A click-based electrocorticographic brain-computer interface enables long-term high-performance switch scan spelling.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Brain-computer interfaces (BCIs) offer communication restoration for individuals with severe motor and speech impairments.
• Single command click detectors represent a fundamental yet effective BCI capability.

Purpose of the Study:

• To evaluate the performance and long-term stability of a click decoding system using electrocorticography (ECoG) in a human clinical trial.
• To assess the feasibility of training a BCI click detector with limited data and minimal retraining.

Main Methods:

• A high-density ECoG BCI system was implanted, covering the sensorimotor cortex of a participant with amyotrophic lateral sclerosis.
• The click detector was trained using less than 44 minutes of data across 4 days, up to 21 days before BCI use, and tested for 90 days without retraining.

Main Results:

• The participant achieved a median spelling rate of 10.2 characters per minute using the click detector with a switch scanning speller.
• Despite transient signal power fluctuations, a newly trained click detector demonstrated comparable performance with even less training data (<15 minutes).

Conclusions:

• A click detector can be trained effectively using a small ECoG dataset, maintaining robust performance over extended periods.
• This demonstrates the potential for BCIs with click detectors to provide functional, text-based communication for users.