Related Experiment Video
Updated: Sep 13, 2025
13:40
Combining Computer Game-Based Behavioural Experiments With High-Density EEG and Infrared Gaze Tracking
Published on: December 16, 2010
16.8K
AMuCS: Affective multimodal Counter-Strike video game dataset
Marios Fanourakis1, Guillaume Chanel2
1Social Intelligence and MultiSensing (SIMS) lab, University of Geneva, Geneva, Switzerland. marios.fanourakis@unige.ch.
Scientific Data
|July 30, 2025
Summary
This study introduces a large dataset of 256 players of Counter-Strike: Global Offensive, collected during live events. The rich, multi-modal data is ideal for training advanced machine learning models in human-computer interaction research.
Area of Science:
- Human-Computer Interaction
- Cognitive Science
- Machine Learning
Background:
- Video games offer complex stimuli for studying human cognition and behavior.
- Existing datasets often lack participant numbers, data modalities, or are confined to lab settings.
Purpose of the Study:
- To address limitations of prior datasets by creating a large-scale, multi-modal dataset.
- To facilitate advanced machine learning model training for analyzing player experiences.
Main Methods:
- Recorded 256 participants playing Counter-Strike: Global Offensive at LAN events.
- Collected diverse data: physiological (ECG, EDA, Respiration), behavioral (facial expressions, eye-tracking, depth, pressure), and interaction (keyboard/mouse, game actions).
- Included stimulus information: gameplay video and game logs.
Main Results:
- The dataset features a high number of participants (256) and a wide array of complementary data modalities.
- Demonstrated the dataset's advantage for training machine learning models.
Conclusions:
- The large-scale, multi-modal dataset collected at live events is a valuable resource.
- This dataset enables more robust machine learning applications for understanding player behavior and cognition in gaming.
