Tone Discrimination in Adolescents Who Have Lost Their Autism Diagnosis: Low-Level Auditory Perceptual Abilities

Background: A variety of previous studies have found an unusually heightened sensitivity to low-level perceptual auditory distinctions in individuals with autism spectrum disorders (ASD). Adults with ASD were able to distinguish tones that differ by just 1% in frequency at a higher rate than controls (Bonnel, et al., 2003). These findings have suggested a bias toward elementary stimuli that do not require cortical or cortical-cortical processing; such a bias has been proposed to reflect a neuroanatomical bias towards local stimulus processing, and could be implicated in language deficits (e.g., if a child is unable to make generalizations about stimulus differences that are within a linguistic category; Minshew & Hobson, 2008). Objectives: Adolescents who had ASD prior to age 5 but currently do not meet criteria (i.e., optimal outcome, OO) appear to have resolved their earlier high-level social and communicative deficits. However, we do not know the extent to which they may still show basic, low-level perceptual processing characteristics (possibly due to neuroanatomical characteristics) that continue to reflect their history of ASD. Tone discrimination provides an excellent domain for assessing whether symptoms of ASD are tightly linked to a particular perceptual bias because there is significant individual variability, and typical individuals do not typically perform at ceiling. Methods: Adolescents with high-functioning autism (HFA; n = 9, mean age 13.5 years) were compared to IQ-matched OO adolescents (n = 17, mean age 12.7) and typically developing children (TD; n = 23, mean age 13.3). Groups did not differ in age (p = .58) or full-scale IQ (p = .71), though they differed in receptive vocabulary (p = .007). Participants completed a tone discrimination task modeled on Bonnel et al., 2003, in which they listened to 40 pairs of tones at each of 3 levels of difficulty: 3% (easy), 2% (medium), 1% (hard). Responses required only a same/different judgment, and all participants completed a training set, with feedback, to ensure task comprehension. Results: Performance in the easiest (3%) condition showed relatively high accuracy (M = 83% correct) relative to the 2% (75% correct) and 1% (68% correct) conditions, suggesting that participants understood the task and that the difficulty manipulation was effective. A repeated-measures MANOVA of response sensitivity (d-prime) by Group (HFA, OO, TD) and Condition (3%, 2%, 1%) with PPVT as a covariate indicated a significant effect of Group (p = .02), with no interaction, such that the HFA and TD groups differed significantly from each other; OO and HFA (p = .09) and OO and TD (p = .28) group differences did not reach significance. Conclusions: These data are consistent with previous work showing a strength in tone discrimination in HFA, a perceptual bias that may have contributed to the autistic symptomatology.