Abstract
Purpose
Investigations into the nature of communication disorders in autistic individuals increasingly evaluate neural responses to speech stimuli. This integrative review aimed to consolidate the available data related to speech and language processing across levels of stimulus complexity (from single speech sounds to sentences) and to relate it to the current theories of autism.
Method
An electronic database search identified peer-reviewed articles using event-related potentials or magnetoencephalography to investigate auditory processing from single speech sounds to sentences in autistic children and adults varying in language and cognitive abilities.
Results
Atypical neural responses in autistic persons became more prominent with increasing stimulus and task complexity. Compared with their typically developing peers, autistic individuals demonstrated mostly intact sensory responses to single speech sounds, diminished spontaneous attentional orienting to spoken stimuli, specific difficulties with categorical speech sound discrimination, and reduced processing of semantic content. Atypical neural responses were more often observed in younger autistic participants and in those with concomitant language disorders.
Conclusions
The observed differences in neural responses to speech stimuli suggest that communication difficulties in autistic individuals are more consistent with the reduced social interest than the auditory dysfunction explanation. Current limitations and future directions for research are also discussed.
Autism spectrum disorder (ASD) is characterized by social communication deficits and restricted and repetitive behaviors (American Psychiatric Association, 2013). Although not required for the diagnosis, many autistic 1 individuals present with concomitant receptive and/or expressive language difficulties that range in severity. Up to 60% of children on the autism spectrum may be diagnosed with a language disorder, and up to 30% plateau at minimal or no spoken language (e.g., Anderson et al., 2007; Tager-Flusberg & Kasari, 2013; Thurm et al., 2015).
The reasons for high incidence of language disorders in ASD are not yet fully understood. The reduced social motivation theory (Chevallier et al., 2012; Dawson, 1991; Dawson et al., 2005) posits that impairments in attention to social stimuli emerging early in life interfere with effective learning experiences in the social domain, including language development. Conversely, language deficits in ASD could also be attributed to more general abnormalities in the processing of low-level sensory information due to weak central coherence or enhanced attention to details at the expense of the whole (e.g., Frith, 1989; Groen et al., 2008; Mottron et al., 2006). More recently, Bayesian (Pellicano & Burr, 2012) and predictive coding (van Boxtel & Lu, 2013) perspectives suggested that many of ASD symptoms arise from abnormalities in both perception and learning. Autistic individuals may prioritize a subset of immediate sensory inputs over the broader context and prior experiences, making it difficult to separate relevant from irrelevant details or form generalizable representations (Haker et al., 2016). To date, findings from studies of speech and language processing in ASD have not been extensively examined in relation to these theories.
Speech is a complex auditory signal that requires several stages of processing, from isolating speech among other environmental sounds to phoneme identification, word recognition, and sentence comprehension. Alterations to any of these steps can disrupt speech and language processing, as has been shown in other clinical groups (see Hämäläinen et al., 2013; McArthur & Bishop, 2005; Tallal & Piercy, 1974, for reviews). Noninvasive recordings of electrical (electroencephalography [EEG]) or magnetic (magnetoencephalography [MEG]) signals in the brain offer a direct measurement of neural activity. Both techniques involve placing sensors on (EEG) or near (MEG) the scalp and record postsynaptic activity of thousands of neurons located mainly in the cortex. The strength of the signal depends on the number of active neurons, the temporal synchrony of their firing, and their location relative to the scalp and to each other. EEG is sensitive to electrical fields generated by extracellular currents and can capture the activity of neural sources oriented radially or tangentially to the scalp. MEG detects the magnetic fields perpendicular to the intracellular electric currents and therefore records data primarily from tangential sources. Differences in conductivity between the brain, skull, and skin affect the electrical signal and thus limit spatial resolution of scalp-recorded EEG (approximately 10 mm). Conversely, magnetic fields pass through unimpeded, supporting MEG's good spatial resolution (2–3 mm) for effective neural source estimation. Practically, EEG recordings can be performed in a variety of settings (e.g., research labs, schools, and homes), whereas MEG requires a magnetically shielded room and specialized superconductor sensors cooled by liquid helium.
The millisecond-level temporal resolution of EEG/MEG provides a unique opportunity to objectively characterize multiple stages of auditory processing. The presentation of a stimulus (e.g., speech sound) causes a transient change in the ongoing brain activity. The resulting waveforms consisting of positive and negative peaks (i.e., event-related potentials [ERPs]/fields) reflect various neural processes elicited by the stimulus, from sensory detection of physical characteristics to comprehension (McWeeny & Norton, 2020). Of particular importance for the study of ASD, these data can be obtained using passive listening paradigms that do not require overt responses, making them optimal for individuals with a variety of developmental and language levels, including minimally verbal participants.
Recent meta-analyses of auditory processing in ASD (e.g., Chen et al., 2020; Schwartz et al., 2018) reported inconsistent findings with regard to the presence of group differences in auditory processing of speech between the autistic and typically developing participants. This could be due, at least in part, to a restricted focus on the sensory discrimination response (mismatch negativity [MMN]) or a small number of included studies that used speech stimuli, or because the analyses did not explicitly distinguish among the findings at the level of vowel, syllable, or whole word.
Therefore, the goal of this integrative review is to summarize the current neural evidence related to speech and language processing in autistic individuals across levels of stimulus complexity (from single speech sounds to sentences) and relate it to the proposed theories of autism to begin identifying the possible reasons for the language disorders in ASD. Auditory processing of nonspeech stimuli in ASD has been reviewed elsewhere (Haesen et al., 2011; O'Connor, 2012; Williams et al., 2020).
Method
The reviewed studies were identified following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (Moher et al., 2010; see Figure 1). A thorough literature search using PubMed, PsycINFO, Google Scholar, and Web of Science database queries with the terms (AUTIS* OR ASD OR ASPERGER) AND (CAEP OR ERP OR MEG) AND (SOUND OR LANGUAGE OR SPEECH) identified articles published prior to April 3, 2020. Criteria for inclusion were empirical human studies of auditory processing involving participants on the autism spectrum, publication in English language in a peer-reviewed journal, and statistical evaluation of contrasting speech or language stimuli. Approximately 30% of the articles were double coded to ensure accuracy, and any disagreements were resolved through discussion.
Figure 1.
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.
The 51 selected articles were organized by increasing stimulus complexity: vowels, consonant–vowel (CV) and multisyllabic stimuli, meaningful words and sentences, and suprasegmental prosodic processing. Within each section, when available, the results were further divided into basic sensory and higher order attentional processing. If additional content themes emerged during review, the associated articles were grouped accordingly (e.g., hemisphere differences). Summary details for the reviewed studies, such as sample sizes and ages, ASD diagnosis procedures, comparison groups, specific paradigms and stimulus types, and the analyzed responses are presented in Tables 1 –4. The review article concludes with a synthesis of the current literature, along with the discussion of possible clinical implications and future directions.
Table 1.
All included articles on vowel processing in autism spectrum disorder (ASD), sample demographics, paradigms, stimuli, measurement type, and analyzed responses.
| Study | Subjects | Age (years) | Paradigm | Speech stimulus type | EEG/MEG | Components |
|---|---|---|---|---|---|---|
| Berman et al. (2016) | 95 ASD (ADOS): 35 ASD+LI, 56 ASD-LI 44 TD |
M
ASD = 10.2 ± 2.6 years M ASD+LI = 9.3 ± 2.5 years M ASD-LI = 10.6 ± 2.5 years M TD = 10.4 ± 2.4 years |
Passive oddball (85/15) | Synthetic English vowels /a/, /u/ | MEG | MMF |
| Ceponiené et al. (2003) | 9 ASD (DSM-IV) 10 TD |
6–12 years M ASD = 8.9 M TD = 8.4 |
Passive oddball (86/7/7) | Semisynthetic vowel (Finnish /ö/); 10% frequency change; duration change (not reported) | EEG | MMN, P3a |
| Kasai et al. (2005) | 9 ASD (DSM-IV) 19 TD (matched on age, gender, and handedness) |
M
ASD = 27.2 ± 7.7 years M TD = 27.3 ± 7.0 |
Passive oddball (90/10) | Natural Japanese vowels /a/, /o/; duration change | MEG | MMF |
| Kemner et al. (1995) | 20 ASD (DSM-III) 20 TD |
M
ASD = 9.8 ± 1.5 years M TD = 10.6 ± 1.2 |
Passive or active oddball (80/10/10) | Synthetic English vowels /oy/, /ay/ | EEG | N1, MMN, P3 |
| Lepistö et al. (2005) | 15 ASD (DSM-IV) 15 TD (age and sex matched) |
7–11 years M ASD = 9.4 M TD = 9.6 |
Passive oddball (76/8/8/8) | Semisynthetic Finnish vowels /a/, /o/; duration, frequency, category change | EEG | MMN, P3a |
| Lepistö et al. (2006) | 10 ASD (DSM-IV) 10 TD |
7–10 years M ASD = 8.11 M TD = 8.10 |
Passive oddball (76/8/8/8) | Semisynthetic Finnish vowels /a/, /o/; duration, frequency, category change | EEG | MMN, P3a |
| Lepistö et al. (2007) | 9 Asperger (DSM-IV) 9 TD (matched on age, hearing, verbal and performance IQ) |
20–41 years M ASD = 22 M TD = 30 |
Passive oddball (76/8/8/8) | Semisynthetic Finnish vowels /a/, /o/; duration, frequency, category change | EEG | MMN, P3a |
| Lepistö et al. (2008) | 10 ASD (DSM-IV) 16 TD |
6–11 years M ASD = 9.1 M TD = 9 |
Passive oddball (85/15) with fixed or varied standard | Semisynthetic Finnish vowels (/a/, /e/, /i/, /o/, /u/, /y/), each with 6 different pitches; frequency or category change | EEG | MMN |
| Matsuzaki et al. (2019) | 57 ASD (ADOS-2): 9 ASD-MVNV, 27 ASD-V, 21 ASD-LI 27 TD |
8–12 years M ASD-MVNV = 9.67 ± 1.41 M ASD-V = 10.55 ± 1.21 M ASD-LI = 10.67 ± 1.20 M TD = 10.14 ± 1.38 |
Passive oddball (85/15) | Synthetic English vowels /a/, /u/ | MEG | MMF |
| Oram Cardy et al. (2005) | 7 ASD (ADOS, ADI) 9 TD |
8–17 years M ASD = 11.9 M TD = 11.9 |
Passive oddball (85/15) | Synthetic English vowels /a/, /u/ | MEG | M50, M100, MMF |
| Whitehouse & Bishop (2008) | 15 ASD (DSM-IV, ADOS) 15 TD (age and nonverbal IQ matched) |
7–14 years M ASD = 10:4 M TD = 10:6 |
Passive and active oddball (80/10/10) | Synthetic English vowels /a/, /i/ | EEG | P3a |
Note. EEG = electroencephalography; MEG = magnetoencephalography; ADOS = Autism Diagnosis Observation Scale; MMF = mismatch field; LI = language impairment; TD = typically developing; DSM = Diagnostic and Statistical Manual of Mental Disorders; MMN = mismatch negativity; MVNV = minimally verbal/nonverbal.
Table 2.
All included articles on syllable processing in autism spectrum disorder (ASD), sample demographics, paradigms, stimuli, measurement type, and analyzed responses.
| Study | Subjects | Age (years) | Paradigm | Speech stimulus type | EEG/MEG | Components |
|---|---|---|---|---|---|---|
| Arnett et al. (2018) | 76 ASD (DSM-5, ADOS, ADI-R) 27 TD |
4–17 years M ASD = 12.13 M TD = 13.0 |
Passive, auditory statistical learning task | Pseudorandom stream of English syllables with statistically related trisyllabic combinations | EEG | P1 |
| Dawson et al. (1986) | 17 ASD (DSM-III) 17 TD (age matched) |
6–18 years M ASD = 13;1 M TD = 13;0 |
Active oddball (80/10/10) | Natural speech, /da/ (presented among clicks and music chords) | EEG | N1 |
| Dawson et al. (1988) | 18 ASD (DSM-III) 17 TD (age matched) |
6–18 years M ASD = 13;1 M TD = 13;1 |
Active oddball (80/10/10) | Natural speech, /da/ (presented among clicks and music chords) | EEG | P3 |
| Dawson et al. (1989) | 10 ASD (DSM-III), 10 dysphasia (language and age matched to ASD) 10 TD (age and sex matched to ASD) |
6–15 years M dysphasia = 10;0 M ASD = 10;4 M TD = 10;4 |
Active oddball (80/10/10) | Natural speech, /da/ (presented among clicks and music chords) | EEG | N1 |
| Erwin et al. (1991) | 11 ASD (DSM-III) 14 TD |
17–39 years M ASD = 25.7 M TD = 23.4 |
Active oddball (80/20) | Natural speech, /ba/, /pa/ | EEG | N1, P2, P3 |
| Finch et al. (2017) | 13 high risk–ASD (ADOS) 39 high-risk–no ASD 44 TD |
M
ASD = 376 ± 14.3 days M HR = 375.03 ± 8.9 M TD = 373.18 ± 9.1 |
Passive double oddball (80/10/10) | Voiced, unaspirated, retroflex stop / a/, native deviant (voiceless, aspirated retroflex palatal stop /ta/), nonnative deviant (voiced, unaspirated dental stop /da/) | EEG | Late negative slow wave |
| Galilee et al. (2017) | 14 ASD (ADOS) 14 TD (verbal ability matched) |
2–6 years M ASD = 61 months M TD = 50 |
Passive listening, paired repetition paradigm (congruent or incongruent pairs of speech and nonspeech stimuli) | Semisynthetic /ba/, /da/, /ga/, nonspeech analogues | EEG | Frontal–central N250 and P350, temporal P250 and N330 |
| Huang et al. (2018) | 18 ASD (DSM-IV, ADOS, ADI-R) 17 TD (age and sex matched) |
M
ASD = 9.8 ± 2.03 years M TD = 9.4 ± 1.85 |
Passive oddball (84/16) | Natural speech, nonsense Chinese syllable (/tý/); deviant lengthened the vowel by 100 ms | EEG | MMN, P3a |
| Jansson-Verkasalo et al. (2003) | 11 Asperger (ICD-10) 11 TD (age and sex matched) |
7–12 years M ASD = 9.1 M TD = 9.6 |
Passive oddball (80/10/10) | Semisynthetic Finnish syllables /taa/, /ta/, /kaa/ (only /taa-kaa/ results reported) | EEG | MMN |
| Key et al. (2016) | 24 ASD (DSM-IV, ADOS, ADI-R; or community Dx) 18 TD |
5–10 years M ASD = 6.71 M TD = 7.14 |
Equiprobable, passive listening | Synthetic speech; /ba/, /da/, /ga/, /bu/, /du/, /gu/ | EEG | 84–308 ms (temporal PCA) |
| Kuhl et al. (2005) | 29 ASD 29 TD (mental age matched) 15 TD (age matched) |
ASD: 32–52 months M ASD = 45.31 TD-mental age: 13–48 months M TD-MA = 27.78 TD-age: 33–70 months M TD = 48.33 |
Passive oddball (85/15) | Synthetic speech /ba/, /wa/ | EEG | MMN |
| Kujala et al. (2010) | 15 Asperger (ICD-10) 13 TD (age, gender, handedness, verbal IQ, performance IQ, full-scale IQ matched) |
8–12 years M ASD = 10;9 M TD = 10;6 |
Multifeature paradigm with five different deviants | Semisynthetic Finnish syllables /te:/, /pi:/, frequency, duration, intensity deviants, consonant change; vowel change | EEG | MMN |
| Magnée et al. (2008) | 12 PDD (DSM-IV, ADOS, ADI-R) 13 TD |
M
ASD = 21.1 ± 4 years M TD = 23 ± 2.9 |
AV integration (judge what was said) | A video of a woman's face producing /aba/, /ada/ with congruent or incongruent audio | EEG | N1, P2 |
| Russo et al. (2009) | 16 ASD (community Dx) 11 TD |
7–13 years M ASD = 9.81 M TD = 9.82 |
Passive listening in quiet and in noise (+5 dB) | Synthetic syllable “da” | EEG | P1 |
| Wang et al. (2017) | 16 ASD (DSM-5; ADOS, ADI-R) 15 TD (age, IQ matched) |
M
ASD = 10.4 ± 1.27 M TD = 10.3 ± 1.55 |
Passive double oddball (80/10/10) | 3 Mandarin Chinese monosyllables; between-categories and within-category stimulus changes on the continuum between naturally spoken /ba2/ and /ba4/ | EEG | MMR, P3a |
| Weismüller et al. (2015) | 18 ASD (DSM-5) 15 TD |
6–15 years M ASD = 9.4 M TD = 10.6 |
Passive oddball (92/8) | Synthetic English phonemes /ba/, /pa/ | EEG | MMN |
| Yoshimura et al. (2016) | 35 ASD (DSM-IV; ADOS) 35 TD |
32–121 months M ASD = 74.7 M TD = 75.5 |
Passive oddball (83/17)— only standards reported | Natural speech, Japanese syllable “ne” with a flat or a high falling tone | MEG | P1m |
Note. EEG = electroencephalography; MEG = magnetoencephalography; ADOS = Autism Diagnosis Observation Scale; DSM = Diagnostic and Statistical Manual of Mental Disorders; MMN = mismatch negativity; ADI-R = Autism Disagnostic Interview–Revised; TD = typically developing; ICD-10 = International Classification of Diseases, 10th revision; Dx = diagnosis; PCA = principal component analysis; AV = audiovisual.
Table 3.
All included articles on word and sentence processing in autism spectrum disorder (ASD), sample demographics, paradigms, stimuli, measurement type, and analyzed response.
| Study | Subjects | Age (years) | Paradigm | Speech stimulus type | EEG/MEG | Components |
|---|---|---|---|---|---|---|
| Barzy et al. (2020) | 24 ASD (DSM-5, ADOS) 24 TD (age, verbal IQ and gender matched) |
M
ASD = 32.58 ± 2.23 years M TD = 31.75 ± 2.21) |
Sentence listening | Natural speech; sentences with a critical word in the medial position, consistent/inconsistent with the speaker's age, sex, class, or the sentence meaning | EEG | 100–200 ms 200–300 ms 300–500 ms 500–700 ms 200–700 ms |
| Cantiani et al. (2016) | 10 nonverbal ASD (community Dx) 10 TD (age and sex matched) |
4–7 years M ASD = 6.28 M TD = 6.33 |
Passive picture–word matching | Natural speech; monosyllabic or bisyllabic nouns | EEG | P1, N400 |
| DiStefano et al. (2019) | 18 Minimally verbal and 15 verbal ASD (community Dx, ADOS) 18 TD (age and sex matched) |
5–11 years M TD = 91.61 months M ASD-V = 88.67 M ASD-NV = 92.42 |
Passive picture–word matching | Natural speech; monosyllabic or bisyllabic nouns | EEG | N400, LNC |
| M. A. Dunn & Bates (2005) | 18 ASD (DSM-IV, ADOS, ADI-R) 18 TD (age, nonverbal IQ matched) |
8–9 and 10–12 years | Passive listening to In-category and out-of-category words | Natural speech; animal and nonanimal words | EEG | N1, N400 |
| M. A. Dunn et al. (1999) | 8 ASD 8 TD |
7–10 years M ASD = 8;10 M TD = 9;1 |
Active detection of in-category words | Natural speech; animal and nonanimal words | EEG | N2, N400 |
| Finch et al. (2018) | 14 ASD (ADOS) 29 high-risk sibs (13 with language/cognitive delay) 31 TD |
36 months | Passive listening | Natural speech; concrete nouns acquired early (known) or late (novel) | EEG | N200, N350 |
| Fishman et al. (2011) | 12 ASD (DSM-IV, ADOS, ADI-R) 18 TD |
17–46 years M ASD = 31 M TD = 30 |
Sentence judgment | Natural speech; sentences with congruent/incongruent final word | EEG | N400 |
| Kuhl et al. (2013) | 24 ASD (DSM-IV, ADOS, ADI) 20 TD |
1–2 years M ASD = 2.1 M TD = 2.1 |
Passive listening | Natural speech; known, unknown, and backward words | EEG | 200–500 ms |
| Ludlow et al. (2014) | 11 ASD (DSM-IV, ADOS) 11 TD |
11–16 years M ASD = 13.0 M TD = 13.7 |
Passive oddball (60/20/20) | Syllables [baj], [paj]; word deviants [bajt], [pajp]; pseudoword deviants [bajp], [pajt] | EEG | MMN |
| Manfredi et al. (2020) | 24 ASD (DSM-IV, ADI-R) 16 TD (age matched) |
9–15 years M ASD = 11.4 M TD = 12.6 |
Sentence judgment | Naturally spoken sentences with congruent/incongruent final word | EEG | N400 |
| McCleery et al. (2010) | 14 ASD (DSM-IV, ADOS, ADI-R) 14 TD |
4–7 years M ASD = 5.8 M TD = 6.0 |
Passive picture–word matching with attention probe | Natural speech; nouns | EEG | N400 |
| Megnin et al. (2012) | 14 ASD (ICD-10, ADOS, ADI-R) 14 TD |
M
ASD = 16.9 ± 0.3 years M TD = 16.9 ±.9 |
AV stimuli with attention probe | Natural speech; monosyllabic words | EEG | N1, P2, N4 |
| Ribeiro et al. (2013) | 7 ASD (ADI-R) 7 TD (age, education, verbal IQ, performance IQ, full IQ matched) |
M
ASD = 13 ± 4 years M TD = 14 ± 5 |
Sentence–picture matching | Spoken sentences | EEG | N400 |
| Sandbank et al. (2017) | 34 ASD (DSM-IV-TR, ADOS) | 2–5 years M ASD = 45.40 months |
Passive listening | Natural speech; words and nonwords | EEG | 200–500 ms |
| Yu et al. (2015) | 18 ASD (DSM-IV, ADOS, ADI-R) 16 TD |
6–12 years M ASD = 9.3 M TD = 9.6 |
Passive oddball (84/16) | Lexical tones in monosyllabic words (bai2, bai4) and nonwords (rai) | EEG | MMN, P3a |
Note. EEG = electroencephalography; MEG = magnetoencephalography; DSM = Diagnostic and Statistical Manual of Mental Disorders; ADOS = Autism Diagnosis Observation Scale: TD = typically developing; Dx = diagnosis; LNC = late negative component; ADI-R = Autism Diagnostic Interview–Revised; MMN = mismatch negativity; ICD-10 = International Classification of Diseases, 10th revision; AV = audiovisual.
Table 4.
All included articles on speech prosody processing in autism spectrum disorder (ASD), sample demographics, paradigms, stimuli, measurement type, and analyzed response.
| Study | Subjects | Age (years) | Paradigm | Speech stimulus type | EEG/MEG | Components |
|---|---|---|---|---|---|---|
| Clumeck et al. (2014) | 7 Asperger (DSM-IV, ADOS) 1 TD |
M
ASD = 16 (13–20 years) M TD = 25 (21–38 years) |
Passive listening to live and recorded passage reading | Natural speech; emotionally neutral text | MEG | Coherence |
| DePriest et al. (2017) | 11 ASD (ADOS or DSM-IV) 17 TD |
21–56 years M ASD = 37.45 M TD = 28.47 |
Sentence evaluation (how natural) | Natural speech; prosodically and syntactically congruous and incongruous sentences | EEG | Closure positive shift |
| Erwin et al. (1991) | 11 ASD (DSM-3) 14 TD |
17–39 years M ASD = 25.7 M TD = 23.4 |
Active oddball (80/20) | Natural speech, “Bob” with linguistic (statement/question) and affective (angry/happy) prosody | EEG | N1, P2, P3 |
| Fan & Cheng (2014) | 20 ASD/Asperger (DSM-IV, ADI-R) 20 TD (age, gender, IQ, handedness matched) |
18–29 years M ASD = 21.5 M TD = 22.0 |
Passive oddball (80/10/10) | Natural speech. Meaningless ‘dada’ with neutral, angry, happy prosody | EEG | MMN, P3a |
| Korpilahti et al. (2007) | 14 Asperger (ADOS, ADI-R), 13 TD (age matched) Fathers: 12 AS, 13 TD (SES matched) |
9–12 years M ASD = 11.2 M TD = 10.8 Fathers: 32–52 years, M ASD = 42.8 M TD = 44.3 |
Passive oddball (85/15) | Natural speech, “Give!” with tender or commanding prosody | EEG | Early and late MMN |
| Kujala et al. (2005) | 8 Asperger (DSM-IV) 8 TD (age and sex matched) |
22–43 years M ASD = 33 M TD = 32 |
Passive oddball (79/7/7/7) | Natural speech, Finnish “Saara!” with neutral, commanding, sad, and scornful prosody | EEG | MMN |
| Lindström et al. (2016) | 15 ASD (13 Asperger–DSM-IV, 2 ASD–DSM-5, ADI-R) 16 TD |
7–12 years M ASD = 10.4 M TD = 10.1 |
Passive oddball (79/7/7/7) | Natural speech, Finnish “Saara!” with neutral, commanding, sad, scornful prosody | EEG | MMN/LDN, P3a |
| Yoshimura et al. (2017) | 47 ASD (DSM-IV, ADOS; 23 with speech onset delay) 46 TD (gender, age, and head circumference matched) |
37–79 months M ASD = 60.4 M TD = 58.4 |
Passive oddball (83/17) | Natural speech, Japanese syllable “ne” with a flat or high falling tone (request) | MEG | MMF |
| Zhang et al. (2018) | 15 ASD (community Dx) 16 TD |
M
ASD = 10.04 ± 1.53 years M TD = 9.48 ± 0.86 |
Passive oddball (80/20) | Synthetic speech, “mother” with the first or second syllable stressed | EEG | MMN |
Note. EEG = electroencephalography; MEG = magnetoencephalography; DSM = Diagnostic and Statistical Manual of Mental Disorders; ADOS = Autism Diagnosis Observation Scale; TD = typically developing; ADI-R = Autism Diagnostic Interview–Revised; MMN = mismatch negativity; AS = Asperger syndrome; SES = socioeconomic status; LDN = late discriminative negativity; MMF = mismatch field; Dx = diagnosis.
Results
Vowels
Vowels are foundational components of the speech signal. They are produced with vocal fold vibration and a lax vocal tract; the shape of the vocal tract determines the frequency components or formants. Among the 11 studies examining vowel processing in ASD (see Table 1), the majority used tasks manipulating the acoustic features of a single vowel to investigate processing of basic speech sound characteristics, such as frequency (i.e., pitch), intensity, and duration. A smaller number of articles reported on categorical differentiation of vowels. Most of the reviewed studies used the oddball paradigm, in which repeated presentations of one stimulus (i.e., the standard) are occasionally interrupted by another sound (i.e., target or deviant), with the latter eliciting larger neural responses (e.g., MMN, P3a, and P3b) than the former if the stimulus change was detected. The auditory P1-N1-P2 response, which overlaps in time with the MMN and indexes processing of stimulus acoustic features and initial stimulus characterization, can also be examined.
Sensory Processing
The interest in how individuals on the autism spectrum respond to basic acoustic features of speech stems from the hypothesis that communication disorders in ASD are due to atypical sensory processing of spoken stimuli (Frith, 1989; Mottron et al., 2006). Sensitivity to physical properties of the speech sounds has been often quantified as stimulus change detection during passive (i.e., task-free) listening. A difference in the ERP amplitude between the deviant and the standard stimulus within 100–300 ms after stimulus onset is known as the MMN response (a mismatch field [MMF] in MEG terminology). The MMN amplitude to vowel formant frequency shifts of 10% was either not significantly different (Ceponiené et al., 2003; Finnish /ö/) or enhanced (Lepistö et al., 2005; Finnish /a/, /o/) in autistic children compared with typically developing controls. Greater-than-typical sensitivity to such frequency changes was reported in adults with Asperger syndrome representing the higher- functioning end of the autism spectrum (Lepistö et al., 2007). Similarly, MMN amplitudes to duration changes from a longer standard to a shorter deviant vowel (e.g., 190 vs. 104 ms) were not significantly different between autistic children and typically developing controls (Lepistö et al., 2005). However, children and adults with Asperger syndrome demonstrated larger-than-typical MMN responses to duration changes (Lepistö et al., 2006, 2007). Together, these results suggest that sensitivity to physical properties of vowels in persons on the autism spectrum is generally not impaired (see further confirmation in a meta-analysis by Chen et al., 2020) but may vary with age and ability (see a meta-analysis by Schwartz et al., 2018).
Categorical Perception
Studies investigating categorical perception of vowels in ASD primarily noted slower-than-typical stimulus differentiation, as suggested by the delayed MMN latency (/a/–/o/ contrast; Lepistö et al., 2005, 2006; see also Chen et al., 2020). Similar prolonged latency for the MMF response in ASD was also observed in MEG studies using an /a/–/o/ contrast in adults (Kasai et al., 2005) and /a/–/u/ contrast in children (Oram Cardy et al., 2005). Among autistic children, vowel processing was delayed more for those with a concomitant language disorder (Berman et al., 2016) and in minimally verbal children compared with those with a language disorder (Matsuzaki et al., 2019). Faster neural response (i.e., shorter latency) in participants on the autism spectrum regardless of the presence of a diagnosed language disorder was associated with better performance on behavioral receptive language measures (i.e., Clinical Evaluation of Language Fundamentals–Fourth Edition [CELF-4]; Berman et al., 2016), suggesting that the observed delays in vowel change detection reflected maladaptive rather than compensatory neural processes.
Categorical changes in vowels were also associated with greater-than-typical MMN amplitude in autistic children during an oddball paradigm with multiple deviants (standard: /a/; deviants: /y, e, o, u, i/; Lepistö et al., 2008). However, the increased sensitivity to categorical differences was no longer observed when the stimulus sequence included changes in the frequency characteristics of the vowels. These findings suggest that when the auditory input becomes highly variable, processing of basic acoustic features (i.e., frequency) may dominate the more linguistically meaningful details (i.e., categorical differences) in individuals on the autism spectrum. As the speech stream in a natural setting often includes multiple, potentially ambiguous contrasts, greater sensitivity to physical variability of speech sound characteristics, along with slower categorical perception of individual sounds, could interfere with comprehension.
Attentional Contributions
Accounting for the possibility that phonetic discrimination may extend beyond detection of the basic stimulus physical features to later postperceptual information processing (e.g., Schwartz et al., 2018), several studies examined neural responses occurring after 200 ms and noted speech-specific alterations in ASD. The frontocentral P3a response elicited by the deviant stimuli in passive listening oddball paradigms and observed between 250 and 400 ms after stimulus onset indexes spontaneous allocation of attention to the unexpected auditory input. The P3a response was absent in autistic children for frequency changes in the vowel but not in the tone stimulus (Ceponiené et al., 2003), suggesting reduced attention allocation to speech. Categorical vowel changes elicited a smaller-than-typical P3a response in children and adults on the autism spectrum (Lepistö et al., 2005, 2007). Additionally, the size of the P3a response was affected by the broader stimulus context. Autistic children demonstrated an intact P3a response to infrequent vowels presented among tones but a diminished P3a amplitude to rare tones occurring among vowels, reflecting selective tuning out of frequent speech input that also reduced attention to the infrequent tones (Whitehouse & Bishop, 2008). In combination, these P3a findings align with the reduced social motivation explanation of communication difficulties in ASD. This is further supported by the observation that when explicitly instructed to attend to the stimuli, the P3a amplitude and latency in response to novel tones among frequent speech stimuli were not significantly different between autistic children and their typically developing peers (Whitehouse & Bishop, 2008). Similarly, in an active vowel discrimination task (/oy/–/ay/), the amplitude of the centroparietal P3b response that indexes voluntary attention was not significantly different between children on the autism spectrum and typically developing controls (Kemner et al., 1995). Intact attentional responses during active tasks suggest that autistic individuals are able to process speech typically, but it may require additional cognitive resources that are available only when attention is deliberately directed to the stimuli. Because incidental learning from overhearing language is crucial to communicative development, reduced spontaneous orienting to speech may result in individuals on the autism spectrum benefiting less from their language environment, affecting vocabulary growth and language acquisition.
Overall, studies of vowel processing largely replicated findings from pure-tone studies (e.g., Roberts et al., 2011). There were no speech-specific deficits in early sensory neural responses of autistic individuals, whereas evidence of increased sensitivity to vowel physical features was limited and not consistent across frequency and duration characteristics. Delayed categorical discrimination and reduced stimulus change detection on tasks with greater variability in the auditory input (e.g., the number and type of contrasting stimuli) could be in line with the predictive coding theories. However, observations that diminished spontaneous orienting to speech sound changes were no longer present when task instructions explicitly directed attention to the stimuli are more consistent with the reduced social motivation: Autistic individuals are capable of typical speech processing but do not spontaneously allocate sufficient resources to do so. Furthermore, differences in the participants' age, language ability, and overall level of functioning affected the extent of group differences between those on the autism spectrum and those with typical development. Also, while the speed of speech sound discrimination is crucial to language processing, to date, only one study (Berman et al., 2016) explicitly documented the associations between the observed delays in neural responses to vowels and performance on standardized behavioral measures of language in ASD.
CV Combinations
CV combinations represent a more complex linguistic stimulus. Compared with the vowels presented in isolation, the addition of a consonant introduces rapidly changing physical characteristics (i.e., formant transitions). Similar to the studies using vowels, CV stimuli have been used in ASD to probe processing of basic acoustic features, such as frequency and duration, as well as categorical discrimination (17 studies; see Table 2).
Sensory Processing
Analysis of the earliest stages of sensory processing indexed by the P1, a positive peak occurring within 100 ms after stimulus onset, revealed slower latencies in response to the meaningless English syllable /da/ presented in quiet and in white noise (+5 dB SNR) in school-age autistic children compared with typically developing controls (Russo et al., 2009). Delayed P1 latencies to speech in noise were associated with lower receptive language skills (CELF-4) in the ASD group. Furthermore, the P1 amplitude in quiet for children with ASD was not significantly different from the typical P1 amplitude in noise, suggesting that, even under the optimal listening conditions, children on the autism spectrum may be less efficient in detecting CV stimulus onset. Conversely, a MEG study using a Japanese CV syllable /ne/ that does have functional significance (i.e., a request for acknowledgement) noted faster-than-typical P1 latency in 2- to 10-year-old children with ASD (Yoshimura et al., 2016). Such accelerated processing could have reflected a more advanced developmental stage; however, only the typically developing group demonstrated the expected reduction in P1 latency with increasing age. No specific pattern of age-related changes was observed in the ASD group due to greater-than-typical interindividual variability, even after accounting for differences in intellectual or verbal skills.
In a multifeature passive oddball paradigm that included several different types of deviants at the same time, children with Asperger syndrome demonstrated greater-than-typical sensitivity to intensity changes (55 ± 6 dB) but reduced detection of frequency changes (± 8%) in CV stimuli (Finnish /te/, /pi/) as indexed by MMN amplitudes (Kujala et al., 2010). Conversely, the long-to-short vowel duration contrasts (170 vs. 100 ms) in the context of CV stimuli revealed no significant group differences in the MMN amplitude, contradicting the reports of increased sensitivity to duration changes for vowels alone (e.g., Lepistö et al., 2006, 2007). Processing of short-to-long duration changes (250 vs. 350 ms) was also not significantly different between autistic children and typically developing controls (Chinese syllable /tý/; Huang et al., 2018). However, in that study, the ASD group demonstrated smaller-than-typical P3a response, indicative of difficulties with attentional switch to the unexpected longer deviant. The apparent reduction in the magnitude of the frequency change detection responses for the CV compared with the vowel stimuli suggests that increasing complexity of the auditory inputs may reduce efficiency of early sensory processing of speech in ASD, while stimulus duration processing remains intact.
Hemisphere Differences
Atypical neural specialization for speech stimuli emerged as a possible explanation for the observed language difficulties in ASD following several investigations of CV processing. In a passive listening study examining implicit learning of trisyllabic meaningless CV combinations, children and adolescents on the autism spectrum did not demonstrate the typical pattern of smaller P1 response over the left hemisphere than the right hemisphere (Arnett et al., 2018). Instead, smaller bilateral P1 amplitude to familiarized versus novel stimuli was associated with higher receptive language scores (Peabody Picture Vocabulary Test–Fourth Edition [PPVT-4]; L. M. Dunn & Dunn., 2007) even after controlling for age, ASD symptom severity, and nonverbal ability. In the absence of significant group differences in the P1 amplitude for stimulus condition or hemisphere, the authors attributed the lack of left hemisphere specialization in the ASD group to a combination of atypical bottom-up and top-down auditory perceptual processes.
Alterations in the hemisphere involvement were also noted in a series of active stimulus discrimination studies. Autistic adolescents compared with typically developing controls elicited smaller left hemisphere N1 amplitude (a negative peak occurring approximately 100 ms after stimulus onset and indexing activity of the primary and secondary auditory cortex) and shorter right hemisphere N1 latency in response to infrequent CV syllable (/da/) among frequent clicks (Dawson et al., 1986). Further analysis revealed substantial heterogeneity within the ASD group as 41% (seven of 17) of the sample showed the typical N1 amplitude lateralization (larger over the right than the left hemisphere) and 35% (6/17) had the typical asymmetry for N1 latency (longer over the right than the left hemisphere). More typical lateralization of the N1 amplitude was more likely to be observed in older participants on the autism spectrum and associated with higher scores on standardized behavioral measures of articulation (Arizona Articulation and Proficiency Scale; Fudala, 1970), vocabulary (Wechsler Intelligence Scale; Wechsler, 1974), and greater mean length of utterance in a spontaneous language sample. More typical hemisphere distribution of the N1 latency was related to better articulation, receptive and expressive language use (Northwestern Syntax Screening Test; Lee, 1971), comprehension (Wechsler Intelligence Scale), and higher PPVT-4 scores. The atypical pattern of faster N1 responses over the right than the left hemisphere observed in the ASD group was also noted in nonautistic children with language impairment (dysphasia; Dawson et al., 1989). However, for them, shorter N1 latency over the left hemisphere was related to poorer language outcomes, suggesting a left hemisphere dysfunction. Conversely, in autistic children, faster left hemisphere N1 responses correlated with better language performance, implicating atypical right hemisphere processes in the observed lateralization differences.
Altered hemisphere lateralization was also reported for 4- to 6-year-olds on the autism spectrum during passive listening to CV stimuli (/ba/, /da/, and /ga/) and sinusoidal tones presented as either matching (e.g., speech–speech) or mismatching (e.g., speech–nonspeech; Galilee et al., 2017) pairs. Autistic children and typically developing controls differentiated CV inputs from nonspeech at the sensory stage of information processing (i.e., no group differences in the P150 and N250 responses) and also detected the shift from speech to nonspeech sounds, evidenced by larger midline P350 responses compared with the speech–speech pairs. However, lateralized responses to that change were limited to the left hemisphere (temporal N330 response) in the ASD group, while the typically developing controls showed a bilateral effect. These results suggest that autistic children may engage a more limited set of neural mechanisms for speech processing.
Analysis of voluntary attention to CV stimuli in children on the autism spectrum revealed smaller-than-typical P3b amplitudes at left and midline central scalp locations in response to the target (/da/) presented among clicks but not to the nonspeech targets (i.e., musical chords; Dawson et al., 1988), consistent with reduced attention allocation to speech. Brain–behavior correlations noted that larger (more typical) left hemisphere P3b amplitudes to speech were associated with higher scores on the Arizona Articulation and Proficiency Scale. On the other hand, larger right hemisphere P3b responses to speech and to musical chords were predictive of lower receptive vocabulary (PPVT-4) and poorer articulation, raising possibility of a general alteration in auditory attention in autistic individuals.
Emerging data also indicate that atypical hemisphere asymmetry in ASD may be already detectable in infancy. Twelve-month-old infants later diagnosed with ASD elicited a larger left than right anterior late slow wave response (300–700 ms) for CV syllables (/ta/, /da/, and /ɖa/), while the direction of hemisphere differences was reversed in typically developing infants (Finch et al., 2017). However, these individual differences in hemisphere lateralization did not correlate with behavioral performance on receptive or expressive language tasks at 36 months (Mullen Scales of Early Learning; Mullen, 1995), suggesting that they may be indicative of atypical neurodevelopment rather than specific to language difficulties.
Categorical Perception
Categorical perception of CV syllables in ASD was examined in the context of varying vowels, pitch contours, and differences in consonant voice onset time and place of articulation. Increasing stimulus complexity did not interfere with categorical processing of vowels, as there were no group differences between children with Asperger syndrome or ASD and their typically developing peers in the amplitude of the early sensory perceptual P1-N1-P2 or MMN responses for vowel contrasts in CV stimuli (/te/–/ti/, /pi/–/pe/, Kujala et al., 2010; /ba/, /da/, /ga/ vs. /bu/, /du/, /gu/, Key et al., 2016). These results are consistent with the findings from the vowel studies.
Using a 10-step pitch contour (lexical tone) continuum that produced either a within-category variant of a CV syllable /ba/ (Step 1 vs. Step 5) or a meaningful between-categories tone change (Step 5 vs. Step 9), Wang et al. (2017) observed no group differences between Mandarin-speaking autistic children and typically developing controls in the mismatch response for either of the contrast types. However, unlike the typically developing children, participants on the autism spectrum did not demonstrate greater mismatch amplitudes to between-categories than to within-category deviants, suggesting not fully developed phonemic boundaries. The latter could interfere with efficient categorization of individual speech sound variants.
Studies examining consonant discrimination revealed difficulties detecting changes in the place of articulation. Relative to typically developing controls, children with Asperger syndrome exhibited delayed MMN latency in the right hemisphere for the /ta/–/ka/ contrast (Jansson-Verkasalo et al., 2003). Smaller-than-typical P1-N1-P2 amplitude differences were also reported in young autistic children in response to /b/–/g/ and /b/–/d/ contrasts (averaged across vowels; Key et al., 2016). A mediation analysis in that study revealed that reduced consonant discrimination contributed to the increased incidence of language disorders in children on the autism spectrum, defined as the discrepancy between verbal and nonverbal IQ. Conversely, autistic children and adults were not significantly different from typically developing controls in the MMN or P3b responses to the voice onset time contrasts (/ba/–/pa/; Erwin et al., 1991; Weismüller et al., 2015). These findings extend previous results of vowel studies, noting possible alterations in processing of frequency changes but not in sensitivity to basic temporal characteristics (i.e., duration) of the speech stimuli.
Individual differences in consonant discrimination in ASD might be associated with more general preferences for social stimuli. Kuhl et al. (2005) noted that while, at the group level, preschool children on the autism spectrum failed to generate a significant MMN response to a consonant change (/ba/–/wa/), the individual participants who exhibited increased attention to child-directed speech (measured via head turn) demonstrated more typical MMN responses. Conversely, preference for the nonspeech stimulus was associated with diminished MMN amplitudes to the CV contrast in the left hemisphere and increased responses in the right hemisphere. These results were consistent with prior findings in older autistic children noting atypical hemisphere specialization and high interindividual variability (Dawson et al., 1986, 1988, 1989), further highlighting the existence of subgroups within ASD.
A related line of research examined whether consonant processing in ASD is affected by additional cues, such as facial movements. In typically developing individuals, congruent visual information lowers detection thresholds (Grant & Greenberg, 2001) and improves identification of speech sounds (Schwartz et al., 2018). Adults with Asperger syndrome or autism (collectively labeled as pervasive developmental disorder [PDD]) were not significantly different from their typically developing peers in behavioral identification of visual, auditory, or auditory–visual VCV stimuli (/ada/ and /aba/; Magnée et al., 2008). The audiovisual presentation facilitated early sensory processing in both groups as reflected by the faster latency of the N1 and P2 responses, but the PDD group did not show the late audiovisual incongruence response (600–800 ms) that indexed higher order integration of phonological and visual information, indicating a more limited benefit from the synchronous multimodal stimuli (see also a meta-analysis by Zhang, 2019).
In summary, the studies using CV stimuli in individuals on the autism spectrum identified atypical hemisphere specialization for speech processing but no consistent evidence of enhanced sensory processing. Findings of less efficient detection of speech onset, sensitivity to the immediate stimulus context (e.g., speech vs. nonspeech), and possible alterations in phonemic boundaries could be interpreted to fit with the predictive coding difficulties. However, typical-like facilitation of early sensory speech processing by concurrent facial movements suggests that at least some predictive mechanisms are functional. Conversely, frequency-based consonant discrimination (i.e., place of articulation contrasts) emerged as a possible area of weakness associated with language difficulties in ASD. More typical CV stimulus processing in autistic individuals exhibiting greater preference for social stimuli (e.g., child-directed speech) supports the social motivation theory.
Word and Sentence Comprehension
Word- and sentence-level stimuli represent the next tier of auditory stimulus complexity. In addition to being longer and more acoustically diverse than single vowels or meaningless CV stimuli, they convey semantic information. Studies of speech comprehension in ASD contrasted real words with nonwords and examined integration of meaning across multiple auditory and/or visual inputs (15 studies; see Table 3).
Sensory Processing
Expanding the investigation of sensitivity to changes in the acoustic features of complex linguistic stimuli, Yu et al. (2015) reported that autistic children speaking a tonal language (Mandarin Chinese) demonstrated a reduced detection of (smaller-than-typical MMN amplitude) and slower orienting (delayed P3a latencies) to the tonal changes that altered meaning in real words (/bai/ with rising or falling contours). These results contrast with the findings of enhanced or intact MMN responses to pitch changes in pure tones observed in the same study or previously reported for vowel stimuli (Lepistö et al., 2005). In combination with findings of reduced frequency change discrimination for CV stimuli (Kujala et al., 2010), the results of Yu et al. may reflect the detrimental effects of additional processing demands due to greater acoustic complexity of word-level stimuli. The concurrent changes in both pitch and meaning might have exceeded the capacity of the auditory change detection system in ASD.
Similar reduction in sensitivity to semantic changes following acoustic modifications of word-level stimuli was observed in English-speaking autistic adolescents (Ludlow et al., 2014). In a multifeature oddball paradigm, the ASD group was not significantly different from the typically developing controls in response to the standard stimuli (CV syllables /bai/ and /pai/) but demonstrated smaller-than-typical MMN amplitudes to the deviant CVC stimuli created by adding a final consonant that formed real words (/bait/ and /paip/) or nonwords (/pait/ and /baip/). Of note, the standard stimuli were also English words (“bye” and “pie”), and thus, the observed results could potentially reflect reduced sensitivity to the acoustic change in the final consonant, regardless of whether it resulted in a new semantic content (i.e., a different word) or a loss of meaning (i.e., a nonword). However, higher auditory sensory sensitivity scores on the adolescent/adult sensory profile were associated with smaller MMN responses to the deviant stimuli.
Auditory processing of spoken words in ASD was also less sensitive to the facilitating effects of concurrently presented facial cues. High-functioning autistic adolescents demonstrated smaller-than-typical reduction in the P2 amplitude elicited by auditory–visual compared with auditory-only stimuli (Megnin et al., 2012). The two groups were not significantly different in the N1, P2, and N4 responses for the auditory-only or visual-only conditions or in the N1 amplitude suppression for the auditory–visual stimuli, suggesting that the observed lack of P2 modulation indicated difficulties with multisensory integration at the early lexical semantic processing stage. This reduced multisensory facilitation was more pronounced in participants with greater social difficulties (Social Communication Questionnaire).
Known Word Detection
Several studies directly examined whether young children on the autism spectrum differentiate known from unknown words or nonwords. In autistic toddlers who had better social skills (Autism Diagnosis Observation Scale [ADOS]; Lord et al., 1999), known compared with unknown words (10 per condition) elicited more negative left parietal amplitudes within 200–500 ms after stimulus onset (Kuhl et al., 2013), resembling the response topography previously reported in typically developing participants (e.g., Mills et al., 1993). Conversely, participants with more pronounced social deficits demonstrated known versus unknown word differentiation at the temporal and parietal sites in the right hemisphere. In addition, while there were no significant concurrent brain–language associations, the more typical left hemisphere response to known words at 2 years of age was predictive of better receptive language (i.e., Auditory Comprehension subtest of the Preschool Language Scale; Zimmerman et al., 2002), cognitive (i.e., Mullen composite score), and adaptive (i.e., Vineland Adaptive Behavior Scales composite standard score; Sparrow et al., 1984) outcomes at ages 4 and 6 years.
Using a similar paradigm in autistic preschoolers with limited spoken language, Sandbank et al. (2017) observed that typical-like word–nonword differentiation responses (i.e., more negative left temporal amplitudes to words than nonwords) were associated with concurrent parent reports of better receptive language (MacArthur–Bates Communicative Development Inventories; Fenson et al., 2007), but only in participants who knew five or more of the 10 stimulus words. No such correlations were present for data from the right hemisphere sites, and the results were independent of ASD symptom severity (ADOS scores) or intellectual functioning. These findings suggest that greater word experience and comprehension abilities together may contribute to more typical brain lateralization for speech in autistic children, as has been previously observed in typical development (e.g., Mills et al., 1993).
Of note, in a study with a more diverse stimulus set (i.e., 20 early and 20 late acquired words), 3-year-olds on the autism spectrum and typically developing controls demonstrated the expected temporoparietal ERP responses to known versus unknown words, while a reversed pattern with more negative amplitudes for unknown words was reported in high-risk siblings of autistic children who exhibited language or cognitive difficulties (Finch et al., 2018). The ERP markers of known word differentiation did not correlate with the concurrent behavioral measure of verbal development (Mullen Scales of Early Learning), possibly due to those scores representing a combination of receptive and expressive language skills. Together, these results suggest that atypical neural responses to known words may not be specific to ASD but rather reflect language difficulties more broadly.
Word-Level Semantic Processing
Studies examining speech comprehension used a variety of active and passive paradigms to create a particular semantic context and then presented an incongruent stimulus. The violated expectation is associated with the increased amplitude of the centroparietal N400 response after the critical word onset (Kutas & Hillyard, 1980). When semantic context was established using pictures, both the autistic children and their typically developing peers generated significant N400 responses to the incongruent environmental sounds (McCleery et al., 2010). However, unlike the typically developing controls, the ASD group did not differentiate between spoken words matching and mismatching the preceding pictures, suggesting a specific alteration in processing the meaning of speech. However, there were no significant correlations between the N400 responses and the standardized language measures (i.e., Receptive One Word [Gardner, 1990b] and Expressive One Word [Gardner, 1990a] Vocabulary Tests), potentially due to overall high behavioral performance.
In a similar picture–word paradigm, half of the sample (five of 10) of minimally verbal autistic children (3–7 years) elicited the expected N400 response to mismatching words (Cantiani et al., 2016), demonstrating speech comprehension. However, the subgroups with and without the N400 response were not significantly different on language or cognitive ability, and it was unknown whether receptive vocabularies of the individual children included the words used in the task. In a follow-up study, DiStefano et al. (2019) reported no group differences in the N400 amplitude but a delayed latency in verbal and minimally verbal autistic children (5–11 years) compared with typically developing controls. There were no significant relationships between N400 amplitude or latency and cognitive or language measures. Analysis of individual differences noted that the expected N400 response was detected in approximately 80% (27 of 33) of the sample, with comparable representation of verbal and minimally verbal autistic children, suggesting that word-level semantic processing was not directly related to expressive language abilities.
Reduced semantic processing in children on the autism spectrum (7–10 years) was more apparent in a word classification task (M. A. Dunn et al., 1999). When asked to determine whether a spoken word belonged to a specific category (i.e., animals), typically developing controls generated larger N400 responses to out-of-category (i.e., nonanimals) than in-category exemplars. Conversely, the ASD group did not demonstrate significant differences in the N400 amplitude between the target and nontarget words and elicited a diminished P3b amplitude to out-of-category words, suggesting reduced allocation of attention to differences in semantic content. Longer-than-typical latencies of the early sensory responses (i.e., N1, P2) were also noted for the ASD group. This delay could be related to the increased perceptual complexity of word-level stimuli compared with single vowels or syllables. The N400 findings were replicated in a larger sample of 8- to 12-year-olds (M. A. Dunn & Bates, 2005). However, a slower N1 response to the physical characteristics of the stimuli was observed only in 8-year-olds on the autism spectrum, suggesting that the delay in early sensory processing may lessen with age. Correlation analysis revealed no significant association between the N1 and N400 latencies, indicating that the timing of early responses was not directly related to later semantic processing.
Sentence Comprehension
Using a sentence–picture matching paradigm, Ribeiro et al. (2013) found a smaller-than-typical N400 response to semantic violations in autistic adolescents. This finding suggests a weakened semantic network, consistent with the results from picture–word studies (e.g., McCleery et al., 2010). That study also reported increased positive amplitude to semantic mismatches in the 600- to 900-ms time window for individuals on the autism spectrum, potentially reflecting compensatory activity following a diminished N400 response. Relying on a later stage of cognitive processing could result in slower comprehension, which could impact the efficiency of interaction with conversational partners.
Reduced N400 in picture-based paradigms could also reflect the difficulty of combining information across auditory and visual modalities as atypical multisensory integration in ASD has been already reported at the level of syllables (Magnée et al., 2008) and single words (Megnin et al., 2012). However, smaller-than-typical N400 response to incongruous final words was also reported in school-age autistic children when semantic context was established by the spoken sentence alone (Manfredi et al., 2020), suggesting that difficulties with semantic processing in children with ASD may be independent of stimulus modality.
In autistic adults, the presence of the N400 response varied based on task complexity. When asked to judge whether a spoken sentence made sense, the N400 amplitude to incongruent final words was not significantly different between adults on the autism spectrum and typically developing controls (Fishman et al., 2011). The N400 amplitude to semantically anomalous words in the middle of a sentence was smaller than typical in high-functioning autistic adults (Barzy et al., 2020). However, both participant groups detected mismatch between the overall message and the speaker's voice (e.g., not age-appropriate content) by 200 ms after the critical word onset, suggesting effective use of pragmatic auditory context (i.e., child vs. adult speaker) to constrain expectations about forthcoming content. Of note, in typically developing populations, comparable N400 responses have been observed for semantically incongruent critical words in the final or medial sentence positions (see Stowe et al., 2018, for a review), suggesting that sentence wrap-up effects may not explain the difference in the pattern of results for autistic adults.
Overall, studies using spoken words and sentences in ASD noted reduced sensitivity to changes in meaning following alterations in stimulus physical characteristics (e.g., word–final phoneme and tone). Atypical semantic processing emerges early in development (reduced word–nonword discrimination) and persists through adolescence (diminished N400 responses) across levels of language ability. More typical neural markers of word comprehension demonstrated concurrent and predictive associations with higher language and adaptive functioning. Of interest, the atypical N400 responses were pronounced in cross-modal paradigms (picture–word and sentence–picture matching) and in more complex auditory tasks (varied semantic and pragmatic content), suggesting that semantic difficulties, at least in part, may be due to broader challenges integrating multiple content details. Previously, difficulties combining meaning of individual words into a coherent sentence have been reported in a functional magnetic resonance imaging study of individuals on the autism spectrum (Just et al., 2004) and associated with reduced activation in the left inferior frontal gyrus, one of the known neural sources of the N400 (Maess et al., 2006). These findings, along with the evidence of possible higher order compensatory processes, are consistent with the predictive coding perspective on ASD. At the same time, evidence of more typical semantic processing in individuals with more adaptive social functioning, as well as findings of more successful content integration for nonsocial (e.g., environmental sounds) than speech stimuli (McCleery et al., 2010), points to the possible social motivation-based explanation. Also, the majority of comprehension studies included cross-modal stimuli in children with ASD, whereas studies in adults used mainly the auditory stimuli, making it difficult to separate the contributions of developmental and procedural factors.
Prosody Processing
The prosodic characteristics of spoken language, such as patterns of stress, pitch, and intonation, convey additional linguistic and/or affective information relevant to effective processing of speech. Nine studies examined neural responses to prosodic cues in the context of spoken stimuli in ASD (see Table 4). Autistic children demonstrated atypical lateralization of brain responses to changes in lexical stress within a word (e.g., “MOther” vs. “moTHER”) as reflected by a more negative right than left hemisphere MMN response (Zhang et al., 2018), consistent with the evidence of reversed neural asymmetry for other spoken stimuli. Conversely, adolescents with ASD were not different from typically developing control in their responses to the rhythmic patterns in live versus prerecorded passage reading, as reflected by phase coherence measures between the fundamental frequency of spoken sentences and the MEG signal (Clumeck et al., 2014). However, the prosodic contour during reading could be more monotone than the rhythmic variations in spontaneous speech. Indeed, when presented with a variety of grammatically complex spoken sentences, autistic adults demonstrated atypical prosodic phrase boundary processing (DePriest et al., 2017). Although there were no group differences in behavioral performance on the task of evaluating sentence for naturalness, participants on the autism spectrum did not generate the typical ERPs (e.g., a closure positive shift) for the onset of the pause between two intonational phrases or the N400/P600 responses to mismatches between prosodic chunks and syntactic phrasing in “garden path” sentences. In an MEG study using a Japanese CV syllable (/ne/) pronounced with a flat tone (standard) or a high falling tone (deviant) to create a prosodic contrast, Yoshimura et al. (2017) observed reduced early (100–200 ms) mismatch responses in the left hemisphere of 3- to 5-year-olds on the autism spectrum. In addition, autistic children with speech onset delay exhibited enhanced late (200–350 ms) mismatch response compared with typically developing controls and with autistic children without speech delays, suggesting compensatory processing because that response was smaller in participants with better performance on receptive and expressive vocabulary measures (Picture Vocabulary Test–Revised [Ueno et al., 2008] and Kaufman Assessment Battery for Children [Kaufman & Kaufman, 1983]).
Studies contrasting affective and neutral prosody noted reduced sensitivity in autistic adults and children. Adults on the autism spectrum had smaller-than-typical or absent MMN responses to meaningless CVCV stimuli and nonspeech analogs delivered with happy or angry compared with neutral prosody (Fan & Cheng, 2014). Similarly, adults with Asperger syndrome compared with typically developing controls exhibited diminished MMN amplitude to scornful and commanding single-word utterances and no response to sad intonation (Kujala et al., 2005). A follow-up study in autistic children (Lindström et al., 2016) reported similar findings: Only the scornful and sad deviants elicited MMN responses, whereas the P3a was observed only for commanding and sad stimuli. Furthermore, the MMN and P3a amplitudes to the scornful deviant in the ASD group were significantly smaller than typical.
Studies examining differentiation between positive and negative affective prosodic content resulted in mixed findings. Unlike the typically developing controls, autistic adults showed no significant differences in MMN amplitudes between meaningless CVCV stimuli presented with angry versus happy intonation and generated smaller-than-typical P3a amplitude to angry sounds (Fan & Cheng, 2014). When the request “Give!” was spoken with tender (standard) versus angry (deviant) prosody, delayed N1 responses, larger-than-typical amplitude of the early MMN (i.e., 150–350 ms), and shorter latencies for the late MMN (350–800 ms) were recorded in children with Asperger syndrome (Korpilahti et al., 2007). Of note, fathers of children with Asperger syndrome demonstrated accelerated early MMN and delayed late MMN responses compared with typical adults, suggesting that altered prosody processing may be a part of a broader autism phenotype. Conversely, in an active oddball paradigm using the name “Bob” (Erwin et al., 1991), autistic adults were not significantly different from typically developing controls in behavioral performance (i.e., accuracy and response time) or in N1, P2, and P3b responses to affective (i.e., happy vs. angry) or linguistic (i.e., statement vs. question) prosodic targets. The discrepancy in the findings of Korpilahti et al. (2007) and Erwin et al. (1991) could be attributed to the possible compensatory effects of attention actively directed to spoken stimuli, as had been previously reported in a speech sound discrimination study (e.g., Whitehouse & Bishop, 2008). Age differences (adults vs. children) could also potentially reflect improvements in prosody discrimination later in development.
Together, studies of prosodic processing suggest that autistic individuals may not be distinguishable from typically developing controls in their ability to follow the general rhythm of the connected speech. However, they are less sensitive to specific prosodic cues that mark linguistic or affective details in the spoken input. Not being able to benefit as much as their typically developing peers from the additional cues offered by the more global auditory stimulus characteristics, such as stress and intonation, may place individuals on the autism spectrum at a disadvantage during conversational discourse. The majority of the reviewed studies used the passive oddball design, and thus, their results could be interpreted to support the predictive coding perspective on ASD. Pronounced group differences in the studies of affective prosody highlight the possibility that individuals on the autism spectrum may experience difficulty processing affective speech, especially negatively valenced stimuli, consistent with the general weakness in emotion recognition. However, similar to the findings for vowel processing, these group differences were no longer present when the participants were instructed to actively pay attention. Thus, it is possible that autistic individuals are able to detect changes in prosodic features of speech but do not do so spontaneously. These findings would support the reduced social motivation theory of language and communication disorders in ASD.
Discussion
The purpose of this review was to better understand the nature of speech and language difficulties in ASD by aggregating the available brain-based data regarding auditory processing of spoken stimuli across a range of complexity and semantic content and examining the findings in the context of several theories of autism. The results indicate that autistic individuals demonstrate departures from typical speech processing at almost every level of analysis, from detecting and discriminating isolated speech sounds to comprehending words and sentences. However, the specific extent of group differences varied across tasks and stimuli, with greater difficulties observed in younger children and for more complex speech.
Basic Acoustic Feature Processing
Studies of sensitivity to basic acoustic features of speech in autistic persons primarily used single sounds and revealed mostly intact and not consistently enhanced neural responses to changes in frequency or duration of vowel stimuli (Ceponiené et al., 2003; Lepistö et al., 2005, 2007, 2008). Similar findings were reported previously in studies using simple and complex tones (see Haesen et al., 2011, for a review), suggesting that the social nature of speech may not necessarily interfere with basic sensory processing of simple stimuli. However, distinct group differences began to emerge in response to consonants, including reduced sensitivity to signal onset and frequency changes, increased responses to intensity differences, and atypical hemisphere specialization. It is possible that the consonant-specific group differences reflect difficulties processing rapid formant transitions. Previously, challenges with temporal processing of brief (< 20 ms) changes in auditory stimuli were also reported in ASD using complex nonspeech stimuli (e.g., Foss-Feig et al., 2018). It is currently unknown whether the reported alterations in sensory processing of consonants reflect a developmental phase or characterize autistic individuals in general, as the existing data are limited to children.
Categorical Speech Processing
Investigations of categorical speech perception in individuals on the autism spectrum reported slower-than-typical detection of vowel contrasts and reduced consonant discrimination, which may be associated with lower preference for social stimuli (Kuhl et al., 2005) and explain, in part, the increased incidence of language disorders (Key et al., 2016). Evidence of comparable sensitivity to inter- and intracategory speech sound differences in ASD suggests that the typical phonemic boundaries may not be fully established, which could interfere with efficient processing of linguistically relevant contrasts and, in turn, affect receptive and expressive language outcomes. Impaired processing of categorical distinctions could also explain the limited consonant repertoires characteristic of the ASD in the prelinguistic period (e.g., Wetherby et al., 2007). Although studies have investigated the contributions of speech sound production (i.e., vocalization complexity) to later language skills in ASD (e.g., Wetherby et al., 2007; Yoder et al., 2015), the predictive value of speech sound discrimination for expressive language outcomes has not yet been extensively investigated in ASD. Previously, better consonant discrimination was associated with more optimal language outcomes in typical development (Molfese & Molfese, 1985) and in other clinical populations (Guttorm et al., 2005; Maitre et al., 2013). Thus, measures of consonant discrimination may offer new in-roads for understanding language development in ASD (e.g., Cantiani et al., 2016).
Attention to Speech
Reduced spontaneous orienting to acoustic and categorical stimulus changes was consistently observed across vowels, consonants, and word-level stimuli in autistic participants compared with typically developing controls. These results suggest that, despite the sufficient sensory ability to process spoken stimuli, unexpected changes in speech input may not reach full awareness in individuals on the autism spectrum. The evidence of more typical performance in active compared with passive tasks suggests alterations in top-down processes that could be modulated by differences in social motivation. Reduced self-initiated attention to speech inputs may have critical repercussions for the incidental learning of language and the ability to use subtle auditory cues that typically developing peers process automatically.
Semantic Processing of Words and Sentences
Semantic processing frequently appeared to be atypical in ASD and related to language ability (e.g., vocabulary size) and level of social interest. Similar to the findings from consonant discrimination studies, more typical detection of known among unknown words was predictive of better language and adaptive outcomes, suggesting that neural measures of auditory semantic processing may have potential clinical utility. The findings of reduced semantic mismatch detection between spoken words or sentences and pictures in autistic children indicate inefficient use of the available broader context to create an expectation for the probable verbal stimulus. Of note, semantic processing deficits in ASD have also been observed in studies using visual narratives (Coderre et al., 2018) or written sentences (Pijnacker et al., 2010; Ring et al., 2007), suggesting a general deficit in semantic integration independent of age, level of functioning, or stimulus modality.
Prosody
Adolescents and young adults on the autism spectrum were more successful in processing rhythmic prosody of connected speech and linguistic prosody variations (i.e., differentiating between a question and a statement) than affective prosody. Compared with their typically developing peers, autistic individuals were less sensitive to the emotional tone of speech, particularly when the target stimuli conveyed a negative tone, which could be detrimental to success of real-life verbal interactions. However, responses to changes in types of affective prosody were largely intact or enhanced. The contrasts used in these studies involved basic emotions (e.g., angry vs. happy); thus, future studies should examine processing of more nuanced prosodic contrasts. Furthermore, the existing studies on prosody involved participants with average language and cognitive abilities. Given the reported evidence of age and ability affecting responses to other types of speech stimuli, future studies of prosodic processing should include more diverse autistic participants.
Theoretical Explanations of Communication Deficits in ASD
The reviewed studies provide limited and inconsistent evidence of atypical early sensory processing of speech stimuli. More extensive support was observed for the predictive coding and reduced social motivation explanations of language difficulties in ASD. Group differences in speech processing became more prominent with the increase in speech stimulus complexity and context variability, suggesting difficulties selecting relevant stimulus details and forming general representations needed to establish phonemic boundaries or support effective semantic processing. However, evidence of reduced performance on speech but not nonspeech context matching tasks (McCleery et al., 2010) challenges predictive coding as the explanation for all speech and language difficulties in ASD.
Allocation of processing resources to speech inputs, especially in situations of increased perceptual and/or cognitive workload, could also depend on intrinsic motivation to engage in a communicative exchange. Delayed and reduced neural responses to speech in participants on the autism spectrum have been reported across multiple stimulus types in passive listening paradigms. Conversely, active tasks that included external directives to pay attention to speech resulted in more typical neural responses, suggesting that autistic individuals may have the ability to process speech but do not spontaneously engage it, as would be expected based on the reduced social motivation hypothesis. Studies directly characterizing social symptoms in participants on the autism spectrum, although limited in number, noted more typical speech processing in individuals demonstrating greater preference for or proficiency with social stimuli. In a recent behavioral longitudinal study, early social motivation in toddlers with ASD predicted functional expressive language outcomes 24 months later through its effects on intentional communication and receptive language (Su et al., 2020).
Of note, it is possible that reduced social motivation in ASD reflects more general alterations in reward processing (Bottini, 2018; Clements et al., 2018). However, this view is based largely on reviews of studies using pictures of faces without explicitly examining the motivational value of speech or language. Furthermore, the studies of reward learning (i.e., behavior modification based on feedback from prior trials), which could be viewed as the closest approximation of the processes involved in language development, consistently provided support for the social motivation theory (Bottini, 2018).
Current Limitations and Future Directions
Subject Ages and Developmental Changes
To date, speech processing in ASD has been primarily examined in school-age children and adolescents. Expanding research along the developmental timeline to include autistic toddlers (e.g., Kuhl et al., 2013) or younger siblings of children on the autism spectrum (e.g., Finch et al., 2017) began to clarify the role of speech processing in language acquisition in ASD and lead to the identification of early markers of risk for language disorders. Similarly, limited cross-sectional data suggest that at least some aspects of speech processing in ASD may improve with age. Thus, studies in autistic adults are needed to fully characterize the role of speech processing in language development and outcomes. Furthermore, most of the existing studies reported data from a single time point. Consequently, developmental inferences are being made based on cross-sectional data and nonidentical paradigms. A longitudinal investigation of brain responses to speech would help better document the developmental time course of speech and language abilities in ASD.
Subgroup Analysis
Recordings of brain activity require a certain level of cooperation from the participants in order to obtain usable data. Thus, the majority of existing studies included autistic individuals with more typical cognitive and language levels. Nevertheless, even those high-functioning samples demonstrated substantial behavioral and neural heterogeneity on speech processing tasks. More recently, researchers have begun to include lower-functioning and minimally verbal individuals (Cantiani et al., 2016; DiStefano et al., 2019; Matsuzaki et al., 2019) to gain a better understanding of the full range of language and cognitive abilities in ASD. Brain-based measures of speech processing, particularly those not requiring an overt response, may be especially informative for identifying subgroups with distinct neural patterns that could help explain communication strengths and weakness within the broader ASD, which in turn would contribute to more tailored interventions. Single-subject analyses, especially those using the emerging novel statistical methods (e.g., multivariate pattern analyses; Petit et al., 2020), could be another promising approach to characterizing speech and language processing in a highly heterogeneous autistic population.
Stimuli and Tasks
The types of vowel and consonant contrasts used to investigate speech processing in ASD have been relatively limited. For example, frequency contrasts have used a variety of values in the 100- to 200-Hz range, which is important for perception of vowels. Examination of higher frequency contrasts in the 2000- to 4000-Hz range could be relevant to consonant perception. Similarly, CV studies have focused mainly on stop consonants, which are characterized by a short burst of intense acoustic energy optimal for recording cortical responses. It is possible that individuals with ASD have a particular difficulty processing stop consonants because of the brief formant transitions. Expanding the stimuli to include fricatives (e.g., /s, v, ʃ/) and affricates (e.g., /t͡ʃ, d͡ʒ/) that are produced later in development could identify additional acoustic features that make speech sound discrimination more challenging in ASD and clarify the association between speech perception and production.
The reviewed studies also demonstrated the importance of task instructions. Active tasks that explicitly directed subjects' attention to speech elicited more typical responses (e.g., Erwin et al., 1991; Whitehouse & Bishop, 2008), possibly due to the compensatory effects of higher order cognitive processes. Conversely, passive paradigms were more likely to reveal atypical early sensory neural responses in ASD but yielded limited information about the higher order processes. Systematic examination of task-related variability in results could inform design of novel assessment approaches and selection of treatment targets.
Finally, while a number of studies examined semantic processing in ASD, syntactic processing has not yet been investigated in great detail. Evaluation of neural responses to written violations in syntax reported spared performance during active sentence judgment tasks in autistic adults (e.g., Koolen et al., 2013, 2014), but similar tasks might be more challenging in the auditory modality. The latter would add a critical temporal constraint because the detection of syntax violations in spoken sentences would require the individual to attend and integrate rapidly changing linguistic information.
Brain–Behavior Connections
The existing studies have identified a number of differences in the neural responses to speech between autistic persons and their typically developing peers. However, only a subset of the reviewed studies explicitly examined the connection between altered brain processes and behavioral performance on speech and language tasks. Reported brain–behavior associations highlighted the role of various subject characteristics in more optimal speech processing (e.g., preference for social over nonsocial stimuli; Kuhl et al., 2013) and suggested possible new markers of risk for language disorders (vowel or consonant discrimination; Berman et al., 2016; Key et al., 2016). However, many other studies either did not examine or did not find significant connections between the observed neural differences and behavioral performance. Identifying specific associations between neural markers of speech/language processing and clinical symptoms or behavioral performance could be helpful for expanding the assessment batteries, especially for preverbal or minimally verbal participant, as brain responses may indicate specific difficulties that are not reliably observed by other means.
Conclusions
This review demonstrated multiple alterations in neural processing of spoken stimuli in ASD. While sensory responses to basic speech stimuli (e.g., vowels) are mostly preserved, atypical speech processing becomes apparent with the more complex stimuli (e.g., consonants and multisyllabic stimuli) and tasks (e.g., categorical discrimination and semantic comprehension). The specific difficulties often varied based on participants' age, preference for social stimuli, vocabulary size, language ability, and level of adaptive functioning. Reduced social motivation appears to be the most likely explanation of the observed findings, possibly through its interaction with the top-down control mechanisms as more typical neural responses were observed in active compared with passive tasks. However, gaps in knowledge remain, as data are limited for younger children and adults on the autism spectrum, and brain–behavior connections in the communicative domain are not yet fully characterized.
Acknowledgments
“This work was supported in part by Eunice Kennedy Shriver National Institute of Child Health and Human Development Grant P50HD103537 (Vanderbilt Kennedy Center). The opinions expressed herein are those of the authors and do not necessarily represent the official position of the funding agencies.”
Funding Statement
“This work was supported in part by Eunice Kennedy Shriver National Institute of Child Health and Human Development Grant P50HD103537 (Vanderbilt Kennedy Center). The opinions expressed herein are those of the authors and do not necessarily represent the official position of the funding agencies.”
Footnote
The term autistic person is the preferred language for the majority of people diagnosed with autism (Kenny et al., 2016). Therefore, we are using identity-first rather than person-first language in this review article.
References
- American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (DSM-5). Author. https://doi.org/10.1176/appi.books.9780890425596 [Google Scholar]
- Anderson, D. K. , Lord, C. , Risi, S. , DiLavore, P. S. , Shulman, C. , Thurm, A. , Welch, K. , & Pickles, A. (2007). Patterns of growth in verbal abilities among children with autism spectrum disorder. Journal of Consulting and Clinical Psychology, 75(4), 594–604. https://doi.org/10.1037/0022-006X.75.4.594 [DOI] [PubMed] [Google Scholar]
- Arnett, A. B. , Hudac, C. M. , DesChamps, T. D. , Cairney, B. E. , Gerdts, J. , Wallace, A. S. , Bernier, R. A. , & Webb, S. J. (2018). Auditory perception is associated with implicit language learning and receptive language ability in autism spectrum disorder. Brain and Language, 187, 1–8. https://doi.org/10.1016/j.bandl.2018.09.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barzy, M. , Black, J. , Williams, D. , & Ferguson, H. J. (2020). Autistic adults anticipate and integrate meaning based on the speaker's voice: Evidence from eye-tracking and event-related potentials. Journal of Experimental Psychology: General, 149(6), 1097–1115. https://doi.org/10.1037/xge0000705 [DOI] [PubMed] [Google Scholar]
- Berman, J. I. , Edgar, J. C. , Blaskey, L. , Kuschner, E. S. , Levy, S. E. , Ku, M. , Dell, J. , & Roberts, T. P. (2016). Multimodal diffusion-MRI and MEG assessment of auditory and language system development in autism spectrum disorder. Frontiers in Neuroanatomy, 10, 30. https://doi.org/10.3389/fnana.2016.00030 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bottini, S. (2018). Social reward processing in individuals with autism spectrum disorder: A systematic review of the social motivation hypothesis. Research in Autism Spectrum Disorders, 45, 9–26. https://doi.org/10.1016/j.rasd.2017.10.001 [Google Scholar]
- Cantiani, C. , Choudhury, N. A. , Yan, H. Y. , Shafer, V. L. , Schwartz, R. G. , & Benasich, A. A. (2016). From sensory perception to lexical-semantic processing: An ERP study in non-verbal children with autism. PLOS ONE, 11(8), Article e0161637. https://doi.org/10.1371/journal.pone.0161637 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ceponienė, R. , Lepistö, T. , Shestakova, A. , Vanhala, R. , Alku, P. , Näätänen, R. , & Yaguchi, K. (2003). Speech-sound-selective auditory impairment in children with autism: They can perceive but do not attend. Proceedings of the National Academy of Sciences, 100(9), 5567–5572. https://doi.org/10.1073/pnas.0835631100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen, T.-C. , Hsieh, M. H. , Lin, Y.-T. , Chan, P.-Y. S. , & Cheng, C.-H. (2020). Mismatch negativity to different deviant changes in autism spectrum disorders: A meta-analysis. Clinical Neurophysiology, 131(3), 766–777. https://doi.org/10.1016/j.clinph.2019.10.031 [DOI] [PubMed] [Google Scholar]
- Chevallier, C. , Kohls, G. , Troiani, V. , Brodkin, E. S. , & Schultz, R. T. (2012). The social motivation theory of autism. Trends in Cognitive Science, 16, 231–239. https://doi.org/10.1016/j.tics.2012.02.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Clements, C. C. , Zoltowski, A. R. , Yankowitz, L. D. , Yerys, B. E. , Schultz, R. T. , & Herrington, J. D. (2018). Evaluation of the social motivation hypothesis of autism: A systematic review and meta-analysis. JAMA Psychiatry, 75(8), 797–808. https://doi.org/10.1001/jamapsychiatry.2018.1100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Clumeck, C. , Garcia, S. S. , Bourguignon, M. , Wens, V. , de Beeck, M. O. , Marty, B. , Deconinck, N. , Soncarrieu, M. , Goldman, S. , Jousmäki, V. , Van Bogaert, P. , & De Tiège, X. (2014). Preserved coupling between the reader's voice and the listener's cortical activity in autism spectrum disorders. PLOS ONE, 9(3), e92329. https://doi.org/10.1371/journal.pone.0092329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Coderre, E. L. , Cohn, N. , Slipher, S. K. , Chernenok, M. , Ledoux, K. , & Gordon, B. (2018). Visual and linguistic narrative comprehension in autism spectrum disorders: Neural evidence for modality-independent impairments. Brain and Language, 186, 44–59. https://doi.org/10.1016/j.bandl.2018.09.001 [DOI] [PubMed] [Google Scholar]
- Dawson, G. (1991). A psychobiological perspective on the early socio-emotional development of children with autism. In Cicchetti D. & Toth S. (Eds.), Models and integrations (Vol. 3, pp. 207–234). University Rochester Press. [Google Scholar]
- Dawson, G. , Finley, C. , Phillips, S. , & Galpert, L. (1986). Hemispheric specialization and the language abilities of autistic children. Child Development, 57(6), 1440–1453. https://doi.org/10.2307/1130422 [PubMed] [Google Scholar]
- Dawson, G. , Finley, C. , Phillips, S. , Galpert, L. , & Lewy, A. (1988). Reduced P3 amplitude of the event-related brain potential: Its relationship to language ability in autism. Journal of Autism and Developmental Disorders, 18(4), 493–504. https://doi.org/10.1007/BF02211869 [DOI] [PubMed] [Google Scholar]
- Dawson, G. , Finley, C. , Phillips, S. , & Lewy, A. (1989). A comparison of hemispheric asymmetries in speech-related brain potentials of autistic and dysphasic children. Brain and Language, 37(1), 26–41. https://doi.org/10.1016/0093-934X(89)90099-0 [DOI] [PubMed] [Google Scholar]
- Dawson, G. , Webb, S. J. , Wijsman, E. , Schellenberg, G. , Estes, A. , Munson, J. , & Faja, S. (2005). Neurocognitive and electrophysiological evidence of altered face processing in parents of children with autism: Implications for a model of abnormal development of social brain circuitry in autism. Development and Psychopathology, 17(3), 679–697. https://doi.org/10.1017/S0954579405050327 [DOI] [PubMed] [Google Scholar]
- DePriest, J. , Glushko, A. , Steinhauer, K. , & Koelsch, S. (2017). Language and music phrase boundary processing in autism spectrum disorder: An ERP study. Scientific Reports, 7, 1–12. https://doi.org/10.1038/s41598-017-14538-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- DiStefano, C. , Senturk, D. , & Jeste, S. S. (2019). ERP evidence of semantic processing in children with ASD. Developmental Cognitive Neuroscience, 36, 100640. https://doi.org/10.1016/j.dcn.2019.100640 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dunn, L. M. , & Dunn, D. M. (2007). PPVT- 4: Peabody Picture Vocabulary Test. Pearson Assessments. [Google Scholar]
- Dunn, M. A. , & Bates, J. C. (2005). Developmental change in neutral processing of words by children with autism. Journal of Autism and Developmental Disorders, 35(3), 361–376. https://doi.org/10.1007/s10803-005-3304-3 [DOI] [PubMed] [Google Scholar]
- Dunn, M. A. , Vaughan, H., Jr. , Kreuzer, J. , & Kurtzberg, D. (1999). Electrophysiologic correlates of semantic classification in autistic and normal children. Developmental Neuropsychology, 16(1), 79–99. https://doi.org/10.1207/S15326942DN160105 [Google Scholar]
- Erwin, R. , Van Lancker, D. , Guthrie, D. , Schwafel, J. , Tanguay, P. , & Buchwald, J. S. (1991). P3 responses to prosodic stimuli in adult autistic subjects. Electroencephalography and Clinical Neurophysiology/Evoked Potentials Section, 80(6), 561–571. https://doi.org/10.1016/0168-5597(91)90139-O [DOI] [PubMed] [Google Scholar]
- Fan, Y. T. , & Cheng, Y. (2014). Atypical mismatch negativity in response to emotional voices in people with autism spectrum conditions. PLOS ONE, 9(7), Article e102471. https://doi.org/10.1371/journal.pone.0102471 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fenson, L. , Bates, E. , Dale, P. S. , Marchman, V. A. , Reznick, J. S. , & Thal, D. J. (2007). MacArthur–Bates Communicative Development Inventories. Brookes. [Google Scholar]
- Finch, K. H. , Seery, A. M. , Talbott, M. R. , Nelson, C. A. , & Tager-Flusberg, H. (2017). Lateralization of ERPs to speech and handedness in the early development of autism spectrum disorder. Journal of Neurodevelopmental Disorders, 9(1), 4. https://doi.org/10.1186/s11689-017-9185-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Finch, K. H. , Tager-Flusberg, H. , & Nelson, C. A. (2018). Neural responses to linguistic stimuli in children with and without autism spectrum disorder. European Journal of Neuroscience, 47(6), 709–719. https://doi.org/10.1111/ejn.13721 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fishman, I. , Yam, A. , Bellugi, U. , Lincoln, A. , & Mills, D. (2011). Contrasting patterns of language-associated brain activity in autism and Williams syndrome. Social Cognitive and Affective Neuroscience, 6(5), 630–638. https://doi.org/10.1093/scan/nsq075 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Foss-Feig, J. H. , Stavropoulos, K. K. M. , McPartland, J. C. , Wallace, M. T. , Stone, W. L. , & Key, A. P. (2018). Electrophysiological response during auditory gap detection: Biomarker for sensory and communication alterations in autism spectrum disorder? Developmental Neuropsychology, 43(2), 109–122. https://doi.org/10.1080/87565641.2017.1365869 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Frith, U. (1989). A new look at language and communication in autism. International Journal of Language & Communication Disorders, 24(2), 123–150. https://doi.org/10.3109/13682828909011952 [DOI] [PubMed] [Google Scholar]
- Gardner, M. (1990a). The Expressive One Word Picture Vocabulary Test (Revised Edition) . Academic Therapy Publications. [Google Scholar]
- Gardner, M. (1990b). The Receptive One Word Picture Vocabulary Test (Revised Edition) . Academic Therapy Publications. [Google Scholar]
- Fudala, J. B. (1970). Arizona Articulation Proficiency Scale: Revised. Western Psychological Services. [Google Scholar]
- Galilee, A. , Stefanidou, C. , & McCleery, J. P. (2017). Atypical speech versus non-speech detection and discrimination in 4- to 6-yr old children with autism spectrum disorder: An ERP study. PLOS ONE, 12(7), Article e0181354. https://doi.org/10.1371/journal.pone.0181354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Grant, K. W. , & Greenberg, S. (2001). Speech intelligibility derived from asynchronous processing of auditory-visual information. In AVSP 2001-International Conference on Auditory-Visual Speech Processing. ISCA Archive. [Google Scholar]
- Groen, W. B. , Zwiers, M. P. , Vandergaag, R. , & Buitelaar, J. K. (2008). The phenotype and neural correlates of language in autism: An integrative review. Neuroscience & Biobehavioral Reviews, 32(8), 1416–1425. https://doi.org/10.1016/j.neubiorev.2008.05.008 [DOI] [PubMed] [Google Scholar]
- Guttorm, T. K. , Leppänen, P. H. , Poikkeus, A. M. , Eklund, K. M. , Lyytinen, P. , & Lyytinen, H. (2005). Brain event-related potentials (ERPs) measured at birth predict later language development in children with and without familial risk for dyslexia. Cortex, 41(3), 291–303. https://doi.org/10.1016/S0010-9452(08)70267-3 [DOI] [PubMed] [Google Scholar]
- Haesen, B. , Boets, B. , & Wagemans, J. (2011). A review of behavioural and electrophysiological studies on auditory processing and speech perception in autism spectrum disorders. Research in Autism Spectrum Disorders, 5(2), 701–714. https://doi.org/10.1016/j.rasd.2010.11.006 [Google Scholar]
- Haker, H. , Schneebeli, M. , & Stephan, K. E. (2016). Can Bayesian theories of autism spectrum disorder help improve clinical practice? Frontiers in Psychiatry, 7, 107. https://doi.org/10.3389/fpsyt.2016.00107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hämäläinen, J. A. , Salminen, H. K. , & Leppänen, P. H. (2013). Basic auditory processing deficits in dyslexia: Systematic review of the behavioral and event-related potential/field evidence. Journal of Learning Disabilities, 46(5), 413–427. https://doi.org/10.1177/0022219411436213 [DOI] [PubMed] [Google Scholar]
- Huang, D. , Yu, L. , Wang, X. , Fan, Y. , Wang, S. , & Zhang, Y. (2018). Distinct patterns of discrimination and orienting for temporal processing of speech and nonspeech in Chinese children with autism: An event-related potential study. European Journal of Neuroscience, 47(6), 662–668. https://doi.org/10.1111/ejn.13657 [DOI] [PubMed] [Google Scholar]
- Jansson-Verkasalo, E. , Ceponiene, R. , Kielinen, M. , Suominen, K. , Jäntti, V. , Linna, S. L. , Moilanen, I. , & Näätänen, R. (2003). Deficient auditory processing in children with Asperger syndrome, as indexed by event-related potentials. Neuroscience Letters, 338(3), 197–200. https://doi.org/10.1016/S0304-3940(02)01405-2 [DOI] [PubMed] [Google Scholar]
- Just, M. A. , Cherkassky, V. L. , Keller, T. A. , & Minshew, N. J. (2004). Cortical activation and synchronization during sentence comprehension in high-functioning autism: Evidence of underconnectivity. Brain, 127(Pt 8), 1811–1821. https://doi.org/10.1093/brain/awh199 [DOI] [PubMed] [Google Scholar]
- Kasai, K. , Hashimoto, O. , Kawakubo, Y. , Yumoto, M. , Kamio, S. , Itoh, K. , Koshida, I. , Iwananmi, A. , Nakagome, K. , Fukuda, M. , Yamasue, H. , Yamada, H. , Abe, O. , Aoki, S. , & Kato, N. (2005). Delayed automatic detection of change in speech sounds in adults with autism: A magnetoencephalographic study. Clinical Neurophysiology, 116(7), 1655–1664. https://doi.org/10.1016/j.clinph.2005.03.007 [DOI] [PubMed] [Google Scholar]
- Kaufman, A. , & Kaufman, N. (1983). Kaufman Assessment Battery for Children: Administration and Scoring Manual. American Guidance Service. [Google Scholar]
- Kemner, C. , Verbaten, M. N. , Cuperus, J. M. , Camfferman, G. , & van Engeland, H. (1995). Auditory event-related brain potentials in autistic children and three different control groups. Biological Psychiatry, 38(3), 150–165. https://doi.org/10.1016/0006-3223(94)00247-Z [DOI] [PubMed] [Google Scholar]
- Kenny, L. , Hattersley, C. , Molins, B. , Buckley, C. , Povey, C. , & Pellicano, E. (2016). Which terms should be used to describe autism? Perspectives from the U.K. autism community. Autism, 20(4), 442–462. https://doi.org/10.1177/1362361315588200 [DOI] [PubMed] [Google Scholar]
- Key, A. P. , Yoder, P. J. , & Stone, W. L. (2016). Consonant differentiation mediates the discrepancy between non-verbal and verbal abilities in children with ASD. Journal of Intellectual Disability Research, 60(5), 478–490. https://doi.org/10.1111/jir.12286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Koolen, S. , Vissers, C. T. W. , Egger, J. I. , & Verhoeven, L. (2013). Can monitoring in language comprehension in autism spectrum disorder be modulated? Evidence from event-related potentials. Biological Psychology, 94(2), 354–368. https://doi.org/10.1016/j.biopsycho.2013.07.008 [DOI] [PubMed] [Google Scholar]
- Koolen, S. , Vissers, C. T. W. , Egger, J. I. M. , & Verhoeven, L. (2014). Monitoring in language perception in high-functioning adults with autism spectrum disorder: Evidence from event-related potentials. Clinical Neurophysiology, 125(1), 108–123. https://doi.org/10.1016/j.clinph.2013.06.021 [DOI] [PubMed] [Google Scholar]
- Korpilahti, P. , Jansson-Verkasalo, E. , Mattila, M. L. , Kuusikko, S. , Suominen, K. , Rytky, S. , Pauls, D. L. , & Moilanen, I. (2007). Processing of affective speech prosody is impaired in Asperger syndrome. Journal of Autism and Developmental Disorders, 37(8), 1539–1549. https://doi.org/10.1007/s10803-006-0271-2 [DOI] [PubMed] [Google Scholar]
- Kuhl, P. K. , Coffey-Corina, S. , Padden, D. , & Dawson, G. (2005). Links between social and linguistic processing of speech in preschool children with autism: Behavioral and electrophysiological measures. Developmental Science, 8(1), F1–F12. https://doi.org/10.1111/j.1467-7687.2004.00384.x [DOI] [PubMed] [Google Scholar]
- Kuhl, P. K. , Coffey-Corina, S. , Padden, D. , Munson, J. , Estes, A. , & Dawson, G. (2013). Brain responses to words in 2-year-olds with autism predict developmental outcomes at age 6. PLOS ONE, 8(5), Article e64967. https://doi.org/10.1371/journal.pone.0064967 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kujala, T. , Kuuluvainen, S. , Saalasti, S. , Jansson-Verkasalo, E. , Von Wendt, L. , & Lepistö, T. (2010). Speech-feature discrimination in children with Asperger syndrome as determined with the multi-feature mismatch negativity paradigm. Clinical Neurophysiology, 121(9), 1410–1419. https://doi.org/10.1016/j.clinph.2010.03.017 [DOI] [PubMed] [Google Scholar]
- Kujala, T. , Lepistö, T. , Nieminen-von Wendt, T. , Näätänen, P. , & Näätänen, R. (2005). Neurophysiological evidence for cortical discrimination impairment of prosody in Asperger syndrome. Neuroscience Letters, 383(3), 260–265. https://doi.org/10.1016/j.neulet.2005.04.048 [DOI] [PubMed] [Google Scholar]
- Kutas, M. , & Hillyard, S. A. (1980). Reading senseless sentences: Brain potentials reflect semantic incongruity. Science, 207(4427), 203–205. https://doi.org/10.1126/science.7350657 [DOI] [PubMed] [Google Scholar]
- Lee, L. (1971). The Northwestern Syntax Screening Test. Northwestern University Press. [Google Scholar]
- Lepistö, T. , Kajander, M. , Vanhala, R. , Alku, P. , Huotilainen, M. , Näätänen, R. , & Kujala, T. (2008). The perception of invariant speech features in children with autism. Biological Psychology, 77(1), 25–31. https://doi.org/10.1016/j.biopsycho.2007.08.010 [DOI] [PubMed] [Google Scholar]
- Lepistö, T. , Kujala, T. , Vanhala, R. , Alku, P. , Huotilainen, M. , & Näätänen, R. (2005). The discrimination of and orienting to speech and non-speech sounds in children with autism. Brain Research, 1066(1), 147–157. https://doi.org/10.1016/j.brainres.2005.10.052 [DOI] [PubMed] [Google Scholar]
- Lepistö, T. , Nieminen-von Wendt, T. , von Wendt, L. , Näätänen, R. , & Kujala, T. (2007). Auditory cortical change detection in adults with Asperger syndrome. Neuroscience Letters, 414(2), 136–140. https://doi.org/10.1016/j.neulet.2006.12.009 [DOI] [PubMed] [Google Scholar]
- Lepistö, T. , Silokallio, S. , Nieminen-von Wendt, T. , Alku, P. , Näätänen, R. , & Kujala, T. (2006). Auditory perception and attention as reflected by the brain event-related potentials in children with Asperger syndrome. Clinical Neurophysiology, 117(10), 2161–2171. https://doi.org/10.1016/j.clinph.2006.06.709 [DOI] [PubMed] [Google Scholar]
- Lindström, R. , Lepistö-Paisley, T. , Vanhala, R. , Alén, R. , & Kujala, T. (2016). Impaired neural discrimination of emotional speech prosody in children with autism spectrum disorder and language impairment. Neuroscience Letters, 628, 47–51. https://doi.org/10.1016/j.neulet.2016.06.016 [DOI] [PubMed] [Google Scholar]
- Lord, C. , Rutter, M. , DiLavore, P. C. , & Risi, S. (1999). Autism Diagnostic Observation Schedule–WPS (ADOS-WPS) . Western Psychological Services. [Google Scholar]
- Ludlow, A. , Mohr, B. , Whitmore, A. , Garagnani, M. , Pulvermüller, F. , & Gutierrez, R. (2014). Auditory processing and sensory behaviours in children with autism spectrum disorders as revealed by mismatch negativity. Brain and Cognition, 86, 55–63. https://doi.org/10.1016/j.bandc.2014.01.016 [DOI] [PubMed] [Google Scholar]
- Maess, B. , Herrmann, C. S. , Hahne, A. , Nakamura, A. , & Friederici, A. D. (2006). Localizing the distributed language network responsible for the N400 measured by MEG during auditory sentence processing. Brain Research, 1096(1), 163–172. https://doi.org/10.1016/j.brainres.2006.04.037 [DOI] [PubMed] [Google Scholar]
- Magnée, M. J. , De Gelder, B. , Van Engeland, H. , & Kemner, C. (2008). Audiovisual speech integration in pervasive developmental disorder: Evidence from event-related potentials. The Journal of Child Psychology and Psychiatry, 49(9), 995–1000. https://doi.org/10.1111/j.1469-7610.2008.01902.x [DOI] [PubMed] [Google Scholar]
- Maitre, N. L. , Lambert, W. E. , Aschner, J. L. , & Key, A. P. (2013). Cortical speech sound differentiation in the neonatal intensive care unit predicts cognitive and language development in the first 2 years of life. Developmental Medicine & Child Neurology, 55(9), 834–839. https://doi.org/10.1111/dmcn.12191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Manfredi, M. , Cohn, N. , Mello, P. S. , Fernandez, E. , & Boggio, P. S. (2020). Visual and verbal narrative comprehension in children and adolescents with autism spectrum disorders: An ERP study. Journal of Autism and Developmental Disorders, 50, 2658–2672. https://doi.org/10.1007/s10803-020-04374-x [DOI] [PubMed] [Google Scholar]
- Matsuzaki, J. , Kuschner, E. S. , Blaskey, L. , Bloy, L. , Kim, M. , Ku, M. , Edgar, J. C. , Embick, D. , & Roberts, T. P. L. (2019). Abnormal auditory mismatch fields are associated with communication impairment in both verbal and minimally verbal/nonverbal children who have autism spectrum disorder. Autism Research, 3(1235), 1225–1211. https://doi.org/10.1002/aur.2136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McArthur, G. M. , & Bishop, D. V. (2005). Speech and non-speech processing in people with specific language impairment: A behavioural and electrophysiological study. Brain and Language, 94(3), 260–273. https://doi.org/10.1016/j.bandl.2005.01.002 [DOI] [PubMed] [Google Scholar]
- McCleery, J. P. , Ceponiene, R. , Burner, K. M. , Townsend, J. , Kinnear, M. , & Schreibman, L. (2010). Neural correlates of verbal and nonverbal semantic integration in children with autism spectrum disorders. The Journal of Child Psychology and Psychiatry, 51(3), 277–286. https://doi.org/10.1111/j.1469-7610.2009.02157.x [DOI] [PubMed] [Google Scholar]
- McWeeny, S. , & Norton, E. S. (2020). Understanding event-related potentials (ERPs) in clinical and basic language and communication disorders research: A tutorial. International Journal of Language & Communication Disorders, 55(4), 445–457. https://doi.org/10.1111/1460-6984.12535 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Megnin, O. , Flitton, A. , Jones, C. R. G. , de Haan, J. M. , Baldeweg, T. , & Charman, T. (2012). Audiovisual speech integration in autism spectrum disorders: ERP evidence for atypicalities in lexical-semantic processing. Autism Research, 5(1), 39–48. https://doi.org/10.1002/aur.231 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mills, D. L. , Coffey-Corina, S. A. , & Neville, H. J. (1993). Language acquisition and cerebral specialization in 20-month-old infants. Journal of Cognitive Neuroscience, 5(3), 317–334. https://doi.org/10.1162/jocn.1993.5.3.317 [DOI] [PubMed] [Google Scholar]
- Moher, D. , Liberati, A. , Tetzlaff, J. , Altman, D. G. , &. PRISMA Group. (2010). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA statement. International Journal of Surgery, 8(5), 336–341. https://doi.org/10.1371/journal.pmed.1000097 [DOI] [PubMed] [Google Scholar]
- Molfese, D. L. , & Molfese, V. J. (1985). Electrophysiological indices of auditory discrimination in newborn infants: The bases for predicting later language development? Infant Behavior & Development, 8(2), 197–211. https://doi.org/10.1016/S0163-6383(85)80006-0 [Google Scholar]
- Mottron, L. , Dawson, M. , Soulières, I. , Hubert, B. , & Burack, J. (2006). Enhanced perceptual functioning in autism: An update, and eight principles of autistic perception. Journal of Autism and Developmental Disorders, 36(1), 27–43. https://doi.org/10.1007/s10803-005-0040-7 [DOI] [PubMed] [Google Scholar]
- Mullen, E. (1995). Mullen Scales of Early Learning. American Guidance Service Inc. [Google Scholar]
- O'Connor, K. (2012). Auditory processing in autism spectrum disorder: A review. Neuroscience & Biobehavioral Reviews, 36(2), 836–854. https://doi.org/10.1016/j.neubiorev.2011.11.008 [DOI] [PubMed] [Google Scholar]
- Oram Cardy, J. E. , Flagg, E. J. , Roberts, W. , & Roberts, T. P. L. (2005). Delayed mismatch field for speech and non-speech sounds in children with autism. NeuroReport, 16(5), 521–525. https://doi.org/10.1097/00001756-200504040-00021 [DOI] [PubMed] [Google Scholar]
- Pellicano, E. , & Burr, D. (2012). When the world becomes “too real”: A Bayesian explanation of autistic perception. Trends in Cognitive Sciences, 16(10), 504–510. https://doi.org/10.1016/j.tics.2012.08.009 [DOI] [PubMed] [Google Scholar]
- Petit, S. , Badcock, N. A. , Grootswagers, T. , & Woolgar, A. (2020). Unconstrained multivariate EEG decoding can help detect lexical-semantic processing in individual children. Scientific Reports, 10(1), 10849. https://doi.org/10.1038/s41598-020-67407-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pijnacker, J. , Geurts, B. , Van Lambalgen, M. , Buitelaar, J. , & Hagoort, P. (2010). Exceptions and anomalies: An ERP study on context sensitivity in autism. Neuropsychologia, 48(10), 2940–2951. https://doi.org/10.1016/j.neuropsychologia.2010.06.003 [DOI] [PubMed] [Google Scholar]
- Ribeiro, T. C. , Valasek, C. A. , Minati, L. , & Boggio, P. S. (2013). Altered semantic integration in autism beyond language: A cross-modal event-related potentials study. NeuroReport, 24(8), 414–418. https://doi.org/10.1097/WNR.0b013e328361315e [DOI] [PubMed] [Google Scholar]
- Ring, H. , Sharma, S. , Wheelwright, S. , & Barrett, G. (2007). An electrophysiological investigation of semantic incongruity processing by people with Asperger's syndrome. Journal of Autism and Developmental Disorders, 37(2), 281–290. https://doi.org/10.1007/s10803-006-0167-1 [DOI] [PubMed] [Google Scholar]
- Roberts, T. P. L. , Cannon, K. M. , Tavabi, K. , Blaskey, L. , Khan, S. Y. , Monroe, J. F. , Qasmieh, S. , Levy, S. E. , & Edgar, J. C. (2011). Auditory magnetic mismatch field latency: A biomarker for language impairment in autism. Biological Psychiatry, 70(3), 263–269. https://doi.org/10.1016/j.biopsych.2011.01.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Russo, N. , Zecker, S. , Trommer, B. , Chen, J. , & Kraus, N. (2009). Effects of background noise on cortical encoding of speech in autism spectrum disorders. Journal of Autism and Developmental Disorders, 39(8), 1185–1196. https://doi.org/10.1007/s10803-009-0737-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sandbank, M. , Yoder, P. , & Key, A. P. (2017). Word processing in children with autism spectrum disorders: Evidence from event-related potentials. Journal of Speech, Language, and Hearing Research, 60(12), 3441–3455. https://doi.org/10.1044/2017_JSLHR-S-17-0011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schwartz, S. , Shinn-Cunningham, B. , & Tager-Flusberg, H. (2018). Meta-analysis and systematic review of the literature characterizing auditory mismatch negativity in individuals with autism. Neuroscience & Biobehavioral Reviews, 87, 106–117. https://doi.org/10.1016/j.neubiorev.2018.01.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sparrow, S. S. , Balla, D. A. , & Cicchetti, D. V. (1984). Vineland Adaptive Behavior Scales: Interview Edition. American Guidance Service. [Google Scholar]
- Stowe, L. A. , Kaan, E. , Sabourin, L. , & Taylor, R. C. (2018). The sentence wrap-up dogma. Cognition, 176, 232–247. https://doi.org/10.1016/j.cognition.2018.03.011 [DOI] [PubMed] [Google Scholar]
- Su, P. L. , Rogers, S. J. , Estes, A. , & Yoder, P. (2020). The role of early social motivation in explaining variability in functional language in toddlers with autism spectrum disorder. Autism, 25(1), 244–257. https://doi.org/10.1177/1362361320953260 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tager-Flusberg, H. , & Kasari, C. (2013). Minimally verbal school-aged children with autism spectrum disorder: The neglected end of the spectrum. Autism Research, 6(6), 468–478. https://doi.org/10.1002/aur.1329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tallal, P. , & Piercy, M. (1974). Developmental aphasia: Rate of auditory processing and selective impairment of consonant perception. Neuropsychologia, 12(1), 83–93. https://doi.org/10.1016/0028-3932(74)90030-X [DOI] [PubMed] [Google Scholar]
- Thurm, A. , Manwaring, S. S. , Swineford, L. , & Farmer, C. (2015). Longitudinal study of symptom severity and language in minimally verbal children with autism. The Journal of Child Psychology and Psychiatry, 56(1), 97–104. https://doi.org/10.1111/jcpp.12285 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ueno, K. , Nagoshi, S. , & Konuki, S. (2008). Picture Vocabulary Test–Revised. Success/Bell Co Ltd. [Google Scholar]
- van Boxtel, J. J. , & Lu, H. (2013). A predictive coding perspective on autism spectrum disorders. Frontiers in Psychology, 4, 19. https://doi.org/10.3389/fpsyg.2013.00019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang, X. , Wang, S. , Fan, Y. , Huang, D. , & Zhang, Y. (2017). Speech-specific categorical perception deficit in autism: An event-related potential study of lexical tone processing in Mandarin-speaking children. Scientific Reports, 7, 43254. https://doi.org/10.1038/srep43254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wechsler, D. (1974). Wechsler Intelligence Scale for Children–Revised. Psychological Corporation. [Google Scholar]
- Weismüller, B. , Thienel, R. , Youlden, A. M. , Fulham, R. , Koch, M. , & Schall, U. (2015). Psychophysiological correlates of developmental changes in healthy and autistic boys. Journal of Autism and Developmental Disorders, 45(7), 2168–2175. https://doi.org/10.1007/s10803-015-2385-x [DOI] [PubMed] [Google Scholar]
- Wetherby, A. M. , Watt, N. , Morgan, L. , & Shumway, S. (2007). Social communication profiles of children with autism spectrum disorders late in the second year of life. Journal of Autism and Developmental Disorders, 37(5), 960–975. https://doi.org/10.1007/s10803-006-0237-4 [DOI] [PubMed] [Google Scholar]
- Whitehouse, A. J. , & Bishop, D. V. (2008). Do children with autism ‘switch off’ to speech sounds? An investigation using event-related potentials. Developmental Science, 11(4), 516–524. https://doi.org/10.1111/j.1467-7687.2008.00697.x [DOI] [PubMed] [Google Scholar]
- Williams, Z. J. , Abdelmessih, P. G. , Key, A. P. , & Woynaroski, T. G. (2020). Cortical auditory processing of simple stimuli is altered in autism: A meta-analysis of auditory evoked responses. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 1–15. https://doi.org/10.1016/j.bpsc.2020.09.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yoder, P. , Watson, L. R. , & Lambert, W. (2015). Value-added predictors of expressive and receptive language growth in initially nonverbal preschoolers with autism spectrum disorders. Journal of Autism and Developmental Disorders, 45(5), 1254–1270. https://doi.org/10.1007/s10803-014-2286-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yoshimura, Y. , Kikuchi, M. , Hayashi, N. , Hiraishi, H. , Hasegawa, C. , Takahashi, T. , Oi, M. , Remijn, G. B. , Ikeda, T. , Saito, D. N. , Kumazaki, H. , & Minabe, Y. (2017). Altered human voice processing in the frontal cortex and a developmental language delay in 3- to 5-year-old children with autism spectrum disorder. Scientific Reports, 7(1), 1–11. https://doi.org/10.1038/s41598-017-17058-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yoshimura, Y. , Kikuchi, M. , Hiraishi, H. , Hasegawa, C. , Takahashi, T. , Remijn, G. B. , Oi, M. , Munesue, T. , Higashida, H. , Minabe, Y. , & Kojima, H. (2016). Atypical development of the central auditory system in young children with autism spectrum disorder. Autism Research, 9(11), 1216–1226. https://doi.org/10.1002/aur.1604 [DOI] [PubMed] [Google Scholar]
- Yu, L. , Fan, Y. , Deng, Z. , Huang, D. , Wang, S. , & Zhang, Y. (2015). Pitch processing in tonal-language-speaking children with autism: An event-related potential study. Journal of Autism and Developmental Disorders, 45(11), 3656–3667. https://doi.org/10.1007/s10803-015-2510-x [DOI] [PubMed] [Google Scholar]
- Zhang, J. (2019). McGurk effect by individuals with autism spectrum disorder and typically developing controls: A systematic review and meta-analysis. Journal of Autism and Developmental Disorders, 49(1), 34–43. http://doi.org/10.1007/s10803-018-3680-0 [DOI] [PubMed] [Google Scholar]
- Zhang, J. , Meng, Y. , Tong, X. , Yuan, Z. , Wu, C. , & Ieong, S. L. (2018). Exploring the neural correlates of lexical stress perception in English among Chinese–English bilingual children with autism spectrum disorder: An ERP study. Neuroscience Letters, 666, 158–164. http://doi.org/10.1016/j.neulet.2017.12.031 [DOI] [PubMed] [Google Scholar]
- Zimmerman, I. L. , Steiner, V. G. , & Pond, R. E. V. (2002). Preschool Language Scale–Fourth Edition (PLS-IV). Psychological Corporation. [Google Scholar]