optic vesicles and growth of axons from the retinal ganglion cells (see Table 1.1 in Freeman and Vince, 1974; Rogers, 1995). On day 3, the embryo already responds to light stimulation by increasing its activity and its head is turned to the left side with the left eye positioned next to the yolk sac. It is conceivable that light exposure at this time could have asymmetrical effects on the developing visual retina and the primary visual connections from retina to midbrain. At least this is an alternative hypothesis to test in attempting to elucidate the mechanism by which early light exposure affects behavioral asymmetry on the grid-pecking task (i.e., in addition to the proposal that the light activates genes involved in left-right differentiation).

It seems that the effects of light exposure may be more broadly distributed than first expected. In birds at least, the effect of light is not confined to one side of the mid-brain or to one hemisphere: there is evidence that light exposure of the embryo also enhances interhemispheric communication. Manns and Römling (2012) trained each hemisphere of the pigeon on a different transitive inference task of pecking at pairs of colored keys and they did so by occluding first one eye and then the other. The trained birds were tested binocularly and they had to combine the color choices learnt separately by each hemisphere. In other words, they had to integrate the information acquired by each hemisphere. Birds that had been exposed to light as embryos could perform the task much better than those incubated in the dark. Hence, light exposure enhanced the efficiency of information exchange between the hemispheres. This finding raises interest in further research on other abilities requiring interhemispheric communication and it demonstrates one advantage of having a lateralized brain (see below for more discussion of this).

It remains to be seen whether light exposure of the mammalian embryo or fetus has an influence on lateralization. However, recent research by Rao et al. (2013) has shown that, in mice, development of the retina depends on light stimulation during the later stages of embryonic development: light exposure regulates the development of retinal neuron number and retinal vasculation. These researchers found that melanopsin containing retinal ganglion cells respond to light reaching the embryos in utero. Another paper has presented evidence that light can reach the developing human fetus although clearly the amount varies depending on external illumination and thickness of the mother's abdomen (Del Giudice, 2011). Del Guidice (2011) reasons that at least some fetuses could receive ample light exposure to stimulate the visual pathways during the later stages of gestation. Indeed the stage of development when the fetus responds to visual stimulation is generally at 30–32 at weeks of gestation but sometimes as early as 26 weeks (Fulford et al., 2003; Kiuchi et al., 2000). Since the fetus adopts a lateralized position in utero at these stages of development (some two thirds are in the left occiput position; Matsuo et al., 2007), it is a possibility that light stimulation is asymmetrical and may lead to asymmetries in structure of the visual pathways and visual function. Findings made using animal models support the value of looking to see whether this is the case.

A first attempt to address the potential role of visual experience on the development of two particular lateralities in humans, head-turning preference and hand preference, has been made by comparing congenitally blind adults with sighted adults (Nava et al., 2013). The normal subjects showed a preference to turn their head to the right side, whereas the blind subjects turned their head to the left side. Hand preference was the same in both groups, although another study (Ocklenburg et al., 2010) has reported that early postnatal visual experience affects hand preference. Nevertheless, the results indicate that visual experience, at least postnatally, has an effect on laterality of head turning. However, it tells us nothing about the possible effect of light stimulation before birth.

Experience in other sensory modalities might also influence the development of certain types of lateralization. In some of the earliest research on lateralization in rodents, a specific type of early life experience was shown to modulate the presence or absence of lateralization (Denenberg, 1981). Separating rat pups from their mother for a few minutes per day over the first 21 days of life unmasks the specializations of the right hemisphere and alters interhemispheric communication via a change in the size of the corpus callosum: the corpus callosum of handled pups is larger than that of nonhandled pups possibly because it undergoes less attrition of neurons during development (Cowell and Denenberg, 2002). Nothing is known of the specific sensory stimulation that mediates this effect. Although Denenberg (2005) suggested that it could be dependent on the stress caused by maternal separation, tactile stimulation may well be involved since pups that have been handled receive more licking than normal when they are returned to their mother.

So far only the effects of visual experience and handling on the development of lateralization have been investigated in vertebrate species. Olfactory and auditory experience may also influence the development of lateralization. Indeed, it has been shown recently that laterality of turning behavior in the cuttlefish Sepia officinalis is modulated by exposing the eggs to the odor of seabass, a predator of the species (Jozet-Alves and Hébert, 2012). Left side turning bias in the hatched cuttlefish is enhanced by this chemosensory experience. Since visual input triggers the turning response, as indicated by a correlation between asymmetry of optic lobe size and leftward turning bias (Jozet-Alves et al., 2012), there must be a cross modality effect of being exposed to the odor during embryonic development.

Future research is likely to examine how cross-modal interactions might influence the development of lateralization. In humans, Hoefer et al. (2013) have shown that tactile stimulation on the left side of the body can enhance auditory detection on the left side, a result that suggests interaction between input in these two modalities in the right hemisphere. Although this has nothing to do with development, it alerts one to the possibility of intermodal interactions on lateralized development. Such studies could be conducted with precision only using animal models.

Hormones and the Development of Lateralization

Gender differences in laterality in humans have often been hotly debated. Generally, researchers of laterality in humans have been all too ready to assume that any sex difference is caused by action of the sex hormones on the brain, either as an organizing effect during development or an effect of circulating levels of hormone at the time of testing (e.g., Geschwind and Galaburda, 1987). As an example of possible circulating effects of sex hormones, Kemali et al. (1990) noted seasonal changes in the size of habenular nuclei of frogs and suggested that this might depend on changing levels of sex steroid hormones. Changes in functional lateralization in different phases of the menstrual cycle are found in humans and may also depend on changing levels of the sex hormones (e.g., Weis et al., 2008).

It is extremely difficult to separate the potential influences of sex hormones from the influences of experience that differs according to an individual's sex (discussed more generally by Rogers, 2001). In animals, differences in laterality have been reported in males versus females and, again, this is assumed to depend on differences in the sex hormones. In a number of examples, males have been found to be more strongly lateralized than females although both are lateralized in the same direction (examples discussed in Pfannkuche et al., 2009). A recent study found that both male and female zebrafish displayed a left-eye preference to view their image in a mirror but the eye preference was stronger in males than females (Ariyomo and Watt, 2013). Earlier studies of chicks found that males have stronger asymmetry than females in the thalamofugal visual pathway (Rajendra and Rogers, 1993) and in the left-right eye difference in performing the pebble-floor task (Mench and Andrew, 1986).

The latter sex difference may result from an interaction between sex hormone levels and the effects of exposure to light during the final days of incubation, since embryos exposed to high levels (probably above physiological levels) of either testosterone or estrogen during the final days of incubation fail to develop asymmetry of the visual pathway (Rogers and Rajendra, 1993; Schwarz and Rogers, 1992).

Although sex differences established at early stages of development may persist into adulthood, sex differences measured in adults also have the strong possibility of being generated by differences in experience between males and females. Sex differences in hand or paw preference are a typical candidate for the latter. For example, Wells and Millsopp (2009) reported that female cats have a right paw bias, whereas males were more inclined to use their left paw. This sex difference is commonly assumed to be caused by the action of sex hormones. However, experience alone may alter which hemisphere is in dominant control of the forelimbs during manipulation of objects. If, for example, females have a social role that demands vigilance, face recognition, and broad attention to novel stimuli, their right hemisphere might be in more frequent use than their left and along with this they may be biased toward using their left paw in certain tasks. Certainly, we know that left- and right-handed marmosets have different cognitive styles: the left-handed ones are less likely to explore an unfamiliar environment (Cameron and Rogers, 1999) and less readily show social facilitation of feeding responses (Gordon and Rogers, 2010). Although these differences are not associated with sex, they illustrate the fact that hand preference and cognitive style are related. When these two variables do differ between males and females, it could be because experience favors one or the other cognitive style depending on biological sex. In other words, we should not assume that the sex hormones are the unitary cause of the difference.

Other factors may interact with sex-related differences in lateralization. Ward et al. (1993) found that right-hand preference increases with age in lemurs. One explanation these researchers gave for this result was that hand preference might shift along with age-related change in social dominance: older females are dominant over younger females. This could mean that older females use their left hemisphere more than do younger females, who might need to be alert for novel stimuli and social interactions, both of which are specializations of the right hemisphere. The point to stress here is that indirect effects of sex hormones via social behavior need to be taken into account.

At the Individual Level

It has long been assumed by researchers of hemispheric differences in humans that a lateralized brain is more efficient because it allows one hemisphere to assume control (e.g., as in control of language) without competition from the other (Levy, 1969). Some evidence supports this hypothesis: for example, Gotts et al. (2013) have attempted to quantify the degree of laterality in terms of the left hemisphere's specialization to interact within that hemisphere versus the right hemisphere's specialization to interact across both hemispheres. They found that the stronger this measure of lateralization, the better was the subject's verbal and spatial ability. However, a given human brain has many functions, some of which may be strongly lateralized and others less so or not at all. Therefore, how can we test the hypothesis about the benefit of lateralization more precisely? On what measure or measures should it be deduced that one individual is more lateralized than another? Humans cannot be conveniently categorized as “lateralized” and “nonlateralized” and so compared in cognitive abilities. There have been attempts to do this using hand preference as the measure and some evidence indicates that strength of hand preference is weaker amongst people suffering from a range of conditions, including severe depression (Denny, 2009), schizotypy (Shaw et al., 2001) and post-traumatic stress disorder (Spivak et al., 1998). If hand preference is a useful measure of strength of brain lateralization, these results give some support to the hypothesized benefit of brain lateralization but this evidence is by no means convincing in itself.

Whether having a lateralized brain aids cognitive efficiency can be tested using animal models and precisely controlled experiments. However, as in humans, what behavior is used as a measure of direction and/or strength of lateralization remains a problem, especially if only one test is used to assess laterality. In animal models, it is possible to avoid this problem by manipulating the development of laterality in an entire sensory modality (e.g., in visual perception by light versus dark exposure during early development) or by comparing wild-type animals with those with spontaneous reversals of laterality (e.g., Facchin et al., 2009).

In birds, it is possible to take advantage of the influence of light on the development of visual lateralization to manipulate strength and direction of asymmetry. Birds hatched from eggs exposed to light can be compared with birds hatched from eggs that receive no exposure to light during the final stages of incubation. For example, light-exposed and dark-incubated chicks were tested on a dual task in which they performed the pebble-floor task (use of the left hemisphere) while simultaneously monitoring overhead for a model predator, a silhouette of a bird of prey (use of the right hemisphere; Rogers et al., 2004). Chicks that had hatched from eggs exposed to light performed well on both tasks, whereas those incubated in the dark performed poorly. The latter were unable to avoid pecking at the pebbles (they pecked randomly at grain and pebbles) and they were more likely to miss seeing the predator or, if they did catch sight of it, they were so distracted that they took longer to return to searching for food (Dharmaretnam and Rogers, 2005). A similar result was found later in fish, Girardinus falcatus (Dadda and Bisazza, 2006): in the presence of a predator, lateralized fish were able to capture brine shrimps twice as readily as were nonlateralized fish and they used a different eye/hemisphere for each task.

As mentioned above, Manns and Römling (2012) have shown that pigeons with visual lateralization exchange information more efficiently between the hemispheres than do pigeons not exposed to light before hatching. Although each hemisphere is engaged in its own specialized function, on-going performance by the bird requires integration of the processes occurring, and decisions made, in each hemisphere. It has also been shown that chicks hatched from eggs exposed to light are able to use both hemispheres to guide their behavior, whereas those incubated in the dark use only one hemisphere and that depends on which eye is being used when they are tested monocularly (Chiandetti et al., 2005).

Animals with reversed or no asymmetry can also be tested for differences in behavior compared with those with normal asymmetry. As mentioned previously, zebrafish with left-right reversal of the habenular asymmetry are more fearful and respond more strongly to stressors than fish with normal left-right asymmetry. Mice normally display asymmetry in neural circuitry of their hippocampus (Kawakami et al., 2003) but there are mutants in which this asymmetry of the hippocampus is absent. The latter are slower to learn a spatial task and have weaker retention of working memory (Goto et al., 2010 and see later).

Other examples of superior performance of lateralized compared to nonlateralized individuals are summarized in Rogers et al. (2013). Evidence supporting the advantage of being lateralized in performance of dual tasks has not yet found strong support in experiments with humans. For example, Hirnstein et al. (2008) tested humans on a dual task in which they had to respond to images of faces/nonfaces and words/nonwords presented simultaneously and they found better performance in subjects with weaker or no lateralization than in those with strong lateralization. However, the tasks were presented in a lateralized manner to either the left of right visual field and this could be a reason why the results did not correspond to those obtained testing other species: as the researchers noted, in the tests on chicks and fish, the animals could choose which eye/hemisphere they used for each task, whereas the humans could not. Later Hirnstein et al. (2010) found that performance in humans was better with an optimal degree of lateralization rather than with very strong or weak lateralization. This result is not inconsistent with the findings in studies of nonhuman species.

It now seems to be quite clear, in nonhuman animals at least, that any disadvantages of being lateralized, such as being less vigilant of predators on the right side, are counterbalanced by superior cognitive performance. Nevertheless, this does not in itself explain why the majority of individuals are lateralized in the same direction (i.e., as a directional bias within the population or species).

At the Population Level

On finding that groups of light-exposed (lateralized) chicks form more stable hierarchies than do groups of dark-incubated chicks (nonlateralized for visual performance), Rogers and Workman (1989) hypothesized that population biases of asymmetry might be advantageous in social situations (see also Vallortigara and Rogers, 2005). This hypothesis was tested by Ghirlanda and Vallortigara (2004) using computational models: they found that population biases emerged when lateralized individuals had to interact with each other, further suggesting that it would be worth looking for a link between lateralization and social behavior.

If laterality within a population is associated with social behavior, social species should show population biases and solitary species might show individual lateralization but no population bias. First, this was tested by looking at the relationship between schooling behavior, and turning biases in sixteen different species of fish (Bisazza et al., 2000). The results fitted the hypothesis: population biases in turning were common in fish with strong tendencies to school but absent in those with weak or no propensity to school.

The hypothesis is also being tested in bees of different species, comparing solitary with eusocial species. Eusocial bees, including honeybees and three species of stingless bees, have been tested behaviorally using either their left or right antenna in learning and recall of a task requiring them to associate a specific odor with a food reward. The right but not the left antenna is used in learning (Letzkus et al., 2006) and short-term recall of this association (Rogers and Vallortigara, 2008). Bees extend their proboscis in anticipation of a food reward when they detect the odor used in training in association with a reward of sugar. The population bias is close to 100%. No such population-level asymmetry was found in solitary mason bees (Anfora et al., 2010). In addition, the number of olfactory receptors on the left and right antennae of honeybees is biased towards a somewhat higher number on the right in honeybees but mason bees have no significant asymmetry of receptor numbers (Anfora et al., 2010). These studies provide further evidence linking population level asymmetry with sociality, although more species of bee need to be tested.

Another way of testing the hypothesis is to see whether social behavior is itself lateralized at the population level. Indeed, in a wide range of species, agonistic interactions are lateralized in this way: stronger expression of agonistic interactions to conspecifics on the left side has been reported to occur in fish, amphibians, reptiles, birds, and mammals, the latter including baboons and horses (summarized in Rogers, 2002; Rogers et al., 2013a). Population-level bias to view conspecifics or a mirror image of the animal itself is another example of this: for example, a left-eye preference for image viewing has been shown in a species of fish (Perccottus glenii) and it seems to be depend on seeing the image of an eye, which is a pertinent (or releaser) stimulus (Karenina et al., 2013). Left-eye-and-right-hemisphere advantage for recognizing faces, familiar versus unfamiliar conspecifics and social relationships between conspecifics has been shown also in a number of vertebrate species, especially in chicks (summarized by Salva et al., 2012).

Population-level laterality in social behavior is also present in invertebrates. A recent study (Rogers et al., 2013b) showed that pairs of honeybees using their right antennae (the left antenna was removed) display appropriate social behavior, in that they perform positive interactions (extend the proboscis) when they encounter members of their own colony and negative interactions (C-responses, aggressive behavior in which the body is arched into a posture adopted for stinging) when they meet a member of another colony. This was in contrast to inappropriate behavior shown by bees using their left antennae: the latter performed C-responses in encounters with members of their own colony and not with members of another colony.

Hence, the antennae of honeybees are lateralized at the population level and are used differentially in social interactions (see review by Frasnelli et al., 2014). Considered together, all of these studies indicate that social behavior and the presence of a consistent directional bias of behavior go hand-in-hand.

Lateralization of Learning and Memory

Animal models have been invaluable in understanding the biological basis of learning and memory. Although opportunistic research on memory in humans (e.g., examining patients suffering from trauma, epilepsy, etc.) and tachistoscopic presentation of visual stimuli in the lateral visual fields are methods that have provided some evidence of the location of memory traces, outcomes have been limited since they require the precise methods of neurobiology possible only in animal models. From animal models, we know many details of the cascade of molecular changes that correlate with memory formation (Rose, 2000) and, of interest here, lateralized processes have been discovered.

One of the earliest demonstrations of lateralization in birds concerned memory formation and recall. In the 1980s, the young chick emerged as a model to study the neural correlates of learning and memory (see Rose, 1992). The chick was an ideal model for this research because very soon after hatching chicks must rapidly form important memories to ensure their survival. The strength of the imprinting memory was chosen as a model in the investigation of learning and memory because it was expected to involve measureable molecular and structural changes that could be linked to a precise learning event. Indeed cellular changes associated with imprinting were localized to an associative region of the forebrain in each hemisphere now called the intermediate medial mesopallium, IMM (Horn et al., 1973). These changes included increased RNA and protein synthesis (see Horn, 1985) and increased length of synaptic apposition. Lesions of the left and right IMM prevented the chicks from imprinting.

In terms of laterality, the interesting findings concern recall of imprinting memory. Lesions that removed the IMM prevented recall of short-term memory of imprinting and did so identically on the left and right sides (Cipolla-Neto et al., 1982). However, L-R asymmetry was evident in recall of long-term imprinting memory. In this case, lesion of the left but not the right IMM prevented recall. In the right hemisphere, long-term memory had been encoded somewhere outside the IMM. In the left hemisphere, the IMM was the location of both long- and short-term memory. This was the first evidence that short- and long-term memory can be encoded in different regions of the brain.

Further results showed differing locations within the chick brain for short- and long-term memory of a passive avoidance learning task. In this task, chicks are trained to avoid pecking a colored bead by allowing them to peck it once after it has been coated with a noxious tasting fluid, methyl anthranylate (Rose, 1992). Lesions of IMM prevented recall of short-term memory of this task but they did not prevent recall of long-term memory. The latter was prevented by lesions in an entirely different region of the brain known as the lobus parolfactorius and the shift to recall from this region involved lateralized processes (Johnson and Rose, 2002; Rose, 1992). In fact, interaction between memories held in the left and right hemispheres occurs at precise intervals following learning and some aspect of emotional content from the right hemisphere seems to be necessary if the memory is to be retained long-term (Andrew, 1997, 2002b; summarized in Rogers et al., 2013a). Similar interactions between lateralized processes may explain why Goto et al. (2010) found that mutant mice without asymmetry of the hippocampus have weaker retention of working memory.

Recently use of different neural pathways to recall short- and long-term memory has been found in honeybees (Rogers and Vallortigara, 2008). Following on from the discovery that bees are able to learn to associate an odor with a reward of sugar provided that they use their right antenna but not their left (discussed earlier), the bees' ability to recall this association was tested at different times after they had been trained (i.e., trained with both antennae in use). By coating either the left or right antenna with a latex substance, it was possible to limit sensory perception of the odor to the uncoated antenna during tests of recall. In the recall test presentation of the odor used in training elicits extension of the proboscis (proboscis extension response, PER) in the absence of a sugar reward if the memory is retrieved. At 1 h after training, PER occurs when the right antenna is in use (short-term memory). Recall tests at 6 and 24 h after training led to PER only when the left antenna was in use. Hence retrieval of short- and long-term memory uses different neural circuitry, a result in line with the earlier demonstration of lateralized memory processes in the chick. It seems that further studies of these lateralized processes concerning memory consolidation and recall during development would be valuable.

Although we do not know how widespread the location-specific aspects of memory are, finding them in model species so vastly different (chicks and bees) suggests that they may be a common aspect of memory consolidation and retrieval. This is precisely how discoveries in model species can underpin advances in the study of human brains. Until now, the role of the hippocampus, amygdalae, and temporal lobes in memory in humans has been studied in much detail but, as far as I know, there has been little research on lateralization of memory when both hemispheres are trained (i.e., with bilateral presentation of stimuli) or differing locations of short- and long-term memory. Furthermore, lateralized processes in emotion-related learning, which involve the amygdala (Sah et al., 2008), are potentially of importance in understanding and treating anxiety and post-traumatic stress disorders. As mentioned previously, these disorders in humans have been associated with weaker laterality, in some cases as measured by hand preferences. Weaker paw preference in dogs is associated with heightened fear responses to auditory stimuli (Branson and Rogers, 2006), which suggests that similar behavioral disorders may be related to laterality in nonhuman species. Even Xenopus tadpoles with reduced or reversed laterality have been shown to learn more slowly than tadpoles with normal (wild-type) laterality (Blackiston and Levin, 2013).