Text

Study rationale: Individuals with autism spectrum disorder (ASD) are often impaired in the learning and identification of others by face and voice. However, familiar (famous) voices may be an exception, with some evidence for similar identification rates in ASD and neurotypical participants. To further investigate the issue, we developed the Jena Famous Voices Test (JFVT; [1]), a 15 min screening tool with voice samples of 30 German celebrities. The JFVT comes in two parallel versions for which appropriate celebrities are selected relative to a target population, and provides identification performance scores that are corrected for person knowledge (i.e., proportion of voices identified relative to proportion of persons known). We compared identification performance for famous voices across 83 healthy listeners with a low to high autism-spectrum quotient (AQ, Study 1), as well as across 24 listeners with and without an ASD diagnosis (Study 2). To test for voice learning abilities, we further employed the Jena Voice Learning and Memory Test (JVLMT; [2]) to assess unfamiliar voice memory, and for comparison with face processing abilities, we used two analogous identification and learning tests from the face domain (J-BFFT-2022; CFMT).

Results: In Study 1, no correlations (all rs < .13; all ps > .20) were observed between autistic traits (AQ) in the general population and performance in person identification and learning tests, neither for voices nor for faces. Parallelism of the two JFVT versions was verified by non-significant differences in identification performance (p > .05). In Study 2, we found significant impairments in the ASD group compared to controls across all tests (p < .001), with no interactions with test type.

Conclusions: First, the JFVT is an economical and adaptable new tool to assess voice identification abilities for famous speakers, complementing other tests on high-level auditory perception. Second, general impairments of voice recognition in ASD, both for famous and for trained-to-familiar voices, argue against relatively preserved familiar voice identification abilities in ASD for people scoring highly on the autism spectrum, as in individuals with an ASD diagnosis, but not in the general population.

Literatur