Gamers’ involvement detection from EEG data with cGAAM – A method for feature selection for clustering
Study at a glance
- Area
- Other research
- Study type
- Observational study
- Authors
- Rejer I, Twardochleb M
- Journal
- Expert Systems with Applications
- Year
- 2018
- Population
- Adults
- Protocol
- EEG feature selection and clustering
- Equipment
- BrainMaster Discovery
Key findings
- Frontal theta and left-frontal delta best separated engagement levels
- The new method gave 50% more compact clusters than standard feature selection
A study identifying EEG patterns that mark how engaged a player is during a video game, combining raw EEG with experts' labels and a new genetic-algorithm method for choosing features while clustering. Frontal theta in both hemispheres and left-frontal delta best distinguished levels of engagement.