Research / Measuring player experience

Full paper · 2023

ChallengeDetect: Investigating the Potential of Detecting In-Game Challenge Experience from Physiological Measures

Peng, X., Xie, X., Huang, J., Jiang, C., Wang, H., Denisova, A., Chen, H., Tian, F., & Wang, H.

ConferenceCORE A*Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (CHI'23), 2023

Teaser figure: ChallengeDetect. 32 players played Fallout 4 while five physiological signals were recorded (ECG, skin temperature, respiration, EMG and electrodermal activity). After pre-processing, feature extraction (80 features, 24 most informative) and machine learning (13 models), the best model detected cognitive challenge with 86.8% accuracy, emotional and physical challenge with 85.1%, and decision-making challenge with 87.3%.
Open full-size image

In short

Questionnaires only capture challenge after play has ended. ChallengeDetect explores whether perceived challenge can instead be detected from players' physiological signals, moment to moment.

Key findings

  • Physiological data were collected from 32 players across three typical game scenarios.
  • Most of the machine learning methods tested detected challenge experiences with around 80% accuracy.
  • The paper discusses physiological indicators of challenge and AI-supported challenge evaluation.

Read and cite

PDF (accepted version) Publisher's version (DOI)

APA citation

Peng, X., Xie, X., Huang, J., Jiang, C., Wang, H., Denisova, A., Chen, H., Tian, F., & Wang, H. (2023, April). ChallengeDetect: Investigating the Potential of Detecting In-Game Challenge Experience from Physiological Measures. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (CHI'23).

More on this topic