Autistic children with early language delays process sound differently, study finds

Children on the autism spectrum with early language delays process sound differently than their typically developing peers, showing a distinct preference for pitch over timing. These differences in sound perception might actually help some autistic children develop language and cognitive skills through alternative learning pathways. These findings were recently published in the journal Autism Research.

To understand human speech, the brain must quickly interpret both the timing and the pitch of sounds. The timing of sounds, known as temporal properties, involves how long a sound lasts or when it stops and starts. These time-based features are essential for distinguishing consonants and picking up the rapid, natural rhythm of speech. If a child struggles to process temporal information, they often experience difficulties forming stable mental representations of words, which is a pattern frequently observed in language disorders like dyslexia.

Pitch refers to the spectral properties of sound, which involves the specific frequencies that make a voice or a sound wave high or low. The human brain typically divides the labor of processing these two types of acoustic information. In most neurotypical people, the brain’s left hemisphere specializes in tracking rapid temporal changes, while the right hemisphere tends to focus on slower, spectrally based features like pitch contours.

Autistic individuals often perceive the sensory world differently from neurotypical individuals. Past theoretical models propose that autism might involve a basic reorganization of how the brain processes and prioritizes information. This concept suggests a shift away from complex, socially integrated networks and toward more direct, independent sensory perception. While neuroimaging studies of visual processing tend to support this idea, evidence regarding how the autistic brain organizes auditory information remains limited.

Luodi Yu, an associate professor in the Department of Special Education and director of the Autism Research Center at Guangzhou University in China, designed a new experiment to test these auditory differences directly. Yu co-authored the study alongside Laurent Mottron, a psychiatrist and professor in the Department of Psychiatry and Addictology at the University of Montreal, who also authored the book Si l’autisme n’est pas une maladie, qu’est-ce que c’est? (If Autism Is Not a Disease, What Is It?).

Yu noted that past work pointed to an interesting split in how autistic children hear the world. “Both Dr. Mottron’s research and work from my own group, among others, have suggested that pitch perception (the frequency aspect of sound) can be enhanced in people with prototypical autism or autism with early speech delay,” Yu said. “At the same time, we and others have observed that processing fine timing details in sounds, such as consonant contrasts or rapid auditory transitions, can be challenging for autistic children.”

“This contrast was intriguing because frequency and timing are two fundamental dimensions of sound,” Yu explained. “Both are important for speech and language, but they seemed to follow different directions in autism. So we wanted to test this directly by asking whether they show a different auditory bias: a stronger weighting of spectral or frequency information relative to temporal or timing information.”

To measure these auditory abilities, the scientists recruited 44 children between the ages of eight and 12. The sample included 21 autistic children, consisting of 20 boys and one girl. It also included 23 typically developing children, consisting of 22 boys and one girl. Both groups were matched by age to ensure accurate comparisons. All autistic participants had a clinical diagnosis of autism spectrum disorder and a history of delayed speech onset, meaning they did not use two-word phrases by age two.

The researchers used two specialized listening tests to measure the children’s absolute limits of perception. The first test was a gap detection task, designed to measure temporal resolution. Temporal resolution refers to the brain’s ability to detect rapid changes in sound over time. During this task, the children listened to two static noise sounds, which were represented on a computer screen by cartoon televisions. One sound was continuous, while the other had a microscopic split-second of silence hidden in the middle.

The children had to identify the broken television by selecting the sound with the gap. The second test was a frequency modulation detection task, which measured spectral resolution, or the ability to detect subtle shifts in pitch. The children listened to two tones, represented on the screen by cartoon robots. One tone held a steady pitch, while the other wavered slightly in frequency, prompting the participants to identify the broken robot by choosing the wavering tone.

Before the official tests began, researchers methodically trained the children using visual aids and demonstration videos. They ensured each child understood the rules and could pass practice trials before moving forward. During the actual tasks, the testing software used an adaptive staircase procedure. This means the computer automatically adjusted the difficulty based on the children’s answers, making the differences smaller and smaller to find the exact threshold where they could no longer hear a change.

The findings highlighted opposite strengths and weaknesses between the two groups of children. The typically developing children detected much smaller temporal gaps than the autistic children, showing better timing-based auditory processing. The autistic children needed significantly longer moments of silence to notice the interruption in the static noise. This indicates that the autistic participants experienced reduced temporal resolution compared to their peers.

When it came to detecting changes in pitch, the pattern reversed entirely. The autistic children successfully identified much smaller shifts in frequency than the typically developing children. This result provides evidence that the autistic children possessed superior spectral resolution.

Yu was struck by how prominent these differences were in the data. “The size and specificity of the effects were somewhat surprising,” Yu told PsyPost. “In much of the current autism literature, effect sizes are often modest, partly because autism is very heterogeneous and many studies include participants with quite different developmental profiles.”

“Here, the pattern was very clear: autistic children performed better than typically developing children on the frequency task, but worse on the closely matched timing task,” Yu added. “It is rare in autism research to see such a specific dissociation between two related tasks. This reminded us that sensory-perceptual processing may be an important window for detecting the ‘autism signal’, especially when the sample is carefully defined and the task directly measures the perceptual dimension of interest.”

The researchers mathematically combined these two scores into an Auditory Bias Index to quantify each child’s overall sensory preference. This index revealed that the two groups had opposite ways of processing sound. Typically developing children showed a bias toward fine-grained timing perception, which is usually associated with standard language development and left-hemisphere brain activity. In contrast, the autistic children showed a pronounced bias toward pitch perception. The scientists noted that this reversed preference remained stable across the ages tested in the autism group.

The researchers also looked for connections between these auditory thresholds and the children’s cognitive and language skills. In the autistic group, a better ability to detect subtle pitch changes correlated with higher nonverbal intelligence scores. It also correlated with stronger receptive language abilities. Receptive language refers to a person’s ability to understand words and sentences spoken by others.

“Autistic children’s auditory system is not ‘broken.’ They hear differently,” Yu said. “In our study, autistic children with early language delay were less sensitive to very brief gaps in sounds, but more sensitive to subtle changes in pitch or frequency. This means that their auditory profile was a reversal of bias: they showed an advantage in one basic dimension of sound and a difficulty in another.”

Interestingly, gap detection scores improved with age in the autistic group, while typically developing children seemed to have already reached mature timing perception. A better ability to detect timing gaps was also associated with stronger receptive language in the autistic group. The authors suggest that an enhanced sensitivity to pitch might help autistic children learn language through non-traditional pathways.

“This may matter for language learning,” Yu explained. “Spoken language unfolds very quickly over time, so timing information is important for detecting consonants, word boundaries, and syntax. But pitch, sound patterns, letters, numbers, and other stable or regular features may be easier for some autistic children to detect and learn from.”

Yu noted that this could help explain why some autistic children show a strong interest in written symbols, numbers, or even another language, while still having significant delays in spoken language. For example, some autistic individuals develop unexpected bilingual abilities by memorizing the precise pitches of words from videos or digital games, completely bypassing the need for social interaction.

“So rather than viewing auditory differences in autism only as impairments, we should ask how these perceptual differences shape learning, and how education can make better use of autistic children’s strengths,” Yu said.

These associations might lead some readers to assume that heightened pitch sensitivity directly causes better language skills, but the researchers note this would be a misinterpretation. The current study only measured correlations at a single point in time, meaning it cannot prove cause and effect. It remains possible that early language delays actually cause the developing brain to rely more on pitch.

Yu highlighted these restrictions when discussing the study’s scope. “First, our sample was relatively small and intentionally homogeneous,” Yu said. “We studied autistic children with early language delay and early autistic manifestations, so the findings should not be generalized to all people on the spectrum. Autism spectrum is highly heterogeneous, and children without language delay, older individuals, or people with different developmental profiles may show different auditory patterns.”

“Second, the study was cross-sectional,” Yu added. “We found associations between auditory processing and language abilities, but we cannot conclude from this study alone that auditory processing differences cause language differences. Longitudinal studies especially from the beginning of life are needed to understand whether these auditory biases precede, co-develop with, or emerge as adaptations to language delay.”

Additionally, the tests used in this study only measured very specific aspects of sound processing. The frequency modulation task only tested low-pitch sounds, which the ear and brain process differently than high-pitch sounds. The gap detection task only measured one form of timing perception, leaving out other abilities like judging the specific order of rapid sounds. The scientists recommend that future research address these gaps by using a wider variety of auditory tests.

The researchers hope to track how these sensory preferences develop over a lifespan. They plan to continue exploring how these unique sound processing traits influence early development.

“Our next step is to examine autistic auditory processing and language acquisition in greater detail,” Yu said. “We want to understand how differential auditory processing interacts with language learning and how it may shape unconventional learning pathways.”

Yu hopes this knowledge will eventually change how educators approach learning differences. “More broadly, our goal is to move from simply describing autistic perceptual differences to using them constructively,” Yu explained. “If some autistic children prioritize stable, pattern-based information over rapid sound sequences, education should not only try to remediate weaknesses. It should also identify and leverage their differential information processing as potential strengths. In the long term, we hope this work can contribute to more individualized and strengths-based approaches to language and education for autistic children.”

By relying on pitch rather than timing, some autistic children might decode the structure of language through alternative, nonsocial avenues. Exploring these unconventional learning pathways could provide new strategies for supporting communication development in autistic individuals.

“We hope this study contributes to neurodiversity-informed understanding of autism. A difference can be both a challenge and a strength, depending on the learning context,” Yu concluded. “Enhanced pitch or frequency sensitivity may complicate some aspects of real-time spoken language processing, especially when language depends on rapid timing sequences. But the same sensitivity may also support other forms of learning, especially when information is structured, repeated, visually supported, pattern-based, or presented in less socially demanding contexts. For us, the important question is not how to make autistic children hear or learn like typically developing children. The more useful question is how to understand their own learning architecture and build education around it.”

The study, “Autistic Children With Speech Onset Delay Show Reversed Bias in Spectral Versus Temporal Auditory Processing,” was authored by Luodi Yu, Shuyu Xie, Li Wang, and Laurent Mottron.