Individuals on the autism spectrum often fail to respond to voices, a characteristic feature that has been suggested to reflect sensory deficits in voice processing or reduced reward value of human voices (the ‘social motivation theory’ of autism [Dawson et al., Child Dev., 2002; Chevallier et al., Trends Cogn. Sci., 2012]). While some studies have supported the latter hypothesis, a comprehensive brain-imaging study of voice processing in affected individuals has yet to be carried out.

A new study from SFARI Investigator Vinod Menon and colleagues addresses this gap. In work supported by his SFARI Research Award, he reports on the use of functional magnetic resonance imaging to identify regions of the brain that distinguish children with autism spectrum disorder (ASD) and their typically developing peers in response to unfamiliar voices or their mother’s voice. In each instance, the strength of activity of brain systems implicated in auditory, reward and salience detection was correlated with the degree of social communication function. Using a complex and diverse voice-processing brain network identified in a previous study (Abrams et al., Proc. Natl. Acad. Sci. USA, 2013), the authors further showed that multivariate connectivity patterns in this network in response to the mother’s voice accurately predicted case status (ASD vs. typically developing [TD]). They suggest that differences in this voice-processing network — in particular the reward circuitry — may underlie previously documented findings that individuals with ASD do not show a preference for biologically salient voices.