Figure 2 . Materials and Procedures (A: Congruent Condition; B: Incongruent Condition)
EEG Recording
EEG data were recorded from a 64 Ag/AgCl electrode cap (Neuroscan Inc.) according to the extended International 10–20 system. The physical reference electrode was placed approximately 2 cm posterior to the CZ electrode. Vertical electrooculographic (VEOG) recordings are electrodes placed above and below the left eye. Two electrodes were used at the outer canthi of both eyes to record the horizontal electrooculogram (HEOG). EEG signals were amplified in DC mode at 100 Hz and digitized at a rate of 500 Hz. The scalp interelectrode impedances of each subject were guaranteed to drop below 5 kΩ.
Data Analysis
One-way repeated-measures analysis of variance (ANOVA) was conducted on physiological data (SpO2, SBP, DBP, HR and VC) separately using time as a within-subject factor.
A 4 (Time: Test 1, Test 2, Test 3, Test 4) × 2 (Face stimulus : incomplete, complete) within-subject design was conducted. Factors are time and condition. The amplitudes and latencies of P1 and N170 are dependent variables. Offline data from the EEG recordings were analyzed using EEGLAB (http://sccn.ucsd.edu/eeglab/), software version 13.5.4b, an open-source signal processing toolbox implemented in MATLAB (The MathWorks, Inc, Natick, USA, Version 2018a). The data were referenced offline to the algebraic mean of all cortical electrodes and filtered offline with a bandpass filter of 1–90 Hz and a notch filter (50 Hz). Trials with various artifacts (body movements, ocular artifacts, or muscle activity) were rejected, with a criterion of ±75 μV from the EEG signal before proceeding to grand averages. Waveform averages for each individual were calculated. Average waveforms were divided into epochs of 1200 ms in length, including a baseline from -200 to 0 ms before the onset of the first face in each trial. Based on previous research, four posterior electrode sites (O1, O2, PO3, PO4) were considered for the analysis of P1. Four electrode codes (P7, P8, PO7, PO8) from the parietal and parieto-occipital regions were used for the analysis of N170. Peak amplitudes and peak latencies were adopted for statistical analyses for both components.
Three-way, repeated measure analysis of variances was conducted on P1 and N170 amplitudes and latencies separately. Hemisphere (left, right), face stimulus (incomplete, complete) and time (Test 1, Test 2, Test 3, Test 4) are within-subjects factors.
Data were analyzed using SPSS version 22 (Somers, Inc., Chicago) for Windows. All statistical significance levels were two-tailed, with an a priori alpha level for the significance of p < 0.05. To explore interaction effects, simple effect analyses were conducted. Effect sizes were calculated through partial eta-squared or Cohen’sd (d ). The Greenhouse‐Geisser epsilon correction was employed when the data violated the sphericity assumption. A Bonferroni corrected post hoc test was performed to determine pairwise differences.