An arXiv preprint proposes interpretable metrics that independently quantify temporal, spatial and frequency variability in BCI-related brain activity, and investigates the relationship between BCI performance and these metrics using two motor imagery BCI datasets with 133 users through within-user and cross-user classification experiments.
The researchers report negative correlations of -0.2 to -0.4 across most conditions, suggesting that lower variability is associated with higher BCI performance; the metrics also revealed differences in robustness to variability between deep learning and Riemannian-based classifiers, with the former showing weaker correlations. This study is a preprint and has not been peer reviewed.