This is the accepted version of the article: Beaudet, Amélie; Holloway, Ralph; Benazzi, Stefano. «A comparative study of the endocasts of OH 5 and SK 1585 : implications for the paleoneurology of eastern and southern African Paranthropus». Journal of Human Evolution, Vol. 156 (Jul. 2021), art. 103010. DOI 10.1016/j.jhevol.2021.103010 This version is available at https://ddd.uab.cat/record/240645 under the terms of the license 1 A comparative study of the endocasts of OH 5 and SK 1585: Implications for the 2 paleoneurology of eastern and southern African Paranthropus 3 4 Amélie Beaudeta,b,c*, Ralph Hollowayd, Stefano Benazzie,f 5 a 6 Department of Archaeology, University of Cambridge, Cambridge, United Kingdom b 7 School of Geography, Archaeology and Environmental Studies, University of the 8 Witwatersrand, Johannesburg, South Africa c 9 Institut Català de Paleontologia Miquel Crusafont, Universitat Auto` noma de Barcelona, 10 Barcelona, Spain d 11 Department of Anthropology, Columbia University, New York, USA e 12 Department of Cultural Heritage, University of Bologna, Ravenna, Italy f 13 Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, 14 Leipzig, Germany 15 16 *Corresponding author 17 E-mail address: beaudet.amelie@gmail.com (A. Beaudet). 18 19 Acknowledgments 20 We are indebted to S. Potze and M. Tawane (Pretoria), L. Bam, F. de Beer and J. Hoffman 21 (Pelindaba) for collection access and data acquisition. Model-based deformation computation 22 was granted access to the supercomputing centre of CHPC (AB). For scientific contributions 23 and/or discussion, the authors are especially grateful to J. Braga (Toulouse), J. Dumoncel 24 (Toulouse), A. Oettlé (Pretoria), D. Stratford (Johannesburg) and J.F. Thackeray 25 (Johannesburg). We thank the Editor, Associate Editor and three anonymous reviewers for 26 their comments, which contributed to improving the original version of this manuscript. A.B. 27 is funded by the DST-NRF Center of Excellence in Palaeosciences (CoE-Pal) and the 28 University of the Witwatersrand. S.B. is funded by the European Research Council (ERC) 29 under the European Union’s Horizon 2020 research and innovation programme (grant 30 agreement No 724046–SUCCESS, http://www.erc-success.eu). Opinions expressed and 31 conclusions arrived at, are those of the author and are not necessarily to be attributed to the 32 CoE-Pal. 33 34 1 35 Short Communications 36 37 A comparative study of the endocasts of OH 5 and SK 1585: Implications for the 38 paleoneurology of eastern and southern African Paranthropus 39 40 41 Keywords: Brain shape; Early hominins; Olduvai Gorge; Swartkrans; Surface-based 42 comparison 43 44 2 45 1. Introduction 46 The taxonomy, phylogeny and biology of the Paranthropus species have been the center 47 of debates since the earliest discovery of the TM 1517 (now attributed to Paranthropus 48 robustus) cranium from the Plio-Pleistocene site of Kromdraai (South Africa) in 1938 by R. 49 Broom. In particular, whether Paranthropus boisei and P. robustus represent two distinct taxa, 50 and whether the two species emerged from a common ancestor (i.e., hypothesis of a 51 monophyletic group), remain largely unresolved (reviewed in Constantino and Wood, 2004, 52 2007; Wood and Schroer, 2017). Besides the taxonomic and phylogenetic aspects, the presence 53 of two contemporaneous ‘robust’ species in the eastern and southern African hominin fossil 54 records raises critical questions regarding potential occurrences of homoplasies in the hominin 55 clade, with substantial implications for our understanding of early hominin paleobiology 56 (Wood and Schroer, 2017). Moreover, the absence of associated cranial and postcranial 57 remains attributed to Paranthropus complicate the reconstruction of the behaviour of this 58 enigmatic ‘megadont’ hominin genus (reviewed in Constantino and Wood, 2007). 59 In this respect, brain endocasts may have the potential to provide additional insights 60 into the taxonomy, phylogeny and behavior of Paranthropus taxa. Until now, only a few 61 studies have attempted to assess the evolutionary polarity of the cerebral features in 62 Paranthropus, and even fewer have addressed the question of potential similarities and 63 differences in the brain of the two regional variants of Paranthropus. By investigating the 64 endocasts of Paranthropus aethiopicus, P. boisei and P. robustus, Falk et al. (2000) suggested 65 substantial similarities between Paranthropus and extant chimpanzees and gorillas (i.e., similar 66 overall endocranial dimensions, beak-shaped profile frontal lobes, rounded temporal poles) and 67 hypothesized a less derived cerebral condition in Paranthropus as compared to 68 Australopithecus. However, the description of the sulcal pattern in P. robustus (SK 1585) 69 revealed a configuration of the lunate sulcus that is more similar to extant humans than to extant 70 chimpanzees, thus suggesting a possible reorganisation of the cortical pattern (Holloway, 71 1972). However, there is an ongoing debate as to whether the lunate sulcus is homologous 72 between apes and humans (e.g., Allen et al., 2006; Malikovic et al., 2012) and on the position 73 of the lunate sulcus in early hominins (e.g., Falk, 1980, 1983, 2009; Holloway, 1981; Holloway 74 et al., 2004a; Falk et al., 2018; Gunz et al., 2020; but see Holloway et al., 2016). Interestingly, 75 Holloway (1972) identified critical similarities in the overall shape of the endocast between P. 76 boisei (OH 5) and P. robustus (SK 1585; see Figs. 6–8 in Holloway, 1972), notwithstanding 77 differences in the cerebellum with a rounded shape in P. boisei (Omo L338y-6) and a triangular 78 shape for P. robustus (SK 1585; Holloway, 1972, 1981). 3 79 Given the ongoing debate on the evolutionary polarity of the Paranthropus cerebral 80 features, in this paper we (i) test the hypothesis of a nonhuman, ape-like endocast in 81 Paranthropus and a less derived condition as compared to Australopithecus (Falk et al., 2000), 82 and (ii) investigate further potential morphological similarities between the eastern and 83 southern Paranthropus endocasts (Holloway, 1972). Morphological affinities are 84 quantitatively assessed through the application of a deformation-based method (Durrleman et 85 al., 2012; Beaudet et al., 2016, 2018). Because of their remarkable degree of preservation, we 86 focus our study on the endocasts of OH 5 (P. boisei) and SK 1585 (P. robustus). More 87 specifically, we compare OH 5 and SK 1585 to extant humans, common chimpanzees (Pan 88 troglodytes) and bonobos (Pan paniscus) and to the Australopithecus specimen Sts 5, using 89 statistical shape analyses and color maps that render global and local morphological differences 90 and similarities between groups/specimens. 91 92 2. Materials and methods 93 2.1. Samples 94 The two Paranthropus specimens included in our study, i.e., OH 5 and SK 1585, represent 95 the most complete and well-preserved evidence currently available in the fossil record of the 96 P. boisei and P. robustus endocasts, respectively (Holloway, 1972; Benazzi et al., 2011). The 97 OH 5 face and neurocranium were found in 1959 by Leakey (1959) from the Bed I of Olduvai 98 Gorge and dated to 1.85 Ma. This specimen was first published as the holotype of the hominin 99 species Zinjanthropus boisei, which is now included in the genus Paranthropus (Tamrat et al., 100 1995). The face and neurocranium of OH 5 were scanned by computed tomography (CT) at 101 the Radiologie 2 Department at Medizinische Universität Innsbruck (Austria) with a resolution 102 of 0.48 x 0.48 x 1 mm. Benazzi et al. (2011) virtually reconstructed the OH 5 cranium by 103 combining semilandmark-based geometric morphometric methods and computer-aided design 104 techniques and using KNM-ER 406 as a reference, and produced a complete endocast (Fig. 1). 105 SK 1585 represents a natural endocast of P. robustus from the lime miners’ rubble 106 dump at Swartkrans (Brain, 1970). Although the stratigraphic provenance is not clear, the 107 maximum age is estimated to about 2.19 Ma, based on the cosmogenic nuclide burial dating of 108 Member 1 (Gibbon et al., 2014). The endocast preserves the right hemisphere and a very small 109 portion of the left occipital lobe (Holloway, 1972; Fig. 1). SK 1585 was scanned at the South 110 African Nuclear Energy Corporation in Pelindaba (South Africa) at a resolution of 0.066 mm 111 (Fig. 1; Beaudet et al., 2019). We generated a virtual surface of the natural endocast of SK 112 1585 using Avizo v. 9.0 (Visualization Sciences Group Inc., Berlin). For comparative purposes, 4 113 we included the nearly complete endocast of the Australopithecus africanus specimen Sts 5 114 from the Sterkfontein Caves in South Africa (Broom, 1947; for further details see Beaudet et 115 al., 2018). 116 In terms of extant comparative material, we included 10 Homo sapiens, 10 Pan 117 troglodytes and 10 Pan paniscus with equal proportion of fully mature males and females. The 118 sample of Homo was obtained from the human clinical records of the Pasteur Hospital in 119 Toulouse, France, while the Pan samples were obtained from the Royal Museum for Central 120 Africa, Tervuren, Belgium (Beaudet et al., 2018). The human and nonhuman specimens were 121 scanned by medical CT with a resolution of 0.49–0.50 and 0.39–0.80, mm respectively. 122 123 2.2. Methods of analysis 124 To investigate the nature of the morphological differences between the fossil specimens 125 and the extant groups, we used a size-independent and landmark-free registration method based 126 on smooth and invertible surface deformation by using the software Deformetrica v4 127 (Durrleman et al., 2012; Beaudet et al., 2018). In this paper, we follow the method detailed in 128 Beaudet et al. (2018). First, the surfaces were automatically aligned in position, orientation, 129 and scale using one surface randomly selected as a reference. This was performed by using the 130 tool “Align surfaces” available in Avizo v. 9.0 which is based on the iterative closest point 131 algorithm that minimizes the root mean square distance between the points of each specimen 132 to corresponding points on the reference (Besl and McKay, 1992). The option “rigid + uniform 133 scale” was selected to remove the siz . A template was deformed to each extant specimen. From 134 this process, a global mean shape was generated (Durrleman, 2010). Based on the deformation 135 fields from the global mean shape to each extant specimen, we computed taxon-specific mean 136 shapes (i.e., Homo sapiens, Pan troglodytes and Pan paniscus; Fig. 1). Then, the global mean 137 shape, as well as the taxon-specific mean shapes, were deformed to the fossil specimens. The 138 global mean shape and the deformation fields represent an atlas (Durrleman et al., 2012). In 139 addition, we deformed the endocast of Sts 5 to the endocast of OH 5. In this analysis, we used 140 a total of 22,275 control points with a kernel width parameter of 5. 141 We ran a second analysis using only the preserved endocranial region in SK 1585 142 following the method detailed in Dumoncel et al. (2016; see also Fig. 1 in Beaudet et al., 2018). 143 Accordingly, the region corresponding to the missing surface in SK 1585 was virtually 144 removed from each endocast included in our sample and we computed a second atlas as well 145 as global and taxon-specific mean shapes that were deformed to the fossil endocasts (Beaudet 5 146 et al., 2018). Finally, we deformed Sts 5 to SK 1585, and SK 1585 to OH 5. In this analysis, 147 we used a total 17 600 control points with a kernel width parameter of 5. 148 The 3D deformation fields integrating local orientation and the amplitude of the two 149 deformation sets were statistically analyzed by principal component analyses (PCA) using 150 complete (i.e., extant Homo and Pan specimens, OH 5 and Sts 5) and partial (i.e., extant Homo 151 and Pan specimens, OH 5, SK 1585 and Sts 5) endocasts to quantify shape variation. Signs 152 were obtained by comparing the displacements of the direction perpendicular to the tangential 153 surface, and those signs were applied to the magnitude and orientation of the displacements 154 recorded during the deformation process. Signed distances are rendered by color maps from 155 dark blue, indicating negative values and, thus, flatter surfaces, to red, indicating positive 156 values, and thus more bulging surfaces (Beaudet and Bruner, 2017). In this respect, similar 157 colors indicate similar amount and orientation of displacements between the fossil specimens 158 and the mean shapes, or between fossil specimens. The codes (developed with Python and R) 159 used for the post-processing steps (i.e., PCA, signed color maps) are available at 160 https://gitlab.com/jeandumoncel/tools-for-deformetrica (Beaudet et al., 2018). 161 162 3. Results 163 The general shapes of the endocasts of the Paranthropus specimens OH 5 and SK 1585 are 164 visually compared with the Australopithecus specimen Sts 5 and with the extant human, 165 common chimpanzee and bonobo taxon-specific mean shapes in Figure 1. Results of the 166 deformation process are shown in Figure 2 (statistical analyses) and Figures 3–4 (color maps). 167 The statistical analysis of shape variation using scaled complete endocasts (Fig. 2A) 168 indicates that the morphology of the endocast of OH 5 closely approximates the morphology 169 of Pan (and more particularly P. paniscus) along PC1. By contrast, Sts 5 falls in between the 170 Homo and Pan clusters. The statistical analysis of the scaled partial endocast (Fig. 2B) shows 171 that the morphology of the endocasts of the two Paranthropus specimens OH 5 and SK 1585 172 are more similar to Pan than to Homo along both axes, while Sts 5 plots in between Homo and 173 Pan along PC1, as in the previous analysis. 174 The dorsal surface of the parietal lobes in OH 5 and SK 1585 is more flattened 175 dorsoventrally than in Homo and more bulging than in Pan (Fig. 3). Moreover, the dorsal 176 surface of the frontal lobes in both fossil specimens is more flattened dorsoventrally than in 177 Homo and Pan except along the superior sagittal sinus. The temporoparietal and 178 temporooccipital areas in OH 5 and SK 1585, that encompass the inferior parietal lobule where 179 the supramarginal and angular gyri are located, are more pronounced than in Homo but less 6 180 than in Pan. Moreover, the inferior frontal gyrus is less developed in OH 5 and in SK 1585 181 than in Homo, but slightly more developed than in Pan in the case of SK 1585. The cerebellar 182 lobes in the fossil specimens and in Pan protrude more laterally than in Homo. 183 The endocasts of the Paranthropus specimens OH 5 and SK 1585 differ from the 184 endocast of the Australopithecus specimen Sts 5 by displaying more flattened frontal lobes 185 dorsoventrally, temporal lobes and parietooccipital regions that are more laterally expanded, 186 and less protruding cerebellar lobes (Fig. 4). The areas that mainly differ between OH 5 and 187 SK 1585 correspond to cracks visible on the right parietal and temporal regions of the endocast 188 of OH 5 (Figs. 1 and 4). 189 190 4. Discussion 191 Overall, within the limits of our sample, the two Paranthropus specimens share with 192 the extant common chimpanzees and bonobos a laterally broad and dorsoventrally flattened 193 endocast (Figs. 1 and 3), a morphology that has been previously described in Australopithecus 194 through the use of the deformation-based method. However, our statistical analyses and color 195 maps indicate that the A. africanus specimen Sts 5 is intermediate between the extant human 196 and chimpanzee conditions and less laterally narrower and less dorsoventrally flattened than in 197 the Paranthropus specimens (Fig. 2). Even if limited by one specimen per species, our study 198 of shape variation provides additional support for the affinities of the shape of the 199 Paranthropus endocast with nonhuman apes and the hypothesis of a less derived endocast in 200 Paranthropus as compared to Australopithecus (Falk et al., 2000). If we consider the 201 hypothesis of a brain shape in early Homo that is more modern human-like than ape-like (e.g., 202 Tobias, 1991), our results may suggest distinct brain evolutionary histories between the 203 contemporaneous hominin genera Homo and Paranthropus. Moreover, the human-like 204 configuration of the lunate sulcus in Paranthropus, as revealed by Holloway (1972), raises 205 questions of the timing and pattern of the evolution of the brain, and may imply that cortical 206 reorganization occurred without noticeable changes in overall shape of the brain. Investigating 207 a larger sample using this deformation-based approach would be of particular interest to 208 evaluate patterns of intraspecific variation and further address these questions (Neubauer et al., 209 2012). 210 The frontal lobes are of particular interest because of their role in executive functions 211 and language (e.g., Kringelbach and Rolls, 2004; Petrides and Pandya, 1999; Rajkowska and 212 Goldman-Rakic, 1995; Schenker et al., 2010). The frontal lobes in the Paranthropus specimens 213 are relatively flattened as compared to Homo and Pan. Rather than reflecting potential cerebral 7 214 changes, this morphological specificity could be related to masticatory hypertrophy and 215 structural constraints imposed by the integration of the face with the braincase (Wu and Bruner, 216 2016; Beaudet and Bruner, 2017). Our study indicates substantial morphological differences in 217 the area that encompasses the inferior temporal lobule, including the supramarginal and angular 218 gyri, between the two Paranthropus specimens and Homo (i.e., more pronounced in 219 Paranthropus), and between the two Paranthropus specimens and Pan (i.e., less pronounced 220 in Paranthropus). Since this region is part of the somatosensory association cortex 221 (Mountcastle et al., 1974; Celsis et al., 1999; Seghier, 2013), differences in this area between 222 Paranthropus, Homo and Pan might have potential behavioral implications (e.g., locomotion, 223 manipulation). However, these interpretations would need to be further explored in the future 224 by the comparative study of the sulcal pattern in these key cerebral regions in the endocasts of 225 these taxa. The cerebellar lobes of the two Paranthropus specimens protrude more laterally 226 than in extant humans but less than in Sts 5, which confirms previous observations (Holloway, 227 1972). As for the supramarginal and angular gyri, differences between the Paranthropus, 228 Australopithecus and extant specimens might be interpreted as potential evidence suggesting 229 distinct behaviours because of the role of these regions in fundamental motor-related functions 230 (e.g., Kochiyama et al., 2018; Neubauer et al., 2018). 231 Finally, the overall shapes of the endocasts of the P. boisei specimen OH 5 and the P. 232 robustus specimen SK 1585 are relatively similar (Fig. 4), as previously pointed out by 233 Holloway (1972). Similarities between eastern and southern African Paranthropus would need 234 to be confirmed by future studies of more complete endocasts of southern African 235 Paranthropus. 236 237 238 References 239 Allen, J. S., Bruss, J., Damasio, H., 2006. Looking for the lunate sulcus: A magnetic resonance 240 imaging study in modern humans. Anat. Rec. 288, 867–876. 241 Beaudet, A., Bruner, E., 2017. A frontal lobe surface analysis in three archaic African human 242 fossils: OH 9, Buia, and Bodo. C. R. Palevol. 16, 499–507. 243 Beaudet, A., Dumoncel, J., de Beer, F., Duployer, B., Durrleman, S., Gilissen, E., Hoffman, J., 244 Tenailleau, C., Thackeray, J.F., Braga, J., 2016. Morphoarchitectural variation in South 245 African fossil cercopithecoid endocasts. J. Hum. Evol. 101, 65–78. 8 246 Beaudet, A., Dumoncel, J., de Beer, F., Durrleman, S., Gilissen, E., Oettle, A., Subsol, G., 247 Thackeray, J.F., Braga, J., 2018. The endocranial shape of Australopithecus africanus: 248 Surface analysis of the endocasts of Sts 5 and Sts 60. J. Anat. 232, 296–303. 249 Beaudet, A., Clarke, R.J., de Jager, E., Bruxelles, L., Carlson, K.J., Crompton, R., de Beer, F., 250 Dhaene, J., Heaton, J.L., Jakata, K., Jashashvili, T., Kuman, K., McClymont, J., 251 Pickering, T.R., Stratford, D., 2019. The endocast of StW 573 (“Little Foot”) and 252 hominin brain evolution. J. Hum. Evol. 126, 112–123. 253 Benazzi, S., Bookstein, F.L., Strait, D., Weber, G.W., 2011. A new OH5 reconstruction with 254 an assessment of its uncertainty. J. Hum. Evol. 61, 75–88. 255 Besl, P.J., McKay, N.D., 1992. A method for registration of 3-D shapes. IEE Trans. Pattern 256 Anal. 14, 239–256. 257 Brain, C.K., 1970. New finds at the Swartkrans australopithecine site. Nature 225, 1112–1119. 258 Broom, R., 1938. The Pleistocene anthropoid apes of South Africa. Nature 142, 377–379. 259 Broom, R., 1947. Discovery of a new skull of the South African ape-man, Plesianthropus. 260 Nature 159, 672. 261 Celsis, P., Boulanouar, K., Doyon, B., Ranjeva, J.P., Berry, I., Nespoulous, J.L., Chollet, F., 262 1999. Differential fMRI responses in the left posterior superior temporal gyrus and left 263 supramarginal gyrus to habituation and change detection in syllables and tones. 264 Neuroimage 9, 135–144. 265 Constantino P, Wood B. 2004. Paranthropus paleobiology. Miscelánea en homenaje a 266 Emiliano Aguirre. Museo Arqueolócico Regional, Madrid, pp. 136–151. 267 Constantino, P., Wood, B., 2007. The evolution of Zinjanthropus boisei. Evol. Anthropol. 16, 268 49–62. 269 Dumoncel, J., Subsol, G., Durrleman, S., Jessel, J.-P., Beaudet, A., Braga, J., 2016. How to 270 build an average model when samples are variably incomplete? Application to fossil 271 data. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 272 (CVPR) Workshops, pp. 541-548. 273 Durrleman, S., 2010. Statistical models of currents for measuring the variability of anatomical 274 curves, surfaces and their evolution. Ph.D. Dissertation, Université Nice-Sophia 275 Antipolis. 276 Durrleman, S., Pennec, X., Trouvé, A., Ayache, N., Braga, J., 2012. Comparison of the 277 endocranial ontogenies between chimpanzees and bonobos via temporal regression and 278 spatiotemporal registration. J. Hum. Evol. 62, 74–88. 9 279 Falk, D., 1980. A reanalysis of the South African australopithecine natural endocasts. Am. J. 280 Phys. Anthropol. 53, 525–539. 281 Falk, D., 1983. Cerebral cortices of East African early hominids. Science 221, 1072–1074. 282 Falk, D., 2009. The natural endocast of Taung (Australopithecus africanus): Insights from the 283 unpublished papers of Raymond Arthur Dart. Am. J. Phys. Anthropol. 49, 49-65. 284 Falk, D., Redmond Jr., J.C., Guyer, J., Conroy, G.C., Recheis, W., Weber, G.W., Seidler, H., 285 2000. Early hominid brain evolution: A new look at old endocasts. J. Hum. Evol. 38, 286 695–717. 287 Falk, D., Zollikofer, C.P.E., Ponce de León, M., Semendeferi, K., Alatorre Warren, J.L., 288 Hopkins, W.D., 2018. Identification of in vivo sulci on the external surface of eight 289 adult chimpanzee brains: Implications for interpreting early hominin endocasts. Brain 290 Behav. Evol. 91, 45–58. 291 Gibbon, R.J., Pickering, T.R., Sutton, M.B., Heaton, J.L., Kuman, K., Clarke, R.J., Brain, C.K., 292 Granger, D.E., 2014. Cosmogenic nuclide burial dating of hominin bearing Pleistocene 293 cave deposits at Swartkrans, South Africa. Quat. Geochronol. 24, 10–15. 294 Gunz, P., Neubauer, S., Falk, D., Tafforeau, P., Le Cabec, A., Smith, T.M., Kimbel, W.H., 295 Spoor, F., Alemseged, Z., 2020. Australopithecus afarensis endocasts suggest ape-like 296 brain organization and prolonged brain growth. Sci. Adv. 6, eaaz4729. 297 Holloway, R.L., 1972. New australopithecine endocast, SK 1585, from Swartkrans, South 298 Africa. Am. J. Phys. Anthropol. 37, 173–186. 299 Holloway, R., 1981. The endocast of the Omo L338y-6 juvenile hominid: Gracile or robust 300 Australopithecus? Am. J. Phys. Anthropol. 54, 109–118. 301 Holloway, R.L., Clarke, R.J., Tobias, P.V., 2004a. Posterior lunate sulcus in Australopithecus 302 africanus: Was Dart right? C. R. Palevol. 3, 287–293. 303 Holloway, R.L., Broadfield, D.C., Yuan, M.S., 2004b. The Human Fossil Record: Brain 304 Endocasts, the Paleoneurological Evidence. Wiley-Liss, Hoboken. 305 Holloway, R.L., Schoenemann, P.T., Broadfield, D.C., 2016. Why paleoneurology needs the 306 lunate sulcus. Am. J. Phys. Anthropol. Suppl. 159, 175. 307 Kochiyama, T., Ogihara, N., Tanabe, H.C., Kondo, O., Amano, H., Hasegawa, K., Suzuki, H., 308 Ponce de León, M.S., Zollikofer, C.P.E., Bastir, M., Stringer, C., Sadato, N., Akazawa, 309 T., 2018. Reconstructing the Neanderthal brain using computational anatomy. Sci. Rep. 310 8, 6296. 10 311 Kringelbach, M.L., Rolls, E.T., 2004. The functional neuroanatomy of the human orbitofrontal 312 cortex: evidence from neuroimaging and neuropsychology. Prog. Neurobiol. 72, 341– 313 372. 314 Leakey, L.S.B., 1959. A new fossil skull from Olduvai. Nature 201, 967–970. 315 Malikovic, A., Vucetic, B., Milisavljevic, M., Tosevski, J., Sazdanovic, P., Milojevic, B., 316 Malobabic, S., 2012. Occipital sulci of the human brain: Variability and morphometry. 317 Anat. Sci. Int. 87, 61–70. 318 Mountcastle, V.B., Lynch, J.C., Georgopoulos, A., Sakata, H., Acuna, C., 1975. Posterior 319 parietal association cortex of the monkey: Command functions for operations within 320 extrapersonal space. J. Neurophysiol. 38, 871–908. 321 Neubauer, S., Gunz, P., Weber, G.W., Hublin, J.-J., 2012. Endocranial volume of 322 Australopithecus africanus: New CT-based estimates and the effects of missing data 323 and small sample size. J. Hum. Evol. 62, 498-510. 324 Neubauer, S., Hublin, J.-J., Gunz, P., 2018. The evolution of modern human brain shape. Sci. 325 Adv. 4, eaao5961. 326 Petrides, M., Pandya, D.N., 1999. Dorsolateral prefrontal cortex: Comparative 327 cytoarchitectonic analysis in the human and the macaque brain and corticocortical 328 connection patterns. Eur. J. Neurosci. 11, 1011–1036 329 Rajkowska, G., Goldman-Rakic, P.S., 1995. Cytoarchitectonic definition of prefrontal areas in 330 the normal human cortex: II. Variability in locations of areas 9 and 46 and relationship 331 to the Talairach Coordinate System. Cereb. Cortex 5, 323–337 332 Schenker, N.M., Hopkins, W.D., Spocter, M.A., Garrison, A.R., Stimpson, C.D., Erwin, J.M., 333 Hof, P.R., Sherwood, C.C., 2010. Broca’s area homologue in chimpanzees (Pan 334 troglodytes): Probabilistic mapping, asymmetry, and comparison to humans. Cereb. 335 Cortex 20, 730–742. 336 Seghier, M.L., 2013. The angular gyrus: Multiple functions and multiple subdivisions. 337 Neuroscientist 19, 43–61. 338 Tamrat, E., Thouveny, N., Taieb, M., Opdyke, N.D., 1995. Revised magnetostratigraphy of the 339 Plio-Pleistocene sedimentary sequence of the Olduvai formation (Tanzania). 340 Palaeogeogr. Palaeoclimatol. Palaeoecol. 114, 273–283. 341 Tobias, P.V., 1991. Olduvai Gorge. In: The Skulls, Endocasts and Teeth of Homo habilis, Vol. 342 4. Cambridge University Press, Cambridge. 343 Wood, B., Schroer, K., 2017. Paranthropus: Where do things stand? In: Marom, A., Hovers, 344 E. (Eds.), Human Paleontology and Prehistory. Springer, New York, pp. 95–107. 11 345 Wu, X., Bruner, E., 2016. The endocranial anatomy of Maba 1. Am. J. Phys. Anthropol. 160, 346 633–643. 12 Figure legends Figure 1. Virtual rendering of the endocasts of the Paranthropus specimens OH 5 (P. boisei) and SK 1585 (P. robustus) and Sts 5 (Australopithecus africanus), and the taxon-specific mean shapes of extant humans (Hs), common chimpanzees (Pt) and bonobos (Pp) in lateral right (top) and superior (bottom) views. Images not to scale. 13 Figure 2. Principal component analyses of the deformation-based shape comparisons of the A) complete and B) partial Paranthropus (OH 5, SK 1585), Australopithecus (Sts 5), extant human (red triangles), common chimpanzee (green diamonds) and bonobo (blue squares) endocasts. OH 5 and SK 1585 are more similar to Pan than to Homo along both axes, while Sts 5 plots in between Homo and Pan. 14 Figure 3. Comparative maps of morphological deformations from the extant human (Hs), common chimpanzee (Pt) and bonobo (Pp) mean shapes to the Paranthropus specimens OH 5 and SK 1585. Endocasts for OH5 are shown in lateral right (top row), superior (second row) and lateral left (third row) views. Endocasts for SK 1585 are shown in lateral right (fourth row) and superior (bottom row) views. Cumulative displacement variations (in mm) in the direction perpendicular to the tangential surface are rendered by a pseudocolor scale ranging from dark blue (highest negative values, more flatter surfaces) to red (highest positive values, more bulging surfaces) at the fossil surfaces. 15 Figure 4. Comparative maps of morphological deformations from Sts 5 to OH 5 (top), from Sts 5 to SK 1585 (middle) and from SK 1585 to OH 5 (bottom). Endocasts are shown in lateral right, superior and lateral left (OH 5 only) views. Cumulative displacement variations (in mm) in the direction perpendicular to the tangential surface are rendered by a pseudocolor scale ranging from dark blue (highest negative values, more flatter surfaces) to red (highest positive values, more bulging surfaces) at the OH 5 (top and bottom) and SK 1585 (middle) surfaces. 16 17