autism spectrum disorders (ASD) are pervasive neurodevelopmental disorders characterized by difficulties in social interaction. One critical aspect to fluent social interactions is the ability to comprehend others' emotional states through language context. ASD individuals have difficulty attributing emotional labels to pictures or sentences as well as reduced memory for emotional words as compared to typically developing peers (TD). Some theories of autism have speculated that these individuals have difficulty with generating appropriate emotional responses to the situations of other individuals, that is, empathizing. Other theories argue that individuals with ASD have difficulties in making inferences or drawing conclusions. However, the data to this point suggest that individuals with ASD can elicit emotional responses to others in limited contexts, and that, at least with high functioning individuals with ASD and Asperger's Disorder, they can comprehend language not unlike their typically developing peers. Thus, the deficits in understanding the emotional states of others, particularly through language, may be the result of a disconnection between the neural systems for language and emotional processing. In essence, individuals with ASD can comprehend emotional language as a linguistic process, but fail to evoke the appropriate emotional responses from this abstract system. The current proposal aims to identify the brain mechanisms underlying difficulties with emotional language processing in ASD. We plan to collect functional magnetic resonance imaging (fMRI) data from 40 adolescents and young adults (16-22 years): 20 who are typically developing (TD) and 20 with ASD. Participants will listen to stories that require an emotional or neutral inference followed by a statement (e.g., "She felt panic") for which they will be asked to make a congruity judgment based on the previous verbal context. Our hypothesis is that TD individuals, but not individuals with ASD, will recruit emotion-related brain regions (e.g., the amygdala) while listening to stories with emotional content, suggesting a failure to elicit an appropriate emotional response to stories inferring others' emotions. Additionally, we predict that ASD individuals may be less accurate in making congruity judgments about the emotion words (as compared to neutral words) and show reduced activation in regions involved in error monitoring during the presentation of emotion words that are incongruous to the sentence context. These findings would suggest that ASD individuals do not automatically associate a verbal label (e.g., "panic") with context (e.g., sentence describing panic). The second goal of our investigation is to more directly examine the connectivity between the neural systems for language and emotion in individuals with ASD. Our hypothesis, which is consistent with findings of reduced brain connectivity in ASD, is that these individuals will show reduced correlated brain activity (functional connectivity) between emotion and language-related brain regions both when performing the Emotional Inference task in fMRI as well as during a resting-state fMRI (rsfMRI) scan. The analysis of task-based functional connectivity will enable us to determine whether there is less connectivity during the demands of a relevant task compared to their TD peers, whereas the resting-state connectivity measure will enable us to determine whether this lack of coordination between these brain regions exists due to generally weak connectivity overall outside of specific task demands. Findings from the current study will provide a greater understanding of the neural and cognitive mechanisms underlying difficulties with affective language processing in ASD individuals. In the short term, they would provide an understanding of the difficulties comprehending others' emotions through language as a deficit in the interaction of the language and emotion systems, as opposed to a weakness of either system specifically as suggested by prevalent theories of ASD. In the long term, these results will provide a springboard to study the development of these deficits in emotional language processing such that we can begin to determine when in development children with ASD deviate from their TD peers in making appropriate emotional inferences from language. We can then examine whether difficulties are tied to lapses with acquiring emotional words (e.g., happy, sad, afraid) at early ages or whether these difficulties arise from more deficits specific to processing emotion in contextual situations from language (e.g., evoking the feeling of panic when hearing someone lost their keys or iPhone). The broader impact of these findings will not only be to reevaluate the way in which we approach the deficits in ASD, but can directly impact the way in which we develop treatment interventions for individuals with ASD and understand the success of specific intervention approaches. In sum, these findings will provide a critical understanding of a fundamental deficit that negatively impacts the formation and maintenance of close personal friendships in individuals with ASD.