Neural Noise- A Journal of Neural Engineering study targets passive BCIs for mental workload using code-based VEP (c-VEP), motivated by conventional flickering VEP/c-VEP stimuli being distracting and visually uncomfortable.1
- The approach combines near-invisible, texture-based flicker embedded in UI regions of interest with a c-VEP passive-BCI pipeline to infer workload.1
- Experiment (i) used an ecologically oriented multitasking microworld centered on flying; experiment (ii) used a flight simulator with cognitive workload manipulated across three systematic levels.1
- At the group level, visual ERP amplitude was reported significantly lower under higher workload (numeric details truncated in the available summary).1
- The authors position low-visibility stimulation as relevant to neuroadaptive interfaces and operator monitoring without conventional obtrusive VEP presentation.1 1
- Authors are Pietro Cimarosto, Sebastien Velut, Kalou Cabrera Castillos, Juan Jesus Torre-Tresols, Raphaëlle N. Roy, and Frédéric Dehais from the Centre de Neuroergonomie, ISAE-SUPAERO, Toulouse, France (published online 10 March 2026, DOI: 10.1088/1741-2552/ae4ff6; PMID: 41806473).2
- The texture-based flicker approach builds on the lab’s 2024 StAR (Stimuli for Augmented Response) paradigm using Gabor/Ricker patches, which demonstrated visually comfortable stimuli achieving 93.6–96.3% accuracy with only 88 s of calibration data—outperforming plain flickers (65.6%); online asynchronous decoding reached 94.3% accuracy with mean decoding time of 1.68 s.3
- The near-invisible design addresses a key adoption barrier: standard c-VEP paradigms induce visual fatigue rated ~6.4/10 on subjective scales, while reduced-contrast variants lower fatigue to ~3.7 while maintaining >99% accuracy in controlled settings.3
- The passive BCI pipeline derives workload indexes (not just binary classification) sensitive to workload-related modulation, enabling continuous monitoring in operational environments such as aviation and human-machine supervision.1
- Related work shows that Riemannian geometry classifiers remain the state-of-the-art for cross-session passive BCI workload decoding; extending c-VEP-based stimulation to passive paradigms bridges the gap between high-performance reactive BCIs and practical continuous operator monitoring.4