|
nif:isString
|
-
The final sample consisted of 33 healthy, full-term infants who were aged 6 months (n = 11; five boys; age range: 6∶1–6∶23), 8 months (n = 11; four boys; age range: 8∶0–8∶19) and 10 months (n = 11; four boys; age range: 10∶0–10∶19). Five additional 6-month-olds, four additional 8-month-olds, and one additional 10-month-old were tested but not included in the final sample due to distress, fussiness, lack of attention or poor calibration. Additionally, a group of 11 adult participants (6 male; mean age = 37.9 years, SD = 10.7) was tested. An a-priori sensitivity power analysis (G*Power 3 software; [21] revealed that our final sample size (four equal-size groups of 11 participants) is large enough to detect a within-between interaction corresponding to an effect size as small as ηp2 = .1 with a statistical power of (1– β) = .95 (given α = .05). The protocol of the study was approved by the Psychology Research Ethics Committee of the University of Portsmouth, and the study was conducted in accordance with the 1964 Declaration of Helsinki. Before the experiment, each parent and adult participant provided written informed consent.
Test Environment, Apparatus and Stimuli: Both the action observation and action production tasks were conducted in the same testing room. Infants were tested individually with at least one parent present at a time of day when they were alert and in a good mood. For the action observation task, participants’ eye movements were recorded via corneal reflection using a SensoMotoric Instruments RED-X eye-tracker (sampling rate: 50 Hz). The stimuli were presented on a 17″ LCD monitor from a viewing distance of approximately 60 cm. Infants were seated in a safety car seat and adults were seated on a chair. SMI software (Experiment Center™ and iView X™) were used to collect and record calibration, present the stimuli, and record gaze data. At the beginning of the experiment, the infant’s attention was drawn to the monitor by presenting an attractive cartoon video. As soon as participants looked at the screen, they were presented with a standard 5-point calibration procedure, during which a small cartoon face expanded and contracted in synchrony with a sound. The experimental videos (30 fps; 800 × 600 pixels) showed from the side view a female adult (actor) performing a reaching movement towards either a small or a large ball (targets), both located on a table at a distance of approximately 70 cm from the actor’s torso and 10 cm apart from each other (Figure 1). The target of the actor’s reaching movement was not known in advance. In addition, because of the objects’ location and the fact that two different target layouts were used to counterbalance the hand trajectories, the actor’s goal was not even clearly indicated by its spatial location or by the trajectory of the actor’s approaching hand. All the arm movements started with the actor’s hand resting on the table in front of her torso. In half of the videos, the actor performs a reach-to-touch movement with the fist closed (No Shape condition), while in the other half she performs a reach-to-grasp movement during which the pre-shaping of the hand (either a precision or a whole hand grasp, depending on the target) was clearly visible soon after the movement started (Pre-Shape condition) [18], [22], [23], [24], [25]. Therefore, there were four movement types, corresponding to the four experimental conditions, namely No Shape–Large Target, No Shape–Small Target, Pre-Shape–Large Target and Pre-Shape–Small Target. The first 1000 ms of each video depicted the actor’s hand resting on the table in the starting position with a looming cartoon face, which was accompanied by an attention-grabber sound, superimposed on it (fixation phase). Then, the video showed the entire arm movement, i.e. from the earliest detectable movement of the hand to the hand-object contact (movement phase), lasting approximately 2000 ms (mean = 2045 ms; range = 1720–2280 ms). Note that there was no significant correlation between the movement phase duration and the participants’ gaze behavior. Finally, the last 500 ms consisted of the last frame of the stimulus video that was shown as still (contact phase) (Figure 1). Each video was followed by 1500 ms of black screen. After three stimulus videos, attractive animations with sound were shown to keep infants’ attention focused on the monitor.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pone.0067916.g001 Snapshots of stimulus videos.The figure shows the hand movement kinematic for the two targets layouts (panel A and B) in each experimental condition (for each panel, from top to bottom: No Shape–Small Target, Pre-Shape–Small Target, No Shape–Large Target and Pre-Shape–Large Target). The leftmost column depicts the Fixation phase with the Target AOI (white circle) superimposed on the target object. The central columns show the actor’s hand during the Movement phase for the frames corresponding to each quartile of the movement, and the column corresponding to the 100% of the Movement phase shows the actual end of the action, i.e., the last frame of the actual video. The rightmost column depicts the Contact phase. The person depicted in this figure has given written informed consent, as outlined in the PLoS consent form, to publication of their photograph.
For the grasping task, infants were tested individually while they sat on a caregiver’s lap at a wooden table. A video camera (30 fps) filmed the infant’s actions from a frontal perspective. The stimuli used were five objects from the Bayley Scale of Infant Development: the large objects were a handlebar rattle (11.5 cm long), one plastic cylinder (4 cm in height and 4 cm in diameter) and a plastic cube (side: 2.5 cm), and the small objects were a plastic cube (side: 1.2 cm) and a small sugar pill (0.5 cm in diameter).
For approximately 10 min, the infant was allowed to familiarize with the experimenters and the room while one experimenter described the test procedure to the parents before they signed a consent form. During stimulus presentation, the parent and experimenter stood behind the infant avoiding interacting with him/her. Once the infant and the parent seemed comfortable, the calibration procedure was started. Once all the five points were calibrated successfully, participants were presented with the experimental videos. Stimuli were presented in recording blocks (16 trials: four repetitions for each experimental condition). In each block, the stimuli were balanced for both movement type and target layout. The maximum number of trials presented was 64 (i.e., four recording blocks). After successfully completing the action observation task, infants were presented with the grasping task. In order to avoid priming the infant’s attention to the grasping actions in the action observation task, we always conducted the action production task after the observation task was completed. The reverse priming effect is less likely –i.e., that watching a precision grasp would immediately affect the infant’s own grasping style when confronted with new objects in the production task (see also [26]). One experimenter presented each of the five objects (previously placed on the floor out of view of the infant) one at a time to the infant on top of a flat palm to ensure that no grasp demonstration was provided. Objects were presented on the body midline at a comfortable reaching distance in front of the infant. The infants were allowed approximately 60 s to explore each object. If the infant did not react to the test object, the experimenter tried to attract the infant’s attention by moving the test object and giving verbal encouragement; if he/she was still hesitant to grasp the object from the experimenter’s hand, it was released onto the table in front of him/her (mean number of trials with encouragement: large objects = .06,.06 and.03 for 6-, 8- and 10-month-olds, respectively; small objects = .23,.14 and.05 for 6-, 8- and 10-month-olds, respectively). Finally, if the grasping of an object was not clean (i.e., when the actual grasping did not immediately follow the initial contact and there were some exploratory actions or hand repositioning before grasping, or when the object slipped out of the infant’s hands), the trial was omitted and the presentation of that object was repeated [27] (mean number of trials with object re-presentation: large objects = .30,.27 and.09 for 6-, 8- and 10-month-olds, respectively; small objects = .59,.55 and.23 for 6-, 8- and 10-month-olds, respectively).
For each stimulus video, we defined four areas of interest (AOIs) covering the attention–grabber during the fixation phase (Fixation AOI), the actor’s hand (Hand AOI), and the intended target (Target AOI) during the movement and the contact phases. The Hand AOI was a dynamic AOI, i.e., it was manually added frame by frame to match gaze trace with the moving hand. Data were included in the analyses only if the participants fulfilled the following criteria for at least two trials of each condition [13]. Participants’ gaze had to be within the Fixation AOI at the end of the fixation phase, and then participants had to fixate the Target AOI for 200 ms (or until the end of the video) before the video ended. By using the first criterion, we did not consider as predictive the occasional gaze shifts to the objects before the agents had started to move. A fixation was defined by the BeGaze software as a stable gaze (within 0.8 visual degrees) for at least 60 ms. For each valid trial, we calculated the gaze arrival time by subtracting the time when infants first looked inside the Target AOI from the hand-object contact time (i.e., the end of the movement phase). Therefore, if the participant’s gaze arrived at the Target AOI before the end of the actor’s action, the trial was regarded as predictive, and the gaze arrival time took a negative score. It is important here to note that our choice about the threshold for gaze anticipations was quite conservative. Indeed, in line with prior studies on action understanding and goal anticipation [12], [26], we chose a temporal threshold of 0 ms instead of a more liberal criterion incorporating a 200 ms reaction time in anticipations (e.g., [25]; see [28] for a discussion). Therefore, our estimates of participants’ goal anticipations would heavily underestimate the actual degree of their gaze proactivity. For the grasping task, the number of fingers used in grasping and displacing each object from its initial place of presentation was scored as the dependent variable by a coder who was unaware of both the experimental hypotheses and the age of the infants. For each trial, this measure could range from 2 through 10. On trials where the infants refused to grasp the object, a null value was coded and the trial was not considered in the analyses. A second (blind) judge coded 8 videos for each infant age group (73% of the grasping trials). Inter-observer reliability was high for all grasping trials (0.744< r <0.978, all ps <.001), and disagreements were mediated by a third judge. We computed two composite scores reflecting the ability of infants in performing the whole-hand and the precision grasping, i.e., the large- and the small-object grasping score, calculated by averaging infants’ scores across trials involving the three large and the two small objects, respectively.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}