Figure 2 . Schematic representation of the subject’s motor involvement and body position during the three task conditions.
2.3 Behavioral data
The response speed was assessed using the response time (RT) for correctly executed trials. Response accuracy was assessed by summing the erroneous responses to non-target stimuli and the omitted responses to target stimuli, this value has been expressed in percentage (Err%).
2.4 EEG recording and analysis
All participants were tested using a 64-channel EEG system (BrainampTM amplifiers) with active electrodes (ActicapTM) and software (Recorder 1.2 and Analyzer 2.2) all by Brain Products GmbH (Munich, Germany). The electrodes were mounted according to the 10–10 International System and referenced to the averaged mastoids. Horizontal and vertical electrooculograms (EOG) were monitored by additional bipolar recordings. The EEG was digitized at 250 Hz, amplified (bandpass of 0.01–60 Hz including a 50 Hz notch filter), and stored for offline averaging. The signal was then filtered with 0.1 Hz high-pass and 40 Hz low-pass filters.
To investigate the pre-stimulus activities (independently from the stimulus category), the signal was segmented in epochs starting 2500 ms prior to the stimulus onset (time 0) and lasting for 2800 ms. Eye movement artifacts were corrected using the Independent Component Analysis (ICA) ocular correction algorithm: it has been shown that this method, introduced by Jung et al. (2000), revealed better results than other ocular correction methods (e.g., Hoffmann & Falkenstein, 2008). Furthermore, semi-automatic artifact rejection was performed prior to signal averaging to discard epochs contaminated by signals exceeding the amplitude threshold of ±80 μV and about 3.9% of trials were rejected. The artifact-free trials were averaged, and pre-stimulus activities were measured with respect to a −2500/-2300 ms baseline. Given that the stimulus category was unpredictable at the pre-stimulus phase, target and non-target trials were averaged.
For the intervals and electrodes to be included in statistical analysis, the “collapsed localizer” method was utilized (Luck & Gaspelin, 2017). Accordingly, a localizer ERP was obtained by collapsing (averaging) all the considered groups and conditions. The global field power (GFP) was calculated to select the analysis interval. The GFP describes the ERP spatial variability considering all scalp electrodes and allowing a reference-independent descriptor of the ERP. The interval in which the GPF was larger than 70% of its maximum value was used for further analysis. This approach designated two intervals (from -1516 to -770 ms and from -724 to 0 ms) from which the mean amplitude was calculated for statistical analysis. The electrodes with an amplitude larger than 70% of the maximum value in that interval were collapsed in spatial pools and considered for statistical purposes. Four foci of medial activity were present: prefrontal (pN), central (BP), parietal (pBP), and occipital (vN). The pN was therefore represented by a pool of electrodes including Fp1, Fpz, Fp2, AF3, AFz, and AF4. The BP was represented by a pool including C1, Cz, C2, CP1, CPz, and CP2 electrodes (central pool). The pBP was represented by P3, P1, Pz, P2, P4. The vN was represented by PO1, POz, PO2, O1, O2.
To isolate the preparatory activity associated with reaching, the ERP of the Keypress condition was subtracted from the ERP of theReach condition. In addition, to isolate the preparatory activity associated with stepping, the ERP of the Reaching condition was subtracted from the ERP of the Reaching-Stepping condition. Based on these differential ERP activities, the source localization of these waveforms (in the -1500/0 ms whole pre-stimulus period) was realized using the “exact low-resolution brain electromagnetic tomography” (eLORETA) software (freely available at www.uzh.ch/keyinst/loreta.htm) to compute the cortical three-dimensional distribution of current density. This method utilizes a discrete, three-dimensional distributed, linear, weighted minimum-norm inverse solution. The weights applied in eLORETA confer precise localization capabilities to test point sources, generating current density images with exact localization, albeit with a limited spatial resolution (i.e., neighboring neuronal sources exhibit high correlation). Notably, eLORETA demonstrates no localization bias, even in the presence of structured noise, representing an advancement over LORETA (Pascual-Marqui et al., 1994) and its standardized version, sLORETA (Pascual-Marqui, 2002).