Uncanny Valley Hypothesis Research Articles

Perceptual discrimination difficulty and familiarity in the Uncanny Valley: more like a "Happy Valley".

The Uncanny Valley Hypothesis (UVH) predicts that greater difficulty perceptually discriminating between categorically ambiguous human and humanlike characters (e.g., highly realistic robot) evokes negatively valenced (i.e., uncanny) affect. An ABX perceptual discrimination task and signal detection analysis was used to examine the profile of perceptual discrimination (PD) difficulty along the UVH' dimension of human likeness (DHL). This was represented using avatar-to-human morph continua. Rejecting the implicitly assumed profile of PD difficulty underlying the UVH' prediction, Experiment 1 showed that PD difficulty was reduced for categorically ambiguous faces but, notably, enhanced for human faces. Rejecting the UVH' predicted relationship between PD difficulty and negative affect (assessed in terms of the UVH' familiarity dimension), Experiment 2 demonstrated that greater PD difficulty correlates with more positively valenced affect. Critically, this effect was strongest for the ambiguous faces, suggesting a correlative relationship between PD difficulty and feelings of familiarity more consistent with the metaphor happy valley. This relationship is also consistent with a fluency amplification instead of the hitherto proposed hedonic fluency account of affect along the DHL. Experiment 3 found no evidence that the asymmetry in the profile of PD along the DHL is attributable to a differential processing bias (cf. other-race effect), i.e., processing avatars at a category level but human faces at an individual level. In conclusion, the present data for static faces show clear effects that, however, strongly challenge the UVH' implicitly assumed profile of PD difficulty along the DHL and the predicted relationship between this and feelings of familiarity.

How design characteristics of robots determine evaluation and uncanny valley related responses

Against the background of the uncanny valley hypothesis (Mori, 1970) and its conceptual shortcomings the present study aims at identifying design characteristics which determine the evaluation of robots. We conducted a web-based survey (N=151) with standardized pictures of 40 robots which were evaluated. A cluster analysis revealed six clusters of robots. We were able to identify different groups of robots which were rated significantly different on six dimensions. Interestingly, in two clusters robots were rated both low in human-likeness and low in “mechanical” indicating that in future research these adjectives should be used separately. The robots in those clusters with a high mean concerning “Threatening” are almost all bi-pedal robots. In contrast to the humanoid robots (cluster 5), the android robots (clusters 3 and 4) were rated higher on “Familiarity”, and on “Likability” although all three clusters (3, 4, and 5) were rated as “Threatening” which contrasts the uncanny valley hypothesis. Moreover, the data in the present study could not be explained by a cubic function as would be suggested by the graph proposed by Mori, but rather by linear or quadratic relationships.

Empirical evaluation of the uncanny valley hypothesis fails to confirm the predicted effect of motion

The uncanny valley hypothesis states that the acceptability of an artificial character will not increase linearly in relation to its likeness to human form. Instead, after an initial rise in acceptability there will be a pronounced decrease when the character is similar, but not identical to human form (Mori, 1970/2012). Moreover, it has been claimed but never directly tested that movement would accentuate this dip and make moving characters less acceptable. We used a number of full-body animated computer characters along with a parametrically defined motion set to examine the effect of motion quality on the uncanny valley. We found that improving the motion quality systematically improved the acceptability of the characters. In particular, the character classified in the deepest location of the uncanny valley became more acceptable when it was animated. Our results showed that although an uncanny valley was found for static characters, the deepening of the valley with motion, originally predicted by Mori (1970/2012), was not obtained.

The Uncanny in the Wild. Analysis of Unscripted Human–Android Interaction in the Field

Against the background of the uncanny valley hypothesis we investigated how people react towards an android robot in a natural environment dependent on the behavior displayed by the robot (still vs. moving) in a quasi-experimental observational field study. We present data on unscripted interactions between humans and the android robot “Geminoid HI-1” in an Austrian public cafe and subsequent interviews. Data were analyzed with regard to the participants’ nonverbal behavior (e.g. attention paid to the robot, proximity). We found that participants’ behavior towards the android robot as well as their interview answers were influenced by the behavior the robot displayed. In addition, we found huge inter-individual differences in the participants’ behavior. Implications for the uncanny valley and research on social human–robot interactions are discussed.

Perceptual and Category Processing of the Uncanny Valley Hypothesis' Dimension of Human Likeness: Some Methodological Issues

Mori's Uncanny Valley Hypothesis1,2 proposes that the perception of humanlike characters such as robots and, by extension, avatars (computer-generated characters) can evoke negative or positive affect (valence) depending on the object's degree of visual and behavioral realism along a dimension of human likeness (DHL) (Figure 1). But studies of affective valence of subjective responses to variously realistic non-human characters have produced inconsistent findings 3, 4, 5, 6. One of a number of reasons for this is that human likeness is not perceived as the hypothesis assumes. While the DHL can be defined following Mori's description as a smooth linear change in the degree of physical humanlike similarity, subjective perception of objects along the DHL can be understood in terms of the psychological effects of categorical perception (CP) 7. Further behavioral and neuroimaging investigations of category processing and CP along the DHL and of the potential influence of the dimension's underlying category structure on affective experience are needed. This protocol therefore focuses on the DHL and allows examination of CP. Based on the protocol presented in the video as an example, issues surrounding the methodology in the protocol and the use in "uncanny" research of stimuli drawn from morph continua to represent the DHL are discussed in the article that accompanies the video. The use of neuroimaging and morph stimuli to represent the DHL in order to disentangle brain regions neurally responsive to physical human-like similarity from those responsive to category change and category processing is briefly illustrated.

Perceptual and Category Processing of the Uncanny Valley Hypothesis' Dimension of Human Likeness: Some Methodological Issues

Mori's Uncanny Valley Hypothesis1,2 proposes that the perception of humanlike characters such as robots and, by extension, avatars (computer-generated characters) can evoke negative or positive affect (valence) depending on the object's degree of visual and behavioral realism along a dimension of human likeness (DHL) (Figure 1). But studies of affective valence of subjective responses to variously realistic non-human characters have produced inconsistent findings 3, 4, 5, 6. One of a number of reasons for this is that human likeness is not perceived as the hypothesis assumes. While the DHL can be defined following Mori's description as a smooth linear change in the degree of physical humanlike similarity, subjective perception of objects along the DHL can be understood in terms of the psychological effects of categorical perception (CP) 7. Further behavioral and neuroimaging investigations of category processing and CP along the DHL and of the potential influence of the dimension's underlying category structure on affective experience are needed. This protocol therefore focuses on the DHL and allows examination of CP. Based on the protocol presented in the video as an example, issues surrounding the methodology in the protocol and the use in "uncanny" research of stimuli drawn from morph continua to represent the DHL are discussed in the article that accompanies the video. The use of neuroimaging and morph stimuli to represent the DHL in order to disentangle brain regions neurally responsive to physical human-like similarity from those responsive to category change and category processing is briefly illustrated.

June 2013: This Month in JoVE

Here's a look at what's coming up in the June 2013 issue of JoVE: The Journal of Visualized Experiments. Since JoVE first launched in 2006, we have periodically expanded our scope-adding sections devoted to Neuroscience, Immunology & Infection, Clinical & Translational Medicine, Bioengineering, Applied Physics, and Chemistry. We are now pleased to introduce JoVE Behavior, which explores the science behind various behavioral aspects of humans and animals. These include addiction, learning & memory, social interaction, and emotion. One article in this new section addresses a phenomenon where things become more likeable if they look more like humans. For example, this robotic vacuum cleaner is less appealing than these beloved robots. However, when they are almost (but not quite) humanlike, robots suddenly become repulsive before becoming likeable again. This sudden drop is the focus of the Uncanny Valley Hypothesis, which Cheetham and Jancke examine by generating a continuum of images morphing from human faces to avatars (graphical representations). Then, using volunteers to judge which images represent humans or avatars, they define category boundaries and use neuroimaging to identify brain regions sensitive to category change. Another article in JoVE Behavior features magnetoencephalography, or MEG, which can record brain activation in real time. Combining this technique with high-resolution functional magnetic resonance imaging (fMRI), Balderston, et al. model the sources of the MEG signal to show neural activation in regions of interest in the brain during learning tasks. In behavioral neuroscience, human learning and speech can be modeled with birdsong. Ruijssevelt, et al. present a method for performing blood-oxygen-level-dependent fMRI (BOLD fMRI) on zebra finches. Although fMRI is especially challenging in these tiny animals, this method allows successful imaging of neural substrates of auditory stimuli in the zebra finch brain. A lot of behavioral research is done in rats using the Morris water maze, in which the rat must find an escape platform in a pool of water. However, unlike the classic sewer rat, mice evolved in dry areas and aren't always spectacular swimmers like rats. Therefore, Deacon uses the Oxford paddling pool, which has shallow water and is less stressful than the classic water maze. It also has the escape hatch on the side of the pool, which is a more natural escape route. The paddling pool model can be adapted to other behavioral testing setups, such the Y-maze. These shallow-water tests can provide more ideal systems for testing learning and memory in mice. By now, you're probably behaving very excitedly about our new Behavior section! But don't forget the other subjects covered in JoVE. In one video article, Di Pisquale, et al. make embryoid bodies from human induced pluripotent stem cells (iPSCs), and use them as a source of cardiomyocytes that express cardiac-specific markers. In JoVE Bioengineering, we focus on the placenta, which is the interface between a developing fetus and the mother. The placenta separates the fetal and maternal circulation systems, but some drugs and chemicals can cross this barrier. Grafmuller, et al. demonstrate a method for studying the transport of compounds in an actual placenta using an ex vivo perfusion model. You've just had a preview of a few highlights for the month of June. Visit the website to see the full-length articles, plus many more, in JoVE: The Journal of Visualized Experiments.

Category Processing and the human likeness dimension of the Uncanny Valley Hypothesis: Eye-Tracking Data.

The Uncanny Valley Hypothesis (Mori, 1970) predicts that perceptual difficulty distinguishing between a humanlike object (e.g., lifelike prosthetic hand, mannequin) and its human counterpart evokes negative affect. Research has focused on affect, with inconsistent results, but little is known about how objects along the hypothesis’ dimension of human likeness (DHL) are actually perceived. This study used morph continua based on human and highly realistic computer-generated (avatar) faces to represent the DHL. Total number and dwell time of fixations to facial features were recorded while participants (N = 60) judged avatar versus human category membership of the faces in a forced choice categorization task. Fixation and dwell data confirmed the face feature hierarchy (eyes, nose, and mouth in this order of importance) across the DHL. There were no further findings for fixation. A change in the relative importance of these features was found for dwell time, with greater preferential processing of eyes and mouth of categorically ambiguous faces compared with unambiguous avatar faces. There were no significant differences between ambiguous and human faces. These findings applied for men and women, though women generally dwelled more on the eyes to the disadvantage of the nose. The mouth was unaffected by gender. In summary, the relative importance of facial features changed on the DHL’s non-human side as a function of categorization ambiguity. This change was indicated by dwell time only, suggesting greater depth of perceptual processing of the eyes and mouth of ambiguous faces compared with these features in unambiguous avatar faces.

The human likeness dimension of the “uncanny valley hypothesis”: behavioral and functional MRI findings

The uncanny valley hypothesis (Mori, 1970) predicts differential experience of negative and positive affect as a function of human likeness. Affective experience of humanlike robots and computer-generated characters (avatars) dominates “uncanny” research, but findings are inconsistent. Importantly, it is unknown how objects are actually perceived along the hypothesis’ dimension of human likeness (DOH), defined in terms of human physical similarity. To examine whether the DOH can also be defined in terms of effects of categorical perception (CP), stimuli from morph continua with controlled differences in physical human likeness between avatar and human faces as endpoints were presented. Two behavioral studies found a sharp category boundary along the DOH and enhanced visual discrimination (i.e., CP) of fine-grained differences between pairs of faces at the category boundary. Discrimination was better for face pairs presenting category change in the human-to-avatar than avatar-to-human direction along the DOH. To investigate brain representation of physical change and category change along the DOH, an event-related functional magnetic resonance imaging study used the same stimuli in a pair-repetition priming paradigm. Bilateral mid-fusiform areas and a different right mid-fusiform area were sensitive to physical change within the human and avatar categories, respectively, whereas entirely different regions were sensitive to the human-to-avatar (caudate head, putamen, thalamus, red nucleus) and avatar-to-human (hippocampus, amygdala, mid-insula) direction of category change. These findings show that Mori’s DOH definition does not reflect subjective perception of human likeness and suggest that future “uncanny” studies consider CP and the DOH’s category structure in guiding experience of non-human objects.

The Perception of Humanness from the Movements of Synthetic Agents

As technology develops, social robots and synthetic avatars might begin to play more of a role in our lives. An influential theory of the perception of synthetic agents states that as they begin to look and move in a more human-like way, they elicit profound discomfort in the observer--an effect known as the Uncanny Valley. Previous attempts to examine the existence of the Uncanny Valley have not adequately manipulated movement parameters that contribute to perceptions of the humanness or eeriness. Here we parametrically manipulated three different kinematic features of two walking avatars and found that, contrary to the Uncanny Valley hypothesis, ratings of the humanness, familiarity, and eeriness of these avatars changed monotonically. Our results indicate that, when a full gradient of motion parameter changes is examined, ratings of synthetic agents by human observers do not show an Uncanny Valley.

Interpersonal motor coordination

Here, we propose that bidirectionality in implicit motor coordination between humanoid robots and humans could enhance the social competence of human–robot interactions. We first detail some questions pertaining to human–robot interactions, introducing the Uncanny Valley hypothesis. After introducing a framework pertinent for the understanding of natural social interactions, motor resonance, we examine two behaviors derived from this framework: motor coordination, investigated in and informative about human–human interaction, and motor interference, which demonstrate the relevance of the motor resonance framework to describe human perception of humanoid robots. These two lines of investigation are then put together to “close the loop” by proposing to implement a key feature of motor coordination, bidirectionality, in robots’ behavior. Finally, we discuss the feasibility of implementing motor coordination between humanoid robots and humans, and the consequences of this implementation in enhancing the social competence of robots interacting with humans. Keywords: interpersonal interaction, motor resonance, motor coordination, motor interference, social robotics, anthropomorphism