Identification of Brain Electrical Activity Related to Head Yaw Rotations

Area of Science:

• Neuroscience
• Computer Science
• Human-Computer Interaction

Background:

• Automating the identification of brain stimuli during head movements is crucial for advancing human-computer interaction (HCI).
• Applications include assisting severely impaired individuals and enhancing robotics.
• Electroencephalography (EEG) signals offer a non-invasive method to monitor brain activity.

Purpose of the Study:

• To develop and evaluate a neural network-based technique for recognizing head yaw rotations using EEG signals.
• To identify the input-output function linking brain electrical activity to visually-induced head movements.
• To assess the feasibility of using EEG for brain-computer interfaces (BCIs) controlled by head movements.

Main Methods:

• A neural network utilizing the Levenberg-Marquardt backpropagation algorithm was employed.
• EEG signals were recorded from ten participants during experiments involving visual stimuli.
• The study involved analyzing EEG data to correlate brain activity with head turning.

Main Results:

• The proposed approach successfully identified brain electrical stimuli associated with head turning in individual participants.
• High prediction accuracy was observed, with correlations reaching r = 0.98 and Mean Squared Error (MSE) = 0.02 in the best cases.
• Classifier performance significantly decreased when models trained on one participant were tested on others, indicating subject-specific variations.

Conclusions:

• The study demonstrates the potential of EEG-based neural networks for identifying head movements in HCI applications.
• The findings highlight the effectiveness of the proposed method for within-subject analysis.
• Further research is needed to improve cross-subject generalization for broader BCI applications.