Deficient brainstem encoding of pitch in children with Autism Spectrum Disorders

Abstract

Objective: Deficient prosody is a hallmark of the pragmatic (socially contextualized) language impairment in Autism Spectrum Disorders (ASD). Prosody communicates emotion and intention and is conveyed through acoustic cues such as pitch contour. Thus, the objective of this study was to examine the subcortical representations of prosodic speech in children with ASD.

Methods: Using passively evoked brainstem responses to speech syllables with descending and ascending pitch contours, we examined sensory encoding of pitch in children with ASD who had normal intelligence and hearing and were age-matched with typically developing (TD) control children.

Results: We found that some children on the autism spectrum show deficient pitch tracking (evidenced by increased Frequency and Slope Errors and reduced phase locking) compared with TD children.

Conclusions: This is the first demonstration of subcortical involvement in prosody encoding deficits in this population of children.

Significance: Our findings may have implications for diagnostic and remediation strategies in a subset of children with ASD and open up an avenue for future investigations.

Figure 1

Mental (left) and language ability (right) means (standard errors) for TD and ASD groups. Children with ASD demonstrated poorer mental and language abilities, although their mental ability level was within normal limits.

Figure 2

Representative pitch-tracking contours extracted from brainstem responses of TD (left) and ASD (right) individuals. The fundamental frequency contour of the response (red) is plotted against the contour of the stimulus (black). Shown here are data from both the descending (top) and ascending (bottom) /ya/ stimuli. Pitch tracking is more precise in the typically-developing system. Frequency (Hz) is plotted along the y-axis. The x-axis shows the time corresponding to the midpoint of each 40-ms time bin analyzed.

Figure 3

Autocorrelograms of individual TD (left) and ASD (right) brainstem responses to descending (top) and ascending (bottom) /ya/ stimuli. Running autocorrelations quantify the degree of neural phase locking over time. The autocorrelograms (lag versus time) act a means of visualizing periodicity variation over the course of the response. The time indicated on the x-axis refers to the midpoint of each 40-ms time bin analyzed. The y-axis refers to the amount of lag between the signal (each 40-ms time bin) and a time-shifted copy, and the third dimension, Pitch Strength, is plotted using a color continuum from black to white, with brighter colors representing higher correlations, or more robust encoding of the fundamental frequency contour. The TD response indicates more precise phase locking of pitch than the ASD response.

Figure 4

Group means (standard error) for f₀ Frequency Error (Hz), Pitch Strength (autocorrelation r values), H2 Frequency Error (Hz) and Composite Score (z values). Encoding was significantly more precise in TD responses (left, black) as compared to the ASD group as a whole (middle left, dark gray). ASD OUT children (light gray) are those who have pitch tracking composite scores outside of the TD group, while ASD IN children (middle right, white) have scores that are within the normal range. The ASD OUT group (far right, gray) was largely driving the significant group differences, as the ASD IN group demonstrated encoding similar to the TD group.