Autism Symptom Severity Modulates Responsivity In the Mirror Neuron System: A Replication and Extension of Prior Research Findings

Background: A large number of studies – relying on a variety of neuroimaging tools and paradigms – have reported abnormalities in the so-called mirror neuron system (MNS) in individuals with autism. However, a few studies have failed to find significant group differences. These negative findings have been heralded as evidence against the hypothesis that MNS dysfunction may contribute to core deficits in autism.

Objectives: The present study had two main aims. First, we aimed to replicate our prior findings of MNS dysfunction in a considerably larger and independent sample of children and adolescents with ASD. Second, we aimed to further examine how responsivity within the MNS might vary as a function of symptom severity in order to assess how relatively minor differences in sample characteristics may affect study results when comparing individuals with ASD to neurotypical controls.

Methods: While undergoing two fMRI scans, a total of 40 children and adolescents with ASD and 14 typically-developing (TD) control subjects (matched by age, IQ and head motion) passively observed or imitated faces displaying different emotions (angry, fearful, happy, sad, and neutral). As in our prior study (Dapretto et al., 2006), we used a jittered event-related design, where faces were presented every 3 sec (with each face being displayed for 2 sec) according to an optimized random sequence. The order of the two scans was counterbalanced within and between groups. All participants in the larger ASD sample (ASD_ALL) had an autism diagnosis based on the ADI and best clinical judgment; on the ADOS, 14 participants met criteria for full autism (ADOS_AUT), 18 participants only met criteria for ASD (ADOS_ASD), and 8 participants did not meet criteria for either (ADI_ONLY).

Results: When comparing the largest ASD sample (ASD_ALL) to TD controls, we replicated our previous findings showing significantly greater activity in the frontal component of the MNS in the TD group for both the Imitate and Observe conditions. The results held when comparing the most severely affected group (ADOS_AUT) to TD controls; however, when comparing the less affected group (ADOS_ASD and ADI_ONLY) to TD participants, significant between-group differences were observed only for the Imitate condition and at less stringent statistical thresholds. Consistent with our previous findings, a significant negative correlation was also observed between frontal MNS activity and individual scores on the Social and Communication Subscale of the ADOS such that the most impaired individuals showed the least amount of MNS activity.

Conclusions: These results add to a large body of work indicating hyporesponsivity in the MNS in individuals with autism. Importantly, these findings may help explain some of the discrepant results in the literature showing that relatively small differences in symptom severity (as indexed by the ADOS) may determine whether significant group differences are observed. Overall, these findings highlight the need to carefully consider symptom severity as well as sample heterogeneity within and between studies, particularly when making sense of conflicting results.