Motion-Based BCI: c-MVEP Achieves 85.7% Accuracy Without Flickering

A team of researchers from various institutions (Scheppink, Herpers, Thielen, Volosyak) has published a paper on arXiv introducing a novel brain-computer interface (BCI) paradigm called code-modulated motion visual evoked potentials (c-MVEP). Instead of relying on traditional flickering visual stimuli (which can cause eye strain and discomfort), c-MVEP uses pseudo-random sequences of motion to trigger brain responses. The study compared c-MVEP against three established paradigms: code-modulated VEP (c-VEP), steady-state motion VEP (SSMVEP), and steady-state VEP (SSVEP). In offline experiments with single-object stimulation, c-MVEP exhibited similar time-domain characteristics and broadband frequency responses to c-VEP, with a signal-to-noise ratio comparable to c-VEP but concentrated in lower frequencies. Spatially, c-MVEP activation spread across multiple electrodes (mainly at Oz), while c-VEP was more focused. Subjective ratings showed no clear preference for motion over flicker.

The online experiment evaluated a 4-class BCI under all four conditions to test practical feasibility. c-MVEP achieved a mean accuracy of 85.67% with an average selection time of 2.61 seconds—significantly better than SSMVEP (64.91%, 4.18s) but worse than flicker-based c-VEP (97.81%, 1.15s) and SSVEP (93.42%, 1.94s). The authors conclude that c-MVEP shows great potential as a comfortable alternative for users who find flickering stimuli unpleasant, though it currently lags in speed and accuracy. This work paves the way for more user-friendly BCI systems in applications like assistive communication.