A deep-learning-based workflow to assess taxonomic affinity of hominid teeth with a test on discriminating Pongo and Homo upper molars

OBJECTIVES

Convolutional neural network (CNN) is a state-of-art deep learning (DL) method with superior performance in image classification. Here, a CNN-based workflow is proposed to discriminate hominid teeth. Our hope is that this method could help confirm otherwise questionable records of Homo from Pleistocene deposits where there is a standing risk of mis-attributing molars of Pongo to Homo.

METHODS AND MATERIALS

A two-step workflow was designed. The first step is converting the enamel-dentine junction (EDJ) into EDJ card, that is, a two-dimensional image conversion of the three-dimensional EDJ surface. In this step, researchers must carefully orient the teeth according to the cervical plane. The second step is training the CNN learner with labeled EDJ cards. A sample consisting of 53 fossil Pongo and 53 Homo (modern human and Neanderthal) was adopted to generate EDJ cards, which were then separated into training set (n = 84) and validation set (n = 22). To assess the feasibility of this workflow, a Pongo-Homo classifier was trained from the aforementioned EDJ card set, and then the classifier was used to predict the taxonomic affinities of six samples (test set) from von Koenigswald's Chinese Apothecary collection.

RESULTS

Results show that EDJ cards in validation set are classified accurately by the CNN learner. More importantly, taxonomic predictions for six specimens in test set match well with the diagnosis results deduced from multiple lines of evidence, implying the great potential of CNN method.

DISCUSSION

This workflow paves a way for future studies using CNN to address taxonomic complexity (e.g., distinguishing Pongo and Homo teeth from the Pleistocene of Asia). Further improvements include visual interpretation and extending the applicability to moderately worn teeth.

Disentangling isolated dental remains of Asian Pleistocene hominins and pongines

Scholars have debated the taxonomic identity of isolated primate teeth from the Asian Pleistocene for over a century, which is complicated by morphological and metric convergence between orangutan (Pongo) and hominin (Homo) molariform teeth. Like Homo erectus, Pongo once showed considerable dental variation and a wide distribution throughout mainland and insular Asia. In order to clarify the utility of isolated dental remains to document the presence of hominins during Asian prehistory, we examined enamel thickness, enamel-dentine junction shape, and crown development in 33 molars from G. H. R. von Koenigswald's Chinese Apothecary collection (11 Sinanthropus officinalis [= Homo erectus], 21 "Hemanthropus peii," and 1 "Hemanthropus peii" or Pongo) and 7 molars from Sangiran dome (either Homo erectus or Pongo). All fossil teeth were imaged with non-destructive conventional and/or synchrotron micro-computed tomography. These were compared to H. erectus teeth from Zhoukoudian, Sangiran and Trinil, and a large comparative sample of fossil Pongo, recent Pongo, and recent human teeth. We find that Homo and Pongo molars overlap substantially in relative enamel thickness; molar enamel-dentine junction shape is more distinctive, with Pongo showing relatively shorter dentine horns and wider crowns than Homo. Long-period line periodicity values are significantly greater in Pongo than in H. erectus, leading to longer crown formation times in the former. Most of the sample originally assigned to S. officinalis and H. erectus shows greater affinity to Pongo than to the hominin comparative sample. Moreover, enamel thickness, enamel-dentine junction shape, and a long-period line periodicity value in the "Hemanthropus peii" sample are indistinguishable from fossil Pongo. These results underscore the need for additional recovery and study of associated dentitions prior to erecting new taxa from isolated teeth.

Evolutionary Asiacentrism, Peking man, and the origins of sinocentric ethno-nationalism

This paper discusses how the theory of evolutionary Asiacentrism and the Peking Man findings at the Zhoukoudian site stimulated Chinese intellectuals to construct Sinocentric ethno-nationalism during the period from the late 1920s to the early 1940s. It shows that the theory was first popularized by foreign scientists in Beijing, and the Peking man discoveries further provided strong evidence for the idea that Central Asia, or to be more specific, Tibet, Xinjiang, and Mongolia, was the original cradle of humans. Chinese scholars in the late 1930s and 1940s appropriated the findings to construct the monogenesis theory of the Chinese, which designated that all the diverse ethnic groups within the territory of China shared a common ancestor back to antiquity.

Evolution of P3 morphology in Australopithecus afarensis

The Australopithecus afarensis dental sample exhibits a wide range of variation, which is most notable in the morphology of the lower third premolar (P3). P3 morphology in the A. afarensis sample ranges from the primitive sectorial extreme in AL 128-23 to the derived, bicuspid (molarized) extreme in AL 333w-1. In this paper, the degree and patterning of variation of the 20 known A. afarensis P3s are examined and the evolutionary implications are discussed. Initially, a series of dental and mandibular metric criteria are evaluated to determine whether this sample may be analyzed as a single species. From the metrics, it is clear that the single species hypothesis cannot be rejected. Next, a series of morphological criteria is devised to measure P3 molarization. Taken as a whole, the A. afarensis P3 sample displays more variation than a sample of modern hominoids (Pan troglodytes) and shows a slight trend toward increased molarization through time. When separated by sex, the A. afarensis sample still displays greater variation than the chimpanzee sample; however, only the male A. afarensis specimens show a trend toward increased molarization. Additionally, the male A. afarensis P3s are more molarized than the female, a pattern that is seen as well (though less markedly) in the chimpanzee sample. The trend toward increased molarization over time indicates selection for grinding in A. afarensis. The sexual differences parallel those seen in the postcrania (cf. Stern and Susman: Am. J. Phys. Anthropol. 60:279-318, 1983), as the females tend to retain the primitive condition, while the males display the derived morphology. Consequently, a model of sexual differences in niche exploitation, with the females exploiting a more arboreal environment, would seem to be supported by both the dental and postcranial evidence.

The evolution of large body size in orangutans: A model for hominoid divergence

Recent Miocene fossil discoveries of large hominoids resemble orangutans. Since the evolution of large body size was functionally related to a powerful masticatory system in Miocene ape radiations, a better understanding of adaptations in extant orangutans will be informative of hominoid evolution. It is suggested here, based on the behavioral ecology of extant orangutans, that foraging energetics and large body size are tied to a dietary shift that provided access to and utilization of resources not generally available to other primates.

On aspects of skull form in African apes and orangutans, with implications for hominoid evolution

The study of hominoid phylogeny is currently in a state of controversy and debate due to the discovery of new fossil material and reanalysis of the morphology of extant apes. An important key to the resolution of these debates lies in attaining a fuller understanding of the morphological differences in skull form between the African and Asian great apes. In this paper I have analyzed aspects of facial morphology and internal cranial anatomy in the great apes. Results from this study and previous ones suggest that Pongo is characterized by a marked dorsal deflection of the face relative to the basicranium. Many aspects of circumorbital, midfacial, palatal, and mandibular morphology in Pongo may be related to this airorynchous condition. This hypothesis is supported by Enlow's work on form and pattern in the primate and mammalian skull. The position of the face in known Sivapithecus appears to be similar to that seen in Pongo. Although Pongo may be specialized in its marked degree of airorynchy, it seems likely that an important derived feature linking African apes and hominids is a ventral rotation of the splanchnocranium on the neurocranium. The appearance of marked supraorbital tori and ethmofrontal sinuses are probably correlated developments. Additional implications of this work for debates about hominoid phylogeny are discussed.

Biometrical studies upon hominoid teeth: the coefficient of variation, sexual dimorphism and questions of phylogenetic relationship

Sexual dimorphism as a function of variation in hominoid tooth metrics has been investigated for four groups of taxa: Recent great apes (two subfamilies), Dryopiths (one subfamily), Ramapiths (one subfamily) and hominids (one family). Gorilla, and to a lesser extent Pan, appear characterized by very high levels of sexual dimorphism and meet several criteria for statistical outliers. Recent great apes are the only group exhibiting consistently high levels of sexual dimorphism. Ramapiths are the only group characterized by low levels of sexual dimorphism and their relative canine length is most similar to Dryopiths. Both Dryopiths and hominids contain taxa with low and intermediate levels of sexual dimorphism. The Gingerich and Shoeninger hypothesis relating coefficients of variation to occlusal complexity is supported. Non-parametric statistics suggest that homogeneity of coefficient of variation profiles over most of the tooth row is characteristic of only the Dryopiths and a composite data set composed of the Dryopith plus Ramapith tooth measurements. Oxnard's model for the multifactorial basis of multiple sexual dimorphisms is also supported. The Dryopith and hominid patterns of sexual dimorphism are similar, an observation that suggests phylogenetic relationship. At the taxonomic level of subfamily or family, sexual dimorphism is a character of cladistic usefulness and possible phylogenetic valence. Assuming that breeding system and sexual dimorphism are functional correlates as many workers suggest, then Ramapithecus sp. China, Sivapithecus indicus and possibly Australopithecus boisei are good candidates for having possessed monogamous breeding/social structures. All Dryopith taxa, S. sivalensis, Sivapithecus sp. China, A. afarensis, Homo habilis and H. erectus emerge as the best candidates for having possessed a polygynous breeding/social structure. No biometrical affinities of Ramapiths with hominids can be demonstrated and some phylogenetic relationship with Dryopiths is suggested. Kay's interpretation of Ramapith sexual dimorphism and taxonomic affinity is not supported. The lack of control over temporal and geographic range variation is discussed and the loose association of these variables with differences in tooth morphology is noted. The high heritability of tooth size also suggests that assignment of "high" or "low" index values to extinct taxa as a measure that describes evolving clades at discrete points in evolutionary time is appropriate.(ABSTRACT TRUNCATED AT 400 WORDS)

The evolutionary relationships of man and orang-utans

Man shares uniquely few morphological features with either the chimpanzee or the gorilla, whereas there are many features that suggest affinities between man and the orang-utan, to whom the fossil Sivapithecus appears to be closely related. If these are unique features, inclusion of Sivapithecus, man and the orang-utan in a single clade, distinct from that containing the African apes, is justified but contrary to current opinion.