Performance of Brain-Computer Interfacing Based on Tactile Selective Sensation and Motor Imagery

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Engineering

Background:

• Many users struggle to achieve adequate control with current non-invasive brain-computer interfaces (BCIs), a phenomenon termed "BCI-Illiteracy."
• BCI-Illiteracy affects various BCI modalities, limiting their practical application and accessibility.

Purpose of the Study:

• To compare the performance and BCI-illiteracy rates of tactile selective sensation (SS) and motor imagery (MI) based BCIs.
• To evaluate the potential of SS BCIs as a more accessible alternative for a wider user population.

Main Methods:

• Analyzed 80 experimental sessions (57 subjects) using two-class SS protocols for left- and right-hand discrimination.
• Analyzed 63 experimental sessions (43 subjects) using two-class MI protocols for left- and right-hand discrimination.
• Conducted within-subject comparisons for 24 subjects participating in both SS and MI experiments, analyzing performance across different frequency bands (alpha and beta).

Main Results:

• Tactile SS BCI achieved a group average performance of 79.8%, with 75.4% of subjects exceeding the 70% BCI-illiteracy threshold.
• MI BCI achieved a group average performance of 77.2%, with 69.7% of subjects exceeding the 70% threshold.
• SS BCI demonstrated superior performance over MI BCI, particularly in the beta frequency band, with enhanced discriminative information in the somatosensory cortex. Both modalities showed better performance in the upper alpha band.

Conclusions:

• Tactile SS BCI is a promising modality with a lower BCI-illiteracy issue compared to MI BCI.
• SS BCIs outperform existing tactile BCIs and show significant potential for practical applications targeting a large population.
• Functional dissociation observed in alpha bands suggests distinct neural processing mechanisms influencing BCI performance.