Introduction

Very recently, new evidence has emerged thatchimpanzees possess asymmetries in posterior tem-poral lobe areas homologous to language areas in the human brain.1 This report has rekindled interestin the ongoing debate about whether the human brain has evolved specific areas of specialization forlanguage processing or whether there are anatomi-cally specialized homologous language areas in non-human primate brains.2–4

Hemispheric specialization within language-related areas has traditionally been viewed as a unique feature of human brain evolution.5 Clinicaland experimental evidence suggests that the left hemi-sphere is dominant in the perception and productionof linguistic stimuli.6 Furthermore, specific sites have been identified as critical to the perception andproduction of language. Lesions to the left hemi-sphere in either Broca’s area or Wernicke’s area resultin different forms of aphasia, whereas lesions in thecorresponding areas of the right hemisphere do notproduce the same deficits. More recently, studies withnormal subjects using neural imaging techniques suchas functional magnetic resonance imaging (fMRI)have identified these same asymmetric areas as moreactive when subjects are engaged in either naming or

language comprehension tasks.7 Finally, structuralasymmetries in the pars triangularis (part of Broca’sarea) and planum temporale (part of Wernicke’s area)have been documented in human subjects. In bothcases, the left hemisphere structures are larger thanthe corresponding regions in the right hemisphere,particularly in right-handed individuals.8,9

The substantial evidence for language-relatedhemispheric specialization in humans has been thebasis for longstanding claims that humans possess aunique neuroanatomical adaption for linguistic func-tions and that, concordantly, non-human primateslack homologous Broca’s and Wernicke’s areas.10

Nevertheless, similar neuroanatomical asymmetrieshave been reported in human and non-humanprimates over the course of several decades. Bothhumans and chimpanzees have a longer sylvian fissure(SF) on the left than on the right, whereas monkeysdo not.11,12 Differences in SF length presumablyreflect differences in the volume of the planumtemporale (PT), a key site within the posteriorportion of the temporal lobe, but no direct com-parison between measures has been performed. Therecent report by Gannon et al .1 demonstrated a larger left than right PT in 17 of 18 chimpanzeecadaver brains. These results suggest that chim-panzees possess a Wernicke-like asymmetry that

Brain Imaging

1111234567891011112345678920111123456789301111234567894011112345678950111123456111p

0959-4965 © 1998 Lippincott Williams &Wilkins Vol 9No 1224 August 1998 2913

THE planum temporale (PT), a portion of Wernicke’sarea, is important for linguistic functions in humans andis larger in the left compared to the right hemisphere.In this study, we assessed the presence and size of thePT in a sample of non-human primates including 21great apes, four lesser apes, 11 Old World monkeys andeight New World monkeys using magnetic resonanceimaging. The PT was measured in both the sagittal andcoronal planes by use of multiplanar reformatting soft-ware. The PT could only be identified in the sample ofgreat apes and not in the remaining non-human primatespecies. Within the great ape sample, the PT was largerin the left hemisphere than in the right in a statisticalmajority of the subjects. These results are consistent withthe notion that the PT evolved as a definable structureabout 15 million years ago and may have arisen as aresult for selection for greater cortical folding which inturn led to greater gyrification in larger brains.NeuroReport 9: 2913–2918 © 1998 Lippncott Williams &Wilkins.

Key words: Chimpanzee; Language; Neuroanatomicalasymmetry; Planum temporale

Planum temporaleasymmetries in greatapes as revealed bymagnetic resonanceimaging (MRI)

William D. Hopkins,1,4,CA Lori Marino,2

James K. Rilling3 and Leslie A. MacGregor1

1Department of Psychology, Berry College;Departments of 2Biology and 3Anthropology,and 4Living Links Center, Division ofPsychobiology, Yerkes Regional PrimateResearch Center, Emory University, Atlanta, GA 30322, USA

CA,4Corresponding Author and Address

Website publication 16 August 1998 NeuroReport 9, 2913–2918 (1998)

may represent a homologous language-related area.Nevertheless, as Gannon et al.1 pointed out, it iscrucial to determine whether the extant hominoidsand monkeys show similar patterns of PT asymmetryin order to address questions about the evolution ofPT asymmetries and hemispheric specialization ingeneral.

In view of the important implications of thefindings of Gannon et al.1 that non-human pri-mates possess an anatomically homologous humanlanguage-related brain area, the present studyprovides not only a crucial replication of this contro-versial finding but the important awaited extensionof this test to the other great ape species, lesser apes,and monkeys. Furthermore, our study also providesa test of their findings using MRI, a technique thatutilizes brains from living subjects. Whereas cadaverbrains are often subject to deformation, similar find-ings in living tissue would provide an importantconfirmation of the robustness of the findings ofGannon et al.1

We hypothesized that the PT would be identifi-able and larger in the left hemisphere in the greatapes on the basis of previous findings showinghuman-like functional asymmetries13,14 and behav-ioral evidence of at least some language-related capac-ities in this family.15–18 Moreover, we hypothesizedthat the PT would not be identifiable in lesser apesand monkeys primarily on the basis of findingsshowing that these primates do not exhibit the samedegree of language affinity or functional specializa-tion as has been reported in great apes.

Materials and Methods

Subjects: MRIs were collected on a sample of 21adult or sub-adult great apes (11 male, 10 female)including 12 chimpanzees (Pan troglodytes), fourorangutans (Pongo pygmaeus), two gorillas (Gorillagorilla gorilla) and three bonobos (Pan paniscus).Scans were also collected in a sample of four adultlesser apes (two male, two females; Hylobates lar),11 adult Old World monkeys (eight male, threefemale) including seven rhesus monkeys (Macacamulatta), two baboons (Papio papio) and two sootymangabeys (Cercocebus torquatus), and eight adultNew World monkeys (five male, three female)including four capuchin (Cebus apella) and foursquirrel (Samiri sciureus) monkeys. After collectingthe MRI scans, it became clear that HG could notbe identified in either the Old or New World monkeyspecimens primarily because the transverse temporalsulcus, which forms the anterior border of HG, couldnot be seen on the scans (see Fig. 1). Therefore, thelack of an identifiable HG precluded measurement

of the PT in these species. There was some evidencefor the appearance of a less well-developed HG inthe lesser apes but it was not apparent in everyspecimen or visible in both hemispheres (see Fig. 1).Thus, we report only the quantitative data from thegreat apes for which the PT could be measured.

Procedure: The subjects were first immobilized byketamine injection (10 mg/kg) and subsequentlyanesthetized with propofol (2–6 mg/kg/h) followingstandard procedures at Yerkes Regional PrimateResearch Center (YRPRC). Subjects were trans-ported by van to the MRI facility at EmoryUniversity Hospital. The subjects remained anes-thetized for the duration of the scans as well as thetime needed to travel between YRPRC and Emoryhospital (total time ~2 h). After completing the MRI, the subjects were returned to YRPRC andtemporarily housed in a single cage for 6–12 h toallow the effects of the anesthesia to wear off, afterwhich they were returned to their home cage andcagemates. At the MRI facility, the subjects wereplaced in the scanner chamber and their heads were fitted inside the head coil. This project involvedusing two MRI machines (Phillips, Model 51), eachwith 1.5-Tesla super conducting magnets. For allsubjects, T1-weighted images were collected in thetransverse plane using a gradient echo protocol (pulserepetition = 19.0 ms, echo time = 8.5 ms, slice thick-ness = 1.2 mm, slice overlap = 0.6 mm, number ofsignals averaged = 8 and a 256 ´ 256 matrix). Thesescan parameters were used based on preliminarystudies and provided excellent resolution of the brainareas of interest to this study. The archived data werestored on optical diskettes and transported to a SunSparc work station for post-image processing.

Area measurements of planum temporale (PT): Inorder to test the robustness of our findings, weadopted two different procedures for measuring thePT (described below). Identification and definitionof the regions of interest were based on publishedatlases of the human brain derived from bothcadavers19 and MRI scans20 and from publishedpartial atlases of chimpanzee and monkey cadaverbrains.21

For measurement of the PT in the sagittal plane,the MRI scans were re-formatted (or re-sliced) usingmultiplanar formatting software (Easy Vision) in thesagittal plane with 1 mm slice thickness. In succes-sive slices, two raters determined the presence orabsence of HG in the left and right cerebral hemi-spheres. If HG was present, then the area of the PTwas measured following the procedure and utilizingthe anatomical landmarks outlined by Foundas et al.8,9 The PT is the flat plane that forms the surface

W. D. Hopkins et al.

1111234567891011112345678920111123456789301111234567894011112345678950111123456111p

2914 Vol 9 No 12 24 August 1998

of the superior temporale gyrus caudal to HG. Theanterior border was defined as the transverse sulcuscaudal to HG and the posterior border was definedas the caudal extent of the posterior horizontal ramus(PHR) of the sylvian fissure as it bifurcates into the posterior ascending ramus (PAR) and the poste-rior descending ramus (PDR) (see Fig. 2). When theposterior border was ambiguous owing to lack of thePDR, the ‘knife cut’ method was used. This methodis analogous to that described by Geschwind andLevitsky22 and has been used recently by other inves-tigators when measuring the PT on MRI scans. Usingthis method, a ‘cut’ was made in the plane of thesylvian fissure at the most caudal extent of the PHR.The posterior limit of the PT excluded tissue alongthe PAR. If multiple HG were present, the anteriorborder of the PT was defined as the transverse sulcusof Heschl (first transverse sulcus). This designationof the anterior border of the PT has been used byothers in postmortem studies and in recent MRIstudies. The length of the PT was measured onsuccessive sagittal scans in which HG could beanatomically defined. Planar length was measured tothe closest 0.1 nm using a mouse-driven computer-guided cursor. An estimate of the PT area was derivedby summing the cumulative planar lengths measures

for each slice within a hemisphere. The cumulativeplanar length was then added to the number of slicesin which HG could be identified to derive the esti-mated PT area for each hemisphere (SAG-AREA).

To measurement the PT in the coronal plane, theMRI scans were aligned in the coronal, sagittal andhorizontal planes and cut into 1 mm coronal slicesusing multiplanar reformatting software. On succes-sive 1 mm slices, two raters determined the absenceor presence of HG in the left and right cerebral hemi-spheres. If HG was present, then the area of the PTwas measured following the procedures used byLarsen et al.23 with human subjects in which thedepth of the sylvian fissure was measured from thesurface of the brain to the lateral portion of HG (or the transverse temporal sulcus, see Fig. 2). Aswith the sagittal data, an estimate of the PT area wasderived by summing the cumulative width measuresfor each slice within a hemisphere. This value wasadded to the number of slices in which HG couldbe identified within a hemisphere to derive an esti-mate of the PT area for each hemisphere (COR-AREA).

Data analysis: For each plane, a measurement of PTarea was derived for each hemisphere (i.e., COR-

Neuroanatomical asymmetries in non-human primates

1111234567891011112345678920111123456789301111234567894011112345678950111123456111p

Vol 9 No 12 24 August 1998 2915

FIG. 1. Coronal views of 5 representative non-human primate species used in this study. Although human subjects were not included inthe study, we present one of their brains for the purposes of comparison to the non-human primates. The coronal views were taken fromidentical regions of the posterior portion of the temporal lobe in each species. H+ indicates the presence of Heschl’s gyrus while H– indi-cates the absence of Heschl’s gyrus. Humans = Homo sapiens, Chimpanzee = Pan troglodytes, Gibbon = Hylobates lar, Rhesus monkey =Macaca mulatta, Capuchin monkey = Cebus apella.

AREA and SAG-AREA). From these data, an asym-metry coefficient (AQ) was derived for each planeby subtracting the right hemisphere value from theleft hemisphere, dividing by the sum of the left andright hemisphere values and multiplying by 0.5[(L–R)/(L + R)(0.5)]. The sign of the resulting value(positive or negative) indicated the direction of asym-metry and the absolute value reflected the magnitudeof asymmetry. As has been done in PT studies withhuman subjects, subject’s with AQ scores > 0.025were classified as having a left hemisphere bias (L)while subject’s with values < –0.025 were classifiedas having a right hemisphere bias (R). Subject’s withAQ scores between –0.024 and 0.024 were classifiedas having no bias (N).

Results

Descriptive and quantitative data: Table 1 showsthe mean AQ scores for each great ape and plane ofmeasurement. We were able to reliably identify theHG, and thus PT, in all the great ape specimens inboth the sagittal and coronal planes with the excep-tion of one animal in the coronal plane. Within the

sagittal plane, 16 subjects exhibited a left hemisphereasymmetry, two possessed a right, and three exhib-ited no bias. The number of individuals with lefthemisphere asymmetries was significantly greaterthan the number with either a right hemisphere (z = 3.31, p < 0.01) or no hemispheric bias (z = 2.98,p < 0.01). For the coronal plane, 11 individuals exhib-ited a left hemisphere asymmetry, two possessed a right and seven exhibited no bias. Although thenumber of individuals with a left hemisphereasymmetry was significantly greater than the numberof individuals with a right hemisphere asymmetry (z = 2.50, p < 0.01), they did not significantly differfrom the number of individuals with no bias. Withinthe sagittal plane, the PT was larger in the lefthemisphere in 10 of 12 chimpanzees, two of threebonobos, both gorillas, and two of four orang-utans. In the coronal plane, the left hemisphere waslarger in eight of 11 chimpanzees, one of threebonobos, one of two gorillas, and one of four orangutans. In addition to these qualitative data,directional asymmetries were assessed in each planeusing one-sample t-tests on the AQ scores.Significant directional asymmetries were found forboth the sagittal (t(20) = 3.03, p < 0.01) and coronal

W. D. Hopkins et al.

1111234567891011112345678920111123456789301111234567894011112345678950111123456111p

2916 Vol 9 No 12 24 August 1998

FIG. 2. Outline of the relevant neuroanatomical structures and landmarks used in the measurement of the PT in the sagittal and coronalplanes. HG, Heschl’s gyrus; HS, Heschl’s sulcus; PAR, posterior ascending ramus.

(t(19) = 3.13, p < 0.01) planes, indicating that themean AQ scores deviated from an estimated mean ofzero. Despite the observation of greater lateralizationin the PT when measured in the sagittal contrastedwith coronal plane (see Table 1), a paired t-test failedto reveal significant differences in the AQ scoresbetween planes.

Relationship between PT and subject variables:Independent sample t-tests were performed to testfor sex differences in PT asymmetry within eachplane. None of these values reached statistical signif-icance. We also compared sexes for differences instrength of asymmetry by comparing males andfemales using the absolute values of the PT measures.No sex differences were found. An analysis of therelationship between the age of the subjects and thesagittal and coronal PT measures was performed. Age was negatively correlated with the COR-AREAmeasure (r = –56, p < 0.05) with older subjects havinglarger leftward asymmetries than younger subjects.Finally, neither asymmetries in hemisphere lengthnor overall brain volume correlated with the PTmeasures in either plane.

Discussion

The present study provides the first confirmation thatthe PT is anatomically identifiable in all four great

apes and, as in humans,8,9 that it is larger in the leftthan in the right hemisphere. We also provide thefirst evidence that this morphological pattern ofasymmetry could not be quantified in lesser apes or Old and New World monkeys. These findingsconfirm and extend those of Gannon et al.1 in chim-panzee cadaver specimens and Yeni-Komshian andBenson’s early report of a difference in length of thesylvian fissure in chimpanzees.11 The consistency ofresults across studies, despite the use of differentmethodologies and tissue samples, reflects the robust-ness of the observed morphological asymmetry in theperi-sylvian area of great ape brains. It should beemphasized that the lack of the PT morphologicalasymmetry in lesser apes, Old and New Worldmonkeys does not imply the potential lack of asym-metry in other aspects of neuroanatomy. Specifically,it is possible that architectonic analyses of the cortexin the posterior portion of the temporal lobe wouldreveal neuroanatomical asymmetries in other primatespecies and clearly these studies are warranted in lightof our findings.

Given that a definable PT and asymmetries in thatregion appear to be present in great apes and humans,the question of the evolutionary origin of thisanatomical feature is one which presents itself.Although evolutionary questions cannot be directlyanswered with analyses of modern species, a compar-ative analyses can provide useful data for evaluating

Neuroanatomical asymmetries in non-human primates

1111234567891011112345678920111123456789301111234567894011112345678950111123456111p

Vol 9 No 12 24 August 1998 2917

Table 1. Individual PT data for each plane of measurement.

Sagittal plane Coronal plane

Sex LH RH AQ Bias LH RH AQ Bias

ChimpanzeeAustin M 55.8 52.9 0.013 N 115.3 75.5 0.104 LCarmicheal M 21.6 19.2 0.029 L 54.7 40.8 0.072 LChuck M 26.3 12.0 0.187 L 32.4 28.9 0.028 LDonald M 31.2 21.4 0.093 L 30.2 28.1 0.018 NJimmy Carter M 87.5 52.8 0.124 L 98.8 72.9 0.075 LJeannie F 22.5 15.0 0.100 L 28.0 18.0 0.109 LKengee F 55.5 36.5 0.103 L 31.6 35.0 –0.026 RLana F 65.3 33.0 0.164 LLazarus M 32.1 26.6 0.046 L 73.3 63.6 0.035 LLulu F 33.7 30.1 0.028 L 49.0 32.3 0.102 LMary F 61.2 38.5 0.114 L 36.7 32.2 0.032 LMerv M 31.5 31.8 –0.002 N 36.5 33.9 0.018 N

BonoboBrian M 41.3 43.6 –0.013 N 52.8 53.6 –0.004 NJill F 62.9 23.9 0.224 L 44.4 41.0 0.020 NLorel F 62.5 51.8 0.047 L 35.2 18.0 0.161 L

GorillaKekla F 47.4 39.0 0.049 L 29.1 27.6 0.013 NKinyani F 34.2 23.0 0.097 L 66.3 42.1 0.111 L

OrangutanHati F 53.2 17.2 –0.256 R 79.5 56.8 –0.083 RMentubar M 35.1 39.0 –0.026 R 66.0 67.5 –0.006 NMolek M 43.1 26.2 0.122 L 58.7 56.9 0.008 NMynyak M 28.6 20.3 0.085 L 54.5 46.5 0.039 L

Sex: M = Male, F = female. Bias: L = left hemisphere, R = right hemisphere, N = no bias.

various evolutionary hypotheses. In this caveat, themost parsimonious assumption is that the PT aroseand became lateralized some time between the diver-gence of the lesser apes from the great ape–humanlineage and the emergence of the first great apes (i.e.,15 million years ago).

Although the present findings show purelyanatomical similarities between great apes and humanPT, one may speculate upon the functional aspectsof these similarities. In humans, the posterior portionof the temporal lobe including PT is important inlanguage and speech comprehension.6,7 In light ofrecent studies which have reported remarkable speechcomprehension skills in great apes,24 it is reasonableto postulate that the PT serves a similar function.Although the PT is not morphologically identifiablein monkeys, studies indicate that lesions to the leftbut not the right posterior temporal lobe in Japanesemacaques monkeys lead to a deficit in the discrimi-nation of species-specific vocalizations.25 This area inmonkeys roughly corresponds to the PT in great apesand humans and leads to the question of why the PTcannot be anatomically defined if it has a functionalrole in communicative skills. One possibility may bethat the posterior portion of the left temporal lobein all primates plays an important role in the discrim-ination of auditory signals but only the apes andhumans show a clearly identifiable PT due to therelatively greater degree of gyrification of the brain.Thus, with selection for greater cortical folding in theneocortex in relation to brain size, the PT emergedas a definable structure (due to the emergence ofHG).

Alternatively, the PT could have evolved inresponse to the processing of specific types ofacoustic signals, most notably speech stimuli. Thisevolutionary scenario presents a problem in that great apes do not produce speech and therefore it isdifficult to explain why they would have evolvedspeech comprehension and not speech productionskills. Their capacity to comprehend human speechis therefore counterintuitive to their ability toproduce speech. Perhaps there is acoustic overlap in

speech and non-speech cues that selected for theemergence of the PT in primates, and that the produc-tion rather than comprehension of speech stimuli isuniquely human. If correct, this hypothesis leads tothe prediction that great apes would not possess afunctional or neuroanatomical homologous Broca’sarea, a hypothesis that can be addressed in livingsubjects with continued use of other neuroimagingtechniques.

References

1. Gannon PJ, Holloway RL, Broadfield DC et al Science 279, 220–222 (1997).2. Abotiz F and Garcia RV. Brain Res Rev 25, 381–396 (1997).3. Corballis MC. The Lopsided Brain: Evolution of the Generative Mind. New

York: Oxford University Press, 1992.4. Deacon T. The Symbolic Species. Cambridge, MA: MIT Press, 1997.5. Galaburda AM. Anatomic basis of cerebral dominance. In: Davidson RJ and

Hugdahl K (eds.) Brain Asymmetry. Cambridge, MA: MIT Press, 1995: 51–73.6. Hellige JB. Hemispheric Asymmetry: What’s Right and What’s Left.

Cambridge, MA: Harvard University Press, 1993.7. Just MA, Carpenter PA, Keller TA et al. Science 274, 114–116 (1996).8. Foundas AL, Leonard CM, Gilmore RL et al. Proc Natl Acad Sci 93, 719–722

(1996).9. Foundas AL, Leonard CM and Heilman KM. Arch Neurol 52, 501–508

(1995).10. Killackey HP. Evolution of the human brain: A neuroanatomical perspec-

tive. In: Gazzaniga MS (ed.) The Cognitive Neurosciences. Cambridge, MA:MIT Press, 1995: 1243–1254.

11. Yeni-Komshian G and Benson D. Science 192, 387–389 (1976).12. Heilbronner PL and Holloway RL. Am J Phys Anthropol 76, 39–48

(1988).13. Hopkins WD. (1996) Psychonom Bull Rev 3, 449–457 (1996).14. Hopkins WD, Morris RD and Savage-Rumbaugh ES. J Exp Psychol: Gen

120, 46–56 (1991).15. Gardner RA and Gardner BA. Science 165, 664–672 (1968).16. Rumbaugh DM. Language Learning by a Chimpanzee: The LANA Project.

New York: Academic Press, 1977.17. Savage-Rumbaugh ES. Ape Language: From Conditioned Response to

Symbol. New York: Columbia University Press, 1986.18. Savage-Rumbaugh ES, McDonald K, Sevcik RA et al. J Exp Psych: Gen 3,

211–235 (1986).19. Dublin AB and Dublin WB. Atlas of Neuroanatomy. St Louis, MI: Warren

H. Green, 1982.20. Sartor K. Imaging of the Skull and Brain. Berlin: Springer-Verlag, 1993.21. Bailey P, Bonin VG and McCulloch WS. The Isocortex of the Chimpanzee.

Urbana, IL: University of Illinois Press, 1950.22. Geschwind N and Levitsky W. Science 161, 186–187 (1968).23. Larsen J, Hoien T, Lundberg I et al. Brain Lang 39, 289–301 (1990).24. Savage-Rumbaugh ES, Murphy J, Sevcik RA et al. Monogr Soc Res Child

Dev 58, 1–221 (1993).25. Heffner HE and Heffner RS. Science 226, 75–76 (1986).

ACKNOWLEDGEMENTS: This investigation was supported in part by NIH grantRR-00165 and NS-29574. The supportive services provided by the VeterinaryDepartment and care staff of the Yerkes Center are greatly appreciated. We thank Dr Tom Insel for providing administrative and financial assistance inscanning the subjects.

Received 27 May 1998;

accepted 22 June 1998

W. D. Hopkins et al.

1111234567891011112345678920111123456789301111234567894011112345678950111123456111p

2918 Vol 9 No 12 24 August 1998