Behavioral and brain asymmetries in primates: a preliminary evaluation of two evolutionary hypotheses

Abstract

Contrary to many historical views, recent evidence suggest that species-level behavioral and brain asymmetries are evident in nonhuman species. Here, we briefly present evidence of behavioral, perceptual, cognitive, functional, and neuroanatomical asymmetries in nonhuman primates. In addition, we describe two historical accounts of the evolutionary origins of hemispheric specialization and present data from nonhuman primates that address these specific theories. Specifically, we first discuss the evidence of that genes play specific roles in determining left–right differences in anatomical and functional asymmetries in primates. We next consider and present data on the hypothesis that hemispheric specialization evolved as a by-product of increasing brain size relative to the size of the corpus callosum in different primate species. Lastly, we discuss some of the challenges in the study of hemispheric specialization in primates and offer some suggestions on how to advance the field.

Behavioral and brain asymmetries are key features of the human central nervous system. For instance, it has been known for quite some time that most humans are right-handed^1–5 and, under conditions of typical development, most humans are left hemisphere dominant for language and speech.^6–8 The distinct pattern of right-handedness and left hemisphere dominance for language is robust and consistent enough that atypical patterns of asymmetries are often considered a consequence of early developmental insult or genetic mutation, and thereby often linked to a variety of neurological and neurodevelopmental disorders. For instance, it has been fairly well established that atypical rightward patterns of neuroanatomical and functional asymmetries in the brain, particularly in the planum temporale, and increased incidences of left- or mixed-handedness are associated with schizophrenia.^9–17 Similar patterns of atypical functional and neuroanatomical asymmetries as well as handedness have been reported to be associated with autism,^18–25 stuttering,^26,27 specific language impairment,²¹ and other neurological conditions.

The compelling evidence of atypical behavioral and brain asymmetries and their associations with a variety of neurodevelopmental and neuropsychiatric conditions has led to numerous theories on the origin of hemispheric specialization and, essentially, what processes go wrong during ontogeny to result in these different clinical problems.^9,28,29 Though we lack direct evidence of specific genes that play roles in the determination of left or right hemispheric functional or neuroanatomical asymmetries, many theories start from the premise that cortical organization, including hemispheric specialization, is in some way genetically determined (i.e., dominance for a given function within one hemisphere) and that perturbations, either genetic or non-genetic in form, alter the normal pattern of left–right differentiation.^9,30–37

The role of genes on the expression of population-level asymmetries in humans is not restricted to developmental theories but also has been postulated in the context of evolutionary theory.^30,36 Until about 25 years ago, the prevailing view in the literature was that population-level behavioral, neuroanatomical, and functional asymmetries were unique to humans.^38–40 With some exceptions, this view was primarily supported by a lack of evidence for behavioral and brain asymmetries in vertebrates, although the body of literature at that time was rather limited in terms of (1) the scope of the measures and (2) the diversity of species that had been studied. Further reinforcing this view was the fact that the most robust and well-documented manifestations of hemispheric specialization in humans are related to language functions and handedness. Thus, many theoretical accounts proposed an explicit and causal link between the evolution of hemispheric specialization and the emergence of uniquely human cognitive and motor specialization such as language, bipedalism, and tool use.^41–43

We now know that this perspective was wrong in terms of the assumption that language was a necessary condition for hemispheric specialization. There is now a large body of evidence that population-level behavioral and brain asymmetries can be found in a wide range of vertebrates^44–46 and even in some invertebrates.⁴⁷ For example, toads prefer to use the right forelimb to remove a substrate from the head.⁴⁸ Some fish show right eye preferences for viewing predators⁴⁹ and a number of bird species show left or right hemisphere asymmetries in visual discrimination for a variety of different classes of stimuli.⁵⁰ Several excellent reviews of the evidence of behavioral and functional asymmetries in nonhuman animals have recently been published, and a summary of this extensive literature is not provided here.⁴⁴

Instead, in this review, we briefly summarize the evidence of population- level behavioral, functional and brain asymmetries in nonhuman primates. Primates are of specific interest for several reasons. First, humans are primates and we share a common and more recent evolutionary past with other primates, and therefore the likelihood of identifying the antecedent conditions for homologous lateralized functions in humans is greater by studying nonhuman primates. Second, a distinguishing feature of primates is the evolution of nails instead of claws and the use of an opposable thumb for grasping.⁵¹ Indeed, hands distinguish primates from non-primates, and therefore studies on handedness seem particularly appropriate for different primate species. We follow the summary of behavioral and cognitive asymmetries by discussing two views of the evolution of hemispheric specialization and present recent data from studies in chimpanzees and other primates from our laboratory that address these views.

Primate laterality

Handedness

There have been a plethora of studies that have examined handedness in nonhuman primates, and many studies have differed with respect to the species under investigation, the settings (captive versus wild), and age and sex compositions.^52–55 Based on previous reviews of data on handedness in nonhuman primates, it is now clear that task complexity is a critical factor in the assessment of individual hand preferences in nonhuman primates.^56,57 Notably, all things being equal, tasks that require bimanual and complimentary use of the hands to perform a manual action are significantly more sensitive to detecting individual preferences than tasks that do not.⁵⁸ For instance, historically, and still today, hand preferences are often assessed for unimanual actions such as reaching for a piece of food.⁵⁹ Numerous studies have shown that simple reaching does not elicit a strong and consistent bias in hand use, at least when subjects have a significant degree of freedom in the grasping action (i.e., they are not reaching through a hole or derived device). In contrast, for tasks that require coordination and complimentary in the use of the hands, such as the TUBE task (see Fig. 1), (1) a statistical majority of subjects show a strong and consistent hand preference⁵⁸ and (2) in some species, the hand preferences are significantly shifted leftward or rightward (see Fig. 1).^58,60 Why there are phylogenetic differences in directional biases in hand preference, even between species within the same genus, is unclear. Some have suggested that posture or habitual positional behavior may explain these results, but additional studies are needed to test this hypothesis. More importantly, with the TUBE task, we at least have a measure of handedness test that allows for reasonable comparisons to be made between species.

Figure 1.

Mean HI scores (± standard error (s.e.)) for the TUBE task in different primate genera: red = great apes, blue = Old World monkeys, green = New World Primates. Significant rightward asymmetries were found for Gorilla, Pan, and Papio. Significant leftward asymmetries are found for Cercopithecus and Rhinopithecus.

Perception of species-specific faces and production of facial expressions

Perhaps the most compelling evidence of continuity in behavioral asymmetries among primates comes from research on the perception and production of facial expressions. Indeed, the strongest experimental evidence of functional asymmetries derives from studies in split-brain macaques by Hamilton and Vermeire where they tested more than 20 individuals on a variety of different kinds of tasks.⁶¹ These authors have found that macaque monkeys discriminate species-specific faces better in the right compared to left hemisphere, while performing a line-orientation discrimination better in the left compared to right hemisphere. These patterns of asymmetry were independent of the handedness of the subjects, leading Hamilton and Vermiere⁶¹ to conclude that cognition, not handedness, is lateralized at the population level in macaques. Additional studies from this group investigating the processes underlying this asymmetry have shown that the right but not the left hemisphere is sensitive to the inversion effect;⁶² that is, learning to discriminate faces that are inverted as compared to their normal orientation is more difficult for the right hemisphere, suggesting that this hemisphere is specialized for processing configurational information. Right hemisphere asymmetries in configurational or holistic processing have also been reported in chimpanzees⁶³ and baboons.⁶⁴

In neurologically intact subjects, attempts to characterize asymmetries in the discrimination and production of facial expressions have relied on two general approaches. First, human subjects have been asked to rate the intensity of chimeric facial expressions of different primate species created from images captured during spontaneous interspecies interactions. Second, also using images obtained during typical intraspecies interactions, raters have quantified the degree of difference in the left and right orofacial mouth region during different expressions. From the chimerics, human subjects have reported that the two left halves of the face were judged to be more expressive than the two right halves in chimpanzees,⁶⁵ baboons,⁶⁶ and, to a lesser extent, rhesus macaques.⁶⁷ For the production of species-specific facial expressions, in general, left hemiface asymmetries (indicative of a right hemisphere asymmetry) have been reported in chimpanzees, baboons, and rhesus macaques.^67–72 In baboons, the presence of left hemiface asymmetries have been found in individuals less than 1 year of age, suggesting that they may be present early in development.⁷³

Auditory processing and communicative production

Given the importance of left hemisphere specialization for speech perception in humans, there have not surprisingly been attempts to examine whether lateralization in the processing of species-specific sounds is similarly present in nonhuman primates. Some time ago, Heffner and Heffner⁷⁴ reported that lesions to the left but not the right posterior temporal lobe resulted in a transient deficit in the discrimination of “coo” vocalizations in Japanese macaques. Right ear advantages have also been reported in the discrimination of species-specific calls as well as pure tone discrimination in macaques and baboons.^75–77 Several recent studies have measured orienting asymmetries in macaques and chimpanzees in response to different classes of auditory stimuli, such as species-specific vocalizations. In this paradigm, an acoustic stimulus is presented to a subject when he/she is facing 180° away from the location of the sound source. During or after the stimulus is presented, the experimenter records which direction (left or right) the subjects orient to the cue, as a means of inferring which hemisphere is processing the signal. Both leftward and rightward orienting asymmetries have been reported in bonobos, macaques, and vervet monkeys, depending on the type of cue,^78–81 but some have questioned the validity of this measure.⁸²

Several recent functional imaging studies using positron emission tomography (PET) have also been conducted in macaques and chimpanzees. In macaques, Poremba et al.⁸³ imaged the brain in response to different acoustic stimuli and found several lateralized regions activated, including a leftward asymmetry in the temporal pole. Gil-da-Costa et al.⁸⁴ also used PET to assess cortical activation in macaques in response to species-specific calls including “coos,” “screams,” and non-biological sounds. These authors found significant activation in the superior posterior temporal gyrus and ventral premotor areas in response to both “coos” and “screams” compared to non-biological sounds; however, consistent left–right differences in activation were not found within this sample. Taglialatela et al.⁸⁵ imaged three chimpanzees as they listened to either proximal species-specific sounds, distal species-specific sounds, or chimpanzee calls played backward. These authors found pronounced rightward asymmetries in the processing of the proximal and particularly the distal vocalization when subtracting out PET responses to the time reversed sounds. Most of the significant regions activated in these tasks were in the middle and superior temporal lobe of the right hemisphere. In addition to perception, Taglialatela et al.⁸⁶ also used PET to assess cortical activation during the production of manual gestures accompanied by the use of attention-getting sounds in chimpanzees. In this study, PET images were acquired while chimpanzees pointed to and made attention-directing vocalizations that were directed to a human experimenter. To control for manual motor movements, a second set of images were obtained when the chimpanzees were grasping small objects and handing them back to an experimenter during the uptake period. The result showed a number of significant regions activated in the gesture–vocal condition, including the bilateral prefrontal and posterior cingulate cortex and left hemisphere activation in the dorsal portion of the inferior frontal gyrus (BA44), the region just anterior to the fronto-orbital sulcus (BA45) and the caudate.

Neuroanatomical asymmetries

Over the past 50 years, there have been a number of studies aimed at assessing individual and phylogenetic variation in neuroanatomical asymmetries. These have included linear measures of sulci length from either endocasts, photographs of brains, or by direct measurement, qualitative judgements of cerebral torque or shape asymmetries, and more recently volumetric measure of various cortical regions. We have summarized, in Table 1, the existing data on neuroanatomical asymmetries in nonhuman primates, for those studies that had at least 10 subjects within a given species (i.e., subjects were not lumped into larger taxonomic groups). Further, we have only included studies or data that measured some dimension of cortical asymmetry using direct methods (i.e., they measured them from endocasts or brains and not from photographs or by subjective judgment). Linear measures of sulci length or surface area are by far the most common means of quantifying asymmetry. From Table 1, on balance, evidence of population-level asymmetries is more prevalent in chimpanzees than in Old or New World monkeys. In particular, chimpanzees show leftward asymmetries in the planum temporale surface area and gray matter volume, SF length, and SF and FO surface area. Rightward asymmetries were found for the inferior postcentral sulcus surface area and the volume of the hippocampus.

Table 1.

Description of neuroanatomical asymmetries in nonhuman primates

Lobe	Species	n	Finding
Temporal Lobe
Sylvian fssure (SF) length
Planum temporale169	Pan troglodytes	18	L > R
Planum temporale170	P. troglodytes	189	L > R
Planum temporale171	P. troglodytes	10	L > R
SF172	P. troglodytes	28	L > R
SF173	P. troglodytes	30	L > R
SF173	Macaca mulatta	30	L = R
SF174	M. mulatta	18	L = R
SF175	M. fascicularis	30	L > R
SF175	M. mulatta	29	L > R
SF176	M. mulatta	108	L = R
SF175	Saguinus oedipus	27	L > R
SF175	Saimiri sciureus	20	L = R
SF175	Callithrix jacchus	26	L > R
SF177	Cebus apella	17	L = R
Sylvian fissure (posterior region only)
SF length (posterior only)172	P. troglodytes	28	L > R
Gray matter volume164	P. troglodytes	82	L > R
Gray matter volume164	M. mulatta	21	L = R
Gray matter volume164	M. radiata	15	L = R
SF length (posterior only)175	M. fascicularis	30	L = R
SF length (posterior only)175	M. mulatta	29	L = R
Gray matter volume164	Chlorocebus aethiops sabaeus	43	L = R
Length of ascending ramus178	P. troglodytes	24	R > L
Length of descending ramus178	P. troglodytes	24	L = R
Superior temporal sulcus
Depth of STS sulcus161	P. troglodytes	70	L = R
Total STS surface area162	P. troglodytes	127	L = R
Caudal STS length176	M. mulatta	73	L = R
Rostral STS length176	M. mulatta	271	L = R
Rostral middle temporal176	M. mulatta	248	L = R
Total STS length179	M. mulatta	29	L = R
Total STS length179	M. fascicularis	30	L = R
Frontal lobe
Central sulcus surface area162	P. troglodytes	127	L = R
Precentral inferior surface area162	P. troglodytes	127	L = R
Superior precentral area162	P. troglodytes	127	L = R
Inferior frontal sulcus area162	P. troglodytes	127	L = R
Fronto-orbital area162	P. troglodytes	127	L > R
Fronto-orbital area 180	P. troglodytes	60	L > R
Inferior frontal gyrus (GM)	P. troglodytes	212	L = R
Central sulcus surface area162	M. mulatta	21	L = R
Central sulcus surface area162	M. radiata	16	L = R
Superior precentral area162	M. mulatta	21	L = R
Fronto-orbital area162	M. mulatta	21	L = R
Superior precentral area162	M. radiata	16	L = R
Fronto-orbital area162	M. radiata	16	L = R
Arcuate length176	M. mulatta	104	L = R
Arcuate surface area162	M. mulatta	21	L = R
Arcuate surface area162	M. radiata	16	L = R
Rectus length176	M. mulatta	109	R > L
Orbitofrontal ength176	M. mulatta	24	L = R
Lateral orbital length176	M. mulatta	197	R > L
Fork of medial orbital176	M. mulatta	227	L = R
Medial orbital length176	M. mulatta	104	L = R
Central sulcus length176	M. mulatta	125	L > R
Arcuate sulcus (horizontal)179	M. mulatta	24	L = R
Arcuate sulcus (vertical)179	M. fascicularis	25	L = R
Arcuate sulcus (horizontal)179	M. mulatta	25	L = R
Arcuate sulcus (vertical)179	M. fascicularis	28	L = R
Principal sulcus179	M. mulatta	26	L = R
Principal sulcus179	M. fascicularis	24	L = R
Principal surface area162	M. mulatta	21	L = R
Principal surface area162	M. radiata	16	L = R
Parietal lobe
Intraparietal sulcus area162	P. troglodytes	127	L = R
Inferior postcentral sulcus area162	P. troglodytes	127	R > L
Superior postcentral sulcus area162	P. troglodytes	127	L = R
Lunate surface area162	P. troglodytes	127	L = R
Parietale operculum181	P. troglodytes	99	L > R
Intraparietal sulcus area162	M. mulatta	21	L = R
Intraparietal sulcus area162	M. radiata	16	L = R
Intraparietal sulcus length179	M. mulatta	17	L = R
Lunate surface area162	M. mulatta	21	L = R
Lunate surface area 162	M. radiata	16	L = R
Planum parietale length178	P. troglodytes	23	R > L
Planum parietale length182	P. troglodytes	57	R > L
Limbic and subcortical regions
Hippocampus volume183	P. troglodytes	60	R > L
Amygdala volume183	P. troglodytes	60	L = R
Anterior cingulate gray matter volume184	P. troglodytes	70	L > R
Posterior cingulate gray matter volume184	P. troglodytes	70	R > L
Anterior cingulate area179	M. mulatta	28	L = R
Anterior cingulate area179	M. fascicularis	30	L = R
Posterior cingulate area179	M. mulatta	28	L = R
Posterior cingulate area179	M. fascicularis	30	L = R

Mechanisms underlying the evolution of behavioral and brain asymmetries: two theoretical frameworks

Saltational genetic views

Despite the evidence of population-level asymmetries in nonhuman species, the mechanisms underlying individual and phylogenetic variation in hemispheric specialization remain poorly understood, particularly in human and nonhuman primates. There are essentially two theoretical views on the evolution of hemispheric specialization. One view holds that the processes and mechanisms underlying hemispheric specialization in humans are fundamentally different than in other primates and indeed all other animals. ^39,87–89 For instance, some have hypothesized that genetic mechanisms underlie the expression of behavioral and brain asymmetries in humans, whereas asymmetries in nonhumans are attributable to other, nongenetic factors. Specifically, both single- and two-allele genetic models of behavioral asymmetries have been proposed for humans and they are explicitly linked to the evolution of language and speech.^1,33,90 For example, the right shift (rs) theory proposed by Annett¹ postulates that a single allele codes for left hemisphere dominance for language. Homozygotic or heterozygotic individuals with the rs⁺ allele become left hemisphere dominant, whereas language dominance in rs⁻ individuals is randomly determined. In this model, because right-handedness is a consequence of left hemisphere dominance for language, from a phenotypic perspective, they are the same. Thus, this simple genetic model accounts for both the prevalence of human right-handedness and left hemisphere dominance or language.

Further support for theories proposing genetic mechanisms for hemispheric specialization in humans comes from the evidence that behavioral and neuroanatomical asymmetries are heritable.^34,91–97 For instance, family studies and data from twins clearly demonstrate a significant heritability in both direction (which hand you prefer) and strength (how far you deviate from exhibiting no hand preference) of handedness. The majority of studies on heritability in brain morphology have focused on twins, and there is strong evidence that lobar and regional variation in surface area, volume, and cortical thickness are more heritable in one hemisphere than the other.^91,98–101 The difference in heritability between homologous regions of the hemispheres is interpreted to reflect inherent variation in coding for left–right asymmetry,¹⁰² but direct tests of heritability of asymmetry are essentially absent in studies with humans. That is to say, heritability estimates of laterality or asymmetry quotients are typically not performed or provided in the published works.

Beyond the heritability results, some have proposed specific genetic differences between humans and other primates that explicitly account for species variation in the expression of behavioral and brain asymmetries.^28,89 For instance, Williams et al. have reported that the Protocadherin 11X/Y gene differs between humans and chimpanzees, and these authors postulate that this evolutionary genetic change accounts for the emergence of hemipsheric specialization for language.

Evidence of heritability in behavioral and brain asymmetries in nonhuman animals including primates are few in number.^103,104 Early attempts to test for a genetic basis for laterality in animals largely focused on comparing strain differences in paw preference or selective breeding studies in mice.^105,106 Despite more than 30 generations of selective breeding for directional paw preference, no significant evidence was found. In other words, two right-pawed mice were equally likely to produce a left- or right-pawed offspring.¹⁰⁷ Interestingly, strength of paw preference could be selectively bred in mice. Thus, selectively breeding mice that showed weak or strong paw preferences resulted in offspring that exhibited either weak or strong paw preference.¹⁰⁷ Additional studies showed that paw preference in rodents largely conformed to asymmetries in the environment and could even be socially learned.^108,109 In contrast to mice, scientists have more recently shown that eye preferences for viewing predators can be selectively bred in fish.¹¹⁰ Moreover, selective breeding for eye dominance resulted in a suite of other behavioral asymmetries that conformed to the same directional breeding biases.¹¹¹

Several studies in chimpanzees and other primate species have shown that hand preferences run in families and can be modified by certain life history variables.^112,113 More recently, Hopkins et al.¹¹⁴ adopted a more quantitative approach to the assessment of potential genetic determinants of handedness in chimpanzees. Specifically, Hopkins et al.¹¹⁴ took advantage of the well-documented pedigrees of three captive chimpanzee populations and used quantitative genetics to estimate heritability for three measures of handedness including simple reaching, the bimanual TUBE task, and manual gestures (Table 2). Significant heritability was found for each measure for both direction and strength of handedness. For comparison, we have presented findings from two studies with humans that quantified heritability in handedness in an extended pedigree instead of twins.^95,97 As can be seen, the estimates for overall handedness are generally higher in humans compared to chimpanzees; however, heritability for specific tasks in humans, as reported by Warren et al., is generally lower than those reported for the three measures used with chimpanzees. Additionally, in both humans and chimpanzees, strength of handedness was more heritable than direction, not unlike the findings from selective breeding in mice.

Table 2.

Heritability estimates for handedness in humans and chimpanzees

Human (1)	Direction	Strength
Overall *	0.52	0.67
Human (2)
Overall *	0.57	NR
Write*	0.16	NR
Draw*	0.16	NR
Throw*	0.17	NR
Scissors*	0.11	NR
Toothbrush	0.07	NR
Knife	0.06	NR
Spoon*	0.16	NR
Match *	0.12	NR
Open box	0.07	NR
Chimpanzee
Reaching*	0.36	0.67
Manual gesture*	0.49	0.46
TUBE*	0.24	0.44

^*significant heritabilities.

NR, not reported.

In primates, several quantitative genetic studies have demonstrated that different dimensions of cortical organization are significantly heritable, such as brain size, corpus callosum surface area, and gyrification.^{103,104,115,116} In an early study using brain endocasts, Cheverud et al.¹⁰³ reported that petalia asymmetries in the frontal lobe (h² = 0.33) were significantly heritability in rhesus macaques, but the occipital petalia asymmetry (h² = 0.16) was not. In the sole published study on heritability in brain asymmetries in specific cortical areas in primates, Fears et al.¹¹⁷ measured five dimensions of brain asymmetry, including hemisphere volume, left and right prefrontal and occipital width, cerebral torque in the transverse and coronal planes, and the terminal posterior position of the cingulate sulcus. Fears et al.¹¹⁷ found significant rightward asymmetries in occipital width and a rightward shift in the posterior position of the cingulate sulcus. They also found a rightward occipital/leftward prefrontal torque asymmetry in the transverse plane. Fears et al. also reported that heritability in the asymmetry scores were significant for hemisphere volume (h² = 0.18), transverse torque (h² = 0.29), and particularly the terminal posterior position of the cingulate sulcus (h² = 0.32).

Our laboratory has similarly used quantitative genetics to estimate heritability in gray matter volume and asymmetry in a sample of 223 chimpanzees for two brain regions, including the posterior superior temporal gyrus (p_STG) (which includes a portion of Wernicke’s area) and the inferior frontal gyrus (IFG). The p_STG and IFG were of interest because they are the chimpanzee homologues to Wernicke’s and Broca’s area in the human brain, and, given the theoretical interest in these regions, they seemed appropriate. The methods and landmarks used to define the p_STG and IFG have been described in detail elsewhere.¹¹⁸ For the IFG and p_STG gray matter volumes, asymmetry quotients (p_STG_AQ, IFG_AQ) were derived following the formula AQ = (R − L)/((R + L) × 0.5) where R and L represent the right and left hemisphere volumes. Positive AQ values represented right hemisphere asymmetries and negative values reflected left hemisphere biases. We also classified subjects as left-, right- or non-lateralized (no bias) based on the sign and strength of their AQ values. Following cut points used in the human literature,¹¹⁹ we classified chimpanzees with AQ values ≥ 0.025 or AQ ≤ −0.025 as right- and left-lateralized, respectively. Subjects with AQ scores > −0.025 and < 0.025 were classified as having no bias. The mean AQ values and distribution of asymmetries for the p_STG and IFG gray matter volumes are shown in Table 3. In terms of the AQ scores, significant leftward asymmetries were found for the p_STG but not the IFG gray matter volume. There was no significant difference in the distribution of left–right asymmetries for the IFG gray matter.

Table 3.

Descriptive data on asymmetries in p_STG and IFG GM volume

Measure	AQ	s.e.	t	p	#L	#A	#R
p_STG	−0.082	0.014	−5.82	0.001	144	16	60
IFG_GM	+0.021	0.026	+0.80	0.423	101	14	97

Note that the n is lower for the IFG compared to the p_STG. This is because we could not reliably define this region in 8 individuals. For the distribution of p_STG asymmetries, there were significantly more left-lateralized subjects compared to right-lateralized subjects (X² (1, n = 204) = 34.59, P < 0.001) and subjects with no bias (X² (1, n = 160) = 102.40, P < 0.001) for the p_STG measure.

Consistent with the approach used by Fears et al.,¹¹⁷ to estimate heritability in p_STG and IFG gray matter volume, we used the software package SOLAR.¹²⁰ SOLAR uses a variance-components approach to estimate the polygenic component of variance when considering the entire pedigree (see Refs 104, 177, and 121. Because the AQ values were skewed, the values were transformed using the inormal function within SOLAR. Separate heritability analyses were performed on the raw left and right hemisphere values as well as the AQ scores (Table 4). The quantitative genetic analysis indicated that a significant proportion of variance was attributable to genetic factors for the residual L_p_STG_GM, R_pSTG_GM, L_IFG_GM, and R_IFG_GM volumes (Table 3). For both regions, the heritability estimates were higher for the left compared to the right hemisphere. For the AQ values, significant heritability was found for the p_STG but not the IFG (Table 4). Additionally, Table 4 shows the results of the heritability estimates for the left and right hemisphere p_STG and IFG volumes after adjusting for overall total gray matter volume within each hemisphere. As with the raw gray matter volume data, adjusted gray matter volumes for the left and right p_STG and IFG were significant. Further, the asymmetry for the p_STG but not the IFG_GM was significantly heritable.

Table 4.

Heritability estimates (h²) of p_STG and IFG gray matter volume and asymmetry values

	Left	Right	AQ
p_STG	0.426*** (0.116)	0.322** (0.127)	0.227* (0.126)
IFG_GM	0.477** (0.146)	0.258** (0.118)	0.014 (0.111)
Adjusted_pSTG	0.273*** (0.108)	0.181* (0.119)	0.231* (0.126)
Adjusted IFG_GM	0.457*** (0.159)	0.223* (0.121)	0.000

^***P< 0.001,

^**P < 0.01,

^*P < 0.05. Values in parentheses represent standard errors

Asymmetry as a consequence of increasing brain size in primates

The antithesis of the genetic, saltational views is the argument that behavioral and anatomical asymmetries emerged, essentially, as by-products of evolutionary selection for increasing brain size. Thus, there was no direct selection for hemispheric specialization but, instead, they evolved as a consequence of increasing brain size in relation to physiological and computational constraints on neural transmission time. The primary evidence in support of this hypothesis comes from studies of the corpus callosum (CC) in relation to brain size, both within and between species.

The corpus callosum (CC) is a large collection of white matter fibers that connects homo- and heterotopic regions of the left and right cerebral hemispheres.^122,123 In humans, it has been hypothesized that individual and gender differences in functional and anatomical asymmetries are attributable to variation in the size of the corpus callosum, after adjustment for brain volume.^124–126 Most recent studies in humans have focused on measuring the midsagittal area of the CC from in vivo magnetic resonance imaging (MRI), and the claim is that, after adjustment for brain size, the CC is larger in females than in males, though not all studies have revealed significant gender differences.^127–130 The larger CC in females is interpreted as indicating greater interhemispheric connections between the left and right hemispheres compared to males. This, in turn, has been proposed to explain reports of gender differences in motor and cognitive functions and gender-based differences in the severity as well as the recovery of functions from neurological insults, notably stroke.^131–133 Specifically, females are reported to have better fine motor skill than males and have also been reported to be more right-handed.¹³⁴ It is also often reported that females show less severe neurological problems resulting from stroke, such as aphasic or apraxia, and that they show better recovery of function over the poststroke time period.¹³⁴ Presumably, the increased interhemispheric connectivity in females compared to males results in faster communication between homotopic cortical regions connecting the left and right hemispheres.^{122,135–137}

The relative size of the CC, after adjustment for brain size, has also been posited as an explanation for the evolution of behavioral and brain asymmetries in nonhuman animals, specifically primates.^138,139 It has been shown that, among mammals¹⁴⁰, and within primates,¹³⁹ increases in brain size are associated with smaller CC surface areas (CC:BV). That is to say, species with larger brains have smaller corpus callosi than would be predicted for an organism of their brain size. In primates, humans have the smallest CC:BV ratio, followed by great apes, then Old World and New World monkeys. The assumption of this evolutionary hypothesis is that the speed of neural impulses traversing the CC is constant across different primate species;¹³⁷ however, as the brain became increasingly large, the homo- and heterotopic regions connected via transcallosal fibers grew further apart, thereby resulting in an increase in the time needed to transfer information between two regions. Thus, each hemisphere became increasingly isolated or separated from the other as brain size increased, resulting in the emergence of specialized functions within each hemisphere. Evidence in support of this theory comes from comparative data on axon fiber density and diameter in primates. In macaque monkeys, chimpanzees, and humans, the relative distribution of axon fiber density and the distribution of large-diameter, highly myelinated axons is consistent across different regions of the CC.^141–146 Faster, more myelinated axon fibers are primarily found in the central portion of the CC, the region that connects the primary motor and somatosensory cortices, while smaller, less myelinated fibers are found in CC regions, such as the genu, that connect to the association cortex.

An important aspect of this theory that has not been tested is the assessment of differences in behavioral or brain asymmetries in different primate species. Based on the theory that smaller CC:BV ratios underlie differences in hemispheric specialization in primates, it follows that, all things being equal, humans would show the largest asymmetries, in absolute terms, compared to great apes and monkeys. We have recently performed a preliminary study testing this hypothesis by comparing absolute asymmetries in the surface area of three common sulci in the left and right hemispheres in a variety of primate species. For this study, we used BrainVisa (BV), a software program that extracts cortical sulci and provides measures of the surface area, depth and thickness for cortical folds (Fig. 2). For this analysis, we calculated the left and right surface areas for the central sulcus (CS), sylvian fissure (SF), and superior temporal sulcus (STS)in 153 in vivo and postmortem MRI scans from three main primate taxa including humans (Homo, n = 12), great apes (Pan, n = 35, Gorilla, n = 11, Pongo, n = 10) and Old World monkeys (Papio, n = 12, Macaque, n = 14). For each subject, we computed an absolute asymmetry quotient (ABS_AQ) by taking the absolute difference between the right and left hemisphere surface area values divided by the average surface area. We also calculated the ratio in the size of the total surface area of the corpus callosum divided by the total gray and white matter volume (CC:BV), excluding the cerebellum and brain stem regions. Because the CC:BV and ABS-AQ values are scaled differently, we converted the values to Z-scores, and the results are shown in Figure 3. The results are consistent with the predictions of the CC:BV theory. That is to say, humans had the smallest CC:BV ratio and highest absolute degree of asymmetry while Old World monkeys had the largest CC:BV ratios and the smallest absolute asymmetries. Great apes were in the middle in terms of their CC:BV ratio and ABS_AQ scores.

Figure 2.

Example 3D brain cortex reconstructions and extracted central sulci from representative primate species in the study along a phylogenetic timeline.

Figure 3.

Mean standardized Z-scores (± s.e.) for absolute asymmetry (ABS-AQ) and CC:BV ratio in humans, great apes, and Old World monkeys.

Summary and implications

It is increasingly clear that some nonhuman primates exhibited population-level behavioral and brain asymmetries. In our view, though fewer species and subjects have been tested for cognitive and emotional processes, the most compelling data have come from these areas of research. Indeed, evidence of left–right differences in cognitive and emotional processes may extend to many vertebrates and have a much longer evolutionary history.¹⁴⁷ Far more data are available on handedness in different primate species and, perhaps not surprisingly, the results are more variable even between closely related species. This level of noise in the data has been interpreted as problematic by some,^53,54 but we see them as illustrating the point that population-level handedness could have evolved in some species but not others, depending on a variety of ecological, social or evolutionary factors. Indeed, it is exactly this variability that is necessary in order to examine what factors influence the expression of population-level asymmetries in some species but not others. Importantly, scientists studying nonhuman primate handedness tend to view human handedness as a cumulative outcome of progressive evolutionary factors, but it is important to recognize that Homo sapiens remain the sole surviving species within our genus, and we have very limited ways of comparing handedness in closely related species, beyond exploration of the archeological record. Thus, in our view, the variability in handedness results among different primate species is potentially very important for identifying what evolutionary factors may have selected for the pronounced right-handedness found in modern humans.

Species-level neuroanatomical asymmetries are also evident in some nonhuman primates, particularly great apes. In the realm of quantifying and comparing neuroranatomical asymmetries, one challenge is identifying those approaches that will produce results that lend themselves to fair and rigorous comparative analysis. Because of the variation in the sulci that comprise different primate species brains, this limits most region-of-interest approaches as well as those that measure different dimensions of common sulci such as the BV approach described above (i.e., only those few sulci that are found in all species lend themselves to these analyses). Indeed, measuring asymmetries in highly conserved sulci may constrain the ability to detect those that might have recently evolved and potentially be associated with increasing hemispheric specialization. Nonetheless, the advent of modern anatomical and functional imaging has and will continue to advance this area of research. In particular, the use of either voxel-based or surface-based comparisons of neuroanatomical asymmetries in different primates would provide insightful data on both the direction and strength of asymmetries in the brain. Similarly, there is a significant paucity of data on brain asymmetries at the microstructural level of analysis such as cytoarchitectonics, minicolumn organization, and dendritic arborization, even in the human brain.^148–155 Additional data at this level of analysis would be very useful for understanding the neurobiology of hemispheric specialization at both an individual and species level of analysis. Finally, though not discussed in detail here, a new and potentially important method for assessing asymmetries in different primate species is diffusion tensor imaging (DTI). With DTI, in vivo measures of cortical connectivity through the use of tractography can be made in different primate brains. To date, important similarities and differences in connectivity and asymmetry have been reported in primates,^156–159 and this approach most certainly warrants further study.

The comparative data on variation in the surface area of the CC relative to total brain volume are consistent between studies and different phylogenetic groups (i.e., all mammals or just within primates). In short, it seems well established that species with larger brains have relatively small corpus callosi.^139,140,160 Within primates, these data suggest that humans have increasingly “split” or “disconnected” hemispheres, followed by great apes, then Old World monkeys. Further, the suggestion that increasing separation of the two hemispheres resulted in increased neuroanatomical asymmetries in primates was also supported, at least when limited to measurement of the surface area of the CS, SF, and STS. There are certainly limitations to these initial findings, and it may be the case that, for other sulci¹⁶¹ or possibly other dimensions of cortical organization (i.e., depth of sulci, cortical thickness, white matter connectivity), different patterns of results may be found. Further, we focused on anatomical asymmetries, and a case can be made that the level of analysis should center on functional asymmetries. Lastly, an important but neglected area of research is the possible trade-off between relative CC size and the speed and architecture of the fibers traversing the CC.^141,145,146 It is well known that there are differences in the proportion of large- and small-diameter fibers (as reflected in axon diameter) throughout the CC in human and nonhuman animals but very little data on how these might differ between species. It might be the case that humans have a relatively small CC, but this was accompanied by an increase or decrease in the relative proportion of small- or large-diameter fibers compared to other species.

It is also important to recognize that the CC:BV theory on the evolution of hemispheric specialization is agnostic with respect to predictions regarding directional asymmetry. Recall that our comparative analyses focused on absolute left–right differences in surface area of the CS, SF, and STS. This was intentional, because the CC:BV theory simply posits an inverse association between the relative size of the CC and the degree but not the direction of a given asymmetry. In other words, increasing asymmetries can be either leftward or rightward. The limitation of this theory is that directional asymmetries are present and can vary between species.¹⁶² Indeed, we have previously found that chimpanzees show a small but significant leftward asymmetry in gyrification.¹⁶³ Likewise, we have previously found that chimpanzees show a leftward asymmetry in the gray matter volume of the p_STG, but neither macaques nor vervet monkeys show a group-level bias.¹⁶⁴ Thus, for chimpanzees, absolute asymmetries in sulci surface area or the p_STG essentially reflect the strength of leftward asymmetries (i.e., directional asymmetry), while in macaques or vervet monkeys, absolute asymmetries reflect their negative or positive deviations from zero (i.e., sometimes referred to as antisymmetry). These differences in the distribution of asymmetries create statistical problems as well as challenges in the interpretation of results because “absolute” asymmetries reflect two different dimension of laterality.

Finally, we presented evidence that handedness and, to a lesser extent, neuroanatomical asymmetries are heritable in chimpanzees. Though significant, the heritability estimates are relatively modest, as is the case for the vervet monkeys studied by Fears et al. Research on the role of genetic factors in the development and evolution of behavioral and brain asymmetries in human and nonhuman primates are in their infancy. Recent quantitative genetic analyses combined with voxel-based or surface-based morphometry analyses have proven useful in characterizing heritability in different regions of the brain in human and nonhuman primates. However, one challenge to this area of research is the primary reliance on data from twins in humans, whereas those studies in chimpanzees, baboons, and vervet monkeys primarily utilize extended pedigree information. We mention this issue because, ultimately, it would be ideal to quantify the extent to which genetic factors contribute to individual cortical organization, including asymmetry, in different species. Data and comparisons from twins, though experimentally powerful, likely inflate the estimation of heritability compared to what would be manifest in a typical singleton–birthing species,¹⁶⁵ which most primate species are.

As noted above, early theories posited that human behavioral and brain asymmetries are mediated by single- or double-allele genetic models while nongenetic mechanisms explain nonhuman laterality. This view is clearly challenged by contemporary data in apes and monkeys and, indeed, it could be argued that the role of genes in the determination of asymmetry are stronger in nonhuman compared to human primates. For instance, Gomez-Robles et al.¹⁶⁶ recently examined variability in neuroanatomical morphology in a sample of human and chimpanzee brains and found, with respect to asymmetry, fairly conserved directional asymmetries. For instance, leftward directional asymmetries in the posterior temporal lobe regions were the most robust and consistent between humans and chimpanzees. However, humans showed much greater fluctuating asymmetry, which was interpreted as evidence that the human brain shows greater plasticity with respect to left–right asymmetry (Fig. 4). Though genetic anomalies can induce large fluctuating asymmetry, it is generally agreed that they reflect the extent to which left–right asymmetries are determined by nongenetic perturbations. Thus, the suggestion is that left–right differences in anatomy and function are generally under less specific genetic determination in humans than in other primates. If this is the case, it might be predicted that heritability estimates in cortical organization, and particularly left–right differences, would be significantly lower in humans compared to more distantly related primates. This is actually a very testable hypothesis because heritability can be estimated separately for the left and right hemispheres and the differences in h² coefficients should be smaller in humans than in other primates. Moreover, it would be desirable to determine whether directional or absolute differences in heritability estimates between the two hemispheres correlate with directional or absolute differences in phenotypic asymmetries (i.e., behavior, neuroanatomy, or functional asymmetries). In primates, despite claims for the existence of candidate genes for handedness,³⁷ the evidence of a direct relationship between genetic coding for a directional asymmetry explicitly linked to a specifically lateralized behavioral function does not exist.

Figure 4.

Localization of levels of variation for chimpanzees and humans. Symmetric variation (top) is mainly localized in the posterior perisylvian area in humans, whereas it is more diffusely localized in chimpanzees. The strongest directional asymmetries (middle) are localized in both species at the posterior termination of the Sylvian fissure and of the superior temporal sulcus. Fluctuating asymmetric variation is also localized in the post-Sylvian area in humans, but it is strongest in the inferior frontal region in chimpanzees. Adapted from Ref. 166.

In sum, historical claims of the absence of population-level asymmetries in nonhuman animals are no longer tenable. Additionally, though still new, the evidence increasingly suggests that multiple genes play roles in the determination of behavioral and brain asymmetries, not just in humans but in other species as well. The notion that single genes or alleles account for the expression of asymmetries (i.e., the Protocadherin 11X/Y gene) also seems unlikely, and the lack of candidate genes for asymmetry reinforces this view.^167,168 Instead, the expression of asymmetries is likely influenced by multiple genes that interact with epigenetic, pre-, and postnatal factors to determine individual and phylogenetic variation.¹⁶⁷