SSVEP detection assessment by combining visual stimuli paradigms and no-training detection methods

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Brain-Computer Interfaces (BCI) offer communication potential, especially Steady-State Visually Evoked Potentials (SSVEP) systems.
• SSVEP BCI efficiency relies on stimulation paradigms and detection algorithms.
• Optimizing SSVEP BCI requires balancing performance and user experience.

Purpose of the Study:

• To assess the performance of four visual stimulation paradigms (OOR, OOS, CBR, CBS) combined with three SSVEP detection methods (CCA, FBCCA, MEC).
• To evaluate the classification accuracy of target versus non-target events in SSVEP BCI.
• To investigate user comfort in relation to different visual stimulation paradigms.

Main Methods:

• Collected electroencephalographic (EEG) data from 27 participants under four visual stimulation paradigms (OOR, OOS, CBR, CBS) at five frequencies.
• Applied three SSVEP detection algorithms: Canonical Correlation Analysis (CCA), Filter-Bank CCA (FBCCA), and Minimum Energy Combination (MEC).
• Analyzed the performance of each paradigm-algorithm combination for event classification.

Main Results:

• The combination of On-Off Rectangular (OOR) or On-Off Sinusoidal (OOS) stimuli with Filter-Bank CCA (FBCCA) yielded the highest classification performance.
• Checkerboard stimuli (CBR, CBS) were preferred by 51.9% of participants for visual comfort and focus.
• A discrepancy exists between optimal performance and user-perceived comfort.

Conclusions:

• Filter-Bank CCA with On-Off visual stimuli offers superior performance for SSVEP BCI.
• Checkerboard visual stimuli enhance user comfort, suggesting a need for user-centered design in SSVEP BCI.
• The study provides valuable EEG data for SSVEP BCI algorithm development and evaluation.