Measuring ERP latency shifts across experimental conditions using spatial filtering

For each individual subject, the source of an event-related potential (ERP) component does not vary with different conditions of the EEG experiment. So a common spatial filter can be applied to extract the component of interest in each condition and enhance its signal-to-noise ratio (SNR). In this study, a spatial filter was employed to measure the latency shift of an ERP component across multi-experimental conditions. The latency shift can be meausred using cross-correlation function from the EEG data after spatial filtering. The simulation showed that the proposed method is more accurate than the conventional methods. Even when SNR of ERP decreased to -15dB, the mean error of the proposed method was 1.36ms. The latency difference of N170 under upright face and inverted face stimuli measured by the proposed method was 6.7±0.7ms, whose variance was much smaller than that in previous reports.