Alexithymia, Metarepresentation, and the Dorsomedial Prefrontal Cortex in Autism

Background: Individuals with autism spectrum conditions (ASC) report much difficulty in understanding their own emotions; a trait termed ‘alexithymia'. Research in typical development has shown that the appraisal of subjective emotional experience relies heavily on the engagement of dorsomedial prefrontal cortex (dMPFC). In autism, dMPFC is hypoactive during stimulus-driven emotional self-appraisals (Silani et al., 2008, Soc Neurosci). However, this observation could be due to something inherent about stimulus-driven (i.e. bottom-up) processing, or a more general deficit involved in metarepresentational (i.e. top-down) processing. To disentangle these explanations, we examined whether alexithymic traits would covary with dMPFC activity during a task that has a metarepresentational component (reflecting on one's own mental states), but does not evoke stimulus-driven bottom-up processing. Based on work showing that dMPFC is increasingly engaged during earlier stages of development (Blakemore, 2008, Nat Rev Neurosci) and is increasingly engaged during higher levels of metarepresentation (Coricelli & Nagel, 2009, PNAS), we hypothesized that dMPFC activity would be increased in individuals with higher levels of alexithymia (i.e. a positive correlation). However, because this region is both structurally and functionally atypical in autism, we expected that the hypothesized association between dMPFC activity and alexithymia would either be reversed (i.e. negative correlation) or non-existent (i.e. correlation near 0) in ASC.

Objectives: To assess the association between dMPFC function and individual differences in alexithymia during high-level self-representation.

Methods: 29 adult males (18-45 years old) with a diagnosis of an autism spectrum condition (ASC), and 33 age-, sex-, and IQ-matched neurotypical adults were scanned with fMRI at 3T while making mentalizing or physical judgments about themselves or a non-close other. Regions analyzed were dMPFC, medial orbitofrontal cortex (mOFC), bilateral anterior insula (AI), and bilateral amygdala (Amyg).

Results: Replicating past research, individuals with ASC report more alexithymic traits than neurotypical individuals. In the neurotypical group, Increases in alexithymia were associated with increases in dMPFC activity (r = 0.57, p = 0.0005). However, in the ASC group this association was absent (r = -0.01, p = 0.95). The difference in these correlations between-groups was significant (z = 2.45, p = 0.01). No other region was associated with individual differences in alexithymia in either group (-0.16 < r < 0.21, p > 0.23).

Conclusions: The dMPFC is critical for metarepresentational ability; particularly in appraising subjective emotional experience. In the absence of stimulus-driven processing, increasing alexithymia was strongly associated with increased recruitment of dMPFC during self-mentalizing in the neurotypical group. However, although individuals with ASC report more alexithymia than that of neurotypical individuals, the magnitude of alexithymia was not associated with dMPFC activity. It is likely that the altered neurodevelopment of dMPFC in autism derails the normative role of dMPFC in metarepresentational function. However, the magnitude of alexithymic difficulty in ASC is likely to involve deficits in other aspects of emotion processing besides metarepresentation (e.g., bottom-up emotion processing).