HBO’s science fiction series Westworld is based in an amusement park populated with robotic creatures whose advanced construction and programming make them indistinguishable from humans. The realistic look, feel, and movement of these robotic park hosts allows the human guests to experience elaborately programmed scenarios as if they were interacting with other humans. Despite significant functional and aesthetic advancements in the past few decades, current prostheses are rarely confused with anatomical limbs, but as we anticipate future improvements in all aspects of prosthetic design, we may imagine that as the realism of prosthetic replacements improves, their acceptance by users and observers will improve with it.
In the 1970s, a robotics professor named Masahiro Mori challenged the idea that realistic artificial limbs would inevitably lead to greater acceptance by using the analogy of climbing a mountain to describe the human response to robots with increasingly human features. At times, climbers are required to move in the opposite direction of the intended goal, such as when traversing valleys between the base and the summit. Mori proposed that after a certain point, people may exhibit less affinity for prosthetic restorations designed to look more like human limbs: “[C]limbing toward the goal of making robots appear like a human, our affinity for them increases until we come to a valley…which I call the uncanny valley.”1 That steep drop illustrates a level of discomfort, or lack of affinity, for artificial limbs that too realistically mimic human appearance. Figure 1 shows Mori’s uncanny valley in graphic form. A realistic-looking prosthetic hand may initially appear human to an observer, but “once we realize that the hand that looked real at first sight is actually artificial, we experience an eerie sensation…. When this happens, we lose our sense of affinity, and the hand becomes uncanny…. The appearance of the prosthetic hand is quite humanlike, but the level of affinity is negative, thus placing the hand near the bottom of the valley….”1 We have a long climb ahead of us before we reach the goal of providing patients with prosthetic restorations which, like those in Westworld, look, feel, and move exactly like the anatomical body part they are designed to replace. Mori proposed that an understanding of the negative perceptions of more realistic prosthetic devices should influence prosthetic design, and even suggested that designers attempt “a moderate degree of human likeness and a considerable sense of affinity,” rather than a high degree of human likeness that could lead to low affinity.1 This article reviews several recent studies that have investigated Mori’s uncanny valley thought experiment and offers new perspectives on prosthetic design based on that research.
Rating Eeriness and Affinity
Research conducted by Poliakoff et al. at the University of Manchester, England, investigated the responses of 43 subjects when they viewed still images of 22 different hands and ranked their eeriness and human-likeness. Eeriness “was defined as: mysterious, strange, or unexpected as to send a chill up the spine….”2 The hands were categorized as mechanical, prosthetic, and human, and the realism of the prosthetic hands varied significantly. As the researchers suspected, “In keeping with the notion of an uncanny valley, prosthetic hands that were of intermediate human-likeness were given the highest ratings of eeriness.”2 However, a more complex pattern of responses was found within each category of hands than in Mori’s research, since “ratings of eeriness reduced as human-likeness increased.”2 Poliakoff et al. expanded this research and published a second article in 2018. In this study, the researchers included a variety of prosthetic hands that they had characterized beforehand as more and less human-like. The viewing protocol involved flipping the images vertically (to alternate between first and third person perspective) and horizontally (left and right) to test whether viewing perspective affected ratings of eeriness. The time the participants required to rate each image was measured. This was done to test whether more time would be required to categorize the hand as prosthetic or human, and whether confusion about which category of hand was being viewed would influence the perception of eeriness as the human-likeness of prosthetic hands increased. The researchers report that “static images of less human-like prosthetic hands were consistently rated as eerie relative to more human-like prosthetic, real human, and mechanical hands.”3 The viewing perspective did not influence eeriness ratings. Difficulty categorizing the hands did not result in a higher eeriness rating, since although participants took more time to rate prosthetic hands with more human-like features, they rated them as less eerie than less realistic prosthetic hands. This research supports the uncanny valley hypothesis in one sense, since non-human hands were rated as more eerie than human ones. However, Mori’s suggestion that realism should be avoided when designing prosthetic hands was not supported by this research, since “more realistic prosthetic hands appeared less eerie than less-realistic hands….”3
In 2015, Sansoni et al. published the results of research that investigated participants’ level of attraction toward and preference for realistic or non-realistic prostheses. They also examined whether attraction is related to gender, culture (comparing Italian and British), and the participant having an amputation. Participants viewed images of upper- and lower-limb prostheses and were asked whether they liked the device and whether it looked like a real limb. The results did not support Mori’s uncanny valley hypothesis: “[T]he highest level of attraction was attributed to very realistic prosthetic devices…the more realistic devices are, the more people appreciate them.”4 Rather than a distinct valley in the attractiveness (or familiarity) graph, indicating a high level of eeriness for prosthetic limbs, their results “show a growing attraction for increasingly human-like devices, with a small dip (instead of a ‘valley’) for devices resembling a limb but showing creepy artificial patterns” (Figure 2).4 Additionally, results showed that female participants preferred devices with a high level of realism and men preferred those with robotic features or other clear artificial patterns. Only minor differences were found between Italian and British participants and those with or without amputations.
Relevance to Clinical Practice
Clinical decision making depends on many qualitative and subjective factors, which can significantly challenge the development of research protocols that provide meaningful insight into day-to-day practice. In addition, the vagaries of affinity and eeriness further complicate designing research. How can clinicians use the information in these studies about how prosthetic hands are perceived? First, we should recognize that despite their limitations, these studies provide an important baseline perspective on an issue that affects every encounter with a person following limb loss. Whether they mention this explicitly to us or not, all patients undergo some process of integrating a prosthetic limb into their self-understanding. Concerns related to how others perceive patients based on their absent limbs and their prosthetic replacements are important parts of this process. Any information about how prostheses are perceived by the general population can help us understand our patients better and provide more relevant input when recommending specific treatment options and design features. Second, while visual stimuli is only one way a prosthesis is perceived, it is undoubtedly a good place to begin when investigating this phenomenon. Human beings rapidly process an incredible amount of visual information and often make decisions subconsciously based on that input. Far more people observe a patient wearing and using a prosthesis from a distance than interact with the individual in a more direct manner. They quickly make decisions about whether and how to interact (“Look at it or look away?” “Ask about it or ignore it?”) based on that visual stimuli. How these decisions are made is an intriguing area of inquiry, and at the very least we can assert that how a prosthesis looks is foundational to how it is perceived by most people who encounter someone wearing one.
What Else Don’t We Know?
Future studies investigating other factors that affect affinity and eeriness could improve our understanding of patients. Prosthetists are aware that even though the most artistic prosthetic restoration may be difficult to detect as artificial solely by visual inspection, they can often be easily identified as such due to unnatural patterns of movement associated with their use. Mori suggested that motion “changes the shape of the uncanny valley graph by amplifying the peaks and valleys.”1 Motion can create a greater impression of realism but evokes a stronger negative response when it is perceived as unnatural. More research on how prosthetic hands with and without multiple grip styles and individually articulating fingers are perceived would test this assumption. Most prostheses are intended for functional use, and it would be valuable to know how various upper- and lower-limb prosthetic designs are perceived as they are being worn and used. Since lower-limb prostheses more effectively restore lower-limb function and are more easily disguised by clothing and footwear, it seems likely that upper-limb designs evoke stronger responses. But that assumption is worth testing more robustly. Sansoni et al. suggest that length of time since amputation should be considered in future studies and speculate that individuals “who have fully accepted their loss of limb(s)” will reject more realistic devices, “whereas attraction to realistic prostheses would be found in amputees who are in the early stages after amputation and so have not fully accepted their loss of limb(s).”4 However, they acknowledge that their study does not provide evidence supporting this, and plan future interview-based investigation. There is much to learn about how individuals with limb loss perceive themselves and their prostheses, and what they believe about how others perceive them. Practitioners can use this type of research to develop strategies to help patients make specific design decisions.
There is considerable variation in the preferences of individual patients related to the aesthetic design of upper- and lower-limb prostheses. It is interesting to consider how attitudes toward different prosthetic designs within the broader population may affect these choices. In general, it seems that people in the United States are now more accepting of disabilities that are visually apparent than they were in 1970. The prevalence of stories in the popular media about high-tech prosthetic devices may increase the appreciation for devices whose technological features are visually apparent. This may be contributing to the widespread interest in 3D-printed upper-limb devices, which are presented as high tech because of the method of fabrication and not their realism, function, or technological sophistication. These devices are appealing for reasons other than those which we have assumed are most important to patients, such as functionality, durability, and reliability. It can be frustrating to hear news stories that seem to exaggerate the limitations of the advanced designs we regularly provide, particularly when the rudimentary devices offered as alternatives are presented as more advanced, but it is important to understand how all these options are perceived. Poliakoff et al. suggest in their 2018 article that negative reactions, such as eerie feelings, are caused by a “perceptual mismatch”—features or cues that are inconsistent with what was expected.3 The eerie response may be caused by something that initially appears real but is found to be artificial after closer inspection. Although Mori’s hypothesis has not been confirmed by the research reviewed in this article, what we’ve learned from the 3D-printing experience can give us important information about prosthetic design. Perhaps in our climb toward Westworld we are experiencing a valley of sorts. In moving toward the goal of prosthetic designs that look, move, and feel as realistic as possible, we must be aware of the preferences of individual patients and trends within the population as a whole. Accurately communicating the capabilities and features of various designs and helping patients determine their priorities may reduce the mismatch between expectations and experience and increase their satisfaction with their choices.
John T. Brinkmann, MA, CPO/L, FAAOP(D), is an assistant professor at Northwestern University Prosthetics-Orthotics Center. He has more than 20 years of experience treating a wide variety of patients.
1. Mori, M., K. F. MacDorman, and N. Kageki. 2012. The uncanny valley [from the field]. IEEE Robotics & Automation Magazine 19(2):98-100.
2. Poliakoff, E., N. Beach, R. Best, T. Howard, and E. Gowen. 2013. Can looking at a hand make your skin crawl? Peering into the uncanny valley for hands. Perception 42(9):998-1000.
3. Poliakoff, E., S. O’Kane, O. Carefoot, P. Kyberd, and E. Gowen. 2018. Investigating the uncanny valley for prosthetic hands. Prosthetics and Orthotics International 42(1):21-7.
4. Sansoni, S., A. Wodehouse, A. K. McFadyen, and A. Buis. 2015. The aesthetic appeal of prosthetic limbs and the uncanny valley: The role of personal characteristics in attraction. International Journal of Design 1;9(1).