TY - JOUR T1 - Adaptive spatio-temporal filtering for movement related potentials in EEG-based brain-computer interfaces. JF - IEEE Trans Neural Syst Rehabil Eng Y1 - 2014 A1 - Lu, Jun A1 - Xie, Kan A1 - Dennis J. McFarland KW - Algorithms KW - Artificial Intelligence KW - brain-computer interfaces KW - Data Interpretation, Statistical KW - Electroencephalography KW - Evoked Potentials, Motor KW - Humans KW - Imagination KW - Motor Cortex KW - Movement KW - Pattern Recognition, Automated KW - Reproducibility of Results KW - Sensitivity and Specificity KW - Signal Processing, Computer-Assisted KW - Spatio-Temporal Analysis AB - Movement related potentials (MRPs) are used as features in many brain-computer interfaces (BCIs) based on electroencephalogram (EEG). MRP feature extraction is challenging since EEG is noisy and varies between subjects. Previous studies used spatial and spatio-temporal filtering methods to deal with these problems. However, they did not optimize temporal information or may have been susceptible to overfitting when training data are limited and the feature space is of high dimension. Furthermore, most of these studies manually select data windows and low-pass frequencies. We propose an adaptive spatio-temporal (AST) filtering method to model MRPs more accurately in lower dimensional space. AST automatically optimizes all parameters by employing a Gaussian kernel to construct a low-pass time-frequency filter and a linear ridge regression (LRR) algorithm to compute a spatial filter. Optimal parameters are simultaneously sought by minimizing leave-one-out cross-validation error through gradient descent. Using four BCI datasets from 12 individuals, we compare the performances of AST filter to two popular methods: the discriminant spatial pattern filter and regularized spatio-temporal filter. The results demonstrate that our AST filter can make more accurate predictions and is computationally feasible. VL - 22 UR - http://www.ncbi.nlm.nih.gov/pubmed/24723632 IS - 4 ER -