Detects Steady-State Visual Evoked Potentials using canonical correlation analysis (CCA) or filter-bank CCA (FBCCA). For each candidate stimulus frequency, a CCA is computed between the EEG data and sinusoidal reference signals at the frequency and its harmonics.
Usage
eegSSVEP(
x,
frequencies,
n_harmonics = 3,
method = c("cca", "fbcca"),
assay_name = NULL
)Arguments
- x
A PhysioExperiment object with EEG data. If 3D (time x channels x trials), data is averaged across trials before analysis.
- frequencies
Numeric vector of target stimulus frequencies in Hz.
- n_harmonics
Number of harmonics to include in reference signals (default: 3).
- method
Detection method:
"cca"(canonical correlation analysis) or"fbcca"(filter-bank CCA).- assay_name
Input assay name (default: first assay).
Value
A data.frame with columns: frequency (numeric target
frequency in Hz), correlation (numeric CCA correlation),
snr (numeric signal-to-noise ratio), and
predicted_class (numeric predicted stimulus frequency).
References
Blankertz, B., et al. (2008). Optimizing spatial filters for robust EEG single-trial analysis. IEEE Signal Processing Magazine, 25(1), 41-56.
Norcia, A. M., et al. (2015). The steady-state visual evoked potential in vision research: a review. Journal of Vision, 15(6), 4.
Examples
if (FALSE) { # \dontrun{
pe <- make_eeg_bci(n_trials = 10, n_channels = 8, sr = 256)
result <- eegSSVEP(pe, frequencies = c(10, 12, 15), method = "cca")
} # }