Adaptive Filtering for Improved EEG-Based Mental Workload Assessment of Ambulant Users

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human Factors Engineering

Background:

• Mobile electroencephalography (EEG) enables mental workload assessment in real-world settings.
• Motion artifacts significantly degrade EEG signal quality and reduce assessment accuracy.
• Conventional EEG artifact removal methods are insufficient for mobile, ambulatory scenarios.

Purpose of the Study:

• To develop and validate an adaptive filter for enhancing EEG signals contaminated by motion artifacts.
• To improve the accuracy of mental workload assessment in ambulatory users.
• To investigate the neural correlates of mental workload during locomotion.

Main Methods:

• An adaptive filter utilizing an accelerometer-based referential signal was proposed.
• The adaptive filter was combined with classical EEG artifact removal techniques.
• Data were collected from 48 participants performing the Revised Multi-Attribute Task Battery-II (MATB-II) while walking/jogging or cycling.
• A random forest classifier was employed for mental workload assessment.

Main Results:

• The proposed adaptive filter significantly improved EEG signal quality in ambulatory settings.
• Mental workload classification accuracy reached up to 95% with the enhanced EEG data.
• Increased gamma activity in the parietal cortex was observed during high workload conditions in ambulant users.

Conclusions:

• The adaptive filter effectively removes motion artifacts, enabling accurate mental workload assessment in mobile users.
• The findings suggest a link between sensorimotor integration, attention, and mental workload during ambulation.
• This approach enhances the feasibility of using EEG for workload monitoring in ecological environments.