Measuring Raven's Progressive Matrices Combining Eye-Tracking Technology and Machine Learning (ML) Models.

Area of Science:

• Cognitive psychology
• Artificial intelligence
• Human-computer interaction

Background:

• Extended testing in Raven's Progressive Matrices (RPM) causes fatigue and reduced motivation, potentially impairing cognitive performance.
• Current RPM testing methods may not be optimal for sustained cognitive assessment.
• Need for efficient and objective methods to evaluate cognitive abilities.

Purpose of the Study:

• To explore artificial intelligence (AI) and eye-tracking for improving RPM testing efficiency.
• To identify key eye-tracking metrics predictive of cognitive task performance in RPM.
• To develop a more efficient RPM assessment using machine learning.

Main Methods:

• Combined eye-tracking technology with machine learning (ML) models.
• Trained ten ML models using eye-tracking metrics as features.
• Refined the period of interest and reduced the number of metrics for optimized performance.

Main Results:

• The XGBoost model showed superior performance among the ten ML models.
• Achieved accuracy, precision, and recall above 0.8.
• Effective performance using only 60% of response time and nine eye-tracking metrics.

Conclusions:

• AI and eye-tracking offer a promising approach to enhance RPM testing efficiency.
• Key eye-tracking metrics can significantly predict performance, reducing testing time.
• This methodology provides valuable insights for future cognitive assessment research.