Cognitive load recognition in simulated flight missions: an EEG study.

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Aerospace Engineering

Background:

• Cognitive load recognition (CLR) using electroencephalography (EEG) is advancing.
• Existing methods use simple tasks, limiting real-world applicability.
• Simulated flight missions offer a more operational environment for CLR research.

Purpose of the Study:

• To investigate temporal dynamics of cognitive load during simulated flight missions.
• To assess the effectiveness of deep convolutional neural network (CNN) models for CLR in complex tasks.
• To explore the impact of fatigue and task progression on EEG-based CLR.

Main Methods:

• Thirty-six participants performed simulated flight tasks at low, medium, and high cognitive load levels.
• EEG data, subjective ratings, and performance metrics were collected across three sessions.
• Multiple deep CNN models were applied to raw EEG data for cognitive load classification.

Main Results:

• Significant differences were observed between resting-state and post-fatigue EEG data.
• Shallow CNN models outperformed more complex CNN architectures.
• A decline in CLR performance was noted as simulated missions progressed.

Conclusions:

• Simulated flight missions provide a valuable paradigm for studying cognitive load in operational settings.
• Shallow CNNs show promise for real-time CLR in complex, dynamic environments.
• Temporal dynamics are crucial for understanding cognitive states during prolonged tasks.