AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 114:224–241 (2001) Prefrontal Cortex in Humans and Apes: A Comparative Study of Area 10 Katerina Semendeferi,1* Este Armstrong,2 Axel Schleicher,3 Karl Zilles,3 and Gary W. Van Hoesen4 1Department of Anthropology, University of California, San Diego, La Jolla, California 92093 2Department of Cellular Pathology, Armed Forces Institute of Pathology, Washington, DC 20306 3C. and O. Vogt Institute for Brain Research, Heinrich-Heine University Du¨ sseldorf, Du¨ sseldorf D-40225, Germany 4Departments of Anatomy and Cell Biology and Neurology, University of Iowa, Iowa City, Iowa 52242 KEY WORDS frontal pole; brain evolution; cytoarchitecture; brain mapping; stereology; hominoid; hominid ABSTRACT Area 10 is one of the cortical areas of the slightly across species, including the relative width of frontal lobe involved in higher cognitive functions such as its cortical layers and the space available for connec- the undertaking of initiatives and the planning of future tions. The cortex forming the frontal pole of the gorilla actions. It is known to form the frontal pole of the macaque appears highly specialized, while area 10 in the gibbon and human brain, but its presence and organization in the occupies only the orbital sector of the frontal pole. Area great and lesser apes remain unclear. It is here docu- 10 in the human brain is larger relative to the rest of the mented that area 10 also forms the frontal pole of chim- brain than it is in the apes, and its supragranular layers panzee, bonobo, orangutan, and gibbon brains. Imaging have more space available for connections with other techniques and stereological tools are used to characterize higher-order association areas. This suggests that the this area across species and provide preliminary estimates neural substrates supporting cognitive functions asso- of its absolute and relative size. ciated with this part of the cortex enlarged and became Area 10 has similar cytoarchitectonic features in the specialized during hominid evolution. Am J Phys An- hominoid brain, but aspects of its organization vary thropol 114:224–241, 2001. © 2001 Wiley-Liss, Inc. The prefrontal cortex is located rostral to the mo- initiative” (Sanides, 1964; see also Harlow et al., tor and premotor cortices. It is also called the frontal 1964; Rosvold et al., 1964). Area 10 is located in the association cortex or the frontal granular cortex, frontal pole and is one of the areas involved with the referring to its functional and structural attributes, planning of future actions, the undertaking of initi- respectively. Like the rest of the cortex, it has been atives, and to some extent working memory and subdivided qualitatively into smaller architectonic attention (Okuda et al., 1998; Lepage et al., 2000; regions on the basis of their distinct neuronal orga- Daffner et al., 2000). nization, such as the number and size of the cortical The primate prefrontal cortex, including the fron- layers, the size, shape, and density of the neurons, tal pole, has been the focus of a host of studies for the and the degree of axon myelination. In addition, past century. In recent attempts to map the primate support for this more refined cortical parcellation cortex, the qualitative evaluation has been comple- comes from the distinct connections of each cortical mented by more objective approaches, including the area with the various subdivisions of the mediodor- quantification of different areas (Rajkowska and sal nucleus of the thalamus and other cortical and Goldman-Rakic, 1995a,b; Hof et al., 1995; Semende- subcortical structures (like the temporal and pari- feri et al., 1998). Although monkey (mostly ma- etal lobes, the hypothalamus, the amygdala, and the caque) cytoarchitecture has been described by many hippocampal formation). researchers (e.g., Brodmann, 1909; Walker, 1940; In humans, lesions on the dorsolateral portion of Preuss and Goldman-Rakic, 1991; Carmichael and the prefrontal cortex, including area 10, are associ- ated with impairment in higher-cognitive abilities that facilitate extraction of meaning from ongoing Grant sponsor: Leakey Foundation; Grant number: QM54; Grant experiences, the organization of mental contents sponsor: Wenner-Gren Foundation; Grant number: 5553; Grant spon- that control creative thinking and language, and the sor: National Science Foundation; Grant number: G13. artistic expression and planning of future actions (Damasio, 1985). Frontal lobe impairment produces *Correspondence to: Katerina Semendeferi, Department of Anthro- pology, University of California, San Diego, 9500 Gilman Drive, La the delayed response deficit that has been related to Jolla, CA 92093-0532. E-mail: [email protected] the lack of initiative or, in other words, the impair- ment of “interest and hence sustained attention and Received 1 June 1999; accepted 21 November 2000. © 2001 WILEY-LISS, INC. PREFRONTAL CORTEX IN HUMANS AND APES 225 Price, 1994), studies of the prefrontal cortex of apes only a small portion of the total volume) throughout are scarce. the cortical depth, starting at the pial surface and Here we address the issue of the homology of area ending at the white/gray matter interface. 10 on the basis of quantifiable cytoarchitectonic cri- A characteristic profile for each species was de- teria. We estimate the overall volume of the area in rived by standardizing individual profiles to the hominoids, and we explore aspects of its organiza- same relative width, and then averaging the GLIs. tion, including the relative size of cortical layers, The x-axis represents the relative cortical depth of numbers of neurons, and neuropil space. On the the area, and the y-axis represents the GLI value basis of the comparative information, we reconstruct which shows the percent-area covered by cell bodies aspects of the evolution of area 10 in hominoid and vs. neuropil space (for more details see Semendeferi hominid brains. et al., 1998). Subsequently, mean GLI values were calculated for the entire profile, and for three groups MATERIALS AND METHODS of cortical layers whose boundaries could be identi- Materials fied reliably across species (supragranular layers Subjects of all extant hominoids (humans, chim- II and III, layer IV, and infragranular layers V panzees, bonobos, gorillas, orangutans, and gibbons) and VI). were studied. One Old World monkey, the rhesus Relative size of cortical layers. The above pro- monkey, was also included as an outgroup compar- files were used in the following manner to help de- ison. Two to ten hemispheres per species were in- termine the thickness of the cortical layers and the cluded in the qualitative part of the investigation, mean GLI values within each layer. Area 10 in all and one hemisphere (right) per species was quanti- seven species was measured in the same manner. fied. The quantified specimens were selected on the Microscopic sections of the cortex were projected basis of the quality of staining and of the consistency through a camera lucida over each profile. By this in the fixation protocol and processing procedures. means, profile curves were directly superimposed Included are adult individuals of both sexes, except onto the image of the histological section, enabling for the bonobo, which is represented by a 2-year-old the maxima and minima of the profile to be matched subject (exact age is not known for apes; the human with the cytoarchitectonically and qualitatively de- was 75 years old). None of the subjects died of neu- fined layers. rological disease. All quantified specimens were emerged in 4% formalin within a few hours after the Neuronal counts. The total number of neurons natural death of the subject, and were processed and was determined for area 10 by applying the optical stained in the same manner (Semendeferi et al., fractionator, a modern stereological method for ob- 1998). They were cut at 20 ␮m, and every tenth or taining estimates of total numbers of neurons in a fifteenth section was stained with a modification of cortical area (West and Gundersen, 1990; West et the Gallyas silver stain for neuronal perikarya (Gal- al., 1991; Hyman et al., 1998). Neurons were iden- lyas, 1971; Merker, 1983). tified and counted in a light microscope, using a 60ϫ-oil immersion objective with a numerical aper- Methods ture of 1.4 (Semendeferi et al., 1998). Estimates of volumes. The volume of the cortical gray of area 10 was obtained from histological sec- RESULTS tions with the use of stereological techniques that Cytoarchitecture of area 10 estimate unbiased volume of brain structures and other irregularly shaped objects with a precision The results of the cytoarchitectonic evaluation of better than 5% (Gundersen et al., 1988a,b). area 10 are based on a qualitative description. The size of cortical layers, the stain intensity, and the Grey-level index. The grey-level index (GLI) size of cells are described in relation to each other method detects the percent-area of stained perikarya (e.g., when a cortical layer is identified as “thin,” this vs. neuropil, and demonstrates interareal and inter- means that it is thin in relation to the size of the laminar differences in these densities across species other layers). A quantitative investigation is pre- (Schleicher and Zilles, 1989; Semendeferi et al., 1998). sented later. GLI values are a summation of neuronal cell bodies, glial nuclei, and endothelial cell nuclei divided by the Human. The frontal polar cortex of nine human total space. Thus, the lower the GLI value, the more hemispheres (from five brains) is investigated in space there is for connections (Schlaug et al., 1993), serial sections stained with the Nissl and Gallyas including connections with other cortical and subcor- methods (Fig. 1). In all specimens, the cortex form- tical areas and intrinsic connections. ing the frontal pole has a remarkably homogeneous Selected locations, undistorted by cortical folding, appearance (there is no exaggeration in the appear- were digitized.