Child-robot interactions in different body and emotions oriented tasks: comparison with a human partner

Child-robot interactions in different body and emotions oriented tasks: comparison with a human partner | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Child-robot interactions in different body and emotions oriented tasks: comparison with a human partner Alice Araguas, Adrien Chopin, Arnaud Blanchard, Sébastien Derégnaucourt, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4758583/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The aim of the present study was to compare interactions of children aged between 3 and 6 years, with a NAO robot or an adult partner, in various body-focused tasks: comprehension and recognition of body parts labels, imitation of movements, and recognition of emotions in the postures of the agent. For each task, performances were appreciated through scores levels. We found no effect of the demonstrator type on our results: children of different ages responded similarly to the the human or the robot demonstrator. We found an effect of age, with the olderchildren having higher scores for the comprehension of body parts labels on the demonstrator’s body, the imitation of body parts sequences and the identification of emotional key postures. Results are discussed in light of the implications of the use of social robots such as the NAO one, in interactive and learning situations with typical children. Biological sciences/Psychology Biological sciences/Psychology/Human behaviour children social robots body emotions interaction Figures Figure 1 Figure 2 Figure 3 Introduction During the early stages of their development, children rely on their social learning abilities to acquire essential skills including language and various social behaviours such as initiating conversations, following rules, seeking, and providing help or information, taking turns, and collaborating [ 1 ]. There are different developmental trajectories in acquiring these social skills, and some characteristics of the tutors can influence these trajectories, such as gender [ 2 ], age [ 2 ], familiarity [ 3 ], perceived degree of expertise [ 4 ] or language [ 5 ]. Understanding these dynamics and preferences is crucial for gaining insights into children's social learning processes and how they engage with their social environment. In recent years, there has been a significant increase in the availability of technological media for learning, ranging from computers, tablets, and smartphones to robots. Among these technologies, social robots have gained particular attention. Social robots are designed to interact and communicate with human individuals, either semi-autonomously or autonomously, meaning they can operate with or without real-time human control [ 6 ]. These robots can follow behavioural norms that are typical for human interaction [ 7 ], with their primary objective being to engage in a direct, effective and helpful interaction with people [ 8 ]. Compared to other forms of digital modalities, social robots facilitate interaction with the physical environment and allow for natural engagement with humans, utilizing methods such as touch, gestures and voice [ 9 ]. A key challenge lies in determining whether social robots can serve as relevant tutors for young children. While robots are not designed to replace humans, comparisons between robot and human tutors are useful to explore areas where robots can complement human instructions [ 6 ], and therefore how they can be used as tools in some areas such as teaching and learning. However, there is considerable heterogeneity in terms of thematic areas, experimental protocols and types of robots employed in this context ([6,10 for reviews]). Among the social robots used as teachers, the humanoid robot NAO has emerged as the most frequently utilized (in 48% of 89 articles reviewed by Belpaeme et al., [ 10 ]). Thus, there is a growing interest in examining the potential of social robots, particularly NAO, with the aim of understanding their effectiveness in facilitating learning experiences for children. Social robots hold promises as tools to enhance embodied learning experiences [ 11 ]. The embodied cognition model emerged in the early 1990s and proposes a strong connection between physical actions and cognitive processes, asserting that cognition is cognition is inherently grounded in the body. This perspective highlights the importance of both the body and its interactions with the physical environment in shaping cognition. Moreover, embodied cognition theory emphasizes the significance of bodily and emotional engagement in the learning process, a dimension an aspect frequently neglected in conventional curriculum design [ 12 ]. We conducted a developmental study with children aged from 3 to 6 years old in pre-schools. The aim of this study was to assess how young children, with typical development, would interact with a social robot across multiple body focused tasks. Children were randomly assigned to the robot group, where the demonstrator was the NAO robot, portrayed with a feminine gender. The remaining children were also randomly assigned to the human group, where the demonstrator was an unfamiliar woman. We chose the NAO robot because of its articulate humanoid body that offers the possibility to evaluate body parts knowledge and exhibit emotional key postures [ 13 , 14 ]. According to Raimo and colleagues [ 15 ], bodily representations can be distinguished into three categories. The first, referred to as "body schema" is a dynamic representation of body parts linked both to motor and sensory information received by the individual [ 16 , 17 ]. Indeed, body schema has an important role in setting body movements and in posture control [ 18 ]. The second is termed as the visuo-spatial body map or structural representation of the body. It's a topographical representation of the body primarily derived from visual information concerning the boundaries between each body part and the spatial relationships of proximity [17, 19]. Lastly, the third is the lexicosemantic representation of the body, which includes the names of body parts, their function, and their connection to objects [ 16 , 17 ]. The Bergès and Lézine somatognosia test is frequently used by psychomotor therapists to evaluate knowledges of 34 body part labels in children aged from 3 to 6 years [ 20 ]. The child must point to the body parts named by the adult and then name the body parts pointed out by the adult, both on their own body and on the adult's body. Children show better performance in pointing than in naming, and their abilities improve from ages 3 to 6 in both naming and pointing tasks. Children aged from 2.2 to 3.4 years old can identify and point to 15 to 33 body parts out of 50 tested [ 21 ]. Some of the most recognized body parts are the tongue, the foot, the head, the back, the belly and the mouth. The least known body parts are the temple, the thigh, the armpit and the palm of the hand. A child's knowledge of a body part depends primarily on its location: facial parts and broader body parts are known first, while joints are the least well-identified. It is also linked to the number of sensory representations (but not motor ones) of each body part in the human cortex and the frequency with which it is named by adults [ 21 ]. As in the Bergès and Lézine test, typically developing children of 4.5 years old perform better in the comprehension task of body parts labels (70.44% of correct answers) than in the production task of body parts labels (56.76% of correct answers) [ 22 ]. Body parts knowledge is mostly tested using imitation tasks from a robot and has been mainly assessed to children with Autism Spectrum Disorder [ 23 , 24 , 25 ]. Very few studies compared ASD children performances to typically developing children (aged from 2 to 3.6 years old) whether the demonstrator is a robot or an adult [ 26 ]. In the present study, we decided to evaluate typically developing children performances aged from 3 to 6 years old. We evaluated the comprehension of body parts labels from the child’s body and from the demonstrator’s body, then, we assessed the production of body parts labels following the procedure described by Bergès and Lézine [ 20 ]. We hypothesized that children would perform equally well whether they were facing a NAO robot or an adult, but that their performance would improve with age. Then, we observed the imitation of sequences of pointed-body parts. In their study, Suzuki and colleagues [ 25 ] conducted a comparison between ASD children imitating either the NAO robot or an adult during a dance activity centered around pointing to "head, shoulders, knees, and clap". Their findings revealed that ASD children demonstrated similar levels of imitation whether they were mimicking the robot or the adult. We hypothesized that children would perform equally well whether they were facing a NAO robot or an adult, but that their performance would improve with age. The body is also involved in the expression of emotions. Faces and bodies share common physical properties and convey similar social and emotional information, implying that the encoding of postures and faces follows a common process [27]. For example, when someone is angry, the muscles in their face -especially around the eyebrows- and body -particularly the shoulders and arms- tense up. By the age of 3, children are able to associate facial expressions with the names of the primary six emotions: happiness, sadness, anger, fear, surprise, disgust [ 28 ]. And children aged from 3 to 6 years old are able to name the 6 primary emotions depicted in photographs, both expressed by the face and the body [27]. Concerning emotional body postures only, 9-year-old typically developing children have the same capabilities in recognizing the six primary emotions [ 29 ]. The skills to recognize and name other’s emotions holds even greater significance during childhood, as it is during this period that initial social interactions take place, often before speech has fully developed [30]. Recognition and expression of emotions are two intertwined skills. Indeed, during face-to-face interactions, an individual recognizes the specific emotion expressed by their interlocutor by looking at them, then recreates the emotion through a process of imitation [31]. Only a limited number of studies have examined emotion recognition in preschool children, revealing that the capacity to interpret emotions from facial photographs tends to increase with age, from 3 to 6 years old [32, 33]. Beck and colleagues [ 13 , 14 ] explored the identification of the NAO robot’s emotional postures – defined as static body postures – by adults and 11-to-13-years-old adolescents. They observed that both adults and adolescents succeeded in recognizing basic emotions in the postures of the robot. They concluded that the lack of facial expressions was not a barrier to recognize emotional key postures in some social robots such as the NAO one, which face has no internal movements. In another study, Cohen and colleagues [34] found that children from 8-to-9 years old were also able to identify the NAO robot’s emotional key postures for happiness, sadness, anger, fear and surprise. In the present study, we decided to evaluate the most recognized NAO key postures in Beck and colleagues’ study (corresponding to anger, sadness, fear, happiness) with younger children (i.e., in preschool children). As mentioned in previous studies on human faces [32, 33], we hypothesized that older children would have better scores than younger one, regardless of whether they interacted with the robot or the human demonstrator. Results Comprehension of body parts labels The number of correct body identifications on the child’s body did not vary with our predictors (age (t(57) = -1.22, p = 0.22), type of demonstrator (t(57) = 1.79, p = 0.07), gender ((t(57) = 0.01, p = 0.98), interaction age x demonstrator: t(57) =-1.75, p = 0.08, Fig. 1 A). The number of correct body identifications on the demonstrator’s body increased with age ((t(57) = -2.68, p = 0.001) but it did not vary with other predictors (type of demonstrator (t(57) = 0.37, p = 0.71), gender ((t(57) = 0.41, p = 0.68), interaction age x demonstrator: t(57) =-0.40, p = 0.69, Fig. 1 B). Imitation of body parts sequences The number of body parts correctly labelled (production of body parts labels), significantly increased with age (t(57) = − 3.13, p < 0.001) whereas it did not vary with the type of demonstrator (t(57) = 0.35, p = 0.72), the gender (t(57) = 1.69, p = 0.09) or the interaction age x demonstrator (t(57) =-0.04, p = 0.96, Fig. 1 C). The number of body parts correctly imitated in sequences by children significantly increased with age (t(57) = 3.78, p < 0.001) whereas it did not vary with the type of demonstrator (t(57) = 1.70, p = 0.09), the gender ((t(57) = -0.41, p = 0.68) but the effect of age was stronger for the human demonstrator than for the robot demonstrator (interaction t(57) = − 2.06, p = 0.04, Fig. 1 D). Emotions task The number of correct emotions identified on images did not vary with our predictors (age (t(57) =-0.05, p = 0.95, type of demonstrator (t(57) = 1.38, p = 0.17), gender ((t(57) = -0.03, p = 0.97), interaction age x demonstrator: t(57) =-1.57, p = 0.12). The number of emotions expressed by the child did not vary with our predictors (age ((t(57) = 1.63, p = 0.10, type of demonstrator (t(57) = 0.61, p = 0.54), gender ((t(57) = 1.22, p = 0.22), interaction age x demonstrator: t(57) =-0.64, p = 0.51). We observed an effect of children’s ages for identification of emotional key postures: the mean percentage of emotional key postures recognized by younger children was lower compared to older ones ((t(57) = -2.76, p = 0.007) but not with other predictors as type of demonstrator (t(57) = -1.07, p = 0.28), gender ((t(57) = 0.36, p = 0.17), interaction age x demonstrator: t(57) = 1.02, p = 0.31, see Fig. 2 ). Discussion The present study evidenced that children aged between 3 and 6 years exhibited similar reactions and performances when interacting with a NAO robot as they did with an adult. Interestingly, we observed no significant impact of the demonstrator type on the average scores of correct answers. As we expected, the number of correct answers increased with age for most of the tasks we proposed: in the identification of body parts on the demonstrator’s body, in all subtasks of imitation (the number of body parts labels named by the child and the number of body part imitated), in the identification of emotional key postures. Identification of body parts labels on the child’s body was not influenced by any of the predictors and we observed a ceiling effect. Comprehensive repertoires of children from 2.2 years old to 3.4 years old contained between 15 and 33 body parts (median = 27) on the 50 stimuli presented [ 21 ]. The most correctly located body parts were: tongue, foot, head, back, tummy, mouth, belly, eye, teeth, hand, leg, fingers, arm, ear, bottom, hair, neck, knee, thumb, cheek, shoulder, toes, lips, eyebrow, elbow. There was a positive correlation between children’s comprehension and their age. Therefore, it is possible that by the age of 3, body parts evaluated in our experiment are already known by most of the children, which explains the ceiling effect in our study. However, we found that identification of body parts labels on the demonstrator’s body increased with children’s age. Pointing to body parts of another person is different from pointing body parts on his own body. This is in line with studies which evidenced that patients with autotopoagnosia (i.e., impairment at pointing to parts of their own body named by an experimenter) do not necessarily present heterotopoagnosia (i.e., impairment at pointing to parts of their own body named by an experimenter) [35]. Indeed, spatial cognition theories differentiate two reference frames: on one hand, representation used by the brain to convey the layout of points in space related to two distinct neural circuits [36]. On the other hand, the egocentric frame conveys positions of points using the body as the centre of the surrounding space, the allocentric frame concerns the positions of points using external landmarks. If both frames of reference emerge quite early in development, by the year of 5 [37], further studies showed a gradual development from an egocentric to a more stable map-like representation between 5 and 10 years of age [38]. It can explain the improvement with age of the ability to locate body parts on the demonstrators’ body we found in our study. In the imitation task, we first evaluated the ability to produce labels of body parts. We found that older children gave more correct answers than younger ones. It is concordant with previous studies highlighting that this ability develops later in development compared to the comprehension of body parts labels [ 21 ]. Concerning the imitation of body parts sequences, we also found an effect of age and this effect was stronger for the human demonstrator than for the robot demonstrator. We hypothesis that imitation of pointing body parts was easier as the human demonstrator had a more familiar body for the children than the robot body. Regarding the identification of emotions in the body postures task, we first assessed emotions labelling on images. We observed a ceiling effect: scores were very high and not influenced by the predictors. The ceiling effect is concordant with the study of Widen and Russel [ 28 ] which demonstrated that facial expressions are correctly associated with labels “happiness”, “sadness”, “anger”, “fear”, “surprise” and “disgust”, which correspond to the six primary emotions described by Ekman and Friesen [39], by the year of 3. In another study [31], the ability to recognize sadness, anger, joy and fear on very simple drawings of human faces seems to emerge between 4 and 5 years. Concerning emotions expressed by the child, we found no effect of the predictors and an important interindividual variability. We propose that more than the demonstrator type (human or robot), scores were influenced by the emotional expressiveness of children – the capacity to accurately communicate feelings nonverbally - and the tendency to feel comfortable in the presence of unfamiliar adults, variables that we did not measured. Indeed, a meta-analysis showed that more emotional expressive people also have a more extraverted personality (extraverts are outgoing, talkative, impulsive and uninhibited and have many social contacts) [40]. Finally, we assessed labelling of emotions in the demonstrator key postures. We found no effect of the demonstrator type but an improvement with children’s ages. Correct labels of the demonstrator emotional key postures were lower than in the study of Beck and colleagues in adults [ 13 ] and children from 11 to 13 years old [ 14 ]. We made two assumptions about the effect of children’s age. First, it had been pointed out that emotion labelling implies a more complex cognitive process than emotion recognition, which can be fully mature later in the child’s development [ 41 ]. Secondly, even if emotional key postures can be accurately recognised without any context [ 13 , 14 ], it had been demonstrated that knowledge about the social and environmental context in which the key postures of a teddy bear robot are expressed improve the success of identification [42]. We propose that the context might be more important for younger children than for older children and adults as they are still learning to recognize emotions. Taken together, our study provides encouraging results concerning the use of the social robot NAO with children between 3 and 6 years old to evaluate children knowledge in several domains, from vocabulary learning to prosocial behaviours. This type of social robot is interesting as tools to develop children’s socio-cognitive skills and further research are needed to understand how young children ’s representation of this type of social partner develops through early childhood. Methods Selection and Participation Seventy-three children aged 3 years and 3 months to 6 years and 1 month were evaluated. All children were French native speakers and came from 4 different schools. We obtained parental consent for all children to take part in the study and to be recorded. We also collected verbal assent from children before starting the experiment. On 73 children initially recruited, 9 asked to stop before the experiment was over, whilst in the two other cases, there were technical issues with the camera. In total, 62 child-demonstrator interactions were completed in a between-subject procedure (school A: 7 children, school B: 10 children, school C: 17 children, school D: 28 children). Children were separated into three groups of age: Group 1, “G1”: from 3 years old and one month to 4 years old and one month (mean ± SE = 3.73 ± 0.07); Group 2, “G2”: from four years-old and two months to five years-old and 1 month (mean ± SE = 4.71 ± 0.06) ; Group 3, “G3”: from five years-old and two months to six years-old and one month (mean ± SE = 5.65 ± 0.059). As much as possible, we balanced experimental groups concerning the age and gender of the children. Children were randomly assigned to the experimental groups. Our sample is composed of 62 children (30 in the robot group, 32 in the human group). The Group 1 is composed of 14 children: 8 girls (4 with in the robot group, 4 in the human group) and 6 boys (2 in the robot group, 4 in the human group). The Group 2 is composed of 26 children: 13 girls (7 with in the robot group, 6 in the human group) and 13 boys (6 in the robot group, 7 in the human group). The Group 3 is composed of 22 children: 14 girls (6 with in the robot group, 8 in the human group) and 8 boys (5 in the robot group, 3 in the human group). The protocol was carried out in accordance with the ethical standards of the Declaration of Helsinki (BMJ 1991; 302:1194) and approved by the Ethic Committee of the Department of Psychology of the (CER-PN n°2022-09-01). Experimental procedure General setup The experimental sessions were conducted in a quiet room at the school in the presence of the experimenter and the demonstrator (Fig. 3 ). In each of the four schools, one adult female experimenter was recruited and trained to adhere to the experimental setup. No child had been exposed to the NAO robot prior to the experiment. The experimenter asked the child to sit on the floor in front of the demonstrator and introduced the demonstrator to him/her. The demonstrator was a female adult for half of the children and the NAO robot for the other half. The demonstrator gave instructions to the child. Due to the COVID-19 situation, the experimenter and the human demonstrator wore face masks during the whole experiments. Each individual session lasted approximately 20 min. Two Sony HDR-CX410VE cameras placed on tripods from either side of the demonstrator were used to record the child’s behaviours. Human group: exposure to an adult demonstrator The human demonstrator was always the same adult female and remained as neutral as possible during the experiment: she gave the instructions to the child, but she did not encourage him/her or give him/her feedback. Robot group: exposure to the NAO robot demonstrator Created by Aldebaran Robotics, the NAO robot version 6 is 58 cm high and bipedal. It has 25 degrees of freedom which allows a lot of possibilities of movements. It can manipulate small objects with its three fingers - hands. NAO has two video cameras on its forehead and its mouth. It has vocal capacities both for recognition and synthesis: it is equipped with a stereo broadcast system made up of two loudspeakers, on its ears and 4 omnidirectional microphones (2 on the top of its head and 2 on the back of its head). It also has two ultrasonic sensors (or sonars) which allows it to estimate the distance to obstacles in its environment. Moreover, it has contact and tactile sensors: tactile head and hands, chest button and feet bumpers. The NAO robot can be programmed to carry out autonomously a set of tasks. However, for our experiment, the robot was fully tele-operated to enable the robot to act contingently [43]: the experimenter controlled the robot during the session, via a touch- screen tablet, an iPad mini 4 (20.32 x 13.48 x 0.61 cm). The setup involved connecting the tablet to the robot’s wifi access point. The robot hosted a web server that could be accessed through any browser on the tablet. An HTML page was used to the interface and execute the experiment's sequence using javascript (« vue.js » for the graphical interface, « LibQi » for controlling the robot). As the human demonstrator was a female, we attributed a feminine gender to the NAO robot and a feminine name (“Naomie”). We will use the pronoun “her” in the following text. To ensure that the two demonstrators pronounced the same sentences in both conditions, vocal recordings of the female adult were played back by the robot. Sentences were recorded while she was alone in a silent room and as she was addressing instructions to a child. Experimental tasks and scores Before starting the experimental tasks, there was a warm-up phase to allow the child to see how the demonstrator, robotic or human depending on the group, spoke and moved before the beginning of the experiment. The demonstrator said her name to the child and asked the child to introduce herself/himself. Then, she sang a popular rhyme with hand gestures and encouraged the child to follow it. Then, each child was tested in different situations: 1) the comprehension of body parts labels on their own body and on the demonstrator’s body, 2) the production of body parts labels and an imitation task and 3) the identification of emotional key postures. Comprehension of body parts labels: We evaluated the comprehension of eleven body parts labels: face, eye, nose, mouth, ear, shoulder, elbow, hand, belly, knee, foot. Each body part was randomly chosen in this list. First, the demonstrator asked the child “Show me your [body part]” and repeated it for the eleven possibilities. Then she asked, “Show me my [body part]” and repeated it for the eleven body parts. Imitation of body parts sequences: We evaluated body parts labels production: the demonstrator randomly showed one of her body parts, for example, her eye and asked “What is the name of this body part?” and a second one, for example her belly, and said “And this one?”. As it was complicated for the NAO robot to point her nose and its ear, we decided to keep 9 body parts for this sub-task (face, eye, mouth, shoulder, elbow, hand, belly, knee and foot in Supplementary Information, Fig. 1 ). Then, the demonstrator placed her hands on her knees and asked the child to repeat the body parts sequence. The demonstrator performed two sequences with two body parts (“sequence#1”, “sequence#2”) and a third sequence with three body parts (“sequence#3”). As the majority of children had difficulties to understand the instructions in this task, the first two-body parts sequence (sequence#1) was used by the experimenter to explain the instructions again in order to help the child to understand it. Emotions task Four images of a child expressing an emotion (joy, sadness, anger, fear. Supplementary Information, Fig. 2 ) were placed in front of the child. The demonstrator asked “Do you know what an emotion is? Show me the image in which the child feels [emotion].” Then, the demonstrator and the child stood up. The demonstrator asked “Show me how do you express [emotion] with your whole body”. She waited for the child to express the emotion. The instruction was repeated with the four emotions. As the NAO robot does not have facial expressions, we investigated emotions recognition when emotions were expressed with the body. We selected the four key postures exhibited by the NAO robot that were the most successfully identified by adults and children in previous studies [ 13 , 14 ]: joy, fear, anger and sadness. The demonstrator struck each of these four key postures and between each pose, she went back to a neutral pose (Supplementary Information, Fig. 3 ). For this task, the demonstrator said, “Let me show you what I am doing when I am feeling an emotion”. The demonstrator reproduced one of the four emotional postures and the experimenter asked the child “Which emotion do you think it is?”. After the child replied, the demonstrator returned to the neutral pose. She repeated this sequence for the three other body postures. At the end of the session, the demonstrator thanked the child for his/her participation and told him/her goodbye. Video recordings and analysis All sessions were video recorded. Cameras allowed to have a global view of the setup which encompassed the face and the body of the child. Informed written consents were obtained from the human demonstrator and from the parents of the children for video recordings and to publish their information/images in an online open-access publication. Statistical analysis We conducted models trying to predict each variable with the demonstrator type (robot or human), the child’s age, the child’s age in interaction with the demonstrator’s type and the child’s gender as predictors. We started each model including school and experimenter as random effects of a GLME. We adapted the link and distributions to increase the model goodness-of-fit using AIC and % of variance explained (squared R) as indicators. We removed high p-values factors when it increased the goodness-of-fit and switched to a GLM (without random effects), following the same rationale (Table 1). We conducted specific analysis for the imitation of body parts sequences: we decided to remove sequence#1 from the statistical analysis as it was necessary to use it to explain again instructions to children. We kept the second two-body parts sequence (sequence#2) and the three-body parts sequence only (sequence#3). As results in sequence 2 and 3 followed the same patterns, we calculated a number of correct answers adding answers from sequence 2 and 3 and we conducted GLM on these data. Table 1 summarizes the statistical model and is available as a supplementary material. Declarations Author Contribution A.A. collected and analyzed the data. B.G. and S.D. co-supervised the study. A.C. conducted the statistical analyzes. A.B. programmed the robot. A.A., B.G., and A.C. participated equally to te writing of the manuscript. Acknowledgement We would like to thank all the children who participated as well as schools’ directors who coordinated the study. We also thank the Institut Universitaire de France (IUF) and the Descartes Program from CNRS@Create for their support. Data Availability We declare that data collected will be available upon request to Alice Araguas. References Aksoy, P., & Baran, G. (2010). Review of studies aimed at bringing social skills for children in preschool period. Procedia - Social and Behavioral Sciences, 9, 663 669. https://doi.org/10.1016/j.sbspro.2010.12.214 Perloff, R. M. (1982). Gender constancy and same-sex imitation: A developmental study. The Journal of Psychology: Interdisciplinary and Applied, 111(1), 81‑86. https://doi.org/10.1080/00223980.1982.9923516 Olson, K. R., & Spelke, E. S. (2008). Foundations of cooperation in young children. Cognition, 108(1), 222‑231. https://doi.org/10.1016/j.cognition.2007.12.003 Zmyj, N., Buttelmann, D., Carpenter, M., & Daum, M. M. (2010). The reliability of a model influences 14-month-olds’ imitation. Journal of Experimental Child Psychology, 106(4), 208‑220.https://doi.org/10.1016/j.jecp.2010.03.002 Kinzler, K. D., Dupoux, E., & Spelke, E. S. (2007). The native language of social cognition. Proceedings of the National Academy of Sciences of the United States of America, 104(30), 12577‑12580. https://doi.org/10.1073/pnas.0705345104 van den Berghe, R., Verhagen, J., Oudgenoeg-Paz, O., van der Ven, S., & Leseman, P. (2019). Social Robots for Language Learning : A Review. Review of Educational Research, 89(2), 259-295. https://doi.org/10.3102/0034654318821286 Bartneck, C., & Forlizzi, J. (2004). A design-centred framework for social human-robot interaction. 591‑594. https://doi.org/10.1109/ROMAN.2004.1374827 Sheridan, T. B. (2020). A review of recent research in social robotics. Current Opinion in Psychology. doi: 10.1016/j.copsyc.2020.01.003. Robinson, N.L., Cottier, T.V., Kavanagh, D.J. (2019); Psychosocial health interventions by social robots: Systematic review of randomized controlled trials. Journal of Medical Internet Research 21(5): e13203. Belpaeme, T., Kennedy, J., Ramachandran, A., Scassellati, B., & Tanaka, F. (2018). Social robots for education : A review. Science Robotics, 3(21), eaat5954.https://doi.org/10.1126/scirobotics.aat5954 Kim, Y., & Tscholl, M. (2021). Young children’s embodied interactions with a social robot. Educational Technology Research and Development, 69, 2059–2081. https://doi.org/10.1007/s11423-021-09978-3 Glenberg, A. M. (2010). Embodiment as a unifying perspective for psychology. Wiley Interdisciplinary Reviews: Cognitive Science, 1, 586–596. Beck, A., Cañamero, L., & Bard, K. (2010). Towards an Affect Space for robots to display emotional body language. 464‑469. https://doi.org/10.1109/ROMAN.2010.5598649 Beck, A., Cañamero, L., Hiolle, A., Damiano, L., Cosi, P., Tesser, F., & Sommavilla, G. (2013). Interpretation of Emotional Body Language Displayed by a Humanoid Robot : A Case Study with Children. International Journal of Social Robotics, 5(3), 325‑334.https://doi.org/10.1007/s12369-013-0193-z Raimo, S., Iona, T., Di Vita, A., Boccia, M., Buratin, S., Ruggeri, F., Iosa, M., Guariglia, C., Grossi, D., & Palermo, L. (2021). The development of body representations in school-aged children. Applied Neuropsychology: Child, 10(4), 327339.https://doi.org/10.1080/21622965.2019.1703704 Schwoebel, J., Boronat, C. B., & Branch Coslett, H. (2002). The man who executed « imagined » movements : Evidence for dissociable components of the body schema. Brain and Cognition, 50(1), 116. https://doi.org/10.1016/s0278-2626(02)00005-2 Schwoebel, J., & Coslett, H. B. (2005). Evidence for multiple, distinct representations of the human body. Journal of Cognitive Neuroscience, 17(4), 543553. https://doi.org/10.1162/0898929053467587 Picelli, A., Negrini, S., Zenorini, A., Iosa, M., Paolucci, S., & Smania, N. (2016). Do adolescents with idiopathic scoliosis have body schema disorders? A cross-sectional study. Journal of Back and Musculoskeletal Rehabilitation, 29(1), 89–96. https://doi.org/10.3233/BMR-150602 Sirigu, A., Grafman, J., Bressler, K., & Sunderland, T. (1991). Multiple representations contribute to body knowledge processing. Evidence from a case of autotopagnosia. Brain: A Journal of Neurology, 114 (Pt 1B), 629642. https://doi.org/10.1093/brain/114.1.629 [20] Bergès, J., & Lézine, I. (1963). Test of imitation of gestures: Methods for plotting body image and movements of children from 3 to 6 years of age. Masson & Cie Camões-Costa, V., Erjavec, M., & Horne, P. J. (2011). The impact of body-part-naming training on the accuracy of imitative performances in 2- to 3-year-old children. Journal of the Experimental Analysis of Behavior, 96(3), 291315. https://doi.org/10.1901/jeab.2011.96-291 Russo, L., Craig, F., Ruggiero, M., Mancuso, C., Galluzzi, R., Lorenzo, A., Fanizza, I., & Trabacca, A. (2018). Exploring Visual Perspective Taking and body awareness in children with Autism Spectrum Disorder. Cognitive Neuropsychiatry, 23(4), 254-265. https://doi.org/10.1080/13546805.2018.1486182 Costa, S., Lehmann, H., Dautenhahn, K., Robins, B., & Soares, F. (2014). Using a Humanoid Robot to Elicit Body Awareness and Appropriate Physical Interaction in Children with Autism. International Journal of Social Robotics, 7, 114. https://doi.org/10.1007/s12369-014-0250-2 Suzuki, R., & Lee, J. (2016). Robot-play therapy for improving prosocial behaviours in children with Autism Spectrum Disorders. 2016 International Symposium on Micro-NanoMechatronics and Human Science (MHS). https://doi.org/10.1109/MHS.2016.7824238 Suzuki, R., Lee, J., & Rudovic, O. (2017). NAO-Dance Therapy for Children with ASD. 295296. https://doi.org/10.1145/3029798.3038354 Cao, H. L., Simut, R. E., Desmet, N., De Beir, A., Van De Perre, G., Vanderborght, B., & Vanderfaeillie, J. (2020). Robot-assisted joint attention: A comparative study between children with autism spectrum disorder and typically developing children in interaction with NAO. IEEE Access, 8, 223325-223334. Parker, A. E., Mathis, E. T., & Kupersmidt, J. B. (2013). How is this child feeling? Preschool-aged children’s ability to recognize emotion in faces and body poses. Early education and development, 24(2), 188-211. https://doi.org/10.1080/10409289.2012.657536 [28] Widen, S. C., & Russell, J. A. (2003). A closer look at preschoolers’ freely produced labels for facial expressions. Developmental Psychology, 39(1), 114-128. https://doi.org/10.1037//0012- 1649.39.1.114 Peterson, C. C., Slaughter, V., & Brownell, C. (2015). Children with autism spectrum disorder are skilled at reading emotion body language. Journal of Experimental Child Psychology, 139, 35-50. https://doi.org/10.1016/j.jecp.2015.04.012 Izard, C., Fine, S., Schultz, D., Mostow, A., Ackerman, B., & Youngstrom, E. (2001). Emotion Knowledge as a Predictor of Social Behavior and Academic Competence in Children at Risk. Psychological Science, 12(1), 18-23. https://doi.org/10.1111/1467-9280.00304 Denham, S. A., Blair, K. A., DeMulder, E., Levitas, J., Sawyer, K., Auerbach–Major, S., & Queenan, P. (2003). Preschool Emotional Competence : Pathway to Social Competence? Child Development, 74(1), 238-256. https://doi.org/10.1111/1467-8624.00533 Covic, A., von Steinbüchel, N., & Kiese-Himmel, C. (2020). Emotion recognition in kindergarten children. Folia Phoniatrica et Logopaedica:International Journal of Phoniatrics, Speech Therapy and Communication Pathology, 72(4), 273–281. https://doi.org/10.1159/000500589 Schneider, J., Sandoz, V., Equey, L., Williams-Smith, J., Horsch, A., Bickle Graz M. (2022). The Role of Face Masks in the Recognition of Emotions by Preschool Children. JAMA Pediatr. doi: 10.1001/jamapediatrics.2021.4556. PMID: 34779832; PMCID: PMC8593832. Cohen, I., Looije, R., & Neerincx, MA. (2012). Child's recognition of emotions in robot's face and body. In A. Billard et al (Ed.), Proceedings of the 6th international conference on Human-robot interaction (pp. 123-124). ACM DL. https://doi.org/10.1145/1957656.1957692 Pisella, L., Havé, L., & Rossetti, Y. (2019). Body awareness disorders : Dissociations between body- related visual and somatosensory information. Brain : a journal of neurology, 142, 2170‑2173. https://doi.org/10.1093/brain/awz187 Klatzky, R. L. (1998). Allocentric and Egocentric Spatial Representations : Definitions, Distinctions, and Interconnections. In C. Freksa, C. Habel, & K. F. Wender (Éds.), Spatial Cognition : An Interdisciplinary Approach to Representing and Processing Spatial Knowledge (p. 117). Springer. https://doi.org/10.1007/3-540-69342-4_1 Newcombe, N., & Huttenlocher, J. (2003). Making Space : The Development of Spatial Representation and Reasoning. MIT Press. Bullens, J., Iglói, K., Berthoz, A., Postma, A., & Rondi-Reig, L. (2010). Developmental time course of the acquisition of sequential egocentric and allocentric navigation strategies. Journal of Experimental Child Psychology, 107(3), 337350. https://doi.org/10.1016/j.jecp.2010.05.010 Ekman, P., & Friesen, W. V. (1976). Measuring facial movement. Environmental Psychology and Nonverbal Behavior, 1(1), 5675. https://doi.org/10.1007/BF01115465 Riggio, H. R., & Riggio, R. E. (2002). Emotional Expressiveness, Extraversion, and Neuroticism : A Meta-Analysis. Journal of Nonverbal Behavior, 26(4), 195-218. https://doi.org/10.1023/A:1022117500440 Izard, C. E. (1971). The face of emotion. New York: Appleton Century-Crofts Educational Division, Meredith. Li, J., & Chignell, M. (2011). Communication of Emotion in Social Robots through Simple Head and Arm Movements. International Journal of Social Robotics, 3(2), 125142. https://doi.org/10.1007/s12369-010-0071-x Kennedy, J., Lemaignan, S., Montassier, C., Lavalade, P., Irfan, B., Papadopoulos, F., Senft, E., & Belpaeme, T. (2017). Child Speech Recognition in Human-Robot Interaction : Evaluations and Recommendations. Proceedings of the 2017 ACM/IEEE International Conference on HumanRobot Interaction, 82-90. https://doi.org/10.1145/2909824.3020229 Additional Declarations No competing interests reported. Supplementary Files SupplementarymaterialAAACABSDBG.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4758583","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":338158520,"identity":"96a7a4d9-e6d5-47b2-b754-c1d5d620ecee","order_by":0,"name":"Alice Araguas","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+0lEQVRIiWNgGAWjYDACdsYGJF4FEDMzN2BXCgPMKFrOYIhg04LMYWwDk/i18DczN374wGCXz9+/xvDjz3m10fztQC0/Krbh1CJxmLFZcgZDsuWMG2+MpXm3Hc+dcZixgbHnzG3c1hxmbGPmYWA2YLhxxkCacdux3AagFmbGNtxa5EFa/jDUG8jfOGP88+ecY7nzCWkxAGkB2mVgcL7HTIK3oSZ3AyEthiC/9BgcNzC8wVZmzXPsQO5GoJaD+Pwid7z94YcfFdUGcucPb775o6Yud975wwcf/KjA432I84BYIgEcGmD+AQLqoYAfrK6OOMWjYBSMglEwogAAZfFZrckdTt0AAAAASUVORK5CYII=","orcid":"","institution":"Laboratoire Ethologie Cognition Développement","correspondingAuthor":true,"prefix":"","firstName":"Alice","middleName":"","lastName":"Araguas","suffix":""},{"id":338158523,"identity":"eb8e3699-dfaa-4eba-9e2e-4a51ee4ed1e4","order_by":1,"name":"Adrien Chopin","email":"","orcid":"","institution":"Smith-Kettlewell Eye Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Adrien","middleName":"","lastName":"Chopin","suffix":""},{"id":338158529,"identity":"45f45752-a230-4be0-8338-69a9f99dcd46","order_by":2,"name":"Arnaud Blanchard","email":"","orcid":"","institution":"University Cergy","correspondingAuthor":false,"prefix":"","firstName":"Arnaud","middleName":"","lastName":"Blanchard","suffix":""},{"id":338158532,"identity":"8319147a-cff9-4057-a2d4-b8a9ed33831a","order_by":3,"name":"Sébastien Derégnaucourt","email":"","orcid":"","institution":"Laboratoire Ethologie Cognition Développement","correspondingAuthor":false,"prefix":"","firstName":"Sébastien","middleName":"","lastName":"Derégnaucourt","suffix":""},{"id":338158533,"identity":"fed077f2-e0c0-41c9-a695-6aea4de2279a","order_by":4,"name":"Bahia Guellai","email":"","orcid":"","institution":"Laboratoire Cognition, Langues, Langage, Ergonomie","correspondingAuthor":false,"prefix":"","firstName":"Bahia","middleName":"","lastName":"Guellai","suffix":""}],"badges":[],"createdAt":"2024-07-17 19:56:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4758583/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4758583/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":62661091,"identity":"57b05341-ab49-4e34-9fc3-236264e28a56","added_by":"auto","created_at":"2024-08-17 02:38:20","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":158237,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of correct answers (in %) in the body parts’ task depending on the age of the child for the robot group (blue circle) and the human group (red triangle) on: (A) the child’s body and (B) on the demonstrator’s body. (C) Distribution of mean correct answers (in %) in naming body parts depending on the age of the child for the robot group (blue circle) and the human group (red triangle). (D) Distribution of mean correct answers (in %) in imitating body parts depending on the age of the child for the robot group (blue circle) and the human group (red triangle).\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4758583/v1/265ae17994cb9c3daae8b27d.png"},{"id":62661088,"identity":"ef462e40-4a2d-4601-a60a-c1d8ae156326","added_by":"auto","created_at":"2024-08-17 02:38:20","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":23348,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of the percentage of emotional postures correctly identified depending on the age of the child for the robot group (blue circle) and the human group (red triangle).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4758583/v1/ae8c6a1bfa8b840f42c4e29d.png"},{"id":62661728,"identity":"b5ffee7f-d518-4932-8653-9701f4c03682","added_by":"auto","created_at":"2024-08-17 02:46:20","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":44990,"visible":true,"origin":"","legend":"\u003cp\u003eTop view of the experimental setup for the two groups: (A) with the robotic demonstrator and (B) with the human demonstrator (inspired by Kennedy et al., [20]). The child and the demonstrator sat in front of each other and the experimenter was on the right side of the demonstrator.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4758583/v1/b112d330abc0989c31f31390.jpeg"},{"id":74448435,"identity":"6e003e61-4511-4a8c-987b-792c2ba27082","added_by":"auto","created_at":"2025-01-22 11:31:06","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":603012,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4758583/v1/99c0f41d-748d-4d7d-92f6-ca5fc60e820a.pdf"},{"id":62661089,"identity":"107f2cca-2a5d-4bc7-865e-2c6cf29a0b5c","added_by":"auto","created_at":"2024-08-17 02:38:20","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":297996,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementarymaterialAAACABSDBG.docx","url":"https://assets-eu.researchsquare.com/files/rs-4758583/v1/671737e44104459e50c36207.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Child-robot interactions in different body and emotions oriented tasks: comparison with a human partner","fulltext":[{"header":"Introduction","content":"\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eDuring the early stages of their development, children rely on their social learning abilities to acquire essential skills including language and various social behaviours such as initiating conversations, following rules, seeking, and providing help or information, taking turns, and collaborating [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. There are different developmental trajectories in acquiring these social skills, and some characteristics of the tutors can influence these trajectories, such as gender [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], age [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], familiarity [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], perceived degree of expertise [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] or language [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Understanding these dynamics and preferences is crucial for gaining insights into children's social learning processes and how they engage with their social environment.\u003c/p\u003e \u003cp\u003eIn recent years, there has been a significant increase in the availability of technological media for learning, ranging from computers, tablets, and smartphones to robots. Among these technologies, social robots have gained particular attention. Social robots are designed to interact and communicate with human individuals, either semi-autonomously or autonomously, meaning they can operate with or without real-time human control [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. These robots can follow behavioural norms that are typical for human interaction [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], with their primary objective being to engage in a direct, effective and helpful interaction with people [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Compared to other forms of digital modalities, social robots facilitate interaction with the physical environment and allow for natural engagement with humans, utilizing methods such as touch, gestures and voice [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. A key challenge lies in determining whether social robots can serve as relevant tutors for young children. While robots are not designed to replace humans, comparisons between robot and human tutors are useful to explore areas where robots can complement human instructions [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], and therefore how they can be used as tools in some areas such as teaching and learning. However, there is considerable heterogeneity in terms of thematic areas, experimental protocols and types of robots employed in this context ([6,10 for reviews]). Among the social robots used as teachers, the humanoid robot NAO has emerged as the most frequently utilized (in 48% of 89 articles reviewed by Belpaeme et al., [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]). Thus, there is a growing interest in examining the potential of social robots, particularly NAO, with the aim of understanding their effectiveness in facilitating learning experiences for children.\u003c/p\u003e \u003cp\u003eSocial robots hold promises as tools to enhance embodied learning experiences [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The embodied cognition model emerged in the early 1990s and proposes a strong connection between physical actions and cognitive processes, asserting that cognition is cognition is inherently grounded in the body. This perspective highlights the importance of both the body and its interactions with the physical environment in shaping cognition. Moreover, embodied cognition theory emphasizes the significance of bodily and emotional engagement in the learning process, a dimension an aspect frequently neglected in conventional curriculum design [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eWe conducted a developmental study with children aged from 3 to 6 years old in pre-schools. The aim of this study was to assess how young children, with typical development, would interact with a social robot across multiple body focused tasks. Children were randomly assigned to the robot group, where the demonstrator was the NAO robot, portrayed with a feminine gender. The remaining children were also randomly assigned to the human group, where the demonstrator was an unfamiliar woman. We chose the NAO robot because of its articulate humanoid body that offers the possibility to evaluate body parts knowledge and exhibit emotional key postures [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAccording to Raimo and colleagues [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], bodily representations can be distinguished into three categories. The first, referred to as \"body schema\" is a dynamic representation of body parts linked both to motor and sensory information received by the individual [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Indeed, body schema has an important role in setting body movements and in posture control [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. The second is termed as the visuo-spatial body map or structural representation of the body. It's a topographical representation of the body primarily derived from visual information concerning the boundaries between each body part and the spatial relationships of proximity [17, 19]. Lastly, the third is the lexicosemantic representation of the body, which includes the names of body parts, their function, and their connection to objects [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. The Berg\u0026egrave;s and L\u0026eacute;zine somatognosia test is frequently used by psychomotor therapists to evaluate knowledges of 34 body part labels in children aged from 3 to 6 years [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. The child must point to the body parts named by the adult and then name the body parts pointed out by the adult, both on their own body and on the adult's body. Children show better performance in pointing than in naming, and their abilities improve from ages 3 to 6 in both naming and pointing tasks. Children aged from 2.2 to 3.4 years old can identify and point to 15 to 33 body parts out of 50 tested [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Some of the most recognized body parts are the tongue, the foot, the head, the back, the belly and the mouth. The least known body parts are the temple, the thigh, the armpit and the palm of the hand. A child's knowledge of a body part depends primarily on its location: facial parts and broader body parts are known first, while joints are the least well-identified. It is also linked to the number of sensory representations (but not motor ones) of each body part in the human cortex and the frequency with which it is named by adults [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. As in the Berg\u0026egrave;s and L\u0026eacute;zine test, typically developing children of 4.5 years old perform better in the comprehension task of body parts labels (70.44% of correct answers) than in the production task of body parts labels (56.76% of correct answers) [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBody parts knowledge is mostly tested using imitation tasks from a robot and has been mainly assessed to children with Autism Spectrum Disorder [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Very few studies compared ASD children performances to typically developing children (aged from 2 to 3.6 years old) whether the demonstrator is a robot or an adult [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. In the present study, we decided to evaluate typically developing children performances aged from 3 to 6 years old. We evaluated the comprehension of body parts labels from the child\u0026rsquo;s body and from the demonstrator\u0026rsquo;s body, then, we assessed the production of body parts labels following the procedure described by Berg\u0026egrave;s and L\u0026eacute;zine [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. We hypothesized that children would perform equally well whether they were facing a NAO robot or an adult, but that their performance would improve with age. Then, we observed the imitation of sequences of pointed-body parts. In their study, Suzuki and colleagues [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] conducted a comparison between ASD children imitating either the NAO robot or an adult during a dance activity centered around pointing to \"head, shoulders, knees, and clap\". Their findings revealed that ASD children demonstrated similar levels of imitation whether they were mimicking the robot or the adult. We hypothesized that children would perform equally well whether they were facing a NAO robot or an adult, but that their performance would improve with age.\u003c/p\u003e \u003cp\u003eThe body is also involved in the expression of emotions. Faces and bodies share common physical properties and convey similar social and emotional information, implying that the encoding of postures and faces follows a common process [27]. For example, when someone is angry, the muscles in their face -especially around the eyebrows- and body -particularly the shoulders and arms- tense up. By the age of 3, children are able to associate facial expressions with the names of the primary six emotions: happiness, sadness, anger, fear, surprise, disgust [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. And children aged from 3 to 6 years old are able to name the 6 primary emotions depicted in photographs, both expressed by the face and the body [27]. Concerning emotional body postures only, 9-year-old typically developing children have the same capabilities in recognizing the six primary emotions [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. The skills to recognize and name other\u0026rsquo;s emotions holds even greater significance during childhood, as it is during this period that initial social interactions take place, often before speech has fully developed [30]. Recognition and expression of emotions are two intertwined skills. Indeed, during face-to-face interactions, an individual recognizes the specific emotion expressed by their interlocutor by looking at them, then recreates the emotion through a process of imitation [31].\u003c/p\u003e \u003cp\u003eOnly a limited number of studies have examined emotion recognition in preschool children, revealing that the capacity to interpret emotions from facial photographs tends to increase with age, from 3 to 6 years old [32, 33]. Beck and colleagues [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] explored the identification of the NAO robot\u0026rsquo;s emotional postures \u0026ndash; defined as static body postures \u0026ndash; by adults and 11-to-13-years-old adolescents. They observed that both adults and adolescents succeeded in recognizing basic emotions in the postures of the robot. They concluded that the lack of facial expressions was not a barrier to recognize emotional key postures in some social robots such as the NAO one, which face has no internal movements. In another study, Cohen and colleagues [34] found that children from 8-to-9 years old were also able to identify the NAO robot\u0026rsquo;s emotional key postures for happiness, sadness, anger, fear and surprise. In the present study, we decided to evaluate the most recognized NAO key postures in Beck and colleagues\u0026rsquo; study (corresponding to anger, sadness, fear, happiness) with younger children (i.e., in preschool children). As mentioned in previous studies on human faces [32, 33], we hypothesized that older children would have better scores than younger one, regardless of whether they interacted with the robot or the human demonstrator.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eComprehension of body parts labels\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe number of correct body identifications on the child\u0026rsquo;s body did not vary with our predictors (age (t(57) = -1.22, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.22), type of demonstrator (t(57)\u0026thinsp;=\u0026thinsp;1.79, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.07), gender ((t(57)\u0026thinsp;=\u0026thinsp;0.01, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.98), interaction age x demonstrator: t(57) =-1.75, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.08, Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eA).\u003c/p\u003e\n \u003cp\u003eThe number of correct body identifications on the demonstrator\u0026rsquo;s body increased with age ((t(57) = -2.68, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001) but it did not vary with other predictors (type of demonstrator (t(57)\u0026thinsp;=\u0026thinsp;0.37, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.71), gender ((t(57)\u0026thinsp;=\u0026thinsp;0.41, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.68), interaction age x demonstrator: t(57) =-0.40, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.69, Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eImitation of body parts sequences\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe number of body parts correctly labelled (production of body parts labels), significantly increased with age (t(57)\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;3.13, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) whereas it did not vary with the type of demonstrator (t(57)\u0026thinsp;=\u0026thinsp;0.35, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.72), the gender (t(57)\u0026thinsp;=\u0026thinsp;1.69, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.09) or the interaction age x demonstrator (t(57) =-0.04, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.96, Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eC).\u003c/p\u003e\n \u003cp\u003eThe number of body parts correctly imitated in sequences by children significantly increased with age (t(57)\u0026thinsp;=\u0026thinsp;3.78, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) whereas it did not vary with the type of demonstrator (t(57)\u0026thinsp;=\u0026thinsp;1.70, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.09), the gender ((t(57) = -0.41, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.68) but the effect of age was stronger for the human demonstrator than for the robot demonstrator (interaction t(57)\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;2.06, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.04, Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eD).\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eEmotions task\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe number of correct emotions identified on images did not vary with our predictors (age (t(57) =-0.05, p\u0026thinsp;=\u0026thinsp;0.95, type of demonstrator (t(57)\u0026thinsp;=\u0026thinsp;1.38, p\u0026thinsp;=\u0026thinsp;0.17), gender ((t(57) = -0.03, p\u0026thinsp;=\u0026thinsp;0.97), interaction age x demonstrator: t(57) =-1.57, p\u0026thinsp;=\u0026thinsp;0.12).\u003c/p\u003e\n \u003cp\u003eThe number of emotions expressed by the child did not vary with our predictors (age ((t(57)\u0026thinsp;=\u0026thinsp;1.63, p\u0026thinsp;=\u0026thinsp;0.10, type of demonstrator (t(57)\u0026thinsp;=\u0026thinsp;0.61, p\u0026thinsp;=\u0026thinsp;0.54), gender ((t(57)\u0026thinsp;=\u0026thinsp;1.22, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.22), interaction age x demonstrator: t(57) =-0.64, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.51).\u003c/p\u003e\n \u003cp\u003eWe observed an effect of children\u0026rsquo;s ages for identification of emotional key postures: the mean percentage of emotional key postures recognized by younger children was lower compared to older ones ((t(57) = -2.76, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.007) but not with other predictors as type of demonstrator (t(57) = -1.07, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.28), gender ((t(57)\u0026thinsp;=\u0026thinsp;0.36, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.17), interaction age x demonstrator: t(57)\u0026thinsp;=\u0026thinsp;1.02, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.31, see Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe present study evidenced that children aged between 3 and 6 years exhibited similar reactions and performances when interacting with a NAO robot as they did with an adult. Interestingly, we observed no significant impact of the demonstrator type on the average scores of correct answers. As we expected, the number of correct answers increased with age for most of the tasks we proposed: in the identification of body parts on the demonstrator\u0026rsquo;s body, in all subtasks of imitation (the number of body parts labels named by the child and the number of body part imitated), in the identification of emotional key postures.\u003c/p\u003e \u003cp\u003eIdentification of body parts labels on the child\u0026rsquo;s body was not influenced by any of the predictors and we observed a ceiling effect. Comprehensive repertoires of children from 2.2 years old to 3.4 years old contained between 15 and 33 body parts (median\u0026thinsp;=\u0026thinsp;27) on the 50 stimuli presented [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. The most correctly located body parts were: tongue, foot, head, back, tummy, mouth, belly, eye, teeth, hand, leg, fingers, arm, ear, bottom, hair, neck, knee, thumb, cheek, shoulder, toes, lips, eyebrow, elbow. There was a positive correlation between children\u0026rsquo;s comprehension and their age. Therefore, it is possible that by the age of 3, body parts evaluated in our experiment are already known by most of the children, which explains the ceiling effect in our study.\u003c/p\u003e \u003cp\u003eHowever, we found that identification of body parts labels on the demonstrator\u0026rsquo;s body increased with children\u0026rsquo;s age. Pointing to body parts of another person is different from pointing body parts on his own body. This is in line with studies which evidenced that patients with autotopoagnosia (i.e., impairment at pointing to parts of their own body named by an experimenter) do not necessarily present heterotopoagnosia (i.e., impairment at pointing to parts of their own body named by an experimenter) [35]. Indeed, spatial cognition theories differentiate two reference frames: on one hand, representation used by the brain to convey the layout of points in space related to two distinct neural circuits [36]. On the other hand, the egocentric frame conveys positions of points using the body as the centre of the surrounding space, the allocentric frame concerns the positions of points using external landmarks. If both frames of reference emerge quite early in development, by the year of 5 [37], further studies showed a gradual development from an egocentric to a more stable map-like representation between 5 and 10 years of age [38]. It can explain the improvement with age of the ability to locate body parts on the demonstrators\u0026rsquo; body we found in our study.\u003c/p\u003e \u003cp\u003eIn the imitation task, we first evaluated the ability to produce labels of body parts. We found that older children gave more correct answers than younger ones. It is concordant with previous studies highlighting that this ability develops later in development compared to the comprehension of body parts labels [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Concerning the imitation of body parts sequences, we also found an effect of age and this effect was stronger for the human demonstrator than for the robot demonstrator. We hypothesis that imitation of pointing body parts was easier as the human demonstrator had a more familiar body for the children than the robot body.\u003c/p\u003e \u003cp\u003eRegarding the identification of emotions in the body postures task, we first assessed emotions labelling on images. We observed a ceiling effect: scores were very high and not influenced by the predictors. The ceiling effect is concordant with the study of Widen and Russel [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] which demonstrated that facial expressions are correctly associated with labels \u0026ldquo;happiness\u0026rdquo;, \u0026ldquo;sadness\u0026rdquo;, \u0026ldquo;anger\u0026rdquo;, \u0026ldquo;fear\u0026rdquo;, \u0026ldquo;surprise\u0026rdquo; and \u0026ldquo;disgust\u0026rdquo;, which correspond to the six primary emotions described by Ekman and Friesen [39], by the year of 3. In another study [31], the ability to recognize sadness, anger, joy and fear on very simple drawings of human faces seems to emerge between 4 and 5 years.\u003c/p\u003e \u003cp\u003eConcerning emotions expressed by the child, we found no effect of the predictors and an important interindividual variability. We propose that more than the demonstrator type (human or robot), scores were influenced by the emotional expressiveness of children \u0026ndash; the capacity to accurately communicate feelings nonverbally - and the tendency to feel comfortable in the presence of unfamiliar adults, variables that we did not measured. Indeed, a meta-analysis showed that more emotional expressive people also have a more extraverted personality (extraverts are outgoing, talkative, impulsive and uninhibited and have many social contacts) [40].\u003c/p\u003e \u003cp\u003eFinally, we assessed labelling of emotions in the demonstrator key postures. We found no effect of the demonstrator type but an improvement with children\u0026rsquo;s ages. Correct labels of the demonstrator emotional key postures were lower than in the study of Beck and colleagues in adults [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] and children from 11 to 13 years old [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. We made two assumptions about the effect of children\u0026rsquo;s age. First, it had been pointed out that emotion labelling implies a more complex cognitive process than emotion recognition, which can be fully mature later in the child\u0026rsquo;s development [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. Secondly, even if emotional key postures can be accurately recognised without any context [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], it had been demonstrated that knowledge about the social and environmental context in which the key postures of a teddy bear robot are expressed improve the success of identification [42]. We propose that the context might be more important for younger children than for older children and adults as they are still learning to recognize emotions.\u003c/p\u003e \u003cp\u003eTaken together, our study provides encouraging results concerning the use of the social robot NAO with children between 3 and 6 years old to evaluate children knowledge in several domains, from vocabulary learning to prosocial behaviours. This type of social robot is interesting as tools to develop children\u0026rsquo;s socio-cognitive skills and further research are needed to understand how young children \u0026rsquo;s representation of this type of social partner develops through early childhood.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eSelection and Participation\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eSeventy-three children aged 3 years and 3 months to 6 years and 1 month were evaluated. All children were French native speakers and came from 4 different schools. We obtained parental consent for all children to take part in the study and to be recorded. We also collected verbal assent from children before starting the experiment.\u003c/p\u003e\n \u003cp\u003eOn 73 children initially recruited, 9 asked to stop before the experiment was over, whilst in the two other cases, there were technical issues with the camera. In total, 62 child-demonstrator interactions were completed in a between-subject procedure (school A: 7 children, school B: 10 children, school C: 17 children, school D: 28 children). Children were separated into three groups of age: Group 1, \u0026ldquo;G1\u0026rdquo;: from 3 years old and one month to 4 years old and one month (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SE\u0026thinsp;=\u0026thinsp;3.73\u0026thinsp;\u0026plusmn;\u0026thinsp;0.07); Group 2, \u0026ldquo;G2\u0026rdquo;: from four years-old and two months to five years-old and 1 month (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SE\u0026thinsp;=\u0026thinsp;4.71\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06) ; Group 3, \u0026ldquo;G3\u0026rdquo;: from five years-old and two months to six years-old and one month (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SE\u0026thinsp;=\u0026thinsp;5.65\u0026thinsp;\u0026plusmn;\u0026thinsp;0.059). As much as possible, we balanced experimental groups concerning the age and gender of the children. Children were randomly assigned to the experimental groups. Our sample is composed of 62 children (30 in the robot group, 32 in the human group). The Group 1 is composed of 14 children: 8 girls (4 with in the robot group, 4 in the human group) and 6 boys (2 in the robot group, 4 in the human group). The Group 2 is composed of 26 children: 13 girls (7 with in the robot group, 6 in the human group) and 13 boys (6 in the robot group, 7 in the human group). The Group 3 is composed of 22 children: 14 girls (6 with in the robot group, 8 in the human group) and 8 boys (5 in the robot group, 3 in the human group).\u003c/p\u003e\n \u003cp\u003eThe protocol was carried out in accordance with the ethical standards of the Declaration of Helsinki (BMJ 1991; 302:1194) and approved by the Ethic Committee of the Department of Psychology of the (CER-PN n\u0026deg;2022-09-01).\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eExperimental procedure\u003c/p\u003e\n\u003cp\u003eGeneral setup\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe experimental sessions were conducted in a quiet room at the school in the presence of the experimenter and the demonstrator (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). In each of the four schools, one adult female experimenter was recruited and trained to adhere to the experimental setup. No child had been exposed to the NAO robot prior to the experiment.\u003c/p\u003e\n \u003cp\u003eThe experimenter asked the child to sit on the floor in front of the demonstrator and introduced the demonstrator to him/her. The demonstrator was a female adult for half of the children and the NAO robot for the other half. The demonstrator gave instructions to the child. Due to the COVID-19 situation, the experimenter and the human demonstrator wore face masks during the whole experiments. Each individual session lasted approximately 20 min. Two Sony HDR-CX410VE cameras placed on tripods from either side of the demonstrator were used to record the child\u0026rsquo;s behaviours.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eHuman group: exposure to an adult demonstrator\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eThe human demonstrator was always the same adult female and remained as neutral as possible during the experiment: she gave the instructions to the child, but she did not encourage him/her or give him/her feedback.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eRobot group: exposure to the NAO robot demonstrator\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eCreated by Aldebaran Robotics, the NAO robot version 6 is 58 cm high and bipedal. It has 25 degrees of freedom which allows a lot of possibilities of movements. It can manipulate small objects with its three fingers - hands. NAO has two video cameras on its forehead and its mouth. It has vocal capacities both for recognition and synthesis: it is equipped with a stereo broadcast system made up of two loudspeakers, on its ears and 4 omnidirectional microphones (2 on the top of its head and 2 on the back of its head). It also has two ultrasonic sensors (or sonars) which allows it to estimate the distance to obstacles in its environment. Moreover, it has contact and tactile sensors: tactile head and hands, chest button and feet bumpers. The NAO robot can be programmed to carry out autonomously a set of tasks. However, for our experiment, the robot was fully tele-operated to enable the robot to act contingently [43]: the experimenter controlled the robot during the session, via a touch- screen tablet, an iPad mini 4 (20.32 x 13.48 x 0.61 cm). The setup involved connecting the tablet to the robot\u0026rsquo;s wifi access point. The robot hosted a web server that could be accessed through any browser on the tablet. An HTML page was used to the interface and execute the experiment\u0026apos;s sequence using javascript (\u0026laquo; vue.js \u0026raquo; for the graphical interface, \u0026laquo; LibQi \u0026raquo; for controlling the robot).\u003c/p\u003e\n \u003cp\u003eAs the human demonstrator was a female, we attributed a feminine gender to the NAO robot and a feminine name (\u0026ldquo;Naomie\u0026rdquo;). We will use the pronoun \u0026ldquo;her\u0026rdquo; in the following text. To ensure that the two demonstrators pronounced the same sentences in both conditions, vocal recordings of the female adult were played back by the robot. Sentences were recorded while she was alone in a silent room and as she was addressing instructions to a child.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eExperimental tasks and scores\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eBefore starting the experimental tasks, there was a warm-up phase to allow the child to see how the demonstrator, robotic or human depending on the group, spoke and moved before the beginning of the experiment. The demonstrator said her name to the child and asked the child to introduce herself/himself. Then, she sang a popular rhyme with hand gestures and encouraged the child to follow it.\u003c/p\u003e\n \u003cp\u003eThen, each child was tested in different situations: 1) the comprehension of body parts labels on their own body and on the demonstrator\u0026rsquo;s body, 2) the production of body parts labels and an imitation task and 3) the identification of emotional key postures.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eComprehension of body parts labels:\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eWe evaluated the comprehension of eleven body parts labels: face, eye, nose, mouth, ear, shoulder, elbow, hand, belly, knee, foot. Each body part was randomly chosen in this list. First, the demonstrator asked the child \u0026ldquo;Show me your [body part]\u0026rdquo; and repeated it for the eleven possibilities. Then she asked, \u0026ldquo;Show me my [body part]\u0026rdquo; and repeated it for the eleven body parts.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eImitation of body parts sequences:\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eWe evaluated body parts labels production: the demonstrator randomly showed one of her body parts, for example, her eye and asked \u0026ldquo;What is the name of this body part?\u0026rdquo; and a second one, for example her belly, and said \u0026ldquo;And this one?\u0026rdquo;.\u003c/p\u003e\n \u003cp\u003eAs it was complicated for the NAO robot to point her nose and its ear, we decided to keep 9 body parts for this sub-task (face, eye, mouth, shoulder, elbow, hand, belly, knee and foot in Supplementary Information, Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThen, the demonstrator placed her hands on her knees and asked the child to repeat the body parts sequence.\u003c/p\u003e\n \u003cp\u003eThe demonstrator performed two sequences with two body parts (\u0026ldquo;sequence#1\u0026rdquo;, \u0026ldquo;sequence#2\u0026rdquo;) and a third sequence with three body parts (\u0026ldquo;sequence#3\u0026rdquo;). As the majority of children had difficulties to understand the instructions in this task, the first two-body parts sequence (sequence#1) was used by the experimenter to explain the instructions again in order to help the child to understand it.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eEmotions task\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eFour images of a child expressing an emotion (joy, sadness, anger, fear. Supplementary Information, Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e) were placed in front of the child. The demonstrator asked \u0026ldquo;Do you know what an emotion is? Show me the image in which the child feels [emotion].\u0026rdquo;\u003c/p\u003e\n \u003cp\u003eThen, the demonstrator and the child stood up. The demonstrator asked \u0026ldquo;Show me how do you express [emotion] with your whole body\u0026rdquo;. She waited for the child to express the emotion. The instruction was repeated with the four emotions.\u003c/p\u003e\n \u003cp\u003eAs the NAO robot does not have facial expressions, we investigated emotions recognition when emotions were expressed with the body. We selected the four key postures exhibited by the NAO robot that were the most successfully identified by adults and children in previous studies [\u003cspan class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e14\u003c/span\u003e]: joy, fear, anger and sadness. The demonstrator struck each of these four key postures and between each pose, she went back to a neutral pose (Supplementary Information, Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eFor this task, the demonstrator said, \u0026ldquo;Let me show you what I am doing when I am feeling an emotion\u0026rdquo;. The demonstrator reproduced one of the four emotional postures and the experimenter asked the child \u0026ldquo;Which emotion do you think it is?\u0026rdquo;. After the child replied, the demonstrator returned to the neutral pose. She repeated this sequence for the three other body postures.\u003c/p\u003e\n \u003cp\u003eAt the end of the session, the demonstrator thanked the child for his/her participation and told him/her goodbye.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eVideo recordings and analysis\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eAll sessions were video recorded. Cameras allowed to have a global view of the setup which encompassed the face and the body of the child. Informed written consents were obtained from the human demonstrator and from the parents of the children for video recordings and to publish their information/images in an online open-access publication.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\n \u003ch2\u003eStatistical analysis\u003c/h2\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003eWe conducted models trying to predict each variable with the demonstrator type (robot or human), the child\u0026rsquo;s age, the child\u0026rsquo;s age in interaction with the demonstrator\u0026rsquo;s type and the child\u0026rsquo;s gender as predictors. We started each model including school and experimenter as random effects of a GLME. We adapted the link and distributions to increase the model goodness-of-fit using AIC and % of variance explained (squared R) as indicators. We removed high p-values factors when it increased the goodness-of-fit and switched to a GLM (without random effects), following the same rationale (Table 1). We conducted specific analysis for the imitation of body parts sequences: we decided to remove sequence#1 from the statistical analysis as it was necessary to use it to explain again instructions to children. We kept the second two-body parts sequence (sequence#2) and the three-body parts sequence only (sequence#3). As results in sequence 2 and 3 followed the same patterns, we calculated a number of correct answers adding answers from sequence 2 and 3 and we conducted GLM on these data. Table 1 summarizes the statistical model and is available as a supplementary material.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eA.A. collected and analyzed the data. B.G. and S.D. co-supervised the study. A.C. conducted the statistical analyzes. A.B. programmed the robot. A.A., B.G., and A.C. participated equally to te writing of the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe would like to thank all the children who participated as well as schools\u0026rsquo; directors who coordinated the study. We also thank the Institut Universitaire de France (IUF) and the Descartes Program from CNRS@Create for their support.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eWe declare that data collected will be available upon request to Alice Araguas.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAksoy, P., \u0026amp; Baran, G. (2010). Review of studies aimed at bringing social skills for children in preschool period. Procedia - Social and Behavioral Sciences, 9, 663 669. https://doi.org/10.1016/j.sbspro.2010.12.214\u003c/li\u003e\n\u003cli\u003ePerloff, R. M. (1982). Gender constancy and same-sex imitation: A developmental study. The Journal of Psychology: Interdisciplinary and Applied, 111(1), 81‑86. https://doi.org/10.1080/00223980.1982.9923516\u003c/li\u003e\n\u003cli\u003eOlson, K. R., \u0026amp; Spelke, E. S. (2008). Foundations of cooperation in young children. Cognition, 108(1), 222‑231. https://doi.org/10.1016/j.cognition.2007.12.003\u003c/li\u003e\n\u003cli\u003eZmyj, N., Buttelmann, D., Carpenter, M., \u0026amp; Daum, M. M. (2010). The reliability of a model influences 14-month-olds\u0026rsquo; imitation. Journal of Experimental Child Psychology, 106(4), 208‑220.https://doi.org/10.1016/j.jecp.2010.03.002\u003c/li\u003e\n\u003cli\u003eKinzler, K. D., Dupoux, E., \u0026amp; Spelke, E. S. (2007). The native language of social cognition. Proceedings of the National Academy of Sciences of the United States of America, 104(30), 12577‑12580. https://doi.org/10.1073/pnas.0705345104\u003c/li\u003e\n\u003cli\u003evan den Berghe, R., Verhagen, J., Oudgenoeg-Paz, O., van der Ven, S., \u0026amp; Leseman, P. (2019). Social Robots for Language Learning : A Review. Review of Educational Research, 89(2), 259-295. https://doi.org/10.3102/0034654318821286\u003c/li\u003e\n\u003cli\u003eBartneck, C., \u0026amp; Forlizzi, J. (2004). A design-centred framework for social human-robot interaction. 591‑594. https://doi.org/10.1109/ROMAN.2004.1374827\u003c/li\u003e\n\u003cli\u003eSheridan, T. B. (2020). A review of recent research in social robotics. Current Opinion in Psychology. doi: 10.1016/j.copsyc.2020.01.003.\u003c/li\u003e\n\u003cli\u003eRobinson, N.L., Cottier, T.V., Kavanagh, D.J. (2019); Psychosocial health interventions by social robots: Systematic review of randomized controlled trials. \u003cem\u003eJournal of Medical Internet Research\u003c/em\u003e 21(5): e13203.\u003c/li\u003e\n\u003cli\u003eBelpaeme, T., Kennedy, J., Ramachandran, A., Scassellati, B., \u0026amp; Tanaka, F. (2018). Social robots for education : A review. Science Robotics, 3(21), eaat5954.https://doi.org/10.1126/scirobotics.aat5954\u003c/li\u003e\n\u003cli\u003eKim, Y., \u0026amp; Tscholl, M. (2021). Young children\u0026rsquo;s embodied interactions with a social robot. Educational Technology Research and Development, 69, 2059\u0026ndash;2081. https://doi.org/10.1007/s11423-021-09978-3\u003c/li\u003e\n\u003cli\u003eGlenberg, A. M. (2010). Embodiment as a unifying perspective for psychology. Wiley Interdisciplinary Reviews: Cognitive Science, 1, 586\u0026ndash;596.\u003c/li\u003e\n\u003cli\u003eBeck, A., Ca\u0026ntilde;amero, L., \u0026amp; Bard, K. (2010). Towards an Affect Space for robots to display emotional body language. 464‑469. https://doi.org/10.1109/ROMAN.2010.5598649\u003c/li\u003e\n\u003cli\u003eBeck, A., Ca\u0026ntilde;amero, L., Hiolle, A., Damiano, L., Cosi, P., Tesser, F., \u0026amp; Sommavilla, G. (2013). Interpretation of Emotional Body Language Displayed by a Humanoid Robot : A Case Study with Children. International Journal of Social Robotics, 5(3), 325‑334.https://doi.org/10.1007/s12369-013-0193-z\u003c/li\u003e\n\u003cli\u003eRaimo, S., Iona, T., Di Vita, A., Boccia, M., Buratin, S., Ruggeri, F., Iosa, M., Guariglia, C., Grossi, D., \u0026amp; Palermo, L. (2021). The development of body representations in school-aged children. Applied Neuropsychology: Child, 10(4), 327339.https://doi.org/10.1080/21622965.2019.1703704\u003c/li\u003e\n\u003cli\u003eSchwoebel, J., Boronat, C. B., \u0026amp; Branch Coslett, H. (2002). The man who executed \u0026laquo; imagined \u0026raquo; movements : Evidence for dissociable components of the body schema. Brain and Cognition, 50(1), 116. https://doi.org/10.1016/s0278-2626(02)00005-2\u003c/li\u003e\n\u003cli\u003eSchwoebel, J., \u0026amp; Coslett, H. B. (2005). Evidence for multiple, distinct representations of the human body. Journal of Cognitive Neuroscience, 17(4), 543553. https://doi.org/10.1162/0898929053467587\u003c/li\u003e\n\u003cli\u003ePicelli, A., Negrini, S., Zenorini, A., Iosa, M., Paolucci, S., \u0026amp; Smania, N. (2016). Do adolescents with idiopathic scoliosis have body schema disorders? A cross-sectional study. Journal of Back and Musculoskeletal Rehabilitation, 29(1), 89\u0026ndash;96. https://doi.org/10.3233/BMR-150602\u003c/li\u003e\n\u003cli\u003eSirigu, A., Grafman, J., Bressler, K., \u0026amp; Sunderland, T. (1991). Multiple representations contribute to body knowledge processing. Evidence from a case of autotopagnosia. Brain: A Journal of Neurology, 114 (Pt 1B), 629642. https://doi.org/10.1093/brain/114.1.629\u003c/li\u003e\n\u003cli\u003e[20] Berg\u0026egrave;s, J., \u0026amp; L\u0026eacute;zine, I. (1963). Test of imitation of gestures: Methods for plotting body image and movements of children from 3 to 6 years of age. Masson \u0026amp; Cie\u003c/li\u003e\n\u003cli\u003eCam\u0026otilde;es-Costa, V., Erjavec, M., \u0026amp; Horne, P. J. (2011). The impact of body-part-naming training on the accuracy of imitative performances in 2- to 3-year-old children. Journal of the Experimental Analysis of Behavior, 96(3), 291315. https://doi.org/10.1901/jeab.2011.96-291\u003c/li\u003e\n\u003cli\u003eRusso, L., Craig, F., Ruggiero, M., Mancuso, C., Galluzzi, R., Lorenzo, A., Fanizza, I., \u0026amp; Trabacca, A. (2018). Exploring Visual Perspective Taking and body awareness in children with Autism Spectrum Disorder. Cognitive Neuropsychiatry, 23(4), 254-265. https://doi.org/10.1080/13546805.2018.1486182\u003c/li\u003e\n\u003cli\u003eCosta, S., Lehmann, H., Dautenhahn, K., Robins, B., \u0026amp; Soares, F. (2014). Using a Humanoid Robot to Elicit Body Awareness and Appropriate Physical Interaction in Children with Autism. International Journal of Social Robotics, 7, 114. https://doi.org/10.1007/s12369-014-0250-2\u003c/li\u003e\n\u003cli\u003eSuzuki, R., \u0026amp; Lee, J. (2016). Robot-play therapy for improving prosocial behaviours in children with Autism Spectrum Disorders. 2016 International Symposium on Micro-NanoMechatronics and Human Science (MHS). https://doi.org/10.1109/MHS.2016.7824238\u003c/li\u003e\n\u003cli\u003eSuzuki, R., Lee, J., \u0026amp; Rudovic, O. (2017). NAO-Dance Therapy for Children with ASD. 295296. https://doi.org/10.1145/3029798.3038354\u003c/li\u003e\n\u003cli\u003eCao, H. L., Simut, R. E., Desmet, N., De Beir, A., Van De Perre, G., Vanderborght, B., \u0026amp; Vanderfaeillie, J. (2020). Robot-assisted joint attention: A comparative study between children with autism spectrum disorder and typically developing children in interaction with NAO. IEEE Access, 8, 223325-223334.\u003c/li\u003e\n\u003cli\u003eParker, A. E., Mathis, E. T., \u0026amp; Kupersmidt, J. B. (2013). How is this child feeling? Preschool-aged children\u0026rsquo;s ability to recognize emotion in faces and body poses. Early education and development, 24(2), 188-211. https://doi.org/10.1080/10409289.2012.657536\u003c/li\u003e\n\u003cli\u003e[28] Widen, S. C., \u0026amp; Russell, J. A. (2003). A closer look at preschoolers\u0026rsquo; freely produced labels for facial expressions. Developmental Psychology, 39(1), 114-128. https://doi.org/10.1037//0012- 1649.39.1.114\u003c/li\u003e\n\u003cli\u003ePeterson, C. C., Slaughter, V., \u0026amp; Brownell, C. (2015). Children with autism spectrum disorder are skilled at reading emotion body language. Journal of Experimental Child Psychology, 139, 35-50. https://doi.org/10.1016/j.jecp.2015.04.012\u003c/li\u003e\n\u003cli\u003eIzard, C., Fine, S., Schultz, D., Mostow, A., Ackerman, B., \u0026amp; Youngstrom, E. (2001). Emotion Knowledge as a Predictor of Social Behavior and Academic Competence in Children at Risk. Psychological Science, 12(1), 18-23. https://doi.org/10.1111/1467-9280.00304\u003c/li\u003e\n\u003cli\u003eDenham, S. A., Blair, K. A., DeMulder, E., Levitas, J., Sawyer, K., Auerbach\u0026ndash;Major, S., \u0026amp; Queenan, P. (2003). Preschool Emotional Competence : Pathway to Social Competence? Child Development, 74(1), 238-256. https://doi.org/10.1111/1467-8624.00533\u003c/li\u003e\n\u003cli\u003eCovic, A., von Steinb\u0026uuml;chel, N., \u0026amp; Kiese-Himmel, C. (2020). Emotion recognition in kindergarten children. Folia Phoniatrica et Logopaedica:International Journal of Phoniatrics, Speech Therapy and Communication Pathology, 72(4), 273\u0026ndash;281. https://doi.org/10.1159/000500589\u003c/li\u003e\n\u003cli\u003eSchneider, J., Sandoz, V., Equey, L., Williams-Smith, J., Horsch, A., Bickle Graz M. (2022). The Role of Face Masks in the Recognition of Emotions by Preschool Children. JAMA Pediatr. doi: 10.1001/jamapediatrics.2021.4556. PMID: 34779832; PMCID: PMC8593832.\u003c/li\u003e\n\u003cli\u003eCohen, I., Looije, R., \u0026amp; Neerincx, MA. (2012). Child's recognition of emotions in robot's face and body. In A. Billard et al (Ed.), \u003cem\u003eProceedings of the 6th international conference on Human-robot interaction \u003c/em\u003e(pp. 123-124). ACM DL. https://doi.org/10.1145/1957656.1957692\u003c/li\u003e\n\u003cli\u003ePisella, L., Hav\u0026eacute;, L., \u0026amp; Rossetti, Y. (2019). Body awareness disorders : Dissociations between body- related visual and somatosensory information. Brain : a journal of neurology, 142, 2170‑2173. https://doi.org/10.1093/brain/awz187\u003c/li\u003e\n\u003cli\u003eKlatzky, R. L. (1998). Allocentric and Egocentric Spatial Representations : Definitions, Distinctions, and Interconnections. In C. Freksa, C. Habel, \u0026amp; K. F. Wender (\u0026Eacute;ds.), Spatial Cognition : An Interdisciplinary Approach to Representing and Processing Spatial Knowledge (p. 117). Springer. https://doi.org/10.1007/3-540-69342-4_1\u003c/li\u003e\n\u003cli\u003eNewcombe, N., \u0026amp; Huttenlocher, J. (2003). Making Space : The Development of Spatial Representation and Reasoning. MIT Press.\u003c/li\u003e\n\u003cli\u003eBullens, J., Igl\u0026oacute;i, K., Berthoz, A., Postma, A., \u0026amp; Rondi-Reig, L. (2010). Developmental time course of the acquisition of sequential egocentric and allocentric navigation strategies. Journal of Experimental Child Psychology, 107(3), 337350. https://doi.org/10.1016/j.jecp.2010.05.010\u003c/li\u003e\n\u003cli\u003eEkman, P., \u0026amp; Friesen, W. V. (1976). Measuring facial movement. Environmental Psychology and Nonverbal Behavior, 1(1), 5675. https://doi.org/10.1007/BF01115465\u003c/li\u003e\n\u003cli\u003eRiggio, H. R., \u0026amp; Riggio, R. E. (2002). Emotional Expressiveness, Extraversion, and Neuroticism : A Meta-Analysis. Journal of Nonverbal Behavior, 26(4), 195-218. https://doi.org/10.1023/A:1022117500440\u003c/li\u003e\n\u003cli\u003eIzard, C. E. (1971). The face of emotion. New York: Appleton Century-Crofts Educational Division, Meredith.\u003c/li\u003e\n\u003cli\u003eLi, J., \u0026amp; Chignell, M. (2011). Communication of Emotion in Social Robots through Simple Head and Arm Movements. International Journal of Social Robotics, 3(2), 125142. https://doi.org/10.1007/s12369-010-0071-x\u003c/li\u003e\n\u003cli\u003eKennedy, J., Lemaignan, S., Montassier, C., Lavalade, P., Irfan, B., Papadopoulos, F., Senft, E., \u0026amp; Belpaeme, T. (2017). Child Speech Recognition in Human-Robot Interaction : Evaluations and Recommendations. Proceedings of the 2017 ACM/IEEE International Conference on HumanRobot Interaction, 82-90. https://doi.org/10.1145/2909824.3020229\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"children, social robots, body, emotions, interaction","lastPublishedDoi":"10.21203/rs.3.rs-4758583/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4758583/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe aim of the present study was to compare interactions of children aged between 3 and 6 years, with a NAO robot or an adult partner, in various body-focused tasks: comprehension and recognition of body parts labels, imitation of movements, and recognition of emotions in the postures of the agent. For each task, performances were appreciated through scores levels. We found no effect of the demonstrator type on our results: children of different ages responded similarly to the the human or the robot demonstrator. We found an effect of age, with the olderchildren having higher scores for the comprehension of body parts labels on the demonstrator’s body, the imitation of body parts sequences and the identification of emotional key postures. Results are discussed in light of the implications of the use of social robots such as the NAO one, in interactive and learning situations with typical children.\u003c/p\u003e","manuscriptTitle":"Child-robot interactions in different body and emotions oriented tasks: comparison with a human partner","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-17 02:38:15","doi":"10.21203/rs.3.rs-4758583/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e09cc4d7-bbda-4b1b-9342-ea0453ac42a5","owner":[],"postedDate":"August 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":35825710,"name":"Biological sciences/Psychology"},{"id":35825711,"name":"Biological sciences/Psychology/Human behaviour"}],"tags":[],"updatedAt":"2025-01-22T11:27:57+00:00","versionOfRecord":[],"versionCreatedAt":"2024-08-17 02:38:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4758583","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4758583","identity":"rs-4758583","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}