Monkeys and humans share a strange reaction to eerily realistic bodies, according to a new study. The research reveals that monkeys exhibit signs of the uncanny valley, a phenomenon where people feel unease around near-human robots and lifelike digital characters, when they observe another monkey's body. This finding provides valuable insights into how the human brain processes social signals and offers a controllable digital monkey model for brain scientists to study these processes.
The study's senior author, Professor Martin Giese, emphasizes the importance of understanding the neural mechanisms of social perception for various clinical applications. To achieve this, the researchers developed a digital monkey avatar named MacAction, which can be manipulated to isolate specific cues that the brain reacts to. This avatar is a detailed three-dimensional model with 86 separate joints, allowing for realistic movement.
One of the challenges in creating a realistic monkey avatar is motion capture. Reflective markers, commonly used in film and sports, are ineffective for monkeys due to their tendency to remove or disrupt the markers. Traditional motion capture methods require extensive camera setups and manual frame labeling, making it costly and time-consuming. However, MacAction revolutionizes this process by using a neural network that learns a monkey's pose from a few hand-labeled frames and fills in the gaps, enabling the reconstruction of entire actions from as few as two labeled frames per second.
The study involved eight rhesus macaques at KU Leuven in Belgium. By tracking the monkeys' eye movements, researchers observed that the animals responded similarly to real and animated monkey footage. This response indicates that the animation was realistic enough to be considered a valid substitute for a real monkey in experiments.
To test the uncanny valley effect, the researchers gradually reduced the realism of the avatar, presenting four versions of the walk and turning gesture. The results showed a sharp drop in attention for the versions in the middle range of realism, known as the uncanny valley. This finding aligns with the concept proposed by robotics researcher Masahiro Mori in 1970, suggesting that people experience discomfort when encountering something that closely resembles reality but is not quite perfect.
Interestingly, the uncanny valley effect was not solely due to differences in brightness, contrast, or the number of edges in the images. The study also found that this effect occurred in still images, ruling out motion as the primary cause. Additionally, the dip in attention varied slightly between individual monkeys, similar to the way it occurs with faces.
The implications of this study extend beyond the realm of animation. It suggests that primates perceive social signals in a manner similar to humans, indicating that the mechanisms of social perception are shared between different primate species. This finding supports the use of monkeys as models for studying human social vision.
Furthermore, the development of realistic virtual monkeys for research is gaining momentum. Independent groups are creating digital macaque heads to study social behavior, but caution is advised to avoid the uncanny valley. The study's method, which is cost-effective and accessible to smaller labs, has the potential to enhance our understanding of social perception and has applications in various fields, including digital humans in games and films.
In conclusion, this research highlights the shared reaction of monkeys and humans to eerily realistic bodies, shedding light on the brain's processing of social signals. With the development of advanced digital monkey models, scientists can now explore the neural mechanisms behind this phenomenon, leading to potential advancements in various areas of study.
