Toward brain-computer interface speller with movement-related cortical potentials as control signals.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Brain-Computer Interfaces (BCIs) offer communication alternatives for individuals with severe motor impairments, such as Amyotrophic Lateral Sclerosis (ALS).
• Movement-related cortical potentials (MRCPs) are brain signals associated with executed movements, presenting a potential control signal for BCIs.
• This research explores MRCPs for BCI spellers, distinct from motor imagery-based approaches.

Purpose of the Study:

• To assess the feasibility of using MRCPs as a control signal for a Brain-Computer Interface (BCI) speller in an offline setting.
• To evaluate MRCP performance under varying task demands, including control, phrase spelling, and random letter selection.
• To investigate the effectiveness of signal processing techniques like Laplacian filtering on MRCP detection.

Main Methods:

• Fifteen healthy subjects performed spelling tasks using executed ballistic dorsiflexion of the dominant foot.
• Electroencephalographic (EEG) signals were recorded from 10 sites.
• Three conditions were tested: control (repeated 'O'), phrase spelling ('HELLO IM FINE'), and random letter selection.

Main Results:

• The success rate for MRCP detection was approximately 69% in both control and phrase spelling conditions.
• A slight decrease in success rate was observed in the random condition, attributed to increased task complexity.
• Significant differences in MRCP features were found using Laplacian filtering, but not with single-site Cz recordings.

Conclusions:

• MRCP-based BCI spellers are feasible for spelling tasks, demonstrating potential for communication aids.
• Laplacian filtering enhances the detection of MRCPs for BCI applications.
• Further research is necessary to develop and validate real-time MRCP-based BCI speller systems.