Superior Vocal Identity Memory in Autism

Background:  Faces and voices are the most important social stimuli in the visual and auditory domains. Individuals with autism have difficulty in memorizing faces. Since children with autism exhibit less preference toward mother's voice, and brain activation related to voice perception is less significant in individuals with autism than in neurotypical individuals, it is important to investigate how individuals with autism perceive voices.

Objectives:  This study adopted a quantitative methodology to examine vocal identity recognition in individuals with autism.

Methods: Fourteen individuals with autism (3 females) and 14 age- and IQ-matched neurotypical adults (3 females) participated in this study. The study consisted of three experiments: (1) gender discrimination, (2) vocal identity recognition via naming, and (3) a familiarity test. In the first experiment, the participants identified the genders of the speakers (5 female and 5 male speakers). In the second experiment, the participants learned the correlation between the voices (heard in the first experiment) and their names. In the third experiment, the participants identified familiar voices (from the 10 previously heard speakers) and unfamiliar voices (from another 10 speakers).

Results: In the first experiment, the performance of both groups reached the ceiling level, and there was no significant difference between the groups. In the second experiment, the performance of both groups improved after training, and there was no significant difference between the groups. In the third experiment, the performance of the autism group was significantly better than that of the control group.

Conclusions:  The results show that individuals with autism could recognize speakers' gender and memorize speakers' identity. The superior performance in the familiarity test contrasts with the mixed results in previous studies where the performance of the autism group was similar or worse. When compared with previous studies using long uttered sentences, the short uttered words used in this study provided timbre cues instead of prosody cues for vocal identity recognition. Therefore, the superior performance in the autism group might result from their superior perception of timbre cues in voices.