Decoding of top-down cognitive processing for SSVEP-controlled BMI.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Brain-machine interfaces (BMIs) typically rely on bottom-up visual processing.
• Existing methods often require gaze-shifting, limiting their application.
• Steady-state visual evoked potentials (SSVEPs) are a common EEG signal for BMI.

Purpose of the Study:

• To develop and validate a novel non-invasive BMI system.
• To investigate the decoding of user intention based on top-down visual processing using SSVEPs.
• To assess the efficacy of this paradigm compared to traditional bottom-up approaches and electrooculography (EOG).

Main Methods:

• Participants observed a flickering line array and intentionally attended to specific lines forming letters.
• Steady-state visual evoked potentials (SSVEPs) were recorded using electroencephalography (EEG).
• Regularized linear discriminant analysis was employed for decoding, with Granger causality analysis for neural pathway investigation.

Main Results:

• The SSVEP-BMI achieved decoding accuracies of 35.81% (up to 65.83%), significantly exceeding chance levels (2.05 to 3.77-fold).
• EEG signals were sufficient for decoding; EOG provided no significant additional accuracy.
• Top-down processing demonstrated focalized activation in visual areas, with prefrontal control over early visual processing.

Conclusions:

• The study presents the first neurophysiological evidence for a top-down SSVEP BMI paradigm.
• This novel approach enables multi-class intentional control of EEG-BMIs without gaze-shifting.
• The findings suggest a promising new direction for developing more intuitive and versatile brain-computer interfaces.