The ASME-speller: 30-class auditory brain-computer interface speller using stream segregation and the QWERTY layout.

Area of Science:

• Neuroscience
• Computer Science
• Biomedical Engineering

Background:

• Brain-computer interfaces (BCIs) offer alternative communication pathways for individuals with severe motor impairments.
• Auditory BCIs can provide a vision-free communication method, but often require complex setups or lack intuitive interfaces.

Purpose of the Study:

• To introduce the ASME-speller, a novel 30-class auditory BCI speller system.
• To evaluate the system's performance in facilitating intuitive, vision-free communication using auditory stream segregation and a QWERTY keyboard layout.

Main Methods:

• The ASME-speller presents three distinct auditory streams (low, middle, high frequency) corresponding to QWERTY keyboard rows.
• Users selectively attend to auditory stimuli within specific streams to select letters and symbols.
• Online experiments were conducted with ten healthy participants to assess system performance.

Main Results:

• The ASME-speller achieved an average classification accuracy of 0.76 and an information transfer rate (ITR) of 2.16 bits/min.
• With artifact exclusion and optimized pipelines (LDA with dynamic stopping), performance improved to 0.80 accuracy and 4.76 bits/min ITR.
• The best participant achieved 1.0 accuracy and 14.44 bits/min ITR using a deep learning model (EEGNet4,2).

Conclusions:

• The ASME-speller demonstrates comparable performance to existing auditory BCI spellers.
• The system offers advantages in simplicity, requiring only standard headphones and no visual input.
• These findings support the feasibility of the ASME-speller for practical auditory BCI communication applications.