Neural NoiseSteady-state visually evoked potentials (SSVEPs) are periodic neural responses generated in the visual cortex when a person gazes at a visual stimulus flickering at a constant frequency. The resulting eeg signal contains strong spectral peaks at the stimulus frequency and its harmonics, providing a reliable and high signal-to-noise ratio marker of visual attention. SSVEPs can be detected with minimal training and are robust against artifacts, making them attractive for practical applications.
In bci-and-neural-decoding, SSVEP-based systems achieve among the highest information transfer rates of any non-invasive BCI paradigm. Multiple on-screen targets flicker at distinct frequencies; the user selects a target by directing gaze toward it, and the system identifies the chosen target by analyzing the dominant frequency in the occipital EEG. This approach requires no prior user training and can support large command sets, enabling applications such as speller interfaces, device control, and communication aids for individuals with motor impairments.
Current research directions include joint frequency and phase coding to expand the number of discriminable targets, filter bank canonical correlation analysis and deep learning methods for improved detection accuracy, hybrid paradigms that combine SSVEP with p300 or motor-imagery signals, and stimulus designs that reduce visual fatigue. Efforts to develop code-modulated and aperiodic stimulation schemes aim to improve user comfort and extend SSVEP-BCI use to more naturalistic settings.