Elsevier

Cognition

Volume 153, August 2016, Pages 1-5
Cognition

Original Articles
Similar impressions of humanness for human and artificial singing voices in autism spectrum disorders

https://doi.org/10.1016/j.cognition.2016.04.004Get rights and content

Abstract

People with autism spectrum disorder (ASD) exhibit impairments in the perception of and orientation to social information related to humans, and some people with ASD show higher preference toward human-like robots than other humans. We speculated that this behavioural bias in people with ASD is caused by a weakness in their perception of humanness. To address this issue, we investigated whether people with ASD detect a subtle difference between the same song sung by human and artificial voices even when the lyrics, melody and rhythm are identical. People without ASD answered that the songs sung by a human voice evoked more impressions of humanness (human-likeness, animateness, naturalness, emotion) and more positive feelings (warmth, familiarity, comfort) than those sung by an artificial voice. In contrast, people with ASD had similar impressions of humanness and positive feelings for the songs sung by the human and artificial voices. The evaluations of musical characteristics (complexity, regularity, brightness) did not differ between people with and without ASD. These results suggest that people with ASD are weak in their ability to perceive psychological attributes of humanness.

Introduction

Autism spectrum disorder (ASD) is a developmental disorder characterized by difficulties in social cognition and communication. People with ASD exhibit impairments in orientating to social stimuli, such as human faces (Jemel et al., 2006, Nakano et al., 2010) and speech (Alcantara et al., 2004, Dawson et al., 1998). Instead, some of them show a strong affinity to robots and computers (Diehl, Schmitt, Villano, & Crowell, 2012). A previous study also reported that children with ASD automatically imitated the behaviour of a robot much better than they did that of a human (Pierno, Mari, Lusher, & Castiello, 2008). With regards to this preference for artificial objects rather than humans in autism, Baron-Cohen suggested that individuals with autism are attracted to systems of low variance (such as machines) and less sensitive to systems where there is maximal variance as is the case with human behaviour (Baron-Cohen, 2006).

In general, a human feels a stronger affinity toward another human than they do toward artificial objects. In particular, the subtle differences between a real human and a human-like object evoke an impression of a substantial qualitative difference between them that is accompanied by a negative feeling toward the human-like object. This is a well-known phenomenon known as the “uncanny valley phenomenon” (Mori, 1970). Research verifying the uncanny valley phenomenon clearly demonstrates that people perceive not only a physical difference but also a qualitative difference between humans and artificial objects. Haslam proposed that people have a unique sense of humanness, a special sense of “human nature,” that involves emotion, warmth, and cognitive flexibility and is opposed to “mechanistic” dehumanization (Haslam, 2006). Thus, the dissociation between physical and psychological attributes of humanness evokes a feeling of repulsion as manifested by the “uncanny valley phenomenon”. In contrast, the individuals with ASD do not show a preference bias toward human (Nakano et al., 2010) nor a repulsion toward the artificial objects rather than real humans (Diehl et al., 2012). We speculated that individuals with ASD face difficulties in perceiving the psychological aspects of humanness from human-related information, and thus they were unable to detect subtle differences between human and human-like objects. Therefore, they did not demonstrate any repulsion toward the artificial objects regardless of its degree of human resemblance. To test this hypothesis, the present study examined the impressions of humanness and feeling experienced by people with ASD upon hearing the same song sung by a real human voice and an artificial voice. Although many previous studies have explored the perception of humanness using humanoid robots or computer-generated 3D animation (such as Broadbent et al., 2013, Gray and Wegner, 2012, Thompson et al., 2011), it is technically difficult to create artificial objects with the appearance and movements resembling those of an actual human. To address this problem, we focused on creating an artificial singing voice because the technology for synthesizing an artificial voice has advanced to a point that it can be used in daily life. Our previous study used the same auditory stimuli and demonstrated that a human singing voice evoked a sense of humanness and positive emotion much more strongly than an artificial singing voice did in typically developing adults, even though the lyrics, melody, and rhythm were identical (Tamura, Kuriki, & Nakano, 2015). We expected that if people with ASD had a weakness in their sense of humanness, they would not detect a subtle difference between the stimuli and would have similar impressions of humanness and positive feelings for the human and artificial singing voices.

Section snippets

Participants

Fourteen adults with ASD (10 male, 4 female; mean age, 27.6 years; range, 20–39 years) and fourteen adults without ASD (10 male, 4 female; mean age, 28.9 years; range, 21–37 years) participated in this study. The two groups were matched for sex and approximate age. ASD diagnoses were based on the clinical judgment of medical specialists according to the DSM-V criteria. All ASD participants were high functioning, with their full-scale IQ (Wechsler Adult Intelligence Scale, 3rd edition, WAIS-III)

Results

First, we compared the impressions of the songs sung by either a human or artificial voice for each questionnaire item in each group. In line with our previous study (Tamura et al., 2015), the control participants evaluated the songs sung by the human voice more highly than those sung by the artificial voice on the questionnaire items related to human-likeness and positive feelings but not on the items concerning musical characteristics (mean scores shown in Table 1). Compared with the control

Discussion

The present study confirmed that the human singing voice evokes a sense of humanness and positive feelings to a much greater degree than an artificial singing voice in typically developing adults. In contrast, while adults with ASD showed slightly greater impressions of humanness and positive feelings toward the human singing voice than toward the artificial singing voices, the difference in responses to human and artificial singing voices was much smaller in adults with ASD as compared with

Acknowledgements

This work was supported by the Grant-in-Aid for Scientific Research on Innovative Areas 251195040 “Constructive Developmental Science” from the Ministry of Education, Culture, Sports, Science and Technology, Japan to T.N.

References (26)

  • E. Broadbent et al.

    Robots with display screens: A robot with a more humanlike face display is perceived to have more mind and a better personality

    PLoS ONE

    (2013)
  • G. Dawson et al.

    Children with autism fail to orient to naturally occurring social stimuli

    Journal of Autism and Developmental Disorders

    (1998)
  • L. Gebauer et al.

    Intact brain processing of musical emotions in autism spectrum disorder, but more cognitive load and arousal in happy vs. sad music

    Frontiers in Neuroscience

    (2014)
  • Cited by (0)

    1

    These authors equally contributed to this study.

    View full text