Figure 5

Amplitude threshold estimation. Schematic figure of amplitude threshold estimation. The black curve with a white circle represents the scaled peaks plotted in ascending order. A) The artifact peaks (gray circles) are defined as outliers and are discarded. The candidate pivot point d_n (black circle) is estimated by fitting two straight lines to two datasets (D_n and D_p, see text for details) and by subsequently minimizing the least squares error (LSE) between the fitted lines and the curve. B) The pivot point (d_p) is determined as the point with an amplitude 2% higher than that of d_n. The point d_p divides the data set into noise (D_n) and peaks corresponding to blinks or saccades (D_p).