2.6. Data acquisition and preprocessing

We recorded continuous EEG signal using two 32‐electrode geodesic sensor net arrays and Netstation acquisition software and EGI amplifiers (Electrical Geodesics Inc). Impedance levels were measured as below 50 kΩ with a sampling rate of 500 Hz with no online processing performed. The experimental stimuli were presented using E‐Prime 2.0 software (Psychology Software Tools Inc.) on a 22‐inch LCD screen with a resolution of 1680 × 1050 pixels. Each EEG sensor net had a separate dedicated signal amplifier and acquisition computer.

To synchronize EEG signal acquisition within a dyad, one participant's signal amplifier and acquisition computer were designated as the master clock, with the other set as “slave” to the master clock. This process provided data synchronization within 2 ms (±1 ms). Experimental event markers were synchronized with these data and imported with the use of Noldus' Observer XT and Syncbox (Noldus Information Technology).

EEG raw data were transformed via conversion into a Matlab compatible file using BrainVision Analyzer 2.1.1 (Brain Products GmbH). EEG analysis was then performed using Matlab 2018a (Mathworks) and the EEGLab Toolbox 14.1.2b (Delorme & Makeig, 2004). The EEG signal was first downsampled to 256 Hz, and a bandpass filter of 1–40 Hz (60 Hz notch) was applied. Noisy channels or data segments were identified and marked using artifact subspace reconstruction (SD = 20) and visual inspection (C. Y. Chang et al., 2019; Plechawska‐Wojcik et al., 2018). Independent component analysis (ICA) using the ICA toolbox (Makeig et al., 2000) with extended infomax set to 1 was applied to the resulting signal data for all electrode sources to classify signal variance associated with vertical and horizontal eye‐blinks and heart rate (where applicable), with a maximum of 12 components. When identified, these components were corrected through manual inspection and automated action (Jung et al., 2000). An average of 2.05 components was removed (SD = 0.61).

All EEG signal was re‐referenced offline using the common linked mastoid average reference. An FFT was applied to the EEG data to split the signal into frequency bands, and then the theta (4–7 Hz), alpha (8–12 Hz), and beta (13–30 Hz) frequency bands were extracted; these data were then segmented into 3‐s epochs starting 1 s (–1–2 s) before stimulus presentation.