Can a virtual environment elicit PLAY?

The big idea

Why PLAY matters more than we think

PLAY is the highly pleasurable, fast-paced series of back-and-forth interactions between two or more beings. In Panksepp's affective neuroscience it is one of seven primary emotion systems, and it can be generalised to social joy.

PLAY supports social and cognitive development, and reduced play is linked to conditions such as depression and ADHD. A virtual environment that reliably elicits PLAY could support both research and therapy, yet to our knowledge no prior work has used VR to target it.

What we set out to explore

• Decompose the definition of PLAY into measurable components so that elicitation can be triangulated.
• Propose an affective virtual environment design framework intended to support PLAY elicitation.
• Create and evaluate three virtual environments designed to elicit PLAY.

Three experiments

Three experiments. One shared question

Each team member designed, built and evaluated one environment. The seven parts of PLAY we measured are shown for each; a filled mark means the part was met, a half mark means it was partly met, and an empty mark means it was not met. Each environment's page gives the full results.

• 01

  Apple Interaction

  Zenande Mdludlu

  A woodland meadow where players freely throw and catch apples with an NPC squirrel whose mood-driven finite state machine controls its behaviour, such as offering, chasing and evading, and its expressions.

  2 of 7 parts met

  • Valence (partly met)
  • Arousal (partly met)
  • Dominance (not met)
  • Reciprocity (met)
  • Sociality (not met)
  • Joy (partly met)
  • PLAY (met)
• 02

  Water Balloon Fight

  Devin Marais

  A water balloon fight across a backyard pool with a boy NPC that matches the player's throwing skill, so play passes through winning, losing and fair phases.

  3 of 7 parts met

  • Valence (partly met)
  • Arousal (not met)
  • Dominance (not met)
  • Reciprocity (not met)
  • Sociality (met)
  • Joy (met)
  • PLAY (met)
• 03

  Bunny Rally

  Karla Alaro

  A bat and ball interaction with a humanoid bunny in a forest, where the bunny throws an acorn and the player hits it back.

  2 of 7 parts met

  • Valence (partly met)
  • Arousal (not met)
  • Dominance (partly met)
  • Reciprocity (met)
  • Sociality (met)
  • Joy (partly met)
  • PLAY (not met)

Behind the headset

How we tested the idea

Each environment was evaluated in a within-subjects user study with 25 participants. Every participant first completed a tutorial environment, to absorb the novelty effect of VR and learn the controls, and then the affective environment. Questionnaires were taken at three points: before the tutorial, after the tutorial, and after the affective environment. The post-tutorial reading is the control, so a change between the second and third readings is attributed to the environment itself.

PLAY was decomposed into seven measurable parts, each tied to an instrument: valence, arousal and dominance from the Self-Assessment Manikin, joy from the PANAS-X, PLAY from the Affective Neuroscience State Scale, sociality from the co-presence and social presence scales, and reciprocity from telemetry. We claimed PLAY only if every part was met.

The full method: readings, instruments, hypotheses and the analysis rules

The findings

The signal is promising.

Across all three environments, the framework produced pleasant, positive and socially engaging experiences. No environment met every part of the PLAY definition, so this is partial but promising evidence that PLAY can be elicited in VR.

Where we go next

• Measure physiological data to corroborate self-reported arousal.
• Recruit a larger, more diverse sample for statistical power and generalisability.
• Hold more VR familiarisation sessions before the main session to reduce novelty effects.
• Ensure the tutorial environment is affectively neutral.

Go deeper

Go deeper into the project

Documents produced jointly by the team. Each environment's page links the documents for that environment.

All resources