Figure 3.

Procedure for filtering EEG data using local spatial analysis (LSA). The filter is applied to real laser-evoked event-related potentials (ERPs). Step 1: first, the electrode at which the widespread N2 component is maximal in amplitude is identified in the grand average scalp map series. In this example, the electrode is Cz. Second, the time interval including the widespread signal (i.e., the interval during which most of the electrodes have the same polarity) is selected. In this example, the interval is 130–220 ms. Step 2: LSA exploits the different trial-by-trial variability of the local vs. the widespread component, and thereby removes the activity from the identified electrode (Cz) from all other electrodes. This is done separately for each subject and time point. The last row shows the comparison of the topography at the same time point (in this example, 160 ms) before and after filtering with LSA. Color axes represent voltage (µV).