Anxiety disrupts moral perception

Data from 68 subjects (36 males, aged 19.99 ± 0.94 years [mean ± SD]) were included in the analysis. Electroencephalogram (EEG) signals were continuously recorded from 64 electrodes using a NeuroScan system, digitized at a rate of 1000 Hz, and referenced online to the vertex sensor. The scalp impedance for each electrode was maintained below 5 kΩ. All channels were bandpass filtered online from 0.01 to 100 Hz in AC sampling mode. For offline analysis, we utilized the Matlab-based EEGLAB toolbox (Version 2019b, Mathworks Inc., Natick, USA) EEGLAB (Delorme & Makeig, 2004) 2021.0. After relocating electrode positions, data were re-referenced to the average of the left and right mastoids, and a bandpass filter was applied from 0.1 to 40 Hz. Segments ranged from 200 ms before word onset to 1000 ms after stimulation, with behaviorally erroneous trials excluded. To minimize the impact of electrical shocks on the EEG data, five trials following each shock were deleted in the TOS group. Independent component analysis (ICA)(Makeig et al., 2004; Onton et al., 2005) was employed to correct ocular artifacts, and any remaining artifacts with amplitudes greater than ±100 μV were automatically excluded.