People with Autism Spectrum Disorder (ASD) have difficulties in dealing with social interaction (DSM V, 2013). From long time, it has been reported deficit in recognizing gaze, face and intentions of others. Besides that, in the last decade numerous finds have showed an impairment in coding the human body movements. During my PhD I have explored the ability in recognizing the emotional valence of human movements in people with ASD and with typical development. In order to understand developmental changes, I have initially compared this ability between children and adults. In order to understand the role of intelligence, I have included people with high and low functioning ASD, either in children and adults samples. In the first study, we compared TD children and children with ASD with different level of functioning to investigate i) whether the difficulties in ASD population were associated to emotion comprehension or related to a more basic visual processing of biological motion, and ii) to explore whether this ability improves according to the age and to the non-verbal IQ. To this purposes, we presented point-light (PLDs) and full-light (FLDs) version of human movements with three different emotional valence (Happy, Fearful, Neutral). Our main findings showed that: -TD children were more accurate in recognizing the emotional content of body movements when the shape of the body was visible (i.e. FLDs), compared to the view of pure motion information (i.e. PLDs), but there were no differences in RT across display conditions. This result suggests that school-aged children are able to correctly identify the emotional valence of dynamic bodily expression, also when the form information is minimized by presenting PLDs. -According to the emotional content, we found that the valence had an effect on the performance of TD children. In particular, TD children recognized happy bodily expressions with lower accuracy than fearful and neutral movements. Besides, fearful movements were identified more rapidly. We hypothesized that, from an evolutionary point of view, it is more important to recognize fearful signals when they are far from us. Body movements are particularly relevant to convey other people’s state and intentions from a distance, hence bodies are preferential channels to convey fearful signals. -Comparing typical and atypical children, we found that TD were more accurate and more rapid than children with ASD in recognizing the emotional valence of body movements. Moreover, children with HF ASD outperformed children with LF ASD and their performance was significantly predicted from age and nonverbal IQ. This results suggest that the nonverbal abilities have a role in body movement comprehension, probably underpinning the development of compensatory mechanisms which improve with age. -Compared to TD children, children with ASD were impaired not only in understanding the emotional expressions, but also in recognizing the neutral actions. Irrespectively to their IQ level, children with ASD did not benefit from the richness of visual information, and their ability to recognize BM did not change according to the emotional valence. This finding suggests that the difficulty encountered by individuals with ASD in social interaction could be more generally related to biological motion elaboration, rather than being specific for the emotions comprehension. Hence, the lack in emotion recognition in ASD seems to be attributable to a deficit in elaborating motion cues rather than emotional information. To interpret the body movements, children with ASD may adopt compensatory mechanisms, which acquisition seems to be mediated by the nonverbal IQ and improving with age. If this was the case, we should expect an improvement in the ability of recognizing the BM in adults with ASD, with respect to children with ASD, and this improvement should have been correlated with the non-verbal IQ level. To explore this hypothesis, we run a second experiment, were we asked to a group of TD adults, a group of adults with high functioning Autism matched for age and non-verbal IQ, and a group of adults with low functioning Autism matched for age, to performed the same BM recognition task that children did in our first experiment. Results with adults confirmed what we found in children: the emotional valence and the richness of visual information did not modulate the Accuracy and the velocity in recognizing the BM stimuli in ASD, suggesting that the whole body movement is elaborated differently in ASD than in TD adults and that the impairment in ASD is more related to the processing of body movement, rather than being specific for the emotion comprehension. In a third study, we explored the developmental changes in the ability of recognizing the emotional meaning of human whole-body movements from childhood to adulthood. To this aim, we compared the performance between children and adults in each group of Functioning. Results showed that TD adults were more accurate and faster than TD children; FLDs were recognized with higher accuracy and quicker than PLDs; and fearful and neutral movements were recognized significantly better than happy ones. In participants with ASD, adults were significantly more accurate than children, but not faster. Also, we did not find any difference between displays or emotional categories. This results suggest that the understanding of body movements in individuals with ASD improve with age but the mechanism underpinning this ability works differently. In a fourth experiment we investigate whether the impairment in understanding social relevant signals in individuals with high functioning autism (HF ASD) was restricted to body movements or widespread to other social cues, such as facial expressions. Also, we investigated the role of movement in perceiving bodily expressions. To this aim, we explored whether there was a difference in recognizing the body expressions represented by dynamic or static stimuli. Finally, we investigated whether the vision of the body form, compared to the sight of pure motion information, could influence the identification of the emotional content of body movements. We did not found any group differences in accuracy, but there was a group difference in RTs specific for dynamic stimuli (TD were significantly faster than ASD in recognizing PLDs, and marginally in FLDs). These results suggest that HF ASD adopt compensatory mechanism to understand the meaning of facial and bodily expression that allow them to correctly recognize the emotional valence conveyed by face and body movements. However, this mechanism has a cost in terms of rapidity. In particular, ASD seems to need more time to identify dynamic body stimuli but not images of body expressions, suggesting a deficit in processing the actual body motion. In particular, the comprehension of happy body movements result difficult and time consuming. Indeed, happiness was better recognized from facial expressions and was harder to be identified when expressed by body movements. In the fifth experiment we used TMS-adaptation paradigm to explore the existence of a neural system specialized for the elaboration of emotional body movements. We first behaviourally investigated the existence of an adaptation for emotional PLDs. TD adult participant were adapted with point-light video clips depicting fearful, happy or neutral actions and then asked to recognize point-light with same/different emotional content. Results showed an adaptation after-effect only for incongruent stimuli, suggesting the existence of a neural mechanism for perceiving the body emotion specifically. Subsequently, in a TMS experiment we explore the possible brain location of this mechanism. The sites we stimulated are nodes of the neural network responsible for the human motion understanding, and they are reported to be abnormally activated in ASD. We found a reversed after-effect following TMS over aIPS, while the adaptation was still present after stimulation over pSTS and the control site. These results demonstrate that aIPS contains neurons that specifically code for the emotional body expressions, suggesting that the difficulties encountered by individuals with ASD in understanding the emotional signal during social interaction might rely to deficit in this mirror area.