Non-speech and speech pitch perception among Cantonese-speaking children with autism spectrum disorder: An ERP study

Highlights

• Cantonese-speaking children with ASD manifested impaired speech perception.

• Cantonese children with ASD did not show the advantage in non-speech perception.

• Language background (i.e., Cantonese) might shape the perceptive abilities.

Abstract

The present study compared how Cantonese-speaking children with Autism Spectrum Disorder (ASD) and their typically developing counterparts perceived speech pitch and non-speech pitch information using ERP measurements. Sixteen children with ASD (mean age = 10.42 years, SD = 2.12 years) and sixteen normal controls (mean age = 9.48 years, SD =.86 years) participated in two experiments, in which Cantonese lexical tone contrasts and non-speech pitch variations were presented to children following an oddball paradigm when they watched a silent movie. The results showed that: 1) When processing speech pitch contour, the two groups did not differ in the amplitude of mismatch response (p-MMR), while typically developing controls showed larger mismatch negativity (MMN) responses than children with ASD. In the processing of speech pitch height, more positive p-MMR was observed among children with ASD than among normal controls and stronger MMN was found for typically developing children than for children with ASD. 2) For the processing of non-speech pitch, MMN rather than p-MMR was observed and the two groups did not differ significantly with each other in the amplitudes of MMN. These results indicated that Cantonese-speaking children with ASD manifested impaired ability when processing speech pitch information (i.e., lexical tone), which was in line with previous research. However, they did not show the advantage in processing non-speech or auditory pitch information, which was not in agreement with the previous studies. Results were discussed from the perspective of how language background (i.e., Cantonese) might shape the perceptive abilities of children.

Introduction

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by main deficits in language development and social communication, along with stereotyped behaviors or interests [1]. Although the symptoms severity varies among different people with ASD, there is one common symptom that the majority of the population exhibits hyper or hyposensitivity to sensory information to certain degree [2]. Using the Short Sensory Profile (SSP), Tomchek and Dunn (2007) found that 95% of the children with ASD showed dysfunctions in sensory processing. Given that sensory processing (e.g., auditory processing) skills are fundamental to the development of broader cognitive abilities [3], the atypical sensory symptoms may have an impact on the perception of input signals, such as the perception of speech [2,46,47].

Speech perception serves as the cornerstone for language development and the capabilities to perceive or discriminate speech signals in infancy can predict later language development [4,5]. To perceive speech, typically developing children interpret and integrate multiple information, such as acoustic, visual or even emotional cues from the context [6,7], while for individuals with ASD, they show a bias of processing fine-grained information and fail to focus on the whole picture [6,48]. The preference to local information or detail-focused style in sensory processing among people with ASD was termed as the Weak Central Coherence (WCC) [6,8]. Due to the detail-focused characteristics in sensory processing, people with ASD cannot reach normal level in speech perception, which requires the integration of multiple local information and holistic processing. However, the bias manifested by people with ASD enables them to show the advantages in processing low-level piecemeal information, such as the possession of absolute pitch (AP) [9,10]. AP refers to the inborn capabilities to identify the pitch information of isolated tones, which can be considered as the extreme fine-grained information, without external references [9]. As reported by previous studies, the occurrence of AP among normal population and musicians was about 0.01% [11] and 0.64% [9], respectively. However, its prevalence among people with ASD was as high as 5% [12]. The high prevalence of AP among people with ASD was a manifestation of their bias in locally-oriented perception [9]. Moreover, even individuals with ASD who did not possess AP were better at processing non-speech pitch than normal controls, which was another demonstration of their advantage in processing local features at the expense of processing global information [10]. WCC can be further supported by another theory, the Enhanced Perceptual Functioning (EPF) [13,14], which explains the neural mechanism underlying the detail-focused characteristics of individuals with ASD. According to the EPF, the primary brain cortex involved in low-level perception is over-functioned among individuals with ASD, whereas their connections of brain regions, required by processing high-level perceptual functions, are impaired or under-developed [14]. As a result, people with ASD exhibit stronger sensitivity to piecemeal information, while fail to reach normal level in global processing as compared to their typically developing counterparts.

As mentioned above, the argument that detail-focused style in sensory processing could generate advantages in perceiving non-speech pitch information has been widely demonstrated [6,9,14]. Moreover, the encodings of non-speech pitch could also be shaped by language experience that tonal language speakers were more sensitive to the variation of pitch information than those speaking non-tonal languages [15]. For example, Giuliano et al. [15] compared the detective sensitivity toward the variations of non-speech pitch between Mandarin and English native speakers by adopting both behavioral and event-related potential (ERP) measurements. The results showed that Mandarin speakers were more accurate in detecting pitch changes than their English counterparts and exhibited earlier neural responses than English speakers, indicating that Mandarin speakers were more sensitive to the variation of non-speech pitch information. Furthermore, the results demonstrated that tonal language background could lead to the general enhancement of accuracy in pitch perception and cortical representations of pitch information [15].

Despite of the above evidence that tonal language background could enhance non-speech pitch processing, increasing studies showed that typically developing children whose first language was Mandarin, were less sensitive to the variations of non-speech pitch as compared to Mandarin-speaking children with ASD [16,17]. The pioneering work was intriguing, while such results might be derived from the relative simpler tonal characteristics of Mandarin, as compared to other Chinese dialects, such as Cantonese. Chinese is a tonal language which includes different dialects, such as Mandarin and Cantonese. In Chinese, the variations of lexical tones are realized by altering the fundamental frequency (F0), specifically through the alterations of pitch height and pitch contour of speech [18,19]. Mandarin has four lexical tones and all of them are contour tones, which differ from each other only in pitch direction or contour [20,21]. By contrast, there are six tones in Cantonese, including both contour and level tones [20,22]. Similar to the tones in Mandarin, Cantonese contours tones differ in pitch directions; however, level tones in Cantonese share similar pitch contour and differ with each other mainly in pitch height [21]. Given the more complex tonal feature of Cantonese as compared to that of Mandarin, the exploration among Cantonese speakers allows more room to study the effect of the tonal language background on non-speech pitch processing among Chinese children with ASD.

Therefore, the present study aimed to explore the influence of tonal language background on non-speech pitch processing among Cantonese-speaking children with ASD and their typically developing counterparts by adopting the event-related potentials (ERP) measurements. Specifically, the oddball paradigm was applied and the components of mismatch negativity (MMN) and positive mismatch response (p-MMR) were compared between Cantonese-speaking children with ASD and typically developing controls. Below, we briefly introduce the two ERP components.

MMN is a negative ERP deflection and usually peaks between 150 and 200 ms after the onset of the stimulus, reflecting the automatic detection of the violation (deviant) of auditory regularities (standard) [24]. MMN is generated by subtracting the waveforms of standard stimuli from deviant ones and larger MMN amplitude is an index of stronger discriminative abilities or greater sensitivity toward the violations [25]. MMN can be elicited by both the variations of physical features of non-speech auditory stimuli, such as alterations in pitch, duration, and intensity, as well as the manipulations of complex speech stimuli, such as changes in lexical tone and lexical stress [26]. MMN is widely considered as a stable representation of automatic neural responses [[24], [25], [26]]. However, the latency and cortical distribution of MMN vary across different age ranges and develop with maturation [[21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]]. For example, Shafer et al. [30] found that the peak latency for children between 4- to 7- year-old occurred between 300 and 400 ms and the latency of MMN was shortened by 25 ms per year with the increase of age [30]. In addition, the scalp distributions of MMN also differ between adults and children [32]. For adults, the MMN is dominantly distributed over the frontocentral cortex [25]. While for children, the distribution is broader such that the parietal area is also involved besides frontal and central cortex [33]. Using the MMN topographic evaluation, Martin et al. found that the underlying neural responses of MMN wouldn’t be mature and reach adult level until 11-year-old [34].

Recently, a more age-relevant ERP component, called p-MMR (positive mismatch response), has been reported by numerous studies [21,30,31]. Different from MMN, p-MMR is a positive deflection and can be observed even in infants [35,36]. For example, in the study by Tew et al. [36], the neural response toward the variation of non-speech pitch was compared between adults and infants. A typical MMN between 200 ms and 500 ms was found among adults. While for infants, p-MMR, rather than MMN, was detected within the same time window [36]. In some studies involving elder children, p-MMR co-existed with MMN, while the onset of p-MMR was earlier than MMN [30,31]. For example, Shafer et al. [30] found that for children aged between 4 and 7-year-old, MMN and p-MMR co-existed when processing the deviants varying in vowel. Specifically, the p-MMR was observed between 100 and 200 ms and MMN was observed between 300 and 400 ms. Moreover, the amplitude of p-MMR diminished with age [30]. Previous studies also found that the efforts that children made to discriminate the deviants also had an impact on the amplitude and latency of p-MMR [21,37]. Specifically, the more difficult for the children to discriminate standards and deviants, the larger amplitude or/and longer latency that would be manifested by p-MMR. For example, in the study by Ahmmed et al. [37], the neural responses toward different frequency contrasts (standard: 1000 Hz, 2% deviant: 1020 Hz, 5% deviant: 1050 Hz, 10% deviant: 1100 Hz) were compared between children with specific language impairments (SLI) and typically developing controls [37]. The result showed that for 2% contrast (standard: 1000 Hz, 2% deviant: 1020 Hz), both groups exhibited p-MMR. Moreover, typically developing group exhibited the shift from p-MMR to MMN when the contrasts changed from 2% to 5%. While for children with SLI, a frequency separation of more than 10% was required to generate similar transformation. Taken together, the p-MMR can be considered as the precursor of MMN and the occurrence of p-MMR may be influenced by both the stimulus-related and biological criteria [21,30,31,37].

In summary, the current study aimed to examine how the tonal language background influenced the non-speech pitch processing among Cantonese-speaking children with ASD and typically developing controls in an oddball paradigm using ERP measurement. Two related ERP components (MMN and p-MMR) were compared between the two groups. In addition, given the tonal characteristics of Cantonese, both the processing of speech pitch contour and height could be compared between the two groups. The current study was conducted to answer two research questions: a) whether Cantonese-speaking children with ASD were less sensitive to the variations of both speech pitch height and contour than normal controls. b) whether the advantage in processing non-speech pitch revealed in previous research could still be found among Cantonese-speaking children with ASD when compared with their typically developing counterparts. To answer these questions, two experiments were conducted. Specifically, Experiment 1 was administered to compare the speech pitch processing between children with ASD and typically developing controls with both the speech pitch height and pitch contour manipulated. Experiment 2 was conducted to compare the non-speech pitch perception between the two groups. Given that no such studies have been done among Cantonese speakers, no clear predictions were made prior to the formal tests.