Evaluation of various mental task combinations for near-infrared spectroscopy-based brain-computer interfaces

Area of Science:

• Neuroimaging
• Brain-Computer Interfaces

Background:

• Near-infrared spectroscopy (NIRS) is a promising neuroimaging technique for brain-computer interfaces (BCIs).
• Previous NIRS-BCI research primarily focused on improving classification accuracy for individual mental tasks.

Purpose of the Study:

• To evaluate mental task combinations for optimal performance in customized NIRS-based BCIs.
• To identify specific mental task pairs yielding the highest classification accuracies.

Main Methods:

• Recorded event-related hemodynamic responses from seven participants performing eight distinct mental tasks.
• Estimated classification accuracies for all possible pairs (28 combinations) of the eight mental tasks.

Main Results:

• Mental task combinations including mental multiplication, mental rotation, and right-hand motor imagery demonstrated high classification accuracies.
• Pairs comprising two of these three tasks yielded the highest mean classification accuracies.

Conclusions:

• The identified mental task combinations serve as a valuable reference for NIRS-BCI development.
• These findings can reduce the time required for preliminary testing in creating personalized BCIs.