Comparative analysis of cognitive tasks for modeling mental workload with electroencephalogram

Area of Science:

• Neuroscience
• Cognitive Science
• Human-Computer Interaction

Background:

• Cognitive workload estimation is crucial for understanding human performance and designing adaptive systems.
• Electroencephalogram (EEG) studies have explored various cognitive tasks to measure workload.
• Characterizing these tasks is essential for developing reliable workload classification models.

Purpose of the Study:

• To comparatively analyze the n-back and mental arithmetic tasks for their effectiveness in manipulating cognitive workload.
• To evaluate the consistency, robustness, and efficiency of these tasks in short-term training for workload classification.
• To inform the selection of optimal training tasks for cognitive workload modeling.

Main Methods:

• Experiments were conducted with 7 healthy subjects using Emotiv EPOC EEG system.
• Comparative analysis of the n-back task and the mental arithmetic task.
• Assessment of task consistency, robustness, and efficiency in workload manipulation.

Main Results:

• The mental arithmetic task demonstrated higher consistency and robustness in differentiating high and low cognitive workload levels.
• The mental arithmetic task required less training time for effective workload manipulation.
• No significant task adaptation was observed in users with repeated use of the mental arithmetic task.

Conclusions:

• The mental arithmetic task is a more suitable choice for training cognitive workload classification models due to its consistency and robustness.
• The study quantifies the quality and efficiency of workload modeling based on training task selection.
• Findings provide valuable insights for optimizing the development of EEG-based cognitive workload estimation systems.