Identification of Zhoukoudian Homo erectus brain asymmetry using 3D laser scanning

Endocasts are important materials used for the study of human brain evolution, and allow examination of the external features of brain anatomy from the inside the cranium. Studies examining brain asymmetries in fossil hominids are usually limited to scoring of differences in hemisphere protrusion rostrally and caudally, or to comparing the width of the hemispheres. In the present study, using 3D laser scanning, we examined asymmetries of the hemisphere volumes and surface areas in the Zhoukoudain (ZKD) Homo erectus, dated to 0.4–0.8 Ma. Compared with modern endocasts, we found that the absolute hemisphere volumes and surface areas exhibited no significant asymmetries in the ZKD or in modern specimens. However, the relative hemisphere volumes against surface areas differed between the two groups. When comparing the relative sizes between the left and right hemispheres, the ZKD specimens exhibited a greater variation than in the modern humans; there were no differences in the two hemispheres in the ZKD specimens, while in the modern endocasts the left hemisphere was significantly greater than the right hemisphere. These data suggest that brain asymmetries originated from relative brain sizes rather than absolute brain volumes during human evolution. These anatomical changes are likely related to the origin of human brain lateralization.

It is well established that each hemisphere of the brain specializes in certain activities, a phenomenon termed brain asymmetry, and the organization of the brain is closely related to brain function [1]. In modern humans, numerous anatomical and X-ray, CT, and MRI imaging studies have documented anatomical cerebral asymmetries between the hemispheres, which include the weight of the hemisphere [2], the length of the sigmoid fissure [3], the flow direction of the venous sinuses [4], the protrusion degree of the frontal pole and the occipital pole [5,6], the number of nerve cells [7], the patterns of brain sulcus and gyrus [8], the size of the lateral ventricle [9], and the rates of cerebral maturation [10]. Brain asymmetries have also been correlated with asymmetrical behavioral traits including language, handed-ness, musical talents, and visuospatial abilities [11].
For fossil hominids, the soft brain tissue was removed during fossilization. As such, endocasts are the direct material for the study of brain asymmetries. The endocast is the impression taken directly from the inside of a cranium that retains the surface features of the brain, and produces the most original material for studies of human brain evolution [12]. The structure of the human brain is extremely complex and is composed of multiple sulci and gyri. Even though endocasts reflect the surface of the brain, the sulci are not very clear compared with the brain itself. The most common landmarks that can be seen on endocasts are the coronal suture, sagittal suture, lambdoidal suture, and lateral sulcus, although these endocast landmarks are not very consistent with those seen on the brain. Nevertheless, as we know the variation of the structural organization of the brain, we can determine the general shape of the brain by measuring the endocranial cast. Analysis of fossil hominid endocasts allows inference of functional anatomy, physiology, and phylogeny.
Studies examining cerebral asymmetries in fossil hominids are usually limited to scoring of which hemisphere protruded further rostrally and caudally, or to comparisons of which hemisphere was wider [4,6,13,14]. Paleoanthropologists studying the fossil endocasts of Australopithecus, Homo habilis, Homo erectus, Neanderthals, and Homo sapiens have reported that almost all brain endocasts display distinct cerebral asymmetry. The most common and distinctive pattern is a dominance of the right frontal and left occipital lobes [13,15]. The increase in brain complexity, as seen in advanced tool using and symbolic language, is linked to these cerebral asymmetries [11,16]. It is likely that cerebral asymmetry and handedness have evolved very early in human evolution. Asymmetries seen in comparative studies of human fossil endocasts provide strong evidence for phylogenetic origins of brain lateralization.
The ZKD is the earliest Homo erectus site ever found in China. Since the official excavation in 1927, a group of six nearly complete crania were discovered in ZKD Locality 1. From these, six ZKD Homo erectus endocasts were reconstructed from the original cranium fossils, and are termed: ZKD II, III, V, X, XI and XII (Table 1). From the first reconstruction of the endocast ZKD Homo erectus (skull III), the left inferior frontal gyrus (Broca's area), which is associated with speech functions, was larger than the right hemisphere, suggestive of cerebral asymmetry [19]. Subsequently, the endocasts of ZKDII, ZKD X, ZKD XI and ZKD XII were reconstructed from the original skull fossils, and it was suggested that the ZKD specimens exhibited a larger left hemisphere than the right hemisphere [20,21]. Finally, reconstruction of the endocasts of ZKD V [22] suggested broader frontal and occipital lobes when compared with the other ZKD endocasts [23,24], which are similar to modern humans [13,14].
Currently, the asymmetrical studies on the ZKD specimens are restricted to the comparison of the protruding level or width between the two hemispheres. Although it was suggested that the ZKD Homo erectus might exhibit a larger left hemisphere than the right hemisphere, the subsequent studies did not examine similar parameters. In this study, we used a 3D laser surface technique to scan the ZKD Ho-mo erectus endocasts, from which we reconstructed 3D brain images. Using these methods we were able to calculate absolute and relative volumes and surface areas of two hemispheres.

Materials
Details of the six ZKD endocasts used in this study are shown in Figure 1 and Table 1. ZKD II, ZKD III, ZKD V, ZKD X, ZKD XI, and ZKD XII were from the collections of the Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences. Comparative endocasts of 42 modern humans (21 males and 21 females) were from RLH's collection at Columbia University and the medical department at the University of the Witwatersrand.

3D laser scanning
Using a NextEngine Model 2020i Desktop 3D Scanner (www.NextEngine.com), the six ZKD specimens and 42 modern endocasts were scanned and created into 3D models. The digitizer-based and 3D model-based coordinate measurements had overall standard deviations of ±0.79 and ±1.05 mm, respectively [25]. The scanner is in a portable shape, and can capture full-color 3D images with highmeasurement accuracy. This scanner is an excellent tool for the study of skulls and endocasts. 3D models were imported into Rapidworks™ software program (NextEngine Inc,  Santa Monica, California, US) on a Dell workstation. To standardize the measuring procedure, each endocast was placed on a flat surface with the horizontal plane along the X-axis of the frontal and occipital poles parallel to the flat surface whether rotated for a superior or lateral view (Figure 2(a)), and the coordinate system of the vertical plane along the Y-axis passing through the sagittal center of the two hemispheres (Figure 2(b), (c)). The coordinate system of each model was adjusted so that the axes were made normal to this plane. Along the Y-axis plane, each endocast was split into the left hemisphere (Figure 2(f)) and the right hemisphere (Figure 2(h)).

Measurements and statistics
We determined the absolute value of the left hemisphere volume, right hemisphere volume, left hemisphere surface area, and right hemisphere surface area using Rapidworks software. As the absolute brain size relates to the body size, we used the "hemisphere relative size" to reduce the influence of the predicted body size. By scaling hemisphere volume against surface area, it is possible to make in-ter-specific comparisons between the ZKD Homo erectus and the modern specimens. The formula for the hemisphere relative size is: (1) Left hemisphere relative size (%) = 100 × left hemisphere volume 1/3 /left hemisphere surface area 1/2 ; (2) Right hemisphere relative size (%) = 100 × right hemisphere volume 1/3 /right hemisphere surface area 1/2 . Statistical analyses were performed using SPSS 13.0 (SPSS Inc, Chicago, Illinois, US). Paired samples t-test was used to determine whether the left hemispheres were different from the right hemispheres between the ZKD specimens and the modern endocasts.

Comparisons between left and right hemisphere volumes
Metric data on the ZKD and the modern humans in our samples are shown in Table 2. The right hemisphere volumes were larger than the left hemisphere volumes for ZKD II, ZKD III, ZKD V, ZKD XI and ZKD XII, and the differences between the two hemispheres were 12.2, 5.6, 7.8, 25.6,  . There were no differences between the left versus the right hemisphere volumes among the six ZKD specimens and the modern humans in our study (Table 3).

Comparisons between left and right hemisphere surface areas
The right hemisphere surface area was larger than the left hemisphere for ZKD II, ZKD III, ZKD V, ZKD XI, and ZKD XII, and the differences between the two hemispheres were 4.3, 0.9, 1.1, 10.1, and 0.3 cm 3 , respectively. The left hemisphere surface area for ZKD X was larger than the right hemisphere surface area, and the difference between the two hemispheres was 5.4 cm 3 . For the modern humans, the left hemisphere surface area ranges 280.9-386.8 cm 3 (average, 336.7 cm 3 ), and the right hemisphere volume ranges 280.6-397.2 cm 3 (average, 337.9 cm 3 ). There was no difference between the left versus the right surface area among the six ZKD specimens and the modern humans in our sample (Table 3).

Comparisons between the left and right hemisphere relative sizes
The relative size of the right hemisphere was larger than the left hemisphere for ZKD III, ZKD V, and ZKD XII, and the differences between the two hemispheres were 0.2%, 0.1%, and 0.4%, respectively. The left hemisphere relative size was larger than the right hemisphere for ZKD X and ZKD XI. ZKD II (left, 49.1%; right, 49.1%) and ZKD V (left, 46.9%; right, 47.0%) were higher than the range of the modern humans (left range, 47.1%-48.8%; right range, 47.1%-48.7%). There were no differences between the left and right hemisphere relative sizes among the ZKD specimens, while there was a difference for the modern humans ( Table 3). The relative sizes of the left and right hemispheres of the ZKD and modern humans with a 95% individual prediction interval are shown in Figure 3. The variation range of the hemisphere relative sizes of the ZKD specimens were larger than for the modern humans in our sample.

Figure 3
The left and right hemisphere relative sizes of the ZKD and modern humans (95% individual prediction interval).

Discussion and conclusions
The overall increase in brain size is a critical feature of human evolution. The size of the brain can change in terms of both absolute brain volume and relative brain size. Absolute brain volume is the actual size of a brain, usually expressed by a measure of weight and volume. Relative brain size is the relationship between size of the brain and the body. In the present study, using the 3D laser surface technique, we studied the brain asymmetry of the ZKD Homo erectus, and showed that the brain asymmetry is related to relative size rather than absolute volume. The geological age of the ZKD specimens in our study is 40-80 Ma [17]. ZKD specimens retained a cranial capacity between 915-1225 cm 3 (Table 1), which is in the range of Homo erectus from 600-1251 cm 3 [26], and lower than the average of modern humans at 1390 cm 3 [27]. The cranial capacity of modern humans ranges from 1140-1600 cm 3 [27], and although some individuals have small brain size close to Homo erectus, the morphological features of the modern humans are different from the early fossil humans. The Homo erectus has a low brain height, a low position of the greatest breadth, flat and narrow frontal lobes, depressed parietal lobes, a strong posterior projection of the occipital lobes, and elongated and quite separated cerebella lobes. Compared with the Homo erectus, the height of brain of the modern humans are tall, the frontal lobes are full and wide, the parietal lobes are bulgy, the occipital lobes display a steady reduction, and the cerebella lobes become tighter, more globular, and approach the midline [28]. During human evolution, with the enlargement of brain size, brain morphological features experienced significant changes, but the ratio of the two hemispheres volumes was not modified, and neither the ZKD nor the modern specimens display clear cerebral asymmetry. We found no difference in the relative size of the hemispheres in the ZKD specimens, but did find a difference in modern humans. The variation range and individual differences in the ZKD specimens were greater than that in the modern humans. During the evolution from the ZKD Homo erectus to the modern humans there have been complex changes in the structural asymmetries between the two hemispheres, and this expansion of various brain structures may be functional. The organization of the brain structure is more important than the brain size [29]. Brain asymmetry is not a unique characteristic of humans, with brain asymmetries also present in birds, monkeys, and apes [30]. This lateralization is thought to originate from evolutionary, hereditary, developmental, experience, and pathological factors [31].
Brain enlargement is associated with functional complexity; however, there is no strong direct relationship between brain size and intelligence among modern humans. Intelligence considered mainly influenced by the environment after birth [32]. Interestingly, the increase in cranial capacity does not appear to be a unidirectional trend, with recent studies indicating a decrease in cranial capacities by 95-165 mL for males and 74-106 mL for females in certain populations during the Holocene [33,34]. In studies examining the fossil endocasts of Australopithecus, Homo habilis, Homo erectus, Neanderthals, and Homo sapiens, almost all brain endocasts displayed distinct cerebral asymmetry. The Broca's speech area of apes is not enlarged, while in Australopithecus africanus, Homo hablis, and Homo erectus this area can be clearly observed on the left endocast. It is possible that the speech center has emerged during the time of Australopithecus [16], as some studies have suggested that the brain size of Australopithecus is too small to have language ability. The endocast from the 2.0 Ma KNM-ER 1470 fossil displays a sulcal pattern in the left frontal lobe that is similar to modern humans. However, this does not indicate that Homo habilis used language as we know it. With an enlarged brain, brain function became more complex, which likely provided the ability for humans to communicate with each other using language. It remains unclear whether the Broca's area of Homo erectus was associated with language-like abilities or tool making. At present, there are two anatomical indicators that Homo erectus could not speak: first, respiratory tract reconstruction suggests that Homo erectus had a short larynx and was unlikely to pronounce vowels; second, the thoracic vertebral canal and thoracic spinal nerves of Homo erectus were smaller than that of modern humans. As such, the respiratory system of Homo erectus was not strong enough to control breathing, which is essential for speech [35].
In summary, based on endocast studies, we found that the ZKD and modern specimens exhibit similar symmetries for hemisphere volumes and surface areas, but markedly differ with respect to hemisphere relative sizes. Asymmetries were found for hemisphere relative sizes of the modern humans, but not in the ZKD specimens. The anatomical structures of the ZKD brains likely differ from the modern humans, while that ZKD Homo erectus had no ability to communicate with each other in the form of language. During human evolution, brain anatomical asymmetries experienced marked changes, and certain human abilities including language, intelligence, and cognition are likely related to these asymmetrical changes.