Letswave 6

All EEG processing was performed using letswave 6 (letswave.org/">https://www.letswave.org/) and Matlab 2017 (The Mathworks). EEG data were segmented in 67-s segments (2 s before and 5 s after each sequence), bandpass-filtered (0.1–100 Hz) using a fourth-order Butterworth filter, and down-sampled to 256 Hz. Next, electrodes were visually inspected, and noisy electrodes were linearly interpolated from the 3 spatially nearest electrodes (not more than 5% of the electrodes, i.e., 3 electrodes, were interpolated). All data segments were re-referenced to a common average reference. While in frequency-tagging studies we typically apply blink correction (using ICA) for any participant blinking more than 2 standard deviations above the mean (e.g., [25 (link), 92 (link)], [96 (link)]), in the present study we did not perform any blink correction as none of the participants blinked excessively, i.e., more than two standard deviations above the mean across all participants (0.36 times per second). Note that frequency-tagging EEG yields responses with a high SNR at specific frequency bins, while blink artefacts are broadband and thus do not generally interfere with the responses at the predefined frequency [73 ]. Hence, blink correction (or removal of trials with many blinks) is not systematically performed in such studies (e.g., [33 (link), 78 (link), 108 (link)]).

EMG pre-processing and analysis were performed using Matlab (http://www.mathworks.com) and Letswave 6 (http://nocions.webnode.com) software^{59 (link)}. EMG signals from each participant were high-pass filtered (55 Hz) and full-wave rectified. HBR responses were averaged separately by position, resulting in three average HBR waveforms for each session per participant (i.e., far, near, ultra-near).