Design Considerations for Long Term Non-invasive Brain Computer Interface Training With Tetraplegic CYBATHLON Pilot

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Brain Computer Interface (BCI) technology translates neural signals into control commands, but is often limited to lab settings and healthy individuals.
• Real-world BCI applications for individuals with spinal-cord injury (SCI) require user-friendly, engaging, and long-term usable interfaces.
• The CYBATHLON 2020 BCI race provided a platform to develop and test BCI design strategies for tetraplegic users.

Purpose of the Study:

• To investigate BCI design aspects, specifically calibration paradigms, for long-term use in SCI patients.
• To compare the efficacy of conventional open-loop calibration with real-time closed-loop calibration paradigms.
• To develop a user-friendly and engaging BCI interface for SCI patients.

Main Methods:

• Compared open-loop and closed-loop BCI calibration paradigms using pre-trained decoders.
• Analyzed performance indicators: classification accuracy, game completion time, and brain activation maps.
• Collected subjective feedback from the BCI pilot.

Main Results:

• Closed-loop calibration paradigms with real-time feedback were more engaging for the pilot.
• Closed-loop paradigms showed indications of better online median classification performance (p = 0.0008) compared to conventional methods.
• Stronger, more localized brain activation patterns were observed with the closed-loop paradigm when the interface mimicked the end application.

Conclusions:

• BCI calibration paradigms incorporating active user engagement, such as real-time feedback, can improve user acceptability and performance.
• Closed-loop BCI calibration shows promise for enhancing long-term usability and effectiveness in SCI populations.
• This study highlights the importance of user-centered design in developing practical BCI solutions for individuals with disabilities.