Following the robot's lead: Predicting human and robot movement from EEG in a motor learning HRI task

Area of Science:

• Neuroscience
• Cognitive Science
• Robotics

Background:

• Human behavior is often organized into sequential sensorimotor patterns.
• Learning new sequences involves cognitive functions and specific neural mechanisms.
• Understanding neural underpinnings of motor learning and joint action is crucial for human-robot interaction.

Purpose of the Study:

• To investigate how electroencephalography (EEG) signals reflect behavioral processes during sensorimotor sequence learning.
• To explore neural coding of joint action in a human-robot interaction.
• To determine if EEG can decode movements in a human-robot collaboration.

Main Methods:

• A human-robot interaction scenario involving a robot demonstrating and a human mimicking a pointing sequence.
• Analysis of event-related spectral perturbations (ERSP) in EEG signals during rest, fixation, and movement.
• Application of a Markov-switching linear regression model to decode movements from EEG data.

Main Results:

• Task-related modulation of ERSP was observed, differing for rest, fixation, and movement sequences.
• Motor sequence learning significantly modulated ERSP.
• EEG signals successfully decoded human and robot movements using the developed model.

Conclusions:

• EEG-based neurophysiological activity can reflect motor sequence learning and joint action processes.
• The study provides insights into neural coding for motor performance and collaborative tasks.
• EEG holds potential for decoding movements in human-robot interaction scenarios.