Lower-Limb Motor Imagery-Based Brain-Computer Interface to Control Treadmill Velocities.

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Biomedical Engineering

Background:

• Traditional lower-limb rehabilitation relies on physical therapy.
• Motor imagery (MI)-based brain-computer interfaces (BCIs) offer a novel approach to enhance neuroplasticity by closing the perception-action cycle.
• BCIs can facilitate adaptation and recovery in rehabilitation settings.

Purpose of the Study:

• To develop and evaluate a BCI system for controlling treadmill velocity using kinesthetic motor imagery.
• To establish a closed-loop feedback mechanism for lower-limb rehabilitation.
• To investigate the efficacy of different feature extraction and classification methods for MI-based BCI.

Main Methods:

• A BCI system was designed to translate kinesthetic MI (mu and high-beta rhythms) into treadmill velocity commands.
• Feature extraction techniques including power spectral density (PSD) and Riemannian geometry (RG) were utilized.
• Machine learning classifiers such as Logistic Regression (LR), k-nearest neighbors, Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) were optimized.

Main Results:

• Increased mu and high-beta rhythm modulation was observed at the vertex during MI tasks.
• The online RG+LDA classifier achieved an average accuracy of 72%.
• A pseudo-online RG+LR classifier demonstrated a high accuracy of 95%.

Conclusions:

• The study successfully combined kinesthetic MI with treadmill control, utilizing RG for feature extraction.
• The developed BCI system shows potential for enhancing cortical modulation during lower-limb rehabilitation.
• Further validation in post-stroke patients is required to confirm clinical applicability.