Evaluation of an Asynchronous Tongue-Based Brain–Computer Interface for Two-Dimensional Cursor Control

Background and Objectives: Brain–computer interfaces (BCIs) enable individuals with severe motor impairments to control assistive technologies; however, the performance of asynchronous BCIs is often limited when multiple control classes must be decoded. A potential strategy to improve performance is to combine a binary BCI switch with a state machine. This proof-of-concept study investigated two-dimensional cursor control using BCI-detected tongue movements or tongue motor imagery to select directions determined by a state machine. Methods: Ten able-bodied BCI-naive participants were exposed to two experimental conditions in which the BCI was controlled through detection of sensorimotor rhythms associated with executed or imagined tongue movements. The BCI was combined with a state machine comprising four directional options (up, right, down, and left), enabling two-dimensional cursor control to reach on-screen targets. Movement-related activity was discriminated from idle activity using a linear discriminant analysis classifier combined with common spatial pattern filtering. Results and Conclusions: All participants achieved successful BCI control during executed tongue movements (completion rate: 96.0 ± 4.9%; average completion time per target: 30.8 ± 17.6 s), whereas only seven participants achieved control using motor imagery (completion rate: 41.0 ± 36.2%; average completion time per target: 66.5 ± 26.8 s). These findings demonstrate the feasibility of combining a tongue movement-based BCI with a state machine for two-dimensional cursor control, supporting the development of asynchronous BCI systems for individuals with severe motor impairments.