Oromotor skills in autism spectrum disorder: A scoping review

Abstract

Oromotor functioning plays a foundational role in spoken communication and feeding, two areas of significant difficulty for many autistic individuals. However, despite years of research and established differences in gross and fine motor skills in this population, there is currently no clear consensus regarding the presence or nature of oral motor control deficits in autistic individuals. In this scoping review, we summarize research published between 1994 and 2022 to answer the following research questions: (1) What methods have been used to investigate oromotor functioning in autistic individuals? (2) Which oromotor behaviors have been investigated in this population? and (3) What conclusions can be drawn regarding oromotor skills in this population? Seven online databases were searched resulting in 107 studies meeting our inclusion criteria. Included studies varied widely in sample characteristics, behaviors analyzed, and research methodology. The large majority (81%) of included studies report a significant oromotor abnormality related to speech production, nonspeech oromotor skills, or feeding within a sample of autistic individuals based on age norms or in comparison to a control group. We examine these findings to identify trends, address methodological aspects hindering cross-study synthesis and generalization, and provide suggestions for future research.

Lay Summary

Autistic individuals demonstrate movement abnormalities during tasks including walking, balancing, reaching, and tool use. However, little is known about movements of the mouth despite well-known difficulties with speech and feeding in ASD. Here we (1) summarize studies related to mouth movements in ASD, showing that 81% of these studies indicate significant abnormalities in autistic individuals, (2) discuss common findings in these studies, and (3) make suggestions for future research.

INTRODUCTION

Autism spectrum disorder (ASD) is a developmental disability defined by persistent deficits in social communication and interaction and by restricted, repetitive patterns of behavior, interests, or activities (American Psychiatric Association, 2013). While these core symptoms often manifest as stereotyped or repetitive movements, motor impairments themselves are not considered a diagnostic marker of ASD. Nonetheless, the presence of motor abnormalities among children with ASD is well established, including in infancy (West, 2019), and some have argued that motor impairments constitute a core feature of ASD (Fournier et al., 2010; Hilton et al., 2012). Prevalence estimates of motor impairment in ASD are reported as high as 35%–100% (D. Green et al., 2002; Licari et al., 2020), and these impairments appear to be significantly underdiagnosed (Licari et al., 2020). The reported deficits impact both gross and fine motor skills and include hypotonia, postural instability, and impairments related to balance, gait, object control, manual dexterity, repetitive hand and foot movements, precision grip, and handwriting (Freitag et al., 2007; Garrido et al., 2017; Gong et al., 2020; Jansiewicz et al., 2006; Kushki et al., 2011; Ming et al., 2007; Miyahara et al., 1997; Pan et al., 2009). From a motor control perspective, there is evidence suggesting that these deficits result from difficulties with motor planning (Dewey et al., 2007; Glazebrook et al., 2006; Mari et al., 2003), motor execution (Glazebrook et al., 2009; Rinehart et al., 2006), and the integration of sensory information (Cascio et al., 2012). There is growing interest in understanding motor functioning among individuals with ASD since motor impairments are associated with full-scale IQ (Klupp et al., 2021), executive functioning (Michel et al., 2011), and communicative development (Alcock & Krawczyk, 2010; Bedford et al., 2016; Gernsbacher et al., 2008; West, 2019) in this population.

Oromotor skills (i.e., motor skills involving the mandible, lips, lower face, soft palate, and/or tongue) have received relatively little attention in ASD, which reflects a similar paucity of developmental oromotor research across populations including typically developing children and children with other developmental disabilities. The functional significance of oromotor skills, which provide a foundation for behaviors including sucking, chewing, vocalizing, babbling, and speaking, cannot be overstated. However, many questions remain unanswered regarding oromotor skills in ASD, including whether or when adult-like levels of oromotor function are achieved, how to best assess these skills, how to categorize and interpret observed abnormalities, and whether these skills are linked to the development of other outcomes such as social functioning, nonverbal IQ, ASD symptom severity, and expressive and receptive language.

Oromotor development and function

Typical oromotor development involves the refinement of a variety of functional movements over time, beginning as early as 12 weeks after gestation when sucking behaviors can be observed via ultrasound (de Vries et al., 1982). Precursors to vocal imitation are observed within days after birth (Chen et al., 2004; Meltzoff & Moore, 1983) and the development of adult-like oromotor skills continues past age 10 (see Kent & Vorperian, 1995). There is little information available regarding the developmental course of oromotor skills in ASD, although impairments of oral imitation have been reported as early as 26 months (Rogers et al., 2003), and oromotor differences remain through adulthood (Kissine & Geelhand, 2019; Shriberg et al., 2001). Furthermore, younger siblings of children diagnosed with ASD—who themselves have a heightened likelihood of receiving an ASD diagnosis (Zwaigenbaum et al., 2007)—often demonstrate delayed onset of developmental motor milestones (Iverson & Wozniak, 2007) and significant differences in fine motor skills at 12, 24, and 36 months compared to typically developing controls (Garrido et al., 2017), suggesting that similar deficits may be manifested in the oromotor system.

How oromotor function is assessed

Oromotor function is assessed using a variety of methods, which each provide relative advantages for addressing particular research questions. Perceptual methods include visual-perceptual assessments of structure and function (e.g., an oral mechanism examination) and auditory-perceptual assessments of tasks such as the repetition of speech sounds, spontaneous speech production, and maximum performance tasks (e.g., diadochokinetic rates). Perceptual assessment approaches provide ecologically valid data that are often relatively simple to acquire, although typically require extensive training, are susceptible to multiple sources of rater bias (Kent, 1996), and may be too coarse and/or unreliable to detect subtle differences in motor function (Green, 2015). Instrumental assessment approaches such as acoustic (i.e., based on sound signals) and kinematic (i.e., based on movement signals) analyses, on the other hand, provide highly detailed and reliable data that offer valuable insights into the physiologic development and performance of the oromotor system. However, these methods often require specialized equipment and may be time-consuming to analyze and interpret. Various perceptual and instrumental methods have been used to assess oromotor functioning in ASD, but these methods have not been summarized and reported in prior literature, making it difficult to assess the quality and type of information available regarding this issue.

What is assessed?

Speech

Speech production involves the fine coordination of over 100 muscles across multiple systems responsible for respiration, phonation, resonance, and articulation (Duffy, 2019). Many aspects of speech are studied in typically developing children to characterize normal development and function, and in atypical populations as a tool for clinical stratification, progress monitoring, and the prediction of developmental outcomes.

Perhaps the most widely reported characteristic of speech is articulatory accuracy assessed via auditory-perceptual evaluation. Analysis of speech errors can offer important information about the presence and nature of a motor speech disorder (Pernon et al., 2022), including for differential diagnosis (Allison et al., 2020). Other approaches to assessing the overall adequacy of speech production include measuring intelligibility (i.e., the degree to which the speech signal is understood by a listener) (e.g., Kent et al., 1989) and phonetic inventory (i.e., the number of speech sounds produced correctly) (e.g., Chenausky et al., 2018), which can be used to indirectly assess the capabilities of the oral motor system.

Other aspects of speech production that are commonly assessed include temporal features like the duration of speech sounds and the rate of speech production, which can both provide important information about motor planning and execution abilities. Specific characteristics of the speech signal available via acoustic analysis—including formants (i.e., resonant frequencies corresponding to vocal tract configurations) (see Kent & Vorperian, 2018 for a review) and voice onset time (i.e., the length of time between the release of a stop consonant and the onset of voicing) (Allen et al., 2003)—allow examination of the speed, consistency, and coordination of the speech motor system. Kinematic methods (e.g., electromagnetic articulography or video-based movement tracking) provide highly detailed data regarding movement velocity, acceleration, and magnitude (Nip, 2012; Yunusova et al., 2010), from which researchers can derive information including the level of coordination among articulators (J. R. Green et al., 2002) and the spatiotemporal stability of speech movements (Smith et al., 1995). Speech motor function may also be investigated more directly via real-time recordings of muscle activations via surface electromyography (EMG; e.g., McClean & Tasko, 2003) and imaging of brain activity, such as with electroencephalography (EEG; e.g., Janssen et al., 2020), positron emission tomography (PET; e.g., Price, 2012), functional magnetic resonance imaging (fMRI; e.g., Price, 2012), and magnetoencephalography (MEG; e.g., Heinks-Maldonado et al., 2006).

Nonspeech oromotor skills

Nonspeech oromotor skills are also commonly assessed, including lateralizing and protruding the tongue, spreading and puckering the lips, puffing the cheeks, and opening and closing the mouth. Assessment of these tasks is useful for several reasons. First, it can provide valuable information about the motor functioning of the articulators in isolation, since speakers may be able to compensate for impairments in one articulator by adapting the performance of another during speech (Yunusova et al., 2008). Second, the motor control of speech is domain-specific, characterized by functional neurological organization attuned specifically for the generation of speech sounds (Ziegler & Ackermann, 2013). Therefore, the assessment of single articulators in nonspeech contexts independent of influences from the linguistic system provides unique information about oromotor function (Ballard & Robin, 2002). Third, nonspeech oromotor imitation is particularly useful in the assessment of children with ASD, up to one-third of whom remain minimally verbal by school age (Tager-Flusberg & Kasari, 2013) and therefore have significant difficulty providing speech samples for evaluation. Nonspeech oromotor skills are most commonly assessed visuoperceptually via an oral mechanism examination or standardized test (e.g., the Oral Speech Mechanism Screening Examination; St. Louis & Ruscello, 2000). Alternatively, nonspeech oromotor behaviors such as spontaneous motility of the articulators have been assessed kinematically (Green & Wilson, 2006).

Feeding

Despite the significant functional impacts of speech development, the most critical function of the oromotor system is feeding. There is much research interest in the feeding behaviors of children diagnosed with ASD since they are five times more likely to experience feeding problems compared to their peers (Sharp et al., 2013), leading to a high rate of nutritional deficiencies (Cornish, 1998). The most commonly reported difficulties among children with ASD are restricted food preferences related to sensory aversions, although a limited body of research has investigated feeding-related motor impairments in ASD. Methods of assessment have included perceptual observation as part of an oromotor assessment battery (Amato & Slavin, 1998), observational studies of meals (Peterson et al., 2016), and EMG measurements of muscle activity during a feeding action (e.g., Cattaneo et al., 2007).

Potential correlations between oromotor and other skills in ASD

Oromotor skills may serve as valuable predictors of other skills in ASD, such as expressive and receptive language, social functioning, and nonverbal IQ. Valid predictors of language outcomes in ASD are critical since the mechanisms that underlie severe language impairments in ASD remain unclear and such knowledge would directly inform the development of high-quality language interventions. Predictors of other variables such as ASD symptom severity, nonverbal IQ, and social functioning are also critically needed, since the assessment of such skills currently requires waiting until the typical age of emergence of skills like symbolic play, spoken language, and direction following, leading to a delay in identification of impairments. There is promising evidence that oromotor skills can act as a predictor for these skills; for instance, some studies show that oromotor skills are predictive of expressive language (e.g., Amato & Slavin, 1998; Gernsbacher et al., 2008), including in minimally verbal autistic individuals (Chenausky et al., 2019).

Aims of this scoping review

Due to a paucity of systematic research regarding oromotor skills in ASD, it is currently unclear which oromotor skills have been investigated in ASD, what methods have been used to investigate these skills, and what generalizations (if any) can be made from existing findings. Indeed, there does not even appear to be a consensus on whether oromotor skills are impacted in this population. For instance, while some research suggests that oromotor skills such as articulation are relatively spared among autistic individuals (Dalton et al., 2017; Kjelgaard & Tager-Flusberg, 2001), other studies report overt abnormalities in oromotor functioning such as reduced speech accuracy, longer sound durations, reduced consistency of speech production, and decreased oromotor control during nonspeech tasks (Adams, 1998; Chenausky et al., 2019; Gernsbacher et al., 2008; Kissine & Geelhand, 2019).

For these reasons, a fuller understanding of the oromotor skills of autistic individuals is warranted. Such findings have the potential to advance efforts to improve the early detection of speech and language deficits in ASD, support the use of motor- or articulation-based therapies in conjunction with language-based approaches for autistic individuals, and identify neurobiological mechanisms influencing communication development. The purpose of this scoping review is to summarize and disseminate existing research concerning oromotor skills among autistic individuals—including research correlating these skills to other outcomes—and to address methodological aspects of these studies that hinder cross-study synthesis and generalization to the population. The long-term goal is to motivate future investigations into the presence, nature, and severity of oromotor impairments in ASD and the association of such deficits with other functional skills.

METHODS

This review follows the framework for a scoping review outlined by Arksey and O’Malley (2005), who provided guidelines that were used to structure the methodology reported below. The review also adhered to the PRISMA guidelines for scoping reviews (Tricco et al., 2018).

Research questions

To summarize and disseminate research findings and to identify research gaps in the existing body of literature, we addressed the following three research questions:

(1) What methods have been used to investigate oromotor functioning among individuals with ASD?, (2) Which oromotor behaviors have been investigated?, and (3) What conclusions can be drawn regarding oromotor skills in ASD?

Search procedures

In consultation with a health sciences librarian, a comprehensive search was conducted of the following databases: Cumulative Index to Nursing and Allied Health Literature (CINAHL), Education Resources Information Center (ERIC), MEDLINE, ProQuest Dissertations & Theses Database (PQDT), PsychInfo, ScienceDirect, and Web of Science. Studies pertinent to the outlined research questions were found by searching databases using three combined categories of carefully selected keywords (i.e., autism, motor, and oromotor). The search terms used were: [autis* AND (motor OR articulation) AND ((oral OR oral* OR oro*) OR feeding OR chewing OR (nonspeech OR “nonspeech”) OR speech OR articulation)]. Search syntax was adjusted as necessary to fit the requirements of specific databases.

To be included in the review, a study had to meet the following inclusion criteria: (1) subject-specific or aggregate data that applied only to individuals with ASD (of any age) could be extracted, (2) the diagnosis of ASD was made using an established tool (e.g., the Autism Diagnostic Observation Schedule [ADOS] or the Child Autism Rating Scale [CARS]) according to DSM-IV or DSM-5 criteria (and thus published no earlier than 1994), (3) quantitative data from a behavioral measure (i.e., perceptual or instrumental), parent report, or medical record review of oromotor function was reported, and (4) the study was available in English. A study was excluded from the review if it met any of the following exclusion criteria: (1) the main diagnosis of the ASD participant(s) was a known syndrome (e.g., Fragile X Syndrome, Prader-Willi Syndrome), (2) a comorbid diagnosis with potential to explain oromotor differences (e.g., hearing impairment, cerebral palsy) was reported, or (3) the study was conducted using an animal model.

A total of 16,399 studies were collected for screening (Figure 1). After removing 3855 duplicates that were included in the results from more than one database, 12,544 articles were left to consider. The first author (MFM) read the titles and, as necessary, the abstracts of all remaining articles to remove irrelevant articles (i.e., violated exclusionary criteria within the title or abstract). To assess reliability, the second author (KVC) screened the titles and abstracts of 1500 articles, and the inter-rater reliability was 75%. The first author made the final decision on disagreements during this stage, and 776 studies that could not be excluded based on the title and abstract screening were selected for full-text review. The first and second authors independently read 370 of these studies and achieved an inter-rater agreement of 80%, settling disagreements at this stage by discussion. The first author then reviewed the additional 406 studies. 669 studies were excluded during this stage, with the most common reasons being lack of quantitative oromotor data (512 studies), no ASD-specific findings (34 studies), diagnosis of ASD not made by an approved instrument (32 studies), and being a review paper (18 studies). Reasonable effort was made to contact authors or their associated organizations when papers were unavailable. A total of 107 research papers, book chapters, and theses were ultimately included in the review.

FIGURE 1.

PRISMA flow diagram summarizing the study identification process.

For this scoping review, we define oromotor as referring to skills involving the use of the mandibular, labial, facial, velar, and lingual structures for speech, nonspeech oral movements, and feeding. Studies specific to breath support or phonatory control (e.g., vocal loudness, average fundamental frequency, pitch variation during lexical stress tasks, etc.) were excluded. However, studies investigating the coordination of the articulatory and phonatory subsystems (e.g., measures of voice onset time) were included since they involve oral structures in addition to laryngeal structures. Studies addressing prosody cannot be disregarded when considering the connection between speech and language (for a review of this topic see Loveall et al., 2021) but were generally considered outside the scope of this review unless the duration of speech sounds was reported (e.g., Diehl & Paul, 2013; Grossman et al., 2010; Van Santen et al., 2010). Studies involving infant siblings of children with ASD (who have an elevated likelihood of receiving an ASD diagnosis) were only included if a valid ASD diagnosis was later made and reported in the study. Finally, studies that looked primarily at phonological skills were excluded unless some information about oromotor functioning could be derived from the results.

The following data, if reported, were extracted from each included study and entered into a Microsoft Excel spreadsheet: author(s), title, year of publication, journal, sample size, number of male and female subjects, control group type, participant ages, diagnosis method, NVIQ, expressive language skills, behavior(s) analyzed (e.g., accuracy, duration, etc.), test(s) used, speech unit(s) analyzed (i.e., phoneme, syllable, phrase/sentence, and/or connected/spontaneous speech), domain(s) investigated (i.e., speech, nonspeech, and/or feeding), data type (i.e., perceptual, instrumental, report/questionnaire, and/or record review), and results. Because this is a scoping review of existing studies, it did not require ethics committee approval.

RESULTS

The studies included in this review were heterogeneous in terms of their aims, sample characteristics, and methodology. Table 3 provides a summary of the included literature including a summary of relevant results.

TABLE 3.

Relevant data from included studies.

Author(s)	Domaina	Features	n ASD (M), M:F ratio	ASD age M ± SD (range)	Controlsb	NVIQ M ± SD (range)c	Exp. lang M ± SD (range)d	Methode	Testf	Resultsg
Adams (1998)	S/NS	Accuracy	4 (3), 3:1	9.0 ± 2.3 (6.3–11.3)	TD	LIPS (92.75 ± 10.5)	-	P	KSPT	TD > ASD on oral movements, complex productions, and total score; ASD = TD on simple productions
Akin-Bulbul and Ozdemir (2022)	S/NS	Accuracy	30 (22), 2.8:1	2.7 ± 0.3 (2.3–3)	TD, DD	BSID-III Cognitive m = 24	-	P	-	TD > ASD on meaningful and nonmeaningful vocal imitation
Amato and Slavin (1998)	S/NS/F	Accuracy, feeding difficulty	20 (16), 4:1	(2.5–4.0)	-	-	V and NV	P	OME	Verbal and nonverbal groups had reduced oromotor functioning. V > NV on overall score, eating behaviors, voluntary nonverbal oral ability, and pre-speech/speech behavior; V = NV on musculoskeletal anatomy and basic oral motor functions
Arutiunian et al. (2022)	S	Accuracy	71	9.6 ± 1.4 (7–11.1)	TD	83.1 ± 20.5 (40–125)	Various	P	-	TD > ASD on nonword repetition, TD = ASD on word repetition
Ashley et al. (2020)	F	Feeding difficulty	19 (15), 3.8:1	(1.3–3.0)	TD, HR, non-TD	MSEL-VR (30.9 ± 10.9)	MSEL-EL 25.9 ± 8.1	Q	BPFAS	ASD = TD on oral motor skills for eating
Belmonte et al. (2013)	S/NS	Accuracy	31 (25), 4.2:1	3.4 ± 0.8 (1.8–5.4)	-	-	Various	P	CDOMA	35% of ASD group had an oral motor impairment associated with an EL/RL disparity (EL < RL). Oral motor skills correlated with pre-intervention RL and EL and with learning rates. Oral motor skills varied independently of GM and FM
Biller and Johnson (2019)	S	Accuracy, voicing	5 (4), 4:1	5.3 ± 1.3 (3.6–6.9)	-	MSEL-VR AE 1:8–2:7	MSEL AE 0.9 ± 0.2	P	VMPAC	MV autistic children below age level on oromotor control and number of speech sounds/syllables produced imitatively or spontaneously
Biller and Johnson (2020)	S	Accuracy, voicing	1 (1)	3.3	-	MSEL-NV AE 2:1	MSEL-EL AE 1.8	P	VMPAC	Oromotor control skills in 5th percentile
Biller et al. (2022)	S	Inventory, voicing	1 (1)	4.9	-	MSEL-VR 2:0	Exp. vocab 25 words	P	VMPAC	Limited sound repertoire
Bodison (2015)	NS	Accuracy	32 (25), 3.6:1	7.5 ± 1.4 (5–8.9)	-	-	-	P	SIPT	Reduced imitative oral praxis
Boorom (2018)	S	Accuracy	11 (11)	4.9 ± 0.3 (4.3–5.4)	-	-	TOPEL DV raw 24.1 ± 17.3	P	CTOPP-2	Below average group mean for nonword repetition
Brisson et al. (2012)	F	Motor anticipation	13 (13)	0.3 ± 0.1 (0.3–0.5)	TD	8/13 subjects IQ < 75	-	P	-	TD > ASD on anticipatory mouth opening in response to an approaching spoon
Broome et al. (2021)	S	Accuracy, inventory	22 (20), 10:1	3.9 ± 1.2 (2–5.3)	-	WPPSI-III 99 ± 20.7	PLS-4 EC 65.6 ± 14.2 (50–85)	P	POP	Cluster analysis identified three subgroups: high language/high speech, low EL/low speech/high RL, and low language/low speech
Broome et al. (2022)	S	Inventory	23 (21), 10.5:1	4.4 ± 1.3 (2–7.2)	-	WPPSI-III 99 ± 20.7	PLS-4 EC 65.6 ± 14.2 (50–85)	P	POP, SWPT	Varied speech development trajectories by subgroup: high language/high speech and low EL/low speech/high RL remained stable; low language/low speech was variable
Cattaneo et al. (2007)	F	Muscle activity	8 (7), 7:1	6.2 (5.1–9.0)	TD	WISC-R (98 ± 12.4)	-	I	-	TD > ASD on mylohyoid activity while observing a person eat and while reaching for/grasping food
Chenausky and Schlaug (2018)	S	Accuracy, inventory	30 (25), 5:1	6.4 (3.4–9.7)	-	-	EV < 20 words	P	-	Reduced % syllables approximately correct at baseline
Chenausky and Tager-Flusberg (2017)	S	VOT	11 (7), 1.8:1	2.2 ± 0.6 (1.5–3)	TD, HR−	-	Age 3 MSEL EL 53.6 ± 8.6	P/I	-	ASD = HR− = LRC on VOT mean and standard deviation. ASD significantly less likely to produce acoustically distinct VOT for /b/ and /p/ at 36 months, but not at 18 or 24 months
Chenausky et al. (2016)	S	Accuracy, inventory	30 (27), 9:1	1.5 ± 0.5 (1–2)	-	MSEL-MA (21.4 ± 9.2) (n = 14)	EV < 20 words	P/I	KSPT	Reduced syllables approximated, vowels correct, and consonants correct at baseline
Chenausky, Kernbach, et al. (2017)	S	Accuracy, inventory	30 (25), 5:1	6.3 ± 1.6 (3.4–9.7)	-	MSEL-VR (27.9 ± 9.7)	MSEL EL 11.2 ± 2.4 (6–19)	P	-	White matter integrity accounted for significant variance in % syllable-initial consonants correct, % responses, and % syllable insertions
Chenausky Nelson, and Tager-Flusberg (2017)	S	Inventory	10 (7), 2.3:1	6.3 (3.4–9.7)	TD, HR−	-	Age 2 MSEL EL 48.0 ± 9.2	P	-	Vocalization rate predicted # different consonants for non-autistic, but not autistic, children at 12 months and for both groups at 18 and 24 months. EL predicted # different consonants for only the non-autistic children at 18 and 24 months. Mean # different consonants at 12, 18, and 24 months was not significantly different between groups
Chenausky, Norton, and Schlaug (2017)	S/NS	Accuracy, inventory	4 (4)	5 ± 1.2 (4.1–6.6)	-	MSEL-VR (26.8 ± 13.9)	MSEL EL raw 13.3 ± 4.4	P	KSPT	V > MV on oral movements and simple phoneme/syllables
Chenausky et al. (2018)	S	Accuracy, inventory	38 (31), 4.4:1	6.4 ± 1.6 (3.4–10.7)	-	MSEL-VR (30.1 ± 9.6)	MSEL EL 11.1 ± 2.7	I	-	Reduced % syllables approximately correct at baseline. Baseline phonetic inventory significantly predicted change in % syllables approximately correct, while nonverbal IQ, baseline expressive language, and age did not
Chenausky et al. (2019)	S/NS	Accuracy, rate, coordination, consistency, resonance, voicing	54 (41), 3.2:1	6.5 ± 1.7 (3.4–10.7)	-	LIPS-3 30–115 (67.4 ± 19.5)	NDW 46.3 ± 57.1	P	-	24% of sample exhibited severely disordered speech including >4 signs of CAS, 30% had speech abnormalities inconsistent with CAS, and 24% produced too little speech for analysis. NS oromotor ability was not predictive of EL. Speech imitation predicted # different words in the sCAS and insufficient speech groups
Chenausky et al. (2020)	S	Accuracy, rate, coordination, consistency, resonance, voicing	27 (24), 8:1	6.6 ± 1.4 (4–9.7)	CAS	MSEL-VR raw (29.6 ± 9.0)	MSEL EL raw 11.2 ± 1.7 (8–14)	P	GFTA-2	MV autistic children with CAS exhibited significantly slower rate and more vowel errors, syllable segmentation, groping, difficulty with coarticulation, and additions compared to a verbal CAS group, although fewer consonant distortions
Chenausky et al. (2021)	S	Accuracy, rate, coordination, inventory, consistency, resonance, formants	38 (33), 6.6:1	6.8 ± 1.7 (5–10.7)	-	MSEL-VR 30.3 ± 9.8	MSEL EL raw 11.3 ± 2.7 (6–19)	P	KSPT	A cluster defined by perceptual within-token variability demonstrated more within- and between-token variability and less centralized vowel space than the low variability group and TD
Chenausky, Norton, et al. (2022)	S	Accuracy, inventory	14 (11), 3.7:1	10 ± 3.9 (4.3–18.8)	-	LIPS-3 69.1 ± 8.5	EV < 20 words	P	KSPT	Low % syllables, consonants, and vowels correct; low KSPT scores; small phonetic inventory
Cleland et al. (2010)	S	Accuracy	69	9.5 ± 2.2 (5.0–13.0)	-	RPM (103.0 ± 15.0)	CELF-3 EL 84.1 ± 20.2	P	GFTA-2	12% of sample below average on GFTA, although 33% in the normal range presented with errors. Non-developmental errors observed in both groups
Dalton et al. (2017)	S/NS	Accuracy	10 (6), 1.5:1	4.8 ± 0.5 (3–5.5)	TD, sCAS	MSEL-VR AE 4.8 ± 1.3	MLU > 3.0	P	VMPAC	ASD = TD = sCAS on nonverbal oral, verbal motor, and concurrent verbal motor imitation. NS oral imitation and verbal motor imitation were correlated with joint attention only in ASD group
Demartini et al. (2021)	F	Feeding difficulty	106 (77), 2.7:1	33.2 ± 12.9 (17–67)	TD	WAIS-IV ≥ 70	ADOS communication 4.7 ± 2.0	Q	SWEAA	TD > ASD on motor control for eating
Demopoulos and Lewine (2016)	S	Accuracy	60 (48), 4:1	10.8 ± 3.4 (5.5–18.5)	TD	WISC-IV FSIQ 46–136 (81.7 ± 22.0)	CELF-4 EL 75.5 ± 26.6 (45–128)	P	GFTA-2	ASD = TD on GFTA
Deshmukh (2012)	S/NS	Accuracy, rate, consistency	12 (12)	6.5 ± 1.9 (4–10.3)	TD, MSD	-	EVT 100.0 ± 14.8	P	DEAP, GFTA-2, OSMSE-3	ASD group had normal oral structure/function. 15% of the ASD group below the normal range on the GFTA. ASD = TD on rate and consistency
Diehl and Paul (2013)	S	Duration	24 (16), 2:1	12.3 ± 2.3 (8.0–16.0)	TD, LmD	WASI/D AS (103.61 ± 17.14)	CELF-IV 100.5 ± 16.2	I	PEPS-C	ASD > TD on utterance length when expressing dislike, asking questions, and making statements. ASD = TD when expressing like and when using stress to indicate focus
Ehlen et al. (2020)	S	Rate	32 (18), 1.3:1	37.1 ± 10.7	TD	Incl. criteria >85	-	I	-	ASD = TD on word duration
Espanola Aguirre and Gutierrez (2019)	S/NS	Accuracy	30 (22), 2.8:1	3.6 ± 1.0 (1.3–4.0)	TD	MSEL composite 56.6 ± 11.3	MCDI # of words 155.7 ± 150.1	P	MVIA	ASD = TD on vocal and facial imitation
Franich et al. (2020)	S	Coordination	10 (9), 9:1	(22–31)	TD	TONI-4 (102.33 ± 9.56)	-	I	-	TD > ASD on timing phrases along with a metronome; ASD had longer time interval between two consecutive repetitions of first word in target phrase
Gabig (2008)	S	Accuracy, intelligibility	15 (13), 6.5:1	6.5 ± 0.7 (5–7.9)	TD	DAS 95 ± 10.6	LUL 5.1 ± 2.1	P	TOLD-P:3	53% of ASD group below average on an articulation task
Gal et al. (2022)	F	Feeding difficulty	105 (87), 4.8:1	3.4 ± 1.3 (3–7.9)	TD	-	-	Q	AEQ	TD > ASD on chewing and swallowing function
Gemsbacher et al. (2008)	NS	Accuracy	115 (92), 1.3:1	7.9 ± 3.7 (2.3–18.9)	TD	-	Various	P/Q	KSPT	TD > ASD on parent reports of NS oromotor skills; parent reports distinguished autistic children with minimally, moderately, and highly fluent speech. Reports of oromotor and manual motor skills were correlated. Home videos corroborated parent reports for 97% of subjects. During direct assessment, minimally and highly fluent autistic children were significantly distinguished on most NS oral motor tasks
Gladfelter and Goffman (2018)	S	Accuracy, consistency	12 (9), 3:1	7.8 ± 1.9 (4.6–11.3)	TD	TONI-4 (96.6 ± 6.54)	EVT 95.8 ± 7.6 (79–112)	P/I	-	ASD = TD on oral mechanism exam. ASD = TD on increase in phonetic accuracy during word learning. Although not more stable at baseline, ASD > TD on gains in stability from pre- to post-test
Grossman et al. (2010)	S	Duration	16	12.3 ± 2.3 (7.5–17.0)	TD	KBIT-2 (109.6 ± 19.1)	-	I	-	ASD > TD on length of first- and last-syllable stress items
Heller Murray et al. (2022)	S	Brain activity	15 (12), 4:1	16.7 ± 2.3 (13.8–21.1)	TD	LIPS-3111.9 ± 26.3	NDW/minute 10.3 ± 4.4 (4.9–20.6)	I	-	ASD = TD on average speech activation and inter-subject variability in speech activation; ASD > TD on intra-subject neural variability. Intra-subject variability correlated with autism severity but not number of words
Hubbard and Trauner (2007)	S	Duration	18 (6), 2:1	14.5 (6–21)	TD	-	-	I	-	Non-autistic children and children with AS had longer syllable durations during sad compared to happy and angry utterances, children with autism did not
Karlsson et al. (2013)	F	Feeding difficulty	57 (38), 2:1	18.7 ± 2.9 (15–25)	TD	WISC/WAIS > 70	-	Q	SWEAA	ASD = TD on motor control for eating
Kasthurirathne et al. (2020)	S	Resonance	11 (9), 4.5:1	15.8 ± 1 (14–17)	TD	-	Verbally fluent	I	-	ASD > TD on nasalance
Kim (2014)	S	Accuracy	2 (2)	8.3 ± 1.5 (7.2–9.3)	-	-	-	P	-	Reduced phoneme accuracy at baseline
Kim and Seung (2015)	S/NS	Accuracy, inventory	1 (1)	11	-	-	EVT-2 AE 3.3	P	GFTA-2, KSPT	Normal NS oromotor skills except reduced range of lip movement. Accurately produced all individual vowels, consonants, syllable types, repetitive syllables, and simple monosyllabic words. Age equivalent of 2 years on the GFTA; atypical errors and inconsistent error patterns within and across sessions
Kissine and Geelhand (2019)	S	Consistency, duration, formants	38 (26), 2.2:1	28.1 ± 11.5	TD	WAIS-4 FSIQ (112.0 ± 25.8)	-	I	-	TD > ASD on F1–F3 dispersion, indicating increased articulatory stability in the ASD group. ASD > TD on syllable duration
Kissine et al. (2021)	S	Consistency, formants	20 (20)	31.6 ± 10.7 (17–52)	TD	WAIS-IV 112.06 ± 22.8	-	I	-	ASD > TD on articulatory stability; ASD group’s non-native vowel production was less accurate and more influenced by native vowels than TD group
Kjelgaard and Tager-Flusberg (2001)	S	Accuracy	89 (80), 8.9:1	7.3 ± 2.4 (4–13.9)	-	DAS (90.1 ± 19.6)	CELF-EL 74.86 ± 17.63; EVT 84.89 ± 17.51	P	GFTA	Average GFTA scores for normal, borderline, and impaired language subgroups, although the impaired group scores were significantly lower than both other groups
Koegel et al. (1998)	S	Accuracy, intelligibility	5 (4), 4:1	5.5 ± 1.4 (3.7–7.5)	-	-	Various	P	AAPS	Intelligibility ratings ranged from “mostly not intelligible” to “sometimes intelligible”
Kothare et al. (2021)	S	Speed	22 (12), 1.2:1	11.4 ± 2.5 (8–18)	-	WISC-FSIQ 102.95 ± 19.79	CELF-5 EL 100.1 ± 20.5 (59–131)	I	-	Jaw speed/acceleration correlated with dominant hand speed
Landa et al. (2013)	S	Inventory	54 (44), 4.4:1	(0.5–3.5)	TD, HR−	MSEL-C 85.3 ± 19.0	-	P	-	TD > early-ASD group (first clinical impression by 14 months) on consonant inventory at 14, 18, and 24 months. TD > late-ASD group on consonant inventory at 14 and 24 months. Early ASD = late ASD on consonant inventory
Lau et al. (2022)	S	Rhythm	57 (48), 5.3:1	16.6 ± 8 (6–35)	TD	WASI/WAIS/WISC-IV 106.15 ± 12.97	-	I	-	TD > ASD on speech rhythm measures
Leader et al. (2020)	F	Feeding difficulty	136 (98), 2.6:1	8.4 ± 4.1	-	-	-	Q	STEP-CHILD	60% of sample reported to have chewing problems
Lundin Remnélius et al. (2022)	F	Feeding difficulty	28 (13), 0.9:1	20.3 ± 4.4 (15–31)	TD, LrnD, ID	92.86 ± 21.22	-	Q	SWEAA	Motor control for eating correlated with internalizing conditions but not autistic traits
Lyakso et al. (2016)	S	Duration, formants	25	(5–14)	TD	-	-	I	-	ASD ≠ TD on F3 values during emotional speech, F2–F1 values for /a/ and /u/, and F3–F2 values for /i/. Autistic children who had developmental reversals at age 1.5–3 years ≠ those at developmental risk from birth on formant measures
Lyakso et al. (2017)	S	Duration, intelligibility, formants	30	(5–14)	TD	MA 4:0–7:0	-	P/I	-	TD > ASD on word intelligibility. Listeners falsely identified male autistic subjects as female more often than TD children and tended to underestimate the age of the autistic children. Autistic children with a developmental reversal at age 1.5–3 years ≠ those at developmental risk from birth on vowel duration
Mahler (2012)	S/NS	Accuracy, rate, consistency	7 (7)	7.4 ± 1.7 (5.1–10.3)	TD	-	EVT 105.3	P	DEAP, GFTA-2, OSMSE-3	TD > ASD on DDK accuracy, consistency. TD = ASD on DDK rate. ASD in normal range on GFTA and OSMSE Structure/Function. 86% failed OSMSE DDK task
Mandelbaum et al. (2006)	S/NS	Accuracy, rate, neuro exam	116 (93), 4:1	8.9 ± 1.6 (7–10)	ID, DLD	Various	-	P	-	Non-ASD/low-IQ > ASD/low-IQ on oromotor tasks. ASD/high-IQ > ASD/low-IQ on timed and untimed oromotor tasks. Per neurological exam, ASD/low-IQ > DLD, ASD/low-IQ > ASD/ high-IQ, and non-ASD/low-IQ > ASD/high-IQ on ratings of oromotor apraxia
Manfredonia et al. (2019)	NS	Facial expression	144 (112), 3.5:1	14.6 ± 7.8 (6–54)	TD	KBIT-2 99.2 ± 19.6	-	I	-	TD > ASD on upturned lip comer during happy emotional expression; activations of lip and jaw correlated with social skills
McCann et al. (2007)	S	Accuracy	31 (25), 4.2:1	9.8 (6–13)	TD	RPM 96.4 ± 15.9	CELF-3 EL 90% below normal limits	P	GFTA-2	Within ASD group, 84% scored within the normal range, 10% had a mild impairment and 6% had a more significant impairment on GFTA
McCleery et al. (2006)	S	Accuracy	14 (12), 6:1	3.3 (2.1–6.9)	TD	BSID MA (1:6)	MCDI NDW 7 (0–26)	P	-	ASD = TD on patterns of phoneme acquisition and production
McDaniel et al. (2018)	NS/F	Accuracy	65 (54), 4.9:1	3.6 ± 0.6 (2.7–4.7)	-	MSELDR 36 ± 15	MCDI NDW 17 ± 25 (0–117)	P	OME	Reduced eating and NS oromotor skills. Imitative and nonimitative oral motor performance was not significantly correlated with receptive–expressive vocabulary discrepancy
McKeever et al. (2022)	S	Accuracy, rate, consistency	8 (6), 3:1	10.1 ± 2.1 (6.3–12.5)	TD	LIPS 89.71 ± 11.37	-	I	-	ASD = TD on max rate of syllable production, accuracy of single syllable sequences, and articulatory stability. ASD group was more likely to have different tongue shapes for fast and slow rates
Nadig and Shaw (2011)	S	Rate	15 (13), 6.5:1	10.8 ± 1.5 (8.4–14.4)	TD	WASI PIQ 81–126 (105 ± 15)	-	P/I	-	ASD = TD on speech rate, perceptually and acoustically
Nakaoka et al. (2022)	F	Feeding difficulty	294 (229), 3.5:1	10 ± 4 (3–18)	-	-	SCQ 12.1 ± 7.3	Q	ASD-MBQ	Reduced score on oral motor function for eating. Oral motor function correlated with social skills and sensory profile
Narzisi et al. (2013)	S	Accuracy	22 (22)	9.8 ± 3.7 (5–16)	TD	WISC-III PIQ 72–141 (103.4 ± 16.3)	NEPSY-II Language scores similar to controls	P	NEPSY-II	TD > ASD on production of articulatory sequences and tongue twisters
Noterdaeme (2002)	S/NS	Accuracy	11 (8), 2.7:1	9.8 ± 2.3	TD, ELD, RLD	KABC 103 ± 14	HTLF IS 38 ± 15	P	-	Per neurological exam, ASD = TD on oral motor function
Ochi et al. (2019)	S	Rate	62 (62)	26.9 ± 7	TD	WAIS-III FSIQ 106.4 ± 14.3	-	I	-	ASD = TD on speech rate and speech rate variance
Pang et al. (2016)	S/NS	Brain activity	21 (17), 4.3:1	11.4 ± 3.2 (6–17.6)	TD	WASI NVIQ 96.4 ± 18.1	OWLSII-OE 88.6 ± 24.4	I	-	During nonspeech task, ASD > TD on magnitude and delayed latency in motor control areas and magnitude in an executive control area. During phoneme production, ASD > TD on latency delays in frontal and temporal language processing areas. During oromotor sequencing, ASD > TD on magnitude and delayed latency in a sensory integration area
Parish-Morris et al. (2018)	S	Consistency, diversity	17 (15), 7.5:1	26.9 ± 7.3	TD	WASI-II FSIQ 102.1 ± 19.8	-	I	-	TD > ASD on mouth movement diversity
Parmeggiani et al. (2019)	F	Sucking reflex	105 (82), 3.6:1	1.1 ± 0.8 (0.3–4)	-	Various	-	R	-	Absent sucking reflex in 16.2% of ASD sample
Pascolo and Cattarinussi (2012)	F	Muscle activity	7 (7)	7.3 ± 1.8	TD	WISC-R > 70	-	I	-	ASD = TD on mylohyoid activity and timing of initiation of mouth opening when bringing food to mouth
Patel et al. (2020)	S	Rate	55 (45), 1.2:1	16.6 ± 6.6 (6.5–35.1)	TD	WISC-4 FSIQ 104.22 ± 12.03	-	P/I	-	TD > ASD on speech rate instrumentally, but not perceptually
Paul et al. (2008)	S	Duration	46 (43), 14.3:1	13.2 ± 4.4 (7.3–28.6)	TD	WISC-3 PIQ 95.0 ± 20.5	CELF-III EL 99.7 ± 21.5	I	T-TRIP	No differences in syllable duration among uncombined ASD groups (HFA, AS, PDD-NOS). For the combined ASD group, ASD > TD on difference between duration of stressed and unstressed syllables
Peter et al. (2019)	S	Accuracy	2 (1), 1:1	6.7 ± 3 (4.6–8.8)	TD	RIAS-NII (84)	“Severely delayed”	P	GFTA-2	Two siblings with dx of ASD and CAS demonstrated reduced articulatory accuracy, vowel errors, inconsistent errors, and reduced DDK rates. Shared genetic variants contributing to ASD and CAS were proposed
Peterson et al. (2016)	F	Feeding difficulty	6 (6)	(4–6)	-	-	-	P	-	Reduced ability to clear mouth of food within 30 s for most of ASD sample
Peterson et al. (2019)	F	Feeding difficulty	6 (6)	(3–5)	-	-	-	P	-	Reduced ability to clear mouth of food within 30 s
Petinou (2021)	S	Accuracy, inventory, intelligibility	1 (1)	4	-	-	-	P	-	Reduced phonetic inventory, percent consonants correct, words correct, and intelligibility
Plumb and Wether by (2013)	S	Accuracy	50 (43), 6.1:1	21.3 ± 1.9 (18–26.9)	TD, DD	MSEL NVDQ 76.0 ± 25.8	-	P	-	TD > ASD on proportion of vocalizations containing at least a vowel. ASD > TD on proportion of atypical and distress vocalizations. Within ASD group, vocalizations containing a vowel were correlated with developmental levels, and communicative vocalizations uniquely predicted age 3 EL
Rainsdon (2018)	S	Accuracy	3 (2), 2:1	4 ± 0.2 (3.8–4.2)	-	-	SPELT-P2 54–69	P	GFTA-3	All participants in the below-average range on the GFTA
Rogers et al. (1996)	NS	Accuracy	17 (15), 7.5:1	2.9 ± 0.3 (2.2–3.4)	LrnD, ID, RLD, GenD	WISC-R FSIQ 89.4 ± 12.1	-	P	-	Controls > ASD on non-meaningful sequential facial imitations. ASD = Controls on non-meaningful single facial imitations and single or sequential meaningful facial imitations
Rogers et al. (2003)	NS	Accuracy	24 (20), 5:1	15.5 ± 3.1 (11–23)	TD, DD, GenD	MSEL NVMA 12–44 (23.7 ± 6.3)	-	P	-	TD > ASD, DD > ASD on oral imitation task. In ASD group, oral-facial imitation moderately correlated with ASD severity and joint attention, but not with EL
Samad et al. (2019)	NS	Magnitude	10 (10)	13.5 ± 2.4	TD	Incl. criteria >70	-	I	-	TD > ASD on activation magnitudes of FAUs and correlations between FAUs, although ASD > TD on activation of mouth frown
Schoen et al. (2011)	S	Accuracy, inventory, vocalization quality	30 (23), 3.3:1	2.4 ± 0.4 (1.5–3.0)	TD	-	VABS EL AE 1.2 ± 0.4	P	-	ASD > TD on number of atypical nonspeech vocalizations. TDA > ASD = TDL on number of early, middle, late, and total consonants. TDA > ASD on English consonant blends. ASD > TDA on number of atypical blends
Sheinkopf et al. (2000)	S	Duration, vocalization quality	15 (13), 6.5:1	3.7 ± 0.7	DD	MPSMT MA 22.13 ± 5.07	RDLS EL AE 1.2 ± 0.2	P	-	ASD = TD on proportion of syllables containing vowel sounds and proportion of syllables containing abnormally long vowels
Shriberg et al. (2001)	S	Accuracy, rate, intelligibility	30 (30)	5.8 ± 1.2 (4–7.9)	-	WISC-3 PIQ (89.0 ± 23.8)	HFA TLC:E (8.2 ± 4.0); AS TLC:E (8.3 ± 4.0)	P	-	HFA > TD and AS > TD on residual distortion errors. HFA > TD on “slow articulation/pause time” and “slow/pause time” ratings. HFA > AS on “slow articulation/pause time” ratings
Shriberg et al. (2011)	S	Accuracy, rate, consistency, duration, resonance, pausing	46 (36), 3.6:1	6 ± 1.2	TD, CAS, SSD	WISC-4 PIQ 67–149 (102.8 ± 16.1)	-	P/I	-	Results suggest a modest increase in risk of Speech Delay, substantial increase in risk of Speech Error, and no elevated risk for CAS in verbal ASD. ~55% of ASD sample had lengthened vowels, ~55% had increased percentage of phoneme distortion, and ~25% had slow speaking rate
Shriberg et al. (2019)	S	Accuracy, duration, formants, pausing	42 (33), 3.7:1	21.2 ± 10.7 (10–49)	GenD, TBI	KBIT-2 (104.3 ± 15.7)	-	P/I	-	83.3% of ASD group classified as normal speech acquisition, 16.7% with Speech Delay or Persistent Speech Delay, 0% with Speech Errors or Persistent Speech Errors. 85.7% of ASD group had no motor speech disorder, 14.3% had speech motor delay, 0% had childhood dysarthria or CAS
Spek et al. (2020)	F	Feeding difficulty	89 (53), 1.5:1	38.5 ± 12	TD	-	-	Q	SWEAA	ASD = TD on motor control for eating; within ASD group men had more motor control problems than women
Stevenson et al. (2017)	NS	Accuracy	13 (9), 2.3:1	8 ± 4.1 (3–18)	TD	-	SCQ communication 6.9 ± 2.4	Q	-	TD > ASD on oral motor skills. Oral motor skills were negatively correlated with autistic traits and positively associated with pragmatic language skills
Sullivan et al. (2013)	S	Accuracy, formants, rhythm	39 (29), 2.9:1	1 ± 0.3 (1.5–2.5)	TD, DD	-	MSEL EL 26.9 ± 9.2 (20–56)	I	-	Some of ASD group ≠ TD on a measure associated with place of articulation. Articulatory features were significantly associated with RL in ASD group
Talkar et al. (2020)	S	Formants, coordination	5 (5)	7.2 ± 0.4	TD	FSIQ 105–135 (124)	-	I	-	ASD > TD in the variance of F2 during syllable sequencing and free speech and in the variance of F3 during sustained vowels. Correlation between F0 and formant values during connected speech perfectly discriminated ASD versus TD
Thurm et al. (2007)	S/NS	Accuracy	83 (71), 5.9:1	(2–5)	DD	Various	Age 5 VABS composite EL AE ratio 0.4 ± 0.3	Q	SICD, VABS-II	TD > ASD on parent reports of child imitating sounds of adults immediately after hearing them. Imitation of adult sounds discriminated groups which did and did not acquire EL at age 5
Tierney et al. (2015)	S/NS	CAS dx	11, 3.3:1	(2–4.6)	-	-	-	P	KSPT	63.6% of ASD group met criteria for CAS
Trembath et al. (2019)	S	Vocalization quality	23 (17), 2.8:1	4.1 ± 0.8 (2.7–5.6)	-	MSEL-DQ 64.1 ± 21.9	VABS EL 20.5 ± 11.8 (3–47); MSEL EL 22.9 ± 11.1 (7–46)	P	-	Change in vocalization ratio over time correlated with EL and nonverbal cognition; vocalizations per minute not correlated with language skills
van Dijk et al. (2021)	F	Feeding difficulty	80 (55), 2.2:1	4 ± 1.1 (0–0)	TD	WPPSI/SON-R/BSID 90.32 ± 18.25	-	Q	MCH-FS	ASD = TD on caregiver reports of chewing problems
Van Santen et al. (2010)	S	Duration	26	6.6 ± 1.3 (4–8)	TD	WPPSI-3 PIQ/PRI (117.63 ± 11.48)	-	I	PEPS-C	ASD = TD on duration of syllables during lexical stress, emphatic stress, or contrastive stress tasks
Vashdi et al. (2020)	S	CAS dx	170		-	-	-	R	-	61% of children with CAS or sCAS also had a dx of ASD
Velleman et al. (2010)	S/NS	Accuracy, rate, consistency, duration, formants, pausing	10 (9), 9:1	5.4 ± 0.9 (4.2–6.3)	TD, sCAS	NVIQ 70–90	PLSA-EC (70.8 ± 15.6)	P/I/Q	VMPAC	Parents of 60% of ASD group reported signs of CD, CAS, or both. 6/10 in ASD group exhibited severe focal oromotor control deficits (one moderate), 4/10 exhibited severe speech motor deficits (four moderate). ASD > TD and ASD > sCAS on multiple formant values. ASD > TD on duration of some vowels. TD > ASD on average variation in speech duration
Vissoker et al. (2019)	F	Feeding difficulty	105 (105)	3.4 ± 4.3 (2–7)	TD	-	-	Q	AEQ	ASD > TD on caregiver reports of chewing and swallowing problems
Whitehouse et al. (2008)	S/NS	Accuracy	34 (33), 33:1	10.8 ± 2.8 (7.2–15.8)	DLD	WASI 80–137 (105.0 ± 14.0)	ERRNI MLUw 89.8 ± 13.4	P	NEPSY	ASD with language impairment > ASD with appropriate language on oromotor sequencing. ASD with language impairment > SLI on oromotor sequencing. Both ASD groups > SLI on sentence repetition. ASD with appropriate NWR skills = ASD without appropriate NWR skills on oromotor sequencing
Wynn et al. (2018)	S	Rate	30 (21), 2.3:1	19.2 ± 10.4 (6–40)	TD	Adults >90	Child group CELF-V EL 88.7	I	-	TD adults entrained speech rate; ASD adults, ASD children, and TD children did not
Wynn et al. (2022)	S	Accuracy	30 (21), 2.3:1	19.2 ± 10.4 (6–40)	TD	Adults >90	Adequate to participate in task	I	-	TD > ASD on articulatory precision
Yan et al. (2021)	S	Accuracy	30 (26), 6.5:1	5.7 ± 1.3	TD	PTONI 59.53 ± 12.32	-	P	-	Reduced phonemes correct at baseline
Yoder et al. (2015)	NS/F	Accuracy, inventory, feeding difficulty	87 (71), 4.4:1	2.9 ± 0.6 (1.7–3.9)	-	MSEL MA 1:0 ± 0:5	MCDI-WS 3.7 ± 5.0 (0–18)	P	OME, CSBS	Imitative and non-imitative oromotor skills did not have added predictive value for EL or RL among initially NV children with ASD
Zarokanellou et al. (2022)	S	Accuracy	46 (35), 3.2:1	9.1 ± 1.5 (7–12)	TD	RCPM 104.8 ± 14.1	EOWPVT-R raw score 63.3 ± 14.2	P	TPPD	TD > ASD on nonword repetition, TD > ASD on speech accuracy

Note: Greater than (>) and less than (<) indicate significant difference in performance.

^aF, feeding; NS, nonspeech; S, speech.

^bCAS, childhood apraxia of speech; DD, developmental delay; DLD, Developmental Language Disorder; ELD, expressive language disorder; GenD, genetic disorder; HR, high-risk siblings; HR−, high-risk siblings without ASD; HR+, high-risk siblings with ASD; ID, intellectual disability; LrnD, learning disability; MSD, motor speech disorder; RLD, receptive language disorder; sCAS, suspected childhood apraxia of speech; SSD, speech sound disorder; TBI, traumatic brain injury; TD, typically developing.

^cAE, age equivalent; BSID, Bayley Scales of Infant and Toddler Development; DAS, Differential Ability Scales; DR, developmental ratio; FSIQ, Full-Scale Intelligence Quotient; KABC, Kaufman Assessment Battery for Children; KBIT, Kaufman Brief Intelligence Test; LIPS, Leiter International Performance Scale; MA, mental age; MPSMT, Merrill-Palmer Scale of Mental Tests; MSEL, Mullen Scales of Early Learning; NV, nonverbal; NVDQ, Nonverbal Developmental Quotient; NVIQ, nonverbal IQ; NVMA, Nonverbal Mental Age; PIQ, Performance IQ; PRI, Perceptual Reasoning Index; PTONI, Primary Test of Nonverbal Intelligence; RCPM, The Raven’s Colored Progressive Matrices; RIAS, Reynolds Intellectual Assessment Scales; RPM, The Raven’s Progressive Matrices; SON-R, Snijders-Oomen nonverbal intelligence tests; TONI, Test of Nonverbal Intelligence; VR, Visual Reception; WAIS, Wechsler Adult Intelligence Scale; WASI, Wechsler Abbreviated Scale of Intelligence; WISC, Wechsler Intelligence Scale for Children; WPPSI, Wechsler Preschool and Primary Scale of Intelligence.

^dADOS, Autism Diagnostic Observation Schedule; AE, Age equivalent; AS, Asperger syndrome; CELF, Clinical Evaluation of Language Fundamentals; DV, Definitional Vocabulary; EC, Expressive Communication; EL, Expressive Language; EOWPVT, Expressive One Word Picture Vocabulary Test; ERRNI, Expression, Reception and Recall of Narrative Instrument; EV, expressive vocabulary; EVT, Expressive Vocabulary Test; HFA, high-functioning autism; HTLF, Heidelberg Test of Language Development; IS, Imitation of Grammatical Structures; LUL, longest utterance length; MCDI, MacArthur-Bates Communicative Development Inventory; MLU, mean length of utterance; MLUw, Mean number of words per utterance; MSEL, Mullen Scales of Early Learning; NDW, Number of different words; NV, Nonverbal; OE, Oral Expression; OWLS, Oral and Written Language Scales; PLS, Preschool Language Scale; RDLS, Reynell Developmental Language Scales; SCQ, Social Communication Questionnaire; SPELT-P, Structured Photographic Expressive Language Test-Preschool; TLC, Test of Language Competence; TOPEL, Test of Preschool Early Literacy; V, Verbal; VABS, Vineland Adaptive Behavior Scales; WS, Words said.

^eI, instrumental; P, Perceptual; Q, questionnaire/report; R, medical record review.

^fSee Table 2 for test names.

^gAS, Asperger syndrome; ASD, autism spectrum disorder; CD, childhood dysarthria; DDK, diadochokinetic; dx, diagnosis; EL, expressive language; FAU, facial action unit; FM, fine motor; GM, gross motor; HFA, high-functioning autism; NWR, nonword repetition; PDD-NOS, Pervasive Developmental Disorder-Not Otherwise Specified; RL, receptive language; SLI, specific language impairment; TDA, typically developing age-matched controls; TDL, typically developing language-matched controls.

Aims of included studies

Of the 107 studies included in this scoping review, some were explicitly designed to investigate motor speech skills (21 studies; 20%) or nonspeech oromotor skills (six studies; 6%), or motor aspects of feeding (one study; 1%). However, the majority of studies included oromotor skills as descriptive statistics or an outcome measure while investigating other areas including general eating disturbances (nine studies; 8%), prosody (13 studies; 12%), general speech sound production skills (13 studies; 12%), intervention outcomes (11 studies; 10%), language skills (six studies, 6%), general motor functioning (five studies, 5%), and a variety of other areas (22 studies; 20%) including imitation abilities, early features of ASD, and intention understanding.

Sample characteristics

ASD diagnosis

A valid, documented method of ASD diagnosis or confirmation of diagnosis was an inclusion criterion for this review to ensure that findings pertain specifically to autistic children (i.e., to accurately characterize participants). The Autism Diagnostic Observation Schedule (ADOS or ADOS-2; Lord et al., 2012) was used to confirm an ASD diagnosis in 58 studies (54%); the Autism Diagnostic Interview, Revised (ADI-R; Rutter, Le Couteur, & Lord, 2003) was used in 15 studies (14%); the Child Autism Rating Scale (CARS; Schopler et al., 1986) was used in 14 studies (13%); the Social Communication Questionnaire (SCQ; Rutter, Bailey, & Lord, 2003) was used in 10 studies (9%); and the Children’s Communication Checklist (CCC; Bishop, 1998) was used in two studies (2%). Assessment tools used in a single study were the Autism Spectrum Rating Scales (ASRS; Goldstein & Naglieri, 2009), the Social Responsiveness Scale (SRS; Constantino & Gruber, 2005), the Gilliam Autism Rating Scale (GARS; Gilliam, 2006), and the Screening Tool for Autism in Toddlers and Young Children (STAT; Stone et al., 2004).

Fifty-seven studies (53%) made specific reference to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV or −5; American Psychiatric Association, 2013) criteria for diagnosis of ASD, and eight studies (7%) referred to the World Health Organization’s International Classification of Diseases (ICD; World Health Organization, 2004). The remaining 42 studies (39%) reported only the tool and not the classification system used for diagnosis or confirmation of diagnosis.

Sample size

The number of autistic subjects included in each study ranged from one (Biller et al., 2022; Biller & Johnson, 2020; Kim & Seung, 2015; Petinou, 2021) to 294 (Nakaoka et al., 2022), with nine additional papers including over one hundred autistic subjects (Demartini et al., 2021; Gal et al., 2022; Gernsbacher et al., 2008; Leader et al., 2020; Mandelbaum et al., 2006; Manfredonia et al., 2019; Parmeggiani et al., 2019; Vashdi, 2014; Vissoker et al., 2019). Eighteen studies (17%) included a sample of fewer than 10 autistic subjects, 27 studies (25%) had a sample size of 10–20 autistic subjects, 24 studies (22%) had a sample size of 21–30 autistic subjects, 13 studies (12%) had a sample size of 31–50, and 25 studies (23%) had a sample size of over 50 autistic subjects.

Age range

Means, standard deviations, and ranges of autistic subjects’ ages were extracted from each paper. Among the studies in which mean age was reported or could be calculated (n = 96), nine studies (9%) had a mean subject age of 0–2.9 years, 23 studies (24%) had a mean subject age of 3.0–5.9 years, 25 studies (26%) had a mean subject age of 6.0–8.9 years, 15 studies (16%) had a mean subject age of 9.0–11.9 years, 11 studies (11%) had a mean subject age of 12.0–17.9 years, and 13 studies (14%) had a mean subject age of 18 years or older. Figure 2 provides a visual representation of the mean subject age and, when possible, range in years.

FIGURE 2.

Mean and range of subject ages, ordered by mean (n = 96).

Male-to-female ratio

The male-to-female ratio of children meeting criteria for ASD is often cited to be approximately 4:1, although a systematic review and meta-analysis of prevalence studies found that the ratio may be closer to 3:1 (Loomes et al., 2017). Many studies included in the current scoping review included both male and female subjects and reported the number of each (n = 79), allowing the calculation and comparison of male-to-female ratios.

The smallest male-to-female ratio was 0.9:1 (Lundin Remnélius et al., 2022), the only study with more female than male autistic subjects. The male-to-female ratios of the remaining papers ranged from 1:1 (i.e., an equal number of male and female subjects) to 33:1 (i.e., one female subject for every 33 male subjects). The median ratio was 4:1, the mode was 4:1, and the mean male-to-female ratio was 4.8:1. Of the 79 studies included in this male-to-female ratio analysis, a total of 35 studies (44%) had a male-to-female ratio between 2:1 and 4:1.

Nonverbal IQ/mental age

Seventy-six studies (71%) reported a mean, range, or inclusion cutoff value of nonverbal IQ or mental age for their sample of autistic individuals. There was significant heterogeneity between and within studies, and the assessment tools used to measure IQ or mental age were also varied. Results from a total of 18 different standardized assessment tools were reported, as were various composite scores, subtest scores, and age equivalents from these tests. See Table 3 for IQ data from each included study.

Expressive language abilities

Sixty-one studies (57%) reported data related to the expressive language skills of their sample of autistic individuals, excluding composite language scores measuring both expressive and receptive skills. Overall, 19 different standardized assessment tools were used to quantify expressive language in terms of normative scores or age equivalents. In addition, other methods derived from natural language samples were used (e.g., number of different words, mean length of utterance, and presence/absence of productive syntax), particularly in studies involving minimally verbal autistic children. As with nonverbal IQ, expressive language scores were heterogeneous both within and across studies, and scores are reported in detail in Table 3.

Control samples

A total of 72 studies (67%) included a control group. Most control groups were comprised of individuals considered “typically developing,” “neurotypical,” or simply “non-ASD” (64 studies; 60% of all studies; 89% of studies with a control group). Other control samples included individuals with a developmental delay (six studies; 6% of all studies), suspected or diagnosed CAS (four studies; 4%), borderline/low IQ or intellectual disability (three studies; 3%), genetic disorders (three studies; 3%), an elevated likelihood of receiving an ASD diagnosis due to having an older sibling with ASD (three studies; 3%), learning disabilities (two studies; 2%), and developmental language disorder/specific language impairment (two studies; 2%). Control populations used in a single study were individuals with a motor speech disorder, a speech sound disorder, an expressive language disorder, a receptive language disorder, or a traumatic brain injury.

Research question 1: What methods have been used to investigate oromotor functioning among individuals with ASD?

Methodologies of included studies

Perceptual studies

Sixty-six studies (62%) used perceptual (i.e., auditory perceptual and/or visual perceptual) methods of assessment. Of those 66 perceptual studies, 56 (85%) reported a deficit among autistic individuals on some measure of oromotor functioning. The remaining 10 perceptual studies (15%) did not report a significant oromotor impairment among their ASD sample (Dalton et al., 2017; Demopoulos & Lewine, 2016; Espanola Aguirre & Gutierrez, 2019; McCann et al., 2007; McCleery et al., 2006; Nadig & Shaw, 2011; Noterdaeme, 2002; Sheinkopf et al., 2000; Shriberg et al., 2019; Yoder et al., 2015).

Interviews/questionnaires

Interviews or questionnaires of autistic individuals or their caregivers were used in 14 studies (13%; Ashley et al., 2020; Demartini et al., 2021; Gal et al., 2022; Gernsbacher et al., 2008; Leader et al., 2020; Lundin Remnélius et al., 2022; Nakaoka et al., 2022; Shriberg et al., 2010; Spek et al., 2020; Stevenson et al., 2017; Thurm et al., 2007; van Dijk et al., 2021; Velleman et al., 2010; Vissoker et al., 2019). Nine of these studies (64%) reported a deficit in some measure of oromotor functioning among their autistic sample. Parent questionnaires used in these studies were the Autism Eating Questionnaire (AEQ; Vissoker et al., 2019), the Autism Spectrum Disorder Mealtime Behavior Questionnaire (ASD-MBQ; Nakaoka et al., 2020); the Behavioral Pediatrics Feeding Assessment Scale (BPFAS; Crist & Napier-Phillips, 2001), the Montreal Children’s Hospital Feeding Scale (MCH-FS; Ramsay et al., 2011), the Screening Tool of Feeding Problems for Children (STEP-CHILD; Seiverling et al., 2011), the Sequenced Inventory of Communication Development (SICD; Hedrick et al., 1975), the SWedish Eating Assessment for Autism spectrum disorders (SWEAA; Karlsson et al., 2013), and the item “imitates sounds” from the Vineland Adaptive Behavior Scales (VABS; Sparrow et al., 1984). Three studies used non-standardized questionnaires or caregiver reports: Gernsbacher et al. (2008) and Stevenson et al. (2017) employed a landmark-based parent interview procedure designed to aid recollection of children’s motor skills at specific ages, and Velleman et al. (2010) interviewed parents directly regarding variables associated with signs of dysarthria and apraxia of speech in their children’s speech.

Medical record review

Two studies (2%) conducted retrospective reviews of medical records to identify oromotor deficits among samples of autistic individuals. These studies identified an increased prevalence of an absent sucking reflex in infancy (Parmeggiani et al., 2019) and a high comorbidity of ASD and the motor speech disorder childhood apraxia of speech (CAS; Vashdi et al., 2020).

Instrumental studies

Thirty-six studies (34%) utilized instrumental methods of assessment. Of those 36 studies, 29 (81%) reported a deficit among autistic individuals on some measure of oromotor functioning. Twenty-four studies (67% of instrumental studies) used acoustic analyses, five studies (14%) used facial motion tracking (Gladfelter & Goffman, 2018; Kothare et al., 2021; Manfredonia et al., 2019; Parish-Morris et al., 2018; Samad et al., 2019), two studies (5%) used electromyography to measure muscle activation (Cattaneo et al., 2007; Pascolo & Cattarinussi, 2012), and two studies (5%) used magnetic resonance imaging of the brain (Chenausky, Kernbach, et al., 2017; Heller Murray et al., 2022). Each of the following methods was used in only one study: nasometry (Kasthurirathne et al., 2020), magnetoencephalography (Pang et al., 2016), and ultrasound tongue imaging (McKeever et al., 2022).

Assessment tools used

Less than half of the included studies (n = 51, 48%) used a standardized assessment tool to examine oromotor function. Of these 51 studies, 33 (65%) used a norm-referenced assessment tool. The remaining studies relied on a test under development, a test that had not been norm-referenced, or a test without readily available psychometric data. Table 1 lists the assessment tools used in the included studies, the overall purpose of the test as stated by the test developers, and the number of included studies in which each test was used.

TABLE 1.

Tests used to assess oromotor function in included studies.

Abbreviation	Test name	Test citation	Stated purpose	# Studies
AAPS	Arizona Articulation Proficiency Scale	Fudala and Reynolds (1986)	Articulation/phonology	1
AEQ	Autism Eating Questionnaire	Gal et al. (2022)	Feeding skills	2
ASD-MBQ	Mealtime Behavior Questionnaire for Children with ASD	Nakaoka et al. (2020)	Feeding skills	1
BPFAS	Behavioral Pediatrics Feeding Assessment Scale	Crist and Napier-Phillips (2001)	Feeding skills	1
CDOMA	Com DEALL Oro Motor Assessment	Archana (2008)	Oral motor function	1
CSBS	Communication and Symbolic Behavior Scales	Wetherby and Prizant (2002)	Communication development	1
CTOPP	Comprehensive Test of Phonological Processing	Wagner et al. (2013)	Phonological processing	1
DEAP	Diagnostic Evaluation of Articulation and Phonology	Dodd et al. (2002)	Articulation/phonology	2
GFTA	Goldman Fristoe Test of Articulation	Goldman and Fristoe (2015)	Articulation/phonology	10
KSPT	Kaufman Speech Praxis Test for Children	Kaufman (1995)	Identification of CAS	9
MCH-FS	The Montreal Children’s Hospital Feeding Scale	Ramsay et al. (2011)	Feeding skills	1
MVIA	Motor and Vocal Imitation Assessment	Espanola Aguirre and Gutierrez (2019)	Motor and vocal imitation	1
NEPSY	A Developmental Neuropsychological Assessment	Korkman et al. (2007)	Neuropsychological development	2
OSMSE	Oral Speech Mechanism Screening Examination	St. Louis and Ruscello (2000)	Oral structure and function	2
PEPS-C	Profiling Elements of Prosody in Speech–Communication	Peppé and McCann (2003)	Prosody	2
POME/OME	Pre-School Oral Motor Examination	Sheppard (1990)	Oral motor function	3
POP	Polysyllable Preschool Test	Baker (2013)	Articulation/phonology	2
SICD	Sequenced Inventory of Communication Development	Hedrick et al. (1975)	Communication development	1
SIPT	Sensory Integration and Praxis Tests	Ayres (1996)	Sensory integration	1
STEP-CHILD	Screening Tool of Feeding Problems for Children	Seiverling et al. (2011)	Feeding skills	1
SWEAA	SWedish Eating Assessment for Autism Spectrum Disorders	Karlsson et al. (2013)	Feeding skills	4
SWPT	Single Word Polysyllable Test	Gozzard et al. (2006)	Articulation/phonology	1
TOLD-P	Test of Language Development-Primary	Newcomer and Hammill (1997)	Oral language proficiency	1
TPPD	Greek Test of Phonetic and Phonological Development	Levanti et al. (1995)	Articulation/phonology	1
T-TRIP	Tennessee Test of Rhythm and Intonation Patterns	Koike and Asp (1981)	Prosody	1
VABS	Vineland-II Adaptive Behavior Scales	Sparrow et al. (1984)	Aiding diagnosis of ID/DD	1
VMPAC	Verbal Motor Production Assessment for Children	Hayden and Square-Storer (1999)	Oral motor function	5

Abbreviations: CAS, childhood apraxia of speech; DD, developmental disability; ID, intellectual disability.

Of the assessment tools used, seven were designed to assess articulation and/or phonology or phonological processing (AAPS, CTOPP, DEAP, GFTA, POP, SWPT, and TPPD), six to assess feeding skills (AEQ, ASD-MBQ, BPFAS, MGH-FS, STEP-CHILD, and SWEAA), four to assess the oral speech mechanism (CDOMA, OSMSE, POME, and VMPAC), two to assess prosody (PEPS-C and T-TRIP), and one to aid in the diagnosis of CAS (KSPT). The remaining tests are assessments of expressive and receptive communicative abilities or general development that include subtests useful for assessing oromotor skills or were used to elicit specific types of speech samples for instrumental analysis. For example, the NEPSY is a broad assessment of neuropsychological development that includes the Oromotor Sequences subtest in which a child repeats articulatory sequences and tongue twisters. The specific editions of these tests used in each study are included in Table 3.

Question 2: What oromotor behaviors have been investigated?

As indicated in Figure 3, of the 107 included studies, 83 studies (78%) address the domain of speech production, 28 (26%) address nonspeech oral movements, and 19 (18%) address motor-related feeding issues. Nineteen studies (18%) addressed more than one of these domains (e.g., speech and nonspeech oral movements). Specific information regarding the skills examined in these studies is provided below, organized by methodology (i.e., perceptual, instrumental, parent interview/questionnaire, and medical record review) in descending order of frequency.

FIGURE 3.

Percentage of studies investigating the domains of speech, nonspeech, and feeding.

Perceptual studies

The most common oromotor skill assessed perceptually was speech accuracy (i.e., whether or not a speech stimulus was produced correctly) (52 of 66 perceptual speech studies; 79%). Among studies examining speech accuracy, 31 studies (60%) examined the accuracy of phonemes (i.e., vowels and/or consonants) either in isolation or within productions of words or nonwords, 19 studies (37%) examined the accuracy of syllables, four studies (8%) examined accuracy during the production of phrases or sentences, and seven studies (13%) examined speech accuracy during spontaneous or connected speech (e.g., a picture description task).

Twenty-two perceptual studies (33%) examined the ability to accurately perform nonspeech oral movements (Adams, 1998; Akin-Bulbul & Ozdemir, 2022; Amato & Slavin, 1998; Belmonte et al., 2013; Biller & Johnson, 2019, 2020; Bodison, 2015; Chenausky et al., 2019; Chenausky & Tager-Flusberg, 2017; Dalton et al., 2017; Deshmukh, 2012; Espanola Aguirre & Gutierrez, 2019; Gernsbacher et al., 2008; Kim & Seung, 2015; Mandelbaum et al., 2006; McDaniel et al., 2018; Noterdaeme, 2002; Rogers et al., 2003; Rogers & Pennington, 2021; Tierney et al., 2015; Velleman et al., 2010; Yoder et al., 2015).

Fourteen (21%) measured phonetic inventory either in imitation or from a spontaneous speech sample (Biller et al., 2022; Broome et al., 2021; Chenausky et al., 2016, 2018, 2021; Chenausky, Nelson, & Tager-Flusberg, 2017; Chenausky, Norton, et al., 2022; Chenausky, Norton, & Schlaug, 2017; Chenausky & Schlaug, 2018; Kim & Seung, 2015; Landa et al., 2013; Petinou, 2021; Schoen et al., 2011; Yoder et al., 2015), 11 (17%) examined speech rate or diadochokinetic rate (DDK; rapidly produced sequences of syllables) (Chenausky et al., 2019, 2020, 2021; Deshmukh, 2012; Mahler, 2012; Mandelbaum et al., 2006; Nadig & Shaw, 2011; Patel et al., 2020; Shriberg et al., 2001, 2011; Velleman et al., 2010), seven (11%) examined the consistency or stability of speech production (Chenausky et al., 2019, 2020, 2021; Deshmukh, 2012; Gladfelter & Goffman, 2018; Mahler, 2012; Shriberg et al., 2011), six (9%) examined motor-related feeding/eating behaviors (Amato & Slavin, 1998; Brisson et al., 2012; McDaniel et al., 2018; Peterson et al., 2016, 2019; Yoder et al., 2015), five (8%) examined speech intelligibility (Gabig, 2008; Koegel et al., 1998; Lyakso et al., 2017; Petinou, 2021; Shriberg et al., 2001), five (8%) examined vocalization quality (Chenausky, Nelson, & Tager-Flusberg, 2017; Plumb & Wetherby, 2013; Schoen et al., 2011; Sheinkopf et al., 2000; Trembath et al., 2019), four (6%) examined resonance quality (Chenausky et al., 2019, 2020, 2021; Shriberg et al., 2011), and three (5%) examined coarticulation (Chenausky et al., 2016, 2020, 2021). Speech sound duration (Sheinkopf et al., 2000) and motor anticipation (Brisson et al., 2012) were examined perceptually in one study each.

Instrumental studies

Acoustic analysis

Of the 24 studies that used acoustic analysis, 10 studies (42%) used these methods to measure the duration of vowels, syllables, or words (Diehl & Paul, 2013; Grossman et al., 2010; Hubbard & Trauner, 2007; Kissine & Geelhand, 2019; Lyakso et al., 2016, 2017; Paul et al., 2008; Shriberg et al., 2019; Van Santen et al., 2010; Velleman et al., 2010), nine studies (38%) examined vowel formants (i.e., concentrations of acoustic energy at particular frequencies) (Chenausky et al., 2021; Kissine et al., 2021; Kissine & Geelhand, 2019; Lyakso et al., 2016, 2017; Shriberg et al., 2019; Sullivan et al., 2013; Talkar et al., 2020; Velleman et al., 2010), seven studies (29%) examined speech rate (Ehlen et al., 2020; Nadig & Shaw, 2011; Ochi et al., 2019; Patel et al., 2020; Shriberg et al., 2011; Velleman et al., 2010; Wynn et al., 2018), and five studies (21%) examined the consistency or stability of speech production (Chenausky et al., 2021; Kissine et al., 2021; Kissine & Geelhand, 2019; Shriberg et al., 2011, 2019). Each of the following speech features was examined in a single study: temporal synchrony with a metronome (Franich et al., 2020), speech rhythm (Lau et al., 2022), voice onset time (i.e., the duration of time between the release of a plosive and the onset of voicing) (Chenausky & Tager-Flusberg, 2017), coarticulation (Chenausky et al., 2021), and phoneme accuracy (Chenausky et al., 2021). Two studies employed novel acoustic analysis techniques: Talkar et al. (2020) examined correlations among speech acoustics, videos of facial movements, and handwriting data to investigate correlations across systems, and Sullivan et al. (2013) used spectral analysis to examine three different timescales of acoustic signals and derive information regarding syllabic rhythm, formant transitions, and place of articulation.

Facial motion tracking

Five studies (14% of instrumental studies) used facial motion tracking to examine oromotor performance. Gladfelter and Goffman (2018) used a 3D optical camera to capture signals from infrared light-emitting diodes affixed to children’s faces to quantify speech motor stability during word learning. Markerless facial motion tracking was used in the remaining studies to quantify lip aperture, mouth surface area, jaw velocity, and jaw acceleration (Kothare et al., 2021); the use of the lips and jaw in posed facial expressions of emotions (Manfredonia et al., 2019); movements of the lips and jaw in spontaneous facial actions (Samad et al., 2019); and mouth movement diversity (Parish-Morris et al., 2018).

Electromyography (EMG)

Two studies examined electromyographic activity of the mylohyoid muscle during goal-oriented actions (Cattaneo et al., 2007; Pascolo & Cattarinussi, 2012).

Magnetic resonance imaging (MRI)

Two studies used MRI to examine the relationship between white matter tracts and speech improvement during intervention (Chenausky, Kernbach, et al., 2017) and intra-subject variability in neural activity during speech production (Heller Murray et al., 2022) among autistic individuals.

Magnetoencephalography (MEG)

Pang et al. (2016) used magnetoencephalography to examine differences in brain activity during the performance of increasingly complex oromotor tasks underlying speech production (i.e., a simple oromotor task, a phoneme production task, and a phonemic sequencing task).

Nasometry

Kasthurirathne et al. (2020) used a nasometer to quantify nasalance (i.e., the ratio of nasal resonance and oral resonance during speech) in a group of autistic teenagers.

Ultrasound

McKeever et al. (2022) used ultrasound imaging to examine the rate, accuracy, and consistency of tongue movements during DDK repetitions.

Questionnaire/parent report

Fourteen studies (13%) used questionnaires or parent report to investigate oromotor skills. The majority of these studies (10 studies; 71% of questionnaire/report studies) investigated motor-related feeding behaviors (Ashley et al., 2020; Demartini et al., 2021; Gal et al., 2022; Karlsson et al., 2013; Leader et al., 2020; Lundin Remnélius et al., 2022; Nakaoka et al., 2022; Spek et al., 2020; van Dijk et al., 2021; Vissoker et al., 2019). Four studies (29%) used questionnaires or parent report to address nonspeech skills (Gernsbacher et al., 2008; Stevenson et al., 2017; Thurm et al., 2007; Velleman et al., 2010). Both Gernsbacher et al. (2008) and Stevenson et al. (2017) prompted parents with retrospective questions regarding activities such as puffing cheeks on request at age 2 years. Thurm et al. (2007) prompted parents with one item from the SICD regarding a child’s ability to imitate a raspberry/tongue click, and Velleman et al. (2010) surveyed parents on a variety of oromotor characteristics including muscular weakness of the speech mechanism and problems with the tongue and velum. A questionnaire or parent report was used to investigate speech production in a single study: Velleman et al. (2010) surveyed parents of autistic children with questions regarding speech distortions, abnormal oral motor postures during speech, and the production of unclear consonants.

Medical record review

Two studies (2%) used medical record review to investigate oromotor abnormalities. Parmeggiani et al. (2019) examined the medical records of autistic children to investigate putative early features of ASD including the absence of a sucking reflex in infancy. Vashdi et al. (2020) examined the medical records of children diagnosed with CAS or suspected CAS for the co-occurrence of an ASD diagnosis.

Question 3: What do these studies tell us about oromotor skills in ASD?

A variety of results were reported in the included studies, which are summarized below as well as in Tables 2 and 3. Table 2 displays (1) the most frequently examined behaviors in the included studies, (2) the number of included studies examining that behavior, (3) the number and percentage of studies reporting an abnormality for the behavior among autistic subjects, and (4) and the nature of the abnormality. Table 3 provides detailed information about each study including relevant results.

TABLE 2.

Summary of most frequently examined behaviors in included studies.

Behavior	# Studies	# Abnormal	Type of abnormality
Speech
Accuracy	52	44 (85%)	Reduced accuracy
Consistency/stability	13	11 (85%)	Reduced consistency; increased consistency
Rate	13	7 (54%)	Slower rate; less entrainment
Duration	11	9 (82%)	Longer duration; less difference in duration between speaking conditions
Formant values	8	6 (75%)	Outcome measures too variable to discern a dominant pattern
Coordination/coarticulation	6	6 (100%)	Reduced coordination; difficulty with coarticulation
Vocalization quality	5	3 (60%)	Reduced speechlike vocalizations; increased atypical vocalizations
Intelligibility	4	3 (75%)	Reduced intelligibility
Nonspeech
Nonspeech oral movements	28	20 (71%)	Reduced nonspeech oromotor control; differences in brain activity
Feeding
Eating/feeding skills	19	11 (58%)	Reduced oral control for feeding

Speech

Accuracy

The most commonly investigated aspect of speech motor functioning among the included studies was the accuracy of oral movements during speech, which was investigated in 52 of the 107 studies (49%). Accuracy was quantified using non-standardized perceptual measures (e.g., percent phonemes correct, judgments of distortion), standardized tests like the GFTA and KSPT, and acoustic techniques. Forty-four of these studies (85%) reported abnormal accuracy based on normative data or in comparison to a control group. In all studies reporting an abnormality related to speech accuracy, accuracy was decreased in the ASD group.

Phonetic inventory

Fifteen studies (14%) reported the phonetic inventories of autistic children, all using auditory-perceptual methods. Three of these studies included comparisons to a control group, each of which reported significant differences related to autistic speakers. Two of these studies reported a smaller phonetic inventory in an ASD group (Landa & Garrett-Mayer, 2006; Schoen et al., 2011); Chenausky, Nelson, and Tager-Flusberg (2017) found no significant difference in phonetic inventory between toddlers who were later diagnosed with ASD and those who were not, although observed a significant Age * Group interaction in which toddlers with an elevated likelihood of an ASD diagnosis who did not go on to receive a diagnosis gained a smaller number of phonemes between 18 and 24 months than non-autistic controls and toddlers who were later diagnosed with ASD. The remaining 12 studies (Biller et al., 2022; Broome et al., 2021; Chenausky et al., 2016, 2018, 2021; Chenausky, Kernbach, et al., 2017; Chenausky, Norton, et al., 2022; Chenausky, Norton, & Schlaug, 2017; Chenausky & Schlaug, 2018; Kim & Seung, 2015; Petinou, 2021; Yoder et al., 2015) used phonetic inventory as a baseline measure or outcome measure without direct comparison to a control group or normative values.

Consistency/stability

Thirteen studies (12%) examined the consistency or stability of oromotor movements (six perceptual and five instrumental studies). Eleven of these 13 studies (85%) found significant abnormalities related to consistency or stability in a sample of autistic individuals. Interestingly, of these 11 studies, six reported reduced consistency/stability (Chenausky et al., 2019, 2020, 2021; Mahler, 2012; Shriberg et al., 2011; Talkar et al., 2020) and five reported increased consistency/stability (Gladfelter & Goffman, 2018; Kissine et al., 2021; Kissine & Geelhand, 2019; Parish-Morris et al., 2018; Velleman et al., 2010) in an ASD group. Findings of reduced consistency/stability include (1) higher ratings of “inconsistent errors,” “variable errors,” or “less stable whole word errors” (Chenausky et al., 2019, 2020, 2021; Shriberg et al., 2011); (2) reduced consistency of phoneme accuracy during DDK repetitions (Mahler, 2012); and (3) increased variability in formant values during DDK repetitions, free speech, and sustained vowels (Talkar et al., 2020). Findings of higher consistency/stability include (1) greater improvement in motor stability during a nonword learning task compared to controls (Gladfelter & Goffman, 2018); (2) greater articulatory stability in the production of native vowels compared to non-autistic controls (Kissine et al., 2021; Kissine & Geelhand, 2019); (3) less average variation in speech durations (Velleman et al., 2010); and (4) a restricted repertoire of mouth movements during conversational speech (Parish-Morris et al., 2018).

Rate

Of the 13 studies (12%) that examined speech or DDK rate (i.e., the number of syllables or words per unit of time) among autistic individuals (10 perceptual and six acoustic studies, including two studies using both approaches), eight (62%) reported a significant abnormality in a group of autistic individuals and five (38%) did not report significant group differences. Among the eight studies that reported a rate abnormality in the ASD group, six reported a slower rate among autistic speakers (Chenausky et al., 2019, 2020, 2021; Patel et al., 2020; Shriberg et al., 2001, 2011), one reported general abnormality of rate without indicating a direction (Mandelbaum et al., 2006), and one reported that autistic adults do not entrain their speaking rate (i.e., modify rate to match a communication partner’s rate) while non-autistic adults do (Wynn et al., 2018).

Duration

Of the 11 studies (10%) that examined the duration of phonemes or syllables among autistic individuals (10 instrumental studies and one perceptual study), nine (82%) reported a significant difference between autistic and non-autistic individuals and two (18%) reported similar rates between autistic individuals and controls. Of the eight studies that reported a significant abnormality, four reported longer sound durations in the ASD group (Diehl & Paul, 2013; Grossman et al., 2010; Kissine & Geelhand, 2019; Shriberg et al., 2011; Velleman et al., 2010), one reported a group difference without indicating the direction of the finding (Lyakso et al., 2017), and three studies reported significantly less difference in duration between speaking conditions by autistic speakers (i.e., between emotional states [Hubbard & Trauner, 2007] and between stressed and unstressed syllables [Lyakso et al., 2016; Paul et al., 2008]).

Vowel production/formant values

Eight studies (7%) used acoustic methods to examine the formant values of phonemes (i.e., resonant frequencies corresponding to vocal tract configurations) produced by autistic speakers. Among these studies, six (75%) reported significant group differences between autistic and control groups (Kissine et al., 2021; Kissine & Geelhand, 2019; Lyakso et al., 2016, 2017; Talkar et al., 2020; Velleman et al., 2010), one study examined two autistic children intended to represent subgroups of ASD but without the use of a control group (Chenausky et al., 2021), and one study found no significant difference between groups (Sullivan et al., 2013). Abnormal findings among these studies include (1) reduced F1–F3 dispersion among autistic speakers, suggesting more invariant articulatory gestures during vowel production (Kissine et al., 2021; Kissine & Geelhand, 2019); (2) an increased attraction index when producing non-native vowels, suggesting that autistic speakers produce vowels with formant values closer to those of one of their native vowels than to target non-native vowels compared to controls (Kissine et al., 2021); (3) larger formant triangles, suggesting a larger magnitude of tongue movements during vowel production in autistic speakers (Lyakso et al., 2016, 2017; although note that statistical significance was not reported); (4) increased variance of formants during various speech tasks (Talkar et al., 2020); and (5) higher formant values among autistic speakers during prolonged vowels (Velleman et al., 2010) and during emotional speech (Lyakso et al., 2016). Chenausky et al. (2021) examined the speech of two minimally verbal autistic children hypothesized to belong to different subphenotypes (i.e., motor speech disorder and motor speech disorder + auditory processing disorder). Using formant analyses, they found a similar level of phoneme distortion between the two children, but that the child with only a suspected motor speech disorder had a larger searching articulation index, larger token-to-token variability, and a larger vowel space.

Coordination/coarticulation

Six studies (6%) addressed oromotor coordination among autistic individuals (four perceptual and one instrumental study). Each found significant differences indicating decreased oromotor coordination among autistic individuals. Findings include (1) “difficulty with coarticulation” or “difficulty with initial configurations or transitions” in autistic children or subgroups of autistic children (Chenausky et al., 2019, 2020, 2021), (2) lower performance on a test of oromotor sequences for high-functioning autistic children compared to non-autistic controls (Narzisi et al., 2013), (3) decreased temporal coordination in autistic individuals during a speech task using a metronome (Franich et al., 2020), and (4) a lower complexity in the correlation of formants and facial movements during a reading task, suggesting a decreased independence of movements compared to control subjects (Talkar et al., 2020).

Vocalization quality

Five studies (5%) examined vocalization quality among young autistic children (i.e., whether vocalizations were considered speechlike, nonspeechlike, or otherwise abnormal). Three studies reported significant differences between autistic children and comparison groups including (1) significantly fewer speechlike vocalizations among younger siblings of autistic children who later received an ASD diagnosis compared to younger siblings who did not receive an ASD diagnosis and typically developing controls (Chenausky, Nelson, & Tager-Flusberg, 2017), (2) fewer vocalizations containing a vowel and more vocalizations not containing a vowel in autistic children than typically developing peers (Plumb & Wetherby, 2013), and (3) significantly more atypical vocalizations among autistic children, with high-pitched squeals accounting for much of this difference (Schoen et al., 2011).

Intelligibility

Four studies (4%) examined the intelligibility (i.e., the amount of speech that is understood by a listener) of autistic individuals, all using perceptual methods. Three of these studies (75%) reported low intelligibility among autistic speakers (Koegel et al., 1998; Lyakso et al., 2017; Petinou, 2021), although it should be noted that Petinou (2021) was a case study of one child and Lyakso et al. (2017) did not report the statistical significance of the reduced intelligibility in autistic children compared to their control group.

Nonspeech oromotor skills

Twenty-eight studies (26% of all studies in this review) investigated the nonspeech oromotor skills of autistic individuals, and 25 studies reported whether the skills were normal or abnormal. The remaining three studies provided information regarding performance among subgroups of autistic individuals or regarding correlations between nonspeech oral motor skills and language outcomes rather than providing a comparison with norms or a control group. Twenty studies (80% of those reporting an outcome) reported abnormalities among a sample of autistic individuals and five studies (Dalton et al., 2017; Deshmukh, 2012; Espanola Aguirre & Gutierrez, 2019; Mahler, 2012; Noterdaeme, 2002) reported no significant difference in nonspeech oral motor skills between and ASD group and a control group.

Among the studies reporting a significant difference in nonspeech oromotor skills, a commonly reported finding was that autistic individuals scored in the impaired range or significantly lower than a control group on a test of oromotor imitation such as the KSPT (Adams, 1998; Chenausky et al., 2019; Chenausky, Norton, & Schlaug, 2017; Gernsbacher et al., 2008; Kim & Seung, 2015; Tierney et al., 2015), VMPAC (Biller & Johnson, 2019, 2020; Dalton et al., 2017; Velleman et al., 2010), POME (Amato & Slavin, 1998; McDaniel et al., 2018; Yoder et al., 2015), Com DEALL Oro Motor Assessment (Belmonte et al., 2013), SIPT (Bodison, 2015), or OSMSE (Deshmukh, 2012). These tests assess an individual’s ability to perform nonspeech oral movements such as opening the jaw; elevating and lateralizing the tongue; spreading and puckering lips; performing sequences of oromotor tasks; and performing functional behaviors like sucking and biting.

Other findings include differences in movements of the lips during emotional expression (Manfredonia et al., 2019; Samad et al., 2019) and differences in brain activity during nonspeech oromotor tasks (Pang et al., 2016) including increased magnitude and delayed latency in motor control areas as well as increased magnitude in an executive control area. Of note, Belmonte et al. (2013) found that nonspeech oromotor skills varied independently of gross and fine motor skills among autistic children, offering additional motivation for studying these skills in addition to speech production abilities.

Feeding skills

Nineteen studies (18%) examined feeding-related oromotor skills among autistic individuals—10 using caregiver questionnaires, six using perceptual methods, two using instrumental methods, and one using medical record review. Of these studies, 17 reported whether motor feeding skills were normal or abnormal, and 11 of those 17 (65%) reported feeding-related oromotor deficits (Amato & Slavin, 1998; Brisson et al., 2012; Cattaneo et al., 2007; Demartini et al., 2021; Gal et al., 2022; Leader et al., 2020; McDaniel et al., 2018; Nakaoka et al., 2020; Parmeggiani et al., 2019; Peterson et al., 2016; Vissoker et al., 2019).

Abnormal findings include decreased scores on motor-related portions of the OME (“eating behaviors”; Amato & Slavin, 1998; McDaniel et al., 2018), SWEAA (“motor control”; Demartini et al., 2021), AEQ (“chewing and swallowing problems”; Gal et al., 2022; Vissoker et al., 2019), STEP-CHILD (“chewing problems”; Leader et al., 2020), and ASD-MBQ (“oral motor function”; Nakaoka et al., 2022). Brisson et al. (2012) reported significantly less anticipatory mouth opening in response to an approaching spoon among autistic infants compared to typically developing infants, a finding which may be considered alongside Cattaneo et al. (2007), who reported significantly less mylohyoid muscle activity among autistic children both when observing someone reach for, grasp, and eat food and also when performing the same action themselves. Parmeggiani et al. (2019) used a medical record review of 105 autistic individuals to reveal that 16.2% of the sample had an absent sucking reflex as a newborn. Peterson et al. (2016) observed six autistic children (aged 4–6 years) for “mouth clean” (i.e., the absence of food larger than a grain of rice in the child’s mouth 30 s after a bite entered the mouth) and noted that four of the children (67%) had a mean “mouth clean” of 0% over five attempts.

Associations between oromotor performance and other skills in ASD

Twenty-six studies (24%) reported information regarding associations or correlations between oromotor skills and a variety of linguistic, cognitive, behavioral, or demographic variables among autistic individuals.

Expressive language

The relationship between oromotor skills and expressive language skills was investigated in 12 studies, of which nine (75%) reported an association between abnormal oromotor skills and decreased expressive language abilities. Specifically, it has been reported that verbal autistic children (i.e., those who produce and vocalize consonantvowel syllables and use them during communicative attempts) significantly outperform nonverbal autistic children on measures of eating behaviors, voluntary nonverbal oral skills, and pre-speech/speech behaviors (Amato & Slavin, 1998). Oromotor skills have been shown to accurately differentiate highly, moderately, and minimally fluent autistic children (Gernsbacher et al., 2008), and speech production ability accounts for significant levels of variance in the number of different words produced by minimally verbal autistic individuals (Chenausky et al., 2019). Additionally, oromotor skills have been found to positively correlate with both pre-intervention expressive language and with learning rates during expressive language intervention (Belmonte et al., 2013) and with a caregiver questionnaire assessing communication skills (Nakaoka et al., 2020). Whitehouse et al. (2008) found that autistic children with age-appropriate language outperformed autistic children with language impairment on an assessment of oromotor sequencing. Thurm et al. (2007) reported that performance on the item “imitating sounds of adults immediately after hearing them” from the VABS at age 2 was a significant discriminator of autistic children with and without expressive language deficits at age 5. Trembath et al. (2019) found that children with higher expressive language scores (per the MSEL and VABS) demonstrated greater increases in the vocalization ratio over time. Finally, Sullivan et al. (2013) showed that the number of vocalizations produced by autistic children is related to the variability of acoustic features corresponding to the placement of articulators during speech.

Receptive language

Six studies examined correlations between oromotor skills and receptive language. Five of these studies (83%) reported a significant relationship. As with expressive language, Belmonte et al. (2013) found that oromotor skills are positively correlated with pre-intervention language abilities as well as receptive language learning rates during intervention. Receptive vocabulary differentiates minimally and low verbal autistic children whose speech production is within normal limits from those with suspected CAS, those with non-CAS speech impairment, and those with insufficient speech for perceptual rating (Chenausky et al., 2019). Kjelgaard and Tager-Flusberg (2001) found that word-level speech production ability measured via the GFTA was associated with receptive vocabulary. Thurm et al. (2007) reported that performance on the item “imitating sounds of adults immediately after hearing them” from the VABS at age 2 was a significant discriminator of autistic children with and without receptive language deficits at age 5. Finally, Sullivan et al. (2013) found that acoustic measures of placement of the articulators were associated with receptive language skills among autistic children.

Expressive-receptive disparity

The relationship between receptive and expressive language in ASD is of theoretical interest because of its potential to reveal whether deficits of communicative speech are due to underlying deficits in the use of language or whether they can be attributed to expressive language-specific abilities such as oromotor skills. Of the two studies directly investigating this topic, Belmonte et al. (2013) found that a receptive-expressive language disparity (i.e., receptive language [RL] superior to expressive language [EL]) was associated with oromotor impairments among a subgroup of autistic children, while McDaniel et al. (2018) found that although attention toward a speaker was positively correlated with a receptive-expressive vocabulary discrepancy (EL > RL), oromotor performance was not correlated with a hypothesized receptive-expressive vocabulary discrepancy in the opposite direction (RL > EL).

Social skills

The relationship between oromotor skills and social functioning was explored in three studies, each of which reported significant correlations with measures of social skills. Nakaoka et al. (2022) found that oromotor function was moderately correlated with scores from the Social Communication Questionnaire (SCQ; Rutter, Bailey, & Lord, 2003). Plumb and Wetherby (2013) found moderate correlations between both total vocalizations and transcribable vocalizations (i.e., syllabic vocalizations containing at least a vowel which may also contain a consonant) with scores from the social composite score of the CSBS (Wetherby & Prizant, 2002), and Trembath et al. (2019) found that the ratio of speech to nonspeech vocalizations in autistic children was significantly associated with scores on the SCQ.

Other skills

Studies included in this review have explored associations between oromotor skills and other linguistic, cognitive, behavioral, or demographic variables. In each case, significant associations were reported. Positive associations of various strengths have been found between oromotor abilities and fine motor skills (Belmonte et al., 2013), manual motor skills (Gernsbacher et al., 2008), joint attention (Dalton et al., 2017), nonword repetition skills (Zarokanellou et al., 2022), naming skills (Zarokanellou et al., 2022), developmental outcomes as measured by the Mullen Scales of Early Learning (MSEL; Mullen, 1995) (Plumb & Wetherby, 2013), symbolic behaviors (Plumb & Wetherby, 2013), pragmatic language (Stevenson et al., 2017), nonverbal IQ (Chenausky et al., 2019), sensory profile (Nakaoka et al., 2022), and daily living skills (Nakaoka et al., 2022). Negative associations have been reported between oromotor abilities and autism characteristics (Stevenson et al., 2017) and internalizing conditions (Lundin Remnélius et al., 2022).

DISCUSSION

Our results indicate that despite the widespread notion that speech production is relatively spared—at least when assessed perceptually—among autistic individuals (Kjelgaard & Tager-Flusberg, 2001; Rapin & Dunn, 2003), the majority of available evidence indicates abnormal oromotor control in this population. The percentage of studies reporting oromotor impairments varied by domain (i.e., speech, nonspeech, and feeding) and behavior analyzed. However, the majority of studies, whether conducted with perceptual or instrumental methods, indicated that autistic individuals demonstrate significant abnormalities in oromotor functioning (85% and 81%, respectively). For example, 85% of studies examining speech accuracy and 80% of studies examining nonspeech oral movements among autistic individuals reported deficits. For some less commonly studied features such as coarticulation/coordination, 100% of studies reported an abnormality.

It is critical to note the numerous methodological issues, sampling challenges, and contradictory findings present in the included studies. These factors (discussed in further detail below), along with the relatively small number of studies analyzing particular oromotor behaviors, lead us to conclude that the current level of understanding of oromotor functioning in ASD is poor and that additional rigorous research is necessary before we can draw definitive conclusions. Below we will first discuss the findings of each of our research questions and then explore several limitations associated with the body of literature presented above.

What methods have been used to investigate oromotor functioning among individuals with ASD?

Sixty-one percent of the studies in this review used perceptual (i.e., auditory and/or visual) methods. Perceptual analyses often serve as the gold standard of assessment for a variety of speech features and are used extensively to diagnose the type and severity of motor speech disorders, inform treatment targets, and monitor progress over time (Duffy, 2019). However, the over-representation of perceptual assessments in this review is likely one factor influencing the mixed results across studies. Perceptual assessments typically require intensive clinical training, are vulnerable to rater bias (Borrie et al., 2012; Chenausky, Maffei, et al., 2022; Kent, 1996), and may be too coarse and unreliable for detecting subtle oromotor deficits (Green, 2015). Even experienced listeners are susceptible to top-down linguistic processes during speech perception such as disregarding acoustic errors to comprehend a speech signal more effectively (Bond & Garnes, 1980). Indeed, the differences among perceptual judgments made by expert judges are often larger than the differences needed for diagnostic classification (Kent, 1996). To take one example from this review, nearly all of the assessments of nonspeech oromotor functioning included in this study were obtained via auditory and visual perceptual methods using tools such as the OSMSE or VMPAC. The reliability of such measures is contingent on the assumption that all raters have similar internal standards of parameters such as typical ranges of motion during lip retraction or tongue lateralization. Training materials and procedures are often provided with these assessment tools to improve reliability; however, other sources of bias including listener familiarity with a speaker (Tjaden & Liss, 1995) are known to significantly impact perceptual judgments. Other studies in this review made judgments regarding nonspeech oromotor skills based on novel tasks, modified versions of existing tools, or prototypes of tools still in development, reducing the reliability of these judgments.

Instrumental assessment methods including acoustic analysis, EMG, MEG, MRI, ultrasound imaging, and facial motion tracking were used in only 36 studies (34%). These methods offer objective data, which mitigate the perceptual biases discussed above. Acoustic methods, which are replicable, non-invasive, and widely available, were by far the most commonly used instrumental approach in this review. Some studies included both perceptual and instrumental assessment and draw attention to the differences between these approaches. For example, Patel et al. (2020) found a significant difference between autistic children and their typically developing peers in acoustic—but not perceptual—measures of speaking rate, suggesting that instrumental methods can detect subtle differences unavailable to even the trained human ear.

However, the relationship between an acoustic signal and perceptual speech features is often complex, particularly when measuring gestalt perceptual constructs such as articulatory or prosodic accuracy. For instance, judgments of articulatory precision may not be correlated with a single acoustic feature but rather with a combination of features such as formant transitions and speaking rate (Chiu et al., 2021). In general, there is a need for further research regarding the validity of acoustic measures of constructs such as articulatory function (Allison et al., 2020; Rowe et al., 2021).

Kinematic analyses such as facial motion tracking and ultrasound tongue imaging, used in only six studies in this review, offer highly precise data and crucial physiological context for differences in motor function, although require specialized equipment and time-consuming data analysis. Incidentally, five of the six studies utilizing kinematic technologies in this review found significant abnormalities among individuals with ASD (Gladfelter & Goffman, 2018; Manfredonia et al., 2019; McKeever et al., 2022; Parish-Morris et al., 2018; Samad et al., 2019). There is high potential for kinematic techniques that have been used with other populations to aid the study of oromotor function in ASD, such as lip movement speed and duration (Yunusova et al., 2010), lip shape (Iuzzini-Seigel et al., 2015), temporal coordination of the articulators (Green et al., 2000), jaw motion during chewing (Simione et al., 2016; Wilson et al., 2012), and tongue and jaw movements during swallowing (Perry et al., 2018).

The relative lack of instrumental investigations, which are capable of objectively quantifying perceptual anomalies in disordered speech, appears to be a gap in this literature. It is interesting to note that both perceptual and instrumental methods resulted in nearly the same percentage of studies reporting an oromotor deficit (i.e., 85% of perceptual studies and 81% of instrumental studies). The similarity of these two percentages suggests that both approaches are tapping into real deficits and offer relative methodological strengths. For instance, perceptual methods will likely remain the gold standard for assessing features such as phonetic inventory and vocalization quality, while instrumental measures become increasingly common for measures constructs such as rate, duration, consistency, and certain aspects of accuracy.

Two major trends related to the use of assessment tools were also revealed in this review. First, assessment tools designed specifically to assess oromotor functioning (e.g., the VMPAC or OSMSE) were used in only 11 studies. The interpretation of results regarding oromotor skills obtained using tools designed to assess other domains (e.g., prosody, expressive language, and general development) is significantly limited by these tools’ lack of sensitivity in measuring oromotor function; in fact, some of these assessments include as few as one item related to the complex construct of oromotor functioning. Second, norm-referenced standardized assessment tools were used in 33 total studies. While a control group is often used instead of normative data, this underutilization of norm-referenced assessments underscores a critical need for the development and use of valid and reliable assessment tools in this area.

Sample size is a consideration that must be made carefully when designing a study. Within this review, 39% of studies had a sample size of under 20 individuals with ASD, and 17% had a sample size of fewer than 10 participants. Small samples, including single case studies, offer several advantages such as allowing for the collection of a large amount of data that cannot easily be obtained from a large sample and for initially exploring a novel technique or hypothesis. However, findings of studies with small samples are significantly limited in their statistical power and thus their generalizability to the wider population of individuals with ASD.

What oromotor behaviors (including speech skills) have been investigated?

The majority of studies (78%) in this review examined speech production, compared to 25% of studies examining nonspeech oromotor control and 18% of studies examining oromotor control related to feeding. The study of speech production in this population is critical as researchers investigate the nature of expressive communication deficits among autistic individuals. However, the study of nonspeech oral motor control is imperative. The reason lies in the uniqueness and domain-specificity of speech, which involves distinct neural networks specifically tuned to the motor processes involved in speech production. These processes are distinct from those of nonspeech oral movements in a variety of ways including relative muscular forces, velocities, durations, rhythmic scaffoldings, and contextual embeddings (Ziegler & Ackermann, 2013). Conclusions regarding nonspeech oromotor skills cannot be made based on research studying speech, and vice versa (Weismer, 2006). Studying nonspeech and speech skills in parallel will also shed light on questions related to whether motor impairments in ASD are present globally throughout the motor system or are manifested in the context of cognitively demanding tasks such as language production. Thus, there appears to be a relative dearth of research regarding nonspeech oral motor control among autistic individuals. In the context of examining oromotor skills as potential predictors of language impairment, we cannot ignore the possibility that even sub-perceptual differences in nonspeech oromotor skills may serve as valuable features in predictive models.

This review also allows analysis of which behaviors have been studied perceptually and which have been studied instrumentally. Several features have been examined using both approaches—for instance, speech sound accuracy has been assessed using binary correct/incorrect ratings in perceptual studies and via measures such as formant frequencies or voice onset time measurements in instrumental studies. On the other hand, some features were only examined perceptually, often in cases in which a construct of interest is defined in relation to human listeners (e.g., intelligibility, although note recent work including Jacks et al., 2019 and Gutz et al., 2022 linking automatic speech recognition and intelligibility). However, in the case of other features, a lack of research highlights the need for acoustic measures of these features. For instance, while some work has been done to develop acoustic measures of speech coordination (e.g., Rowe et al., 2021; Weismer et al., 2003), currently there is a need to establish valid and reliable acoustic measures of such features. On the other hand, sophisticated measures of oromotor coordination can be found in studies using facial motion capture technology (Green et al., 2000; Matsuo & Palmer, 2010; van Lieshout & Neufeld, 2014).

What conclusions can be drawn regarding oromotor skills in ASD?

A large majority (81%) of studies included in this review reported significant oromotor abnormalities among autistic individuals. However, there is limited consensus regarding the presence, type, and severity of deficits even within specific domains (for instance, only 54% of studies investigating speech rate found significant differences among individuals with ASD). Additionally, different abnormalities have been reported within the same feature, precluding straightforward synthesis and interpretation. For instance, among the 13 studies investigating speech rate, findings for individuals with ASD included typical rate, abnormally slow rate, an abnormal rate without indication of direction, and a lack of speech rate entrainment. Nonetheless, it is compelling to note that for any given feature—whether related to speech, non-speech oral behaviors, or feeding—at least 50% of the studies addressing that feature indicated a significant abnormality among individuals with ASD.

Several findings from this review can be interpreted in the context of known characteristics of ASD, including the presence of restricted, repetitive patterns of behavior, interests, or activities (American Psychiatric Association, 2013). Of the five instrumental studies investigating oromotor stability in this review, four support the presence of reduced movement variability in the oromotor system among autistic individuals (Gladfelter & Goffman, 2018; Kissine et al., 2021; Kissine & Geelhand, 2019; Parish-Morris et al., 2018), which may be interpreted as a potential extension of “restricted, repetitive patterns of behavior.” Further support for this theme of “restricted” motor behavior can be found in other studies included in this review that did not directly investigate motor stability. Talkar et al. (2020) interpreted their results regarding the coordination of motor acts as suggesting that individuals with ASD may demonstrate higher levels of coupling in their movements than typically developing controls; Velleman et al. (2010) found that their subjects with ASD had less average variation in their speech durations compared to TD controls. Interestingly, the results highlighted above stand in contrast with reports of increased variability in temporal and spatial aspects of other motor activities in ASD, including reaching, pointing, and saccadic movements (Glazebrook et al., 2006; Gowen & Miall, 2005; Stanley-Cary et al., 2010).

Atypical or absent anticipatory behaviors are not a diagnostic marker of ASD but have been observed among autistic children since the earliest descriptions of the disorder (Kanner, 1943). These deficits are widely reported in the literature and include significant difficulty with anticipation during behaviors such as load-lifting (Martineau et al., 2004; Schmitz et al., 2003). For instance, children with ASD demonstrate decreased predictive muscle activity in their left hand when removing a weight from it with their right hand (Schmitz et al., 2003). This and similar results suggest an over-reliance on feedback (reactive) control and an under-reliance on feedforward (predictive) control in the motor systems of individuals with ASD. Several studies in this review examined anticipatory oromotor behaviors; Brisson et al. (2012) reported an anticipation deficit observable at 4–6 months of age among infants later diagnosed with ASD, and Cattaneo et al. (2007) found decreased activation of the mylohyoid muscle when autistic children grasped food and brought it to their mouth. The young age of the subjects in the study by Brisson et al. (2012) study suggests the potential of such motor behaviors as a predictor of outcomes in ASD. Ultimately, a lack of additional studies confirming and expanding on these findings prevents a clear understanding of whether such failures of anticipation are related to the motor system or to other domains (e.g., cognitive processing speed or the interpretation of social cues).

The movement abnormalities observed across domains in individuals with ASD (e.g., reduced variability, reduced accuracy, deficits of anticipatory behavior, and disrupted feedback control) do not appear to neatly fit into a pre-established category of motor disorder (e.g., dystonia, ataxia, hyperkinesia, or apraxia). The existing literature raises the question of whether the motor differences observed in ASD may constitute a distinct category of motor disorder. This notion has specific implications for the characterization of speech error patterns among autistic individuals, which are variable and do not conform to existing categories such as pediatric dysarthria or childhood apraxia of speech in obvious ways. Shriberg et al. (2010) provide a cover term (i.e., Motor Speech Disorder-Not Otherwise Specified; MSD-NOS) for speech, prosody, and voice behaviors that are consistent with motor speech impairment but do not correspond to the main classifications of apraxia or dysarthria. Future research focusing on a holistic picture of the cognitive, motor, and linguistic skills of individuals with ASD should consider impairment profiles that fall outside current diagnostic classification schemes and may be consistent with well-established characteristics of autistic individuals such as behavioral rigidity and anticipatory difficulties.

Methodological limitations

Beyond several trends discussed above, an effective characterization of oromotor functioning among individuals with ASD is hindered by several methodological constraints characterizing the included studies.

First, it must be noted that studies with small sample sizes are often inadequately powered to detect true group differences and to produce generalizable results. For instance, Adams (1998) provided one of the most straightforward accounts of oromotor functioning in ASD and is often cited as providing evidence for oromotor and motor speech deficits in the population. However, the study had a sample of only four children with ASD and four TD controls. Similarly, Tierney et al. (2015) reported the compelling finding that 64% of their sample of children with ASD also had comorbid CAS based on a sample of only 11 children.

A second methodological limitation is the common reliance on perceptual judgments, which were utilized in nearly two-thirds of the studies included in this review. These methods are often considered the gold standard in speech assessment because of their ecological validity but, as discussed above, have known limitations including multiple sources of variability associated with auditory-perceptual assessment (Kent, 1996) and the use of rating schemes that may not be granular enough to detect subtle oromotor differences (Green, 2015). Strategies for attenuating the limitations of perceptual assessment have been proposed, including the use of carefully designed reference samples and increased listener training (Kent, 1996).

Third, many standardized tests of children’s nonverbal oral and speech motor performance are limited by inadequate psychometric development (McCauley & Strand, 2008). Broome et al. (2017) conducted a systematic review specifically examining speech assessments for children with ASD and noted similar psychometric limitations, highlighting the need for standard procedures to aid in cross-study comparison. For some less frequently investigated motor speech features (e.g., consistency, coordination), standardized tests do not currently exist, and these skills are defined and examined in a variety of ways across studies, precluding generalization.

Fourth, assessment tools for examining behaviors like articulation and nonspeech imitation often utilize dichotomous correct/incorrect item scores, which do not make a distinction between errors due to common developmental phonological processes or atypical errors representing disordered development. This provides an obstacle to understanding the nature of speech sound errors in this population, including whether they represent atypical or simply delayed developmental trajectories. Relatedly, specific error types are often not reported, making characterization of articulatory deficits difficult. For instance, Tierney et al. (2015) provided results suggesting a high prevalence of CAS among autistic children, based on “the presence of speech characteristics of apraxia” during administration of the KSPT, but did not discuss these characteristics or their relative presence among their sample.

Finally, the approach of using control groups matched on age, IQ, or mental age in ASD language research, as was done in many papers included in this review, deserves careful consideration for reasons outlined by Tager-Flusberg (2004); in particular, heterogeneity of the ASD population, intellectual disability among ASD participants, developmental changes with age, and the possibility of recruitment bias have significant potential to impact the validity of using matched control samples in language research in ASD.

Sampling challenges

There are numerous significant challenges associated with obtaining an adequate and representative sample of oromotor functioning from children with ASD. First, it is well established that deficits of imitation ability are present among individuals with ASD (see Smith & Bryson, 1994 for a review), although it remains unclear if such deficits are indicative of neurological impairments (e.g., difficulty with action planning), impaired social skills, or deficits in cognitive functions such as forming mental representations of events. Many of the standardized tests used in the studies in this review require imitation of oral movements and speech sounds, which creates a challenge for the assessor who must be willing to utilize alternative methods to elicit attempts and then interpret assessments cautiously. Autistic individuals also often demonstrate deficits in receptive language skills (Kjelgaard & Tager-Flusberg, 2001), with important implications on their ability to follow task instructions, and making many tasks—particularly more complex tasks such as producing sequences of syllables or nonwords—impractical for many autistic individuals. Additionally, as many as one-third of children with ASD can be considered minimally verbal (Tager-Flusberg & Kasari, 2013), meaning that they fail to acquire spoken language beyond a limited set of single words or functional phrases. This provides another significant obstacle to oromotor assessment since these subjects may not produce enough spoken language to assess or may produce unintelligible utterances than cannot easily be compared to target productions. Methods for conducting research with this population have been reviewed and discussed in detail by Chenausky, Maffei, et al. (2022).

The methods used to recruit samples of autistic individuals, often influenced by the study’s research questions, hypotheses, and available resources, also has a significant influence on participant characteristics and thus results. For example, a sample recruited through a speech, language, and feeding clinic will almost certainly be biased toward higher levels of oromotor deficits and/or speech sound errors. On the other hand, a population-based study recruiting through varied sources (e.g., support groups, schools) may be more representative of a range of oromotor strengths and difficulties. Similarly, the research questions addressed by a given study will influence its sample. For instance, studies investigating speech production deficits in minimally verbal autistic children (e.g., Chenausky et al., 2019) or assessing the prevalence of ASD among children with CAS (e.g., Tierney et al., 2015; Vashdi et al., 2020) will likely report different findings than studies in which the participants were more verbal (e.g., Kjelgaard & Tager-Flusberg, 2001).

Children with ASD are known to demonstrate a variety of speech sound errors, including both age-appropriate phonological processes and atypical, age-inappropriate errors (Cleland et al., 2010; Rapin et al., 2009; Shriberg et al., 2011) as well as a potentially higher prevalence of childhood apraxia of speech (Chenausky et al., 2019; Tierney et al., 2015) and abnormalities of rate, coordination, and other features as described in this review. The various classifications of speech sound disorders and motor speech disorders are not always clearly delineated, with significant symptoms (e.g., consonant imprecision, prosodic errors) overlapping between disorders. These similarities necessitate careful and often challenging differential diagnosis (Allison et al., 2020; Iuzzini-Seigel et al., 2022). We excluded studies from this review if subjects had comorbidities with the potential to impact oromotor behaviors (e.g., diagnosed hearing loss, cerebral palsy); however, we cannot account for the real possibility that autistic individuals may have more than one speech disorder, any of which may occur in combination with the others.

Finally, behavioral research regarding ASD requires consideration of the significant heterogeneity known to characterize the population. This heterogeneity has been observed in domains such as language (Rapin, 2006; Tager-Flusberg, 2006) and IQ (Mayes & Calhoun, 2003), as well at the levels of genetics (Szatmari, 1999) and neuroimaging (Martinez-Murcia et al., 2017). Changing diagnostic criteria have led to a large increase in the reported prevalence of ASD over the last several decades, contributing further to the broad range of phenotypes associated with the diagnosis of ASD. This variance has important impacts both within studies (i.e., researchers must ensure a representative sample of individuals with ASD) and across studies. Studies of behavior in ASD often constrain their sample using strict inclusion criteria meant to create clean experimental groups and to answer specific research questions. For instance, subgroups of children with ASD have been selectively researched based on impaired intelligibility (e.g., Koegel et al., 1998), a diagnosis of CAS (e.g., Chenausky et al., 2020), or exposure only to English at home (Schoen et al., 2011). A significant trade-off of this approach is that it limits cross-study comparison by not capturing the full range of behaviors that characterize individuals with ASD.

Future research

The findings of this scoping review are intended to inform and motivate future research. First, there is a relative dearth of instrumental studies compared to perceptual studies investigating oromotor functioning in ASD, despite the wide availability and affordability of acoustic analysis hardware and software. Instrumental methods are objective, replicable, and sensitive to small differences and changes over time compared to perceptual methods, which have known limitations discussed above and which involve intensive listener training. Second, while task selection must be motivated by the specific research questions of a particular project, there appear to be relatively few studies that compare the performance of autistic children on differential tasks (e.g., real words vs. nonwords, speech vs. nonspeech, single words vs. connected speech, etc.). This is particularly important not only to specify how oromotor impairments manifest, but also because tasks often do not generalize to each other. For instance, articulation tests and conversational speech samples can provide significantly different accuracy profiles (Morrison & Shriberg, 1992) and nonspeech oral motor movements may not be associated with speech accuracy (McCauley et al., 2009); future studies should consider task selection in this context. Third, to further investigate the correlations between oromotor performance and cognitive, behavioral, or linguistic skills, further longitudinal work should be performed to examine these relationships over time. Finally, as discussed above, the current lack of consensus on the nature of oromotor impairment among autistic individuals suggests the possibility that the characteristics of such an impairment (or multiple impairment profiles) may fall outside existing classifications of speech disorders. Future research with large samples of autistic individuals may help to further elucidate this ongoing question and contribute to our understanding of this complex topic.

Strengths and limitations

The strengths of this review include the wide-reaching search strategy, which covered multiple large databases and included peer-reviewed publications as well as Ph.D. dissertations and book chapters. The inclusion of only studies that met our criteria for reporting ASD diagnosis also helped focus the findings of this review and increase the generalizability of results to the population of individuals diagnosed with ASD. On the other hand, this strict adherence to our criteria (e.g., excluding studies in which the method of ASD diagnosis was not described, or which examined motor behaviors like phonation which are closely related to oromotor function) may be considered a limitation since numerous relevant studies were excluded and valuable data may have been lost in the process. An additional limitation inherent to this scoping review was the significant heterogeneity in study designs, assessment tools, sample characteristics, and control groups used in the included studies, which made cross-study comparison and synthesis difficult.

Clinical implications

Findings from this scoping review demonstrate that, despite methodological limitations and remaining gaps in knowledge, there appear to be significant differences in oromotor skills between autistic individuals and their peers. This has multiple significant clinical implications related to intervention and assessment in this population. First, findings of existing and future research may motivate the use of motor-based interventions to complement linguistic approaches to expressive language therapy for autistic individuals. Second, this review demonstrates that there exists promising evidence for oromotor functioning as a useful predictor of expressive language, which would have a significant impact on the assessment process of autistic children.

Summary

Research regarding oromotor functioning among autistic individuals is critical to a fuller understanding of ASD, with direct implications for early diagnosis of language impairments, the identification of neurobiological mechanisms influencing communication development, and the creation or modification of language interventions. The majority of studies in this review reported abnormal oromotor skills among autistic individuals, although interpretation and generalization of these findings are impacted by methodological limitations. A subset of available evidence suggests that oromotor skills are correlated with a variety of cognitive and behavioral abilities, particularly expressive and receptive language skills. Suggestions are provided for future research, which may overcome existing limitations and further clarify this important topic.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.