2. Methods

2.1. Stimuli

The stimuli consisted in a set of 120 videos displaying different type of orofacial movements (1- still mouth, 2-syllables, 3-backward played syllables, 4-non-linguistic movements) or non-biological movements (5- non-human). In the first condition, no mouth movements were produced (Baseline). In the second, 3 type of syllables were produced differing in their place of articulation (PoA). Bilabial syllables ([pa] [ba]) requires lip movements whereas alveolar ([da] [ta]) and velar ([ga] [ka]) syllables require upper and lower tongue movements, respectively. These consonants have been chosen because they have the common characteristic of being stop consonants, which mean that they are articulated by closing the airway so as to impede the flow of air, by maintaining airway closed thus generating a slightly pressure because of accumulated air and finally by opening the airway and releasing the airflow producing in that way an audible sound. Importantly, these three kind of syllables have been reported to have different levels of visual salience \cite{Jesse_2010,Paris_2013,van_Wassenhove_2007}. In the third condition, the same syllables were played backward. Because of their particular motoric sequence, stop syllables can not be pronounced backward. In that sense, backward played syllables represent an ideal control condition because this kind of articulatory movements are visually very similar to speech but at the same time they are not pronounceable, they do not belong to our hypothesized motor repertoire. In the forth condition, non-linguistic orofacial movements (e.g., tongue protusion, lip-smacking) were produced. This condition was introduced in order to control for the activity associated to the processing of orofacial movements with no linguistic content.  Finally, in order to control for general movement perception, independently of its biological and facial related nature, a fifth condition was added where movements of different geometrical figures (e.g., ovals, squares, triangles) were shown. These stimuli were generated using PsychoPy toolbox. Importantly, all the videos were silently displayed (i.e., audio removed) and only show the lower part of the actor’s face in order to ensure that his eyes movements could not interfere. Videos were 2 seconds long (M=2052ms and SD=59ms), they started with 10 frames displaying a closed, still mouth.

2.2. Participants

34 right-handed subjects (22 females) with normal or corrected-to-normal vision and hearing and without any history of psychiatric or neurological disorders performed the experiments. Participant's age range from 18 to 36 years old (M=22,8 and SD 4,2 years). The experimental protocol was approved by the Ethics Committee of Pontifical Catholic University of Chile, School of Psychology. Before the experiment started each participant was explained the procedure and signed an informed consent form. Four participants were removed from final analysis because of poor signal-to-noise ratio. 

2.3. Procedure

Participants sat at a distance of approximately 70 cm from the computer screen and were asked to attentively observe or imitate the movements shown in the videos.  Stimuli were displayed on the screen using PsychoPy toolbox. The trial started with a word lasting for 500ms that indicate the instruction, either "Observe" or "Imitate". After 100 to 150ms,  a fixation cross appeared for 250ms. In the observation condition, the video was displayed one time, 1000 to 1500ms after the white cross disappeared. In the imitation condition, the video was displayed a first time and participants were asked to attentively observe in order to co-imitate the orofacial gesture when the video was displayed for the second time. The onset of imitation was cued with a red fixation cross. After video offset, a new trial began within 2 to 3 seconds. In order to study the role of automatic mimicry the very same  experiment was repeated in Experiment 2 (B2) where participants were asked to hold an effector depressor horizontally between their teeth (i.e. in the imitation trials, participants were asked to remove the depressor when the word "Imitate" appeared so they can properly imitate). This procedure allowed to impede the automatic mimicry. The order of Experiment 1 and 2 was counterbalanced between participants.
Each of the 5 conditions [i.e., 1) still mouth, 2) syllables, 3)backward played syllables, 4) non-linguistic mouth movements and 5) non-human movements] consist in 3 repetitions of the 24 video-clips, leading to a total of 72 trials per condition (360 per experiment). The experimental design was an intra-subject 5 (type of movements) X 2 (without/with effector restriction) factorial design.

2.4.  Electroencephalographic recording parameters

Electrophysiological activity was register with a 64-channels (Ag-AgCl) EEG system (Biosemi ® ActiveTwo) with electrodes positioned according to the extended 10-20 international system. The sampling rate was of 2048 Hz (band-passed 0.1 to 100Hz). Four external electrodes were used to monitor eye movements. Two of them were placed in the outer canthi of the eyes in order to record horizontal EOG and the other two were positioned above and below the right eye to record vertical EOG. Two additional external electrodes were placed on bilateral mastoids for re-referenciation. Data pre-processing was performed using MatLab (The Mathworks, Inc.), EEGLAB \cite{Delorme_2004} and ERPLAB toolbox \cite{Lopez_Calderon_2014}. The signal was first down-sampled at a rate of 512 Hz, re-referenced to mastoides and band-passed between 0.1 and 40 Hz for ERP analysis. Then, EEG signal was segmented into epochs from -500ms to 1500ms respective to stimulus onset. Each epoch was visually inspected in order to reject large artifacts due to head movements and muscular artifacts. After Independant Component Analysis (ICA) decomposition and the rejection of components typically associated with eye-blinking, epochs exceeding maximum amplitude of ±100 μV were removed.    + ANALYSIS DE FREQ THETA

2.5. Statistical analysis

The ERP components of interest for statistical analyses were P2, N270, N400 and a positivity around 1000 ms. Using the ERP measurement tool in ERPLAB, mean amplitudes were calculated with respect to a 50 ms prestimulus baseline (un poco chico el baseline, pq tan chico?) for the following selected time windows: P2 [155-185 ms], N270 [245-295 ms], N400 [475-525 ms] and P1000 [975-1025 ms]. After mean amplitudes were extracted for each condition, data were analyzed using a mixed model for each experiement (1 and 2), with conditions as fixed factor and subjects as aleatory factor. Statistical analysis was performed using the lme4 \cite{Bates_2015} package of R. The pairwaise comparision within conditions was optained using emmeans \cite{Lenth_2016} package of R, that permit to compare slopes in a mix model.      + Porqué t distribution (cf. Rodrigo Lagos) 
Agregar Theta como indicador
+ d de cohen reportar 0.5 see G-power