Decoding Emotions From Body Postures: Altered Mirror Neuron Response In Children with Autism

Background:  The ability to interpret others’ body language is a vital skill that helps us infer and understand their thoughts and emotions.  However, individuals with autism have been found to have difficulty in deciphering the body postures and movements of others, perhaps leading to their overarching deficit in processing emotions (Moore et al., 1997; Hubert et al., 2007; Atkinson, 2009; Philip et al., 2010). The present study investigated the neural mechanisms of inferring emotions from body postures. To our knowledge, this is the first neuroimaging study examining body language processing in children with autism, although a previous study examined processing of fearful body language in adults with autism (Hadjikhani et al., 2009).

Objectives:  The primary aim of this functional MRI study was to investigate the neural correlates of emotion and action, in the context of processing body language, in high-functioning children with autism. 

Methods:  fMRI data was acquired from 5 high-functioning children with autism and 7 typically developing controls (data collection in progress) while they made emotion and action judgments about a series of static stick figure characters.  The participants’ task was to view a character’s posture and choose the option, from three alternatives, that best described the action (e.g., pushing) or emotion (e.g., sad) the character was portraying. The stimuli were presented in a blocked design format and data was acquired on a Siemens 3T scanner and analyzed using SPM8.

Results:  The main results are as follows: 1) A between-group comparison of brain activation showed participants with autism recruiting the right inferior frontal gyrus (IFG) and left inferior parietal lobule (IPL) significantly lesser, relative to controls, while inferring emotions from body postures; 2) In an analysis of brain activation for body language in general (emotion and action conditions combined), participants with autism activated bilateral middle temporal and right IFG significantly less than controls. However, they showed significantly greater activation in left cuneus; and 3) A 2 Group (autism vs. control) x 2 Condition (action vs. emotion) mixed ANOVA on the behavioral data showed a significant effect of condition on accuracy, F(1,10)=9.38, p<0.05 with all participants having greater accuracy in the action condition than in the emotion condition. Although there was no significant group effect, we found a significant group x condition interaction for reaction time, F (1,10)=5.93, p<0.05, with the control group being faster in the action condition relative to the emotion condition.

Conclusions:  The IFG and IPL are two main components of the human mirror neuron system (MNS) associated with the understanding of motor actions (Rizzolatti et al., 1996; Iacoboni et al., 1999).  That our participants with autism showed reduced response in these regions while inferring emotions from body postures indicates a possible use of an alternate cortical route in children with autism in decoding body language.  It is possible that the control participants may rely more on motor simulation to infer emotions (activation in IFG and IPL), whereas the participants with autism may be more visuospatial (greater cuneus activation) in their approach.