Auditory evoked potentials in adolescents with autism: An investigation of brain development, intellectual impairment, and neural encoding

Abstract

Limited research has evaluated neural encoding of sounds from a developmental perspective in individuals with autism (ASD), especially among those with intellectual disability. We compared auditory evoked potentials (AEPs) in autistic adolescents with a wide range of intellectual abilities (n = 40, NVIQ 30-160) to both age-matched cognitively able neurotypical adolescent controls (NT-A, n = 37) and younger neurotypical children (NT-C, n = 27) to assess potential developmental delays. In addition to a classic measure of peak amplitude, we calculated a continuous measure of intra-class correlation (ICC) between each adolescent participant's AEP and the age-normative, average AEP waveforms calculated from NT-C and NT-A to study differences in signal morphology. We found that peak amplitudes of neural responses were significantly smaller in autistic adolescents compared to NT-A. We also found that the AEP morphology of autistic adolescents looked more like NT-A peers than NT-C but was still significantly different from NT-A AEP waveforms. Results suggest that AEPs of autistic adolescents present differently from NTs, regardless of age, and differences cannot be accounted for by developmental delay. Nonverbal intelligence significantly predicted how closely each adolescent's AEP resembled the age-normed waveform. These results support an evolving theory that the degree of disruption in early neural responses to low-level inputs is reflected in the severity of intellectual impairments in autism.

Keywords: AEPs; ASD; auditory; autism; autistic; evoked potentials; intellectual disability; intellectual impairment; neural development.

Figure 1.

Schematic of seven fronto-central channels selected from 128-channel EGI cap for primary AEP analyses.

Figure 2.

Histogram of intelligence scores (x axis), with Nonverbal IQ (NVIQ) and Verbal IQ (VIQ), respectively. Bars are binned with an interval width of 5 points along the x axis. Horizontal panels display histograms for autistic (ASD), neurotypical adolescent (NT-A), and neurotypical child (NT-C) groups.

Figure 3:

Evidence of P1-N1-P2-N2 AEP wave morphology in adolescents with ASD. Y axis is amplitude (uV) and x axis is time relative to stimulus onset (ms).

Figure 4:

AEPs across adolescents who are neurotypical (NT-A, n = 38) and who have ASD (n = 40) displayed by A. ERP traces and B. Bar plots of P1-N1 and P2-N2 peak-to-peak amplitudes. Error bars indicate 95% confidence intervals. Significance values based on repeated measures ANOVAs described in detail in Table 2 (*p<0.05, **p<0.01, ***p<0.001).

Figure 5:

Comparison between the AEP waveforms of neurotypical children, ages 3–9 (n = 27) and neurotypical adolescents, ages 10–21 (n = 38). As described in prior research, this morphology changes between the age of 8–12 in neurotypical children, including the formation of a N1-P2 complex.

Figure 6:

Nonverbal IQ (NVIQ) was the single best predictor of age-normed ICC in ASD adolescents (R_adj² = 0.21, B = 0.48, p = 0.002). Pearson’s correlation between age-normed ICC was significant across all adolescents (r(78) = 0.414, p < 0.001), as well as across ASD adolescents alone (r(40) = 0.420, p = 0.007), but was not for NT adolescents (NT-A) alone (NS). From a stepwise linear regression across the full adolescent sample, we found no effect of group or interaction between NVIQ and group; NVIQ was the only significant factor and predicted 16% of the variance (F(1,76) = 15.693, p < 0.001; R²_adj = 0.160, B = 0.014, p < 0.001).