Unusual Auditory Filtering Behaviors in Minimally Verbal ASD: A Mechanism for Regulating Auditory Input?

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
S. Schwartz1, L. Wang2, B. Shinn-Cunningham2 and H. Tager-Flusberg3, (1)Graduate Program for Neuroscience, Boston University, Boston, MA, (2)Biomedical Engineering, Boston University, Boston, MA, (3)Psychological and Brain Sciences, Boston University, Boston, MA
Background: Many people with ASD exhibit behaviors that suggest irregular responses to auditory inputs and a need to actively regulate their auditory environment. Verbally fluent ASD children with high frequency atypical auditory behaviors and overall repetitive and restricted behaviors (including atypical responses to sounds) demonstrate abnormal sensitivity to sound changes in psychoacoustic experiments and have been described as having average nonverbal IQ and delayed language acquisition (Jones et al., 2009; Kargas et. al., 2015). No research has explored whether minimally verbal ASD individuals with frequent auditory behaviors display atypical sensitivity to sound changes, a necessary mechanism for accurately filtering out irrelevant sounds in the environment. In addition, no research has determined whether regulatory behaviors or filtering issues interfere with receptive language.

Objectives: Our objective was to examine the relationship between atypical auditory behaviors and atypical auditory processing, especially in minimally verbal individuals, and the relationship of these factors to receptive language skills.

Methods: Auditory behaviors were coded from ADOS recordings in individuals with ASD aged 5-21. Individuals were classified as either minimally verbal (ASD-MV: single word or phrase speech) or verbally fluent (ASD-V: complex speech). We coded: 1) Prolonging and/or abnormally manipulating sound-making objects, 2) Humming and/or altering ear shape, 3) Covering the ears with hands or fingers or appearing distressed by noise. Nonverbal intelligence and receptive language were measured with the Leiter and PPVT, respectively. Event-related potentials (ERPs) were measured in a passive, single-stream oddball mismatch negativity paradigm designed to quantify automatic detection of sound input change.

Results: ASD-MV individuals displayed a greater incidence of auditory behaviors than ASD-V individuals (χ2(1)=19.66, p<0.01), with the greatest difference in humming and ear shaping (t(10)=-2.24, p=0.04). Significant differences between groups were not observed for the other two behavioral categories (object manipulation (t(10)=-1.83, p=0.09); avoidance/ear covering (t(10)=-1.43, p=0.18)). Because we observed a large range in the frequency of auditory behaviors in the ASD-MV group, we did further within-group analyses. We found that time spent engaging in auditory behaviors significantly correlated with receptive language raw scores (Rho=-0.575, p=0.003; Figure 1A), but not with nonverbal intelligence (Rho=-0.256, p=0.23). Increases in time spent engaging in these behaviors was also associated with reduced early ERP responses to sound change (Rho=0.54, p=0.08; Figure 1B).

Conclusions: This work finds that auditory behaviors like humming and ear shaping are more common in ASD-MV than ASD-V. Within ASD-MV, there is a subset of individuals engaging in auditory behaviors at least 10% of the time. Individuals in this subgroup exhibit extremely low receptive language skills, but within-range nonverbal intellectual functioning, relative to others who are also minimally verbal. There is evidence that this subgroup is less efficient at detecting sound input change and consequently may be less efficient at separating and filtering sounds. Past experience with under-filtered sounds may increase auditory behaviors to serve as a mechanism for regulating auditory input; however, this regulatory activity may also inadvertently block sounds that are important. Future work should continue to investigate the clinical implications of how these behaviors affect receptive language processing.