Optimizing Spatial filters for Robust EEG Single-Trial Analysis

TLDR

EEG volume conduction produces blurred brain‑activity images, so spatial filters are essential to enhance single‑trial signal‑to‑noise ratios, and modern machine‑learning methods enable data‑dependent spatio‑temporal filter optimization beyond fixed sensor‑geometry filters. This study elucidates the theory of the common spatial pattern (CSP) algorithm, reviews its variants, and demonstrates its application in the Berlin BCI project. We employ data‑dependent machine‑learning techniques to optimize CSP‑based spatio‑temporal filters per subject and apply these preprocessing steps to the Berlin BCI dataset. The optimized CSP approach yields powerful performance improvements and provides new theoretical insights, as shown in our Berlin BCI experiments.

Abstract

Due to the volume conduction multichannel electroencephalogram (EEG) recordings give a rather blurred image of brain activity. Therefore spatial filters are extremely useful in single-trial analysis in order to improve the signal-to-noise ratio. There are powerful methods from machine learning and signal processing that permit the optimization of spatio-temporal filters for each subject in a data dependent fashion beyond the fixed filters based on the sensor geometry, e.g., Laplacians. Here we elucidate the theoretical background of the common spatial pattern (CSP) algorithm, a popular method in brain-computer interface (BCD research. Apart from reviewing several variants of the basic algorithm, we reveal tricks of the trade for achieving a powerful CSP performance, briefly elaborate on theoretical aspects of CSP, and demonstrate the application of CSP-type preprocessing in our studies of the Berlin BCI (BBCI) project.