BCIwiki (bciwiki.com) — Visual flicker that steps up or down through a sequence of frequencies can be detected from EEG offline but did not classify better than conventional fixed-frequency flicker, researchers from Walailak University in Thailand and Kyushu University in Japan reported in Symmetry on September 28, 2026.
Steady-state visual evoked potential (SSVEP) brain-computer interfaces (BCIs) work by having a user look at a target flickering at a set frequency; the visual cortex responds at that frequency, letting the system infer the chosen target. The team presented 5 adjacent frequencies spaced 0.2 Hz apart in consecutive 2 s epochs as ascending-frequency flicker patterns (AFFPs) or descending-frequency flicker patterns (DFFPs), in low (6–10 Hz) and high (41–45 Hz) ranges, and compared them with fixed-frequency stimulation.
EEG from 15 participants was analyzed with 8 established methods: FFT, power spectral density (PSD), canonical correlation analysis (CCA), filter bank CCA (FBCCA), dynamic PSD, short-time Fourier transform (STFT), continuous wavelet transform (CWT) and discrete wavelet transform (DWT). Low-frequency stimulation reached an average normalized accuracy of 90.8%, against 80.6% for high-frequency stimulation. CWT had the highest descriptive average accuracy, though post hoc comparisons did not show it was significantly better than the other methods.
Fixed-frequency, AFFP and DFFP conditions averaged 89.7%, 85.6% and 89.5%, respectively. The signal-processing method had a significant main effect; the stimulation pattern did not. Three consecutive commands averaged 91.5 ± 9.6%, and accuracy dropped as command sequences grew longer.
The authors said the small, uniform sample limits how far the findings generalize. They suggested the sequences' main value may be in designing flexible sequential commands, which requires testing in larger, more diverse groups and in an online closed-loop BCI system.