Humanoid Avatars in Individualised Virtual Reality for Supporting Mental Health: Self-Compassion & Cost Minimisation

Abstract

Abstract Avatars are employed within Virtual Reality environments to interact with participants in training and therapeutic settings. Avatars have been introduced with varying quality mainly due to historical technological limitations with respect to computer graphics and hardware. The VR community may assume that with the development of higher quality and therefore ever more lifelike avatars, implies an advance in avatar performance, yielding ever better human interactions. But is that true? Until now this has not yet been investigated fully. The authors of this paper recognised how contextual evidence suggests that ‘lifelike’ avatars may potentially hinder VR therapeutic outcomes. As a result, an individualised virtual reality self-compassion therapy was replicated but replaced all the ‘lifelike’ avatars with cartoony humanoid avatars and re-measured user experience as well as therapeutic benefit. The results show a dramatic increase in user experience on areas pertaining to feelings of dependability, which demonstrates a reduction in ‘eeriness’ likely caused by the Uncanny Valley phenomenon previously invoked by the too realistic ‘lifelike’ avatars. The data therefore suggests that humanoid avatars are an affordable substitute with better interaction.