Using passive BCI for personalization of assistive wearable devices: a proof-of-concept study.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Assistive wearable devices improve mobility but lack personalized assistance.
• Brain-Computer Interfaces (BCI) offer a potential solution for tailoring device support.
• Personalization is key to maximizing the benefits of assistive technologies.

Purpose of the Study:

• To investigate the feasibility of using passive BCI to personalize knee exoskeleton assistance.
• To identify brain activity patterns associated with different levels of exoskeleton support.
• To develop a BCI-based system for adaptive control of wearable robots.

Main Methods:

• Participants performed knee flexion-extension tasks with a powered exoskeleton at varying forces.
• Electroencephalography (EEG) recorded brain activity during exoskeleton use.
• Naive Bayes classification analyzed EEG spectral features to differentiate assistance levels.

Main Results:

• Significant changes in EEG spectral bands (increased δ/θ, decreased α/β) were observed in motor and sensory cortices with higher exoskeleton forces.
• These neural changes correlated with increased attention and motor engagement.
• The Naive Bayes classifier achieved 72% accuracy in distinguishing between low and high exoskeleton force conditions.

Conclusions:

• Passive BCI is a viable approach for personalizing assistance in wearable devices like knee exoskeletons.
• EEG-based BCI can detect user states related to task difficulty and adapt device support.
• Further integration of passive BCI can enhance user experience and effectiveness of assistive wearables.