BCIwiki (bciwiki.com) — A study published in Biomedical Physics & Engineering Express on October 23, 2026, reports a training-free signal enhancement method developed by Washizawa Y of The University of Electro-Communications, Japan. Designed for steady-state visual evoked potential (SSVEP) brain-computer interfaces (BCIs), the method strengthens SSVEP components by repeatedly performing random sampling and pasting of electroencephalography (EEG) signals.
SSVEPs are sustained brain responses to periodic visual stimulation and can be used by BCIs to identify a user's target command. The paper notes that learning-based approaches, including deep neural networks used for command identification, require training data and incur high upfront costs. The proposed approach instead focuses on signal enhancement without model training.
Experiments used synthesized signals and open benchmark datasets. The method outperformed conventional methods at every tested signal length, with an information transfer rate improvement of up to 11.30 bit/min. These results come from synthetic signals and public dataset experiments.