Enhancement of lower limb motor imagery ability via dual-level multimodal stimulation and sparse spatial pattern...

Area of Science:

• Rehabilitation Engineering
• Neuroscience
• Biomedical Engineering

Background:

• Motor imagery brain-computer interfaces (MI-BCIs) are crucial for motor function assistance and rehabilitation.
• Efficient stimulation paradigms and Electroencephalogram (EEG) decoding methods are needed to improve MI-BCI performance.
• Current MI-BCI systems require enhancement for effective lower-limb rehabilitation.

Purpose of the Study:

• To design a multimodal dual-level stimulation paradigm for lower-limb rehabilitation training.
• To propose an advanced EEG decoding method (UTFB-SSP) for improved MI-BCI performance.
• To enhance motor imagery (MI) performance by inducing specific brain rhythms.

Main Methods:

• A multimodal dual-level stimulation paradigm combining visual, auditory, proprioceptive, and functional electrical stimulation for the lower limb was designed.
• The Upper Triangle Filter Bank Sparse Spatial Pattern (UTFB-SSP) method was proposed for automatic selection of optimal frequency sub-bands.
• The system's effectiveness was validated using an in-house experimental dataset and the BCI competition IV IIa dataset.

Main Results:

• The proposed MI-BCI system demonstrated effectiveness in enhancing motor imagery performance.
• The system successfully induced α, β, and γ brain rhythms during lower-limb movement imagery tasks.
• The UTFB-SSP method improved decoding performance by selecting relevant frequency sub-bands.

Conclusions:

• The developed multimodal dual-level stimulation paradigm is effective for lower-limb rehabilitation using MI-BCIs.
• The proposed UTFB-SSP method significantly enhances EEG decoding performance in MI-BCI systems.
• This approach shows promise for advancing motor function assistance and rehabilitation engineering.