2.4 Electrophysiological Recording and Data Processing
The EEG was recorded from the scalp using either 33 or 64 channels from
the Active-Two BioSemi system (BioSemi, Inc., Amsterdam, the
Netherlands) using a modified 10-20 electrode placement system (American
Electroencephalographic Society, 1994). Additional electrodes were
placed on the left and right earlobes and two bipolar electrooculograms
(EOG) electrodes were placed at the supra- and infraorbital regions of
the left eye to measure vertical eye movements as well as on the left
and right outer canthi to measure the horizontal eye movements. All EEG
data were collected using a sampling rate of 1024Hz.
The EEG data were analyzed offline using the software BrainVision
Analyzer2 (Brain Products GmbH, Munich, Germany). The data were
referenced to the averaged signals of bilateral earlobes and then
filtered with a bandpass filter of 0.1-30 Hz with 24 dB/oct. The
continuous data were then segmented into 1200 ms epochs, spanning from
200 ms before stimulus onset to 1000 ms after stimulus onset. Segments
were then baseline-corrected based on the average voltage of -200 to 0
ms of stimulus onset, and eye movement artifacts were removed via a
regression approach based on the vertical EOG (VEOG) channel
(Segalowitz, 1996) then baseline-corrected again using the period of
-200 to 0 ms of stimulus onset. Segments containing voltage greater than
± 100 µV in the midline (e.g. Fz, FCz, Cz, Pz) and VEOG channels were
rejected. The segments were then averaged to create averaged ERPs for
each participant. The number of segments in the averaged ERPs ranged
from 220 to 473. The averaged ERPs obtained for each participant were
scored using a custom-built peak-picking computer program (Gavin, 2013)
written in MATLAB (Mathworks, Natick, MA). This program performs
automatic scoring and visual inspection of ERP components, and, when
necessary, allows for manual marking of components to adjust for any
values that were scored on a slope instead of a peak. In a previous
study (Lin, et al., 2020), for each participant, the averaged ERP
components were measured from baseline-to-peak using the following time
windows: P1 between 0 and 100 ms; N1 between 70 and 150 ms; P2 between
110 and 240 ms; N2 between 170 and 350 ms; and P3 between 320 and 575
ms. The N3 was not measured. All components were measured at the site
Cz. These data were used as the input data for ST measurement for this
current study.