Perturbation-evoked potentials can be classified from single-trial EEG

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Loss of balance control poses significant risks to human well-being and human-machine interaction.
• Balance perturbations elicit a distinct cortical response known as perturbation-evoked potential (PEP).

Purpose of the Study:

• To investigate the classification of perturbation-evoked potentials (PEPs) from ongoing electroencephalography (EEG) signals.
• To determine the feasibility of real-time detection of balance disturbances.

Main Methods:

• Fifteen healthy participants underwent seated whole-body perturbations, involving rapid tilting to the left and right.
• A window-based classification approach was employed to differentiate PEPs from resting EEG.
• Various window lengths and electrode configurations were evaluated.

Main Results:

• Classification accuracies exceeding 85% were achieved in distinguishing PEPs from resting EEG.
• Optimal performance (87.6 ± 8.0% accuracy) was obtained using a 200 ms window and a single Cz electrode.
• Peak classification accuracy correlated with the N1 component of the PEP.

Conclusions:

• Perturbation-evoked potentials (PEPs) can be accurately discriminated from ongoing EEG.
• These findings support the development of systems for online detection of balance perturbations.