Voice Processing in High-Functioning Autism Spectrum Disorder: Neuronal and Behavioral Mechanisms

Background: The ability to recognize the identity of others is an evolutionary conserved process and a key requirement for successful communication. In autism spectrum disorder (ASD), a condition that is associated with difficulties in social interaction, there is good evidence that the ability to identify another person by face is deficient (Weigelt et al., 2012). Previous research suggests that voice identification processing is also impaired in ASD (Schelinski et al., 2014). However, the underlying neuronal and perceptual mechanisms of this voice processing deficit are unknown.

Objectives: Here, we systematically investigated the behavioral and neuronal mechanisms of voice processing in high-functioning ASD.  

Methods: Sixteen adults with high-functioning ASD and sixteen typically developed controls (matched pairwise on age, gender, and IQ) participated in two fMRI experiments and a comprehensive behavioral test battery on voice recognition. In the first fMRI experiment (vocal sound experiment), participants passively listened to blocks of vocal (speech and non-speech) and non-vocal sounds (e.g. musical instruments, nature, animals) (Belin et al., 2000). In the second fMRI experiment (voice recognition experiment), participants performed voice identity recognition and speech recognition tasks on the same stimulus material (von Kriegstein & Giraud, 2004). The behavioral test battery included tests on (i) recognition of newly learned voices, unfamiliar voice discrimination and famous voice recognition, (ii) acoustic processing abilities that are associated with voice recognition (i.e. tests on vocal pitch and timbre discrimination), (iii) control tasks (i.e. tests on musical pitch and timbre perception, and face identity recognition). Results were considered as significant at α = .05 (fMRI: p < .05 family wise error (FWE) corrected for region of interest).

Results: The behavioral results showed that the ASD group had particular difficulties with discriminating, learning, and recognizing unfamiliar voices, while recognizing famous voices was comparable to controls. Difficulties with unfamiliar voices correlated with similar difficulties in recognizing unfamiliar faces. Tests on acoustic processing abilities showed that the ASD group had a specific deficit in vocal pitch discrimination that was dissociable from otherwise intact acoustic processing (i.e. musical pitch, musical timbre, and vocal timbre perception). The vocal sound experiment showed that passive listening to vocal sounds in contrast to non-vocal sounds elicited similar activity in voice-sensitive cortices in the superior temporal sulcus and gyrus (STS/STG) in the ASD and the control group. The voice recognition experiment showed reduced BOLD responses to the voice identity recognition in contrast to the speech recognition task in the ASD compared to the control group in the right posterior STS/STG—a region that has been previously implicated in processing acoustic voice features.

Conclusions: Our results allow, for the first time, a characterization of the voice recognition deficit in ASD: Both the behavioral and neuronal findings indicate that in high-functioning ASD, the ability to recognize voices is impaired because of difficulties with processing acoustic aspects of voices, that is, difficulties with integrating the acoustic characteristics of the voice into a coherent percept. A deficit in voice processing might contribute considerably to the development of communication difficulties - a core hallmark of ASD.