Slower pace in early walking onset is related to communication, motor skills, and adaptive function in autistic toddlers

Abstract

The onset of walking is a major developmental milestone in early childhood and is critical to the development of language and social communication. Delays in walking have been described in individuals with ASD. Yet, less is known about the quality of early gait development in toddlers with ASD and the relationship to motor skills, social communication, and language. Quantitative measures of locomotion can improve our ability to evaluate subtle and specific motor differences in toddlers with ASD and their relationship to other developmental domains. We used quantitative gait analysis to evaluate locomotion in toddlers with ASD (n = 51) and compared these data to a reference chronological aged (CA) and mental aged (MA) matched typically developing (TD) cohort (n = 45). We also examined the relationship of quantitative gait metrics to developmental measures among toddlers with ASD. We found that although toddlers with ASD achieved a typical age range of walking onset, they exhibited a pattern of slower pace compared to the TD cohort when matched by CA and MA. We also found that slower measures of pace were associated with lower developmental scores of communication, motor skills, and adaptive function. Our findings improve characterization of locomotion in toddlers with ASD and the relationship of motor skills to other developmental domains.

Lay Summary

Locomotion is a critical motor milestone and can drive many areas of development such as language. Here, using quantitative gait analysis and developmental measures, we find that toddlers with ASD have atypical gait patterns that differentiate them from TD toddlers. We also found that atypical gait patterns were associated with lower scores on developmental measures of communication, motor skills, and adaptive function. Our findings improve motor characterization in early childhood and ASD.

Introduction

Independent walking is a major developmental milestone and the primary means of locomotion throughout most of a human’s lifespan. The achievement of independent walking requires neurological development, musculoskeletal maturation, and the accumulation of motor experiences [Burnett & Johnson, 1971a,b; Forssberg, 1985; Thelen & Cooke, 1987; Ivanenko et al., 2007]. The onset of walking has been shown to be a critical inflection point for driving social and communicative development [West, 2019; Calabretta et al., 2022; West & Iverson, 2021; Clearfield, 2011; Adolph & Tamis-Lemonda, 2014]. Walking specifically allows infants to travel farther and faster and provides new opportunities for visual input, object interactions, and social engagement. These opportunities lay the foundation for the development of spatial perception, language development, and social communication [Campos et al., 2000; Iverson, 2010; Karasik et al., 2011; LeBarton & Iverson, 2016]. The cascading effects of locomotive ability on broad developmental systems of language, communication, and adaptive function can shape outcomes throughout a lifespan.

Walking clearly represents an important developmental milestone; however, it is not just the age of walking onset that matters, but also the quality of gait that is achieved. While a later onset of walking delays a child’s access to the developmental benefits gained through walking, it is the quality of gait (e.g., speed, variability, and postural control) that cumulatively impacts the child’s access to these benefits over time. Therefore, it is important that gait quality in children at risk for motor impairments be monitored beyond the onset of milestones so that appropriate and timely intervention can be implemented.

Studies of early walking onset and the relationship to social communication and language in autistic toddlers are limited. It is known that children with ASD often achieve independent walking later than neurotypical children and that later age of walking is associated with increased ASD symptom severity [MacDonald et al., 2013; Reindal et al., 2020; Bishop et al., 2016]. Consistent with research in older children [Rinehart et al., 2006], two studies have also demonstrated that toddlers with ASD have higher levels of asymmetry while walking compared to neurotypical toddlers [Esposito et al., 2011; Esposito & Venuti, 2008]. A recent meta-analysis of gait anomalies in older children with autism indicated some shared differences in spatiotemporal gait variables across studies [Lum et al., 2020]. This meta-analysis and other research have also found that age dependent gait related changes may occur in autism [Lum et al., 2020; Li et al., 2021; Manicolo et al., 2019]. Prior work has demonstrated an association between motor skills and adaptive functioning, language development, and communication in children with autism [Macdonald et al., 2013; Bedford et al., 2016; West et al., 2019]. However, to date there are no studies of gait quality or the relationship of gait to social communication, language, and adaptive functioning in autistic children under 4 years of age.

The current study utilizes quantitative spatiotemporal measures of gait to compare early gait development in children ages 12-36 months with and without ASD. We focused on gait domains of pace and postural control. We also evaluate the relationship between spatiotemporal variables of gait quality and standardized measures of language, communication, visual reception, adaptive function, and motor skills in toddlers with ASD. We hypothesized that toddlers with ASD would show differences in gait variables compared to typically developing children – specifically, lower velocity, cadence, and step length as well as greater stride width consistent with studies of older autistic children. We also hypothesized that atypical spatiotemporal gait variables characterized by lower velocity, cadence, and step length and greater stride width would be associated with lower developmental scores on measures of expressive language, communication, fine motor skills, gross motor skills, and adaptive function. The ultimate goal of this work is to expand our understanding of early motor impairments in ASD beyond delays in developmental milestones, and how these motor impairments can have an impact on language, communication, motor skills, and adaptive function. These findings would allow for improved developmental monitoring and help inform motor interventions in toddlers with ASD.

Methods

Participants

Participants were recruited as part of an ongoing Autism Center of Excellence longitudinal study at the University of California, Los Angeles. Gait data for the typically developing (TD) control group was selected from a normative sample collected at the University of North Carolina at Chapel Hill [Dusing & Thorpe, 2007]. A total of 51 toddlers with ASD ages 15-36 months and 45 TD toddlers ages 12-36 months were included in this study. Exclusionary criteria for ASD and TD groups included inability to ambulate without assistance, inability to walk at least 100 feet, muscle, bone, brain, or nerve dysfunction, and any associated physical disorders.

Procedures

The UCLA Institutional Review Board (IRB#17-001269) approved all research. Prior to data collection informed consent was obtained from all families. Cognitive, behavioral, and gait assessments took place at the UCLA Center for Autism Research and Treatment.

Gait Assessment

The Zeno electronic walkway (Protokinetics, Havertown PA) was used in conjunction with the ProtoKinetics Movement Analysis Software (PKMAS) to obtain and analyze quantitative gait data [Egerton et al., 2014; Lynall et al., 2017]. The walkway contains a series of pressure sensors to detect footfalls. We leveraged a TD cohort collected in previous research studies as our normative sample reference group [Dusing & Thorpe, 2007]. We collected data on the ASD group and ensured that the gait data collection protocol was consistent with the TD cohort. Only the active length of the walkway differed between the current study (16 feet) and studies that produced the TD dataset (20 feet). To minimize any differences, the participants in the current study walked eight full lengths of the walkway compared to the six full lengths for the TD studies. All groups walked an additional four feet on either end of the walkway to account for acceleration and deceleration. The gait paradigm that was analyzed was spontaneous self-paced gait in which the subject can apply and control their natural speed. Participants were given a demonstration of the task, allowed to explore the walkway and environment, and practiced a self-paced walk with the assessor and/or caregiver prior to data collection. Trials were repeated when participants stepped off the recordable area of the walkway, did not use a self-paced gait (e.g., running, scooting, shuffling, and toe-walking on walkway), sat down or fell down on the walkway, or when incomplete footfalls occurred. To account for potential behavioral and mood factors that could influence self-paced gait, the ASD group was rated utilizing a mood and attention Likert scale. The scale ranges from 1–5 and evaluates areas such as irritability, hyperactivity, and inattention. A score of 1 indicates that a participant has a content or neutral mood, does not show any verbal or non-verbal protests, is able to be verbally or visually prompted to complete the trial, and is able to be redirected if necessary. The UCLA research team observed all trials for the ASD group and included only those where a self-paced speed was maintained, and the participant displayed a Likert score of 1. The eight self-paced gait trials were then averaged together to provide a global mean of all gait trials.

In this study, we focused on spatiotemporal gait variables within the following two domains to compare between groups:

1. Pace: velocity (centimeters/second), cadence (steps/minutes), step length (centimeters)

2. Postural control: stride width (centimeters)

Measures of gait variability were collected for the ASD group, and we present these data in the manuscript given the paucity of objective gait data in this age range in ASD. These measures include: step length coefficient of variation (%), step time coefficient of variation (%), stride width standard deviation, and gait variability index (GVI). Age of onset of walking was obtained for the ASD group from parent report.

ASD Symptoms.

Evaluation for autism symptoms was completed using The Autism Diagnostic Observation Schedule- Toddler Version (ADOS-T; Luyster et al., 2009). The ADOS-T is a semi-structured, standardized assessment of communication, social interaction, play, and imaginative use of materials. The ADOS-T is designed to assess for a clinical presentation of ASD or other pervasive developmental disorders in young infants and toddlers. Diagnostic algorithms can be used to calculate a total score, where higher scores indicate more autism symptoms. The ADOS-T was used to determine eligibility for the study with those meeting clinical cut-off scores included in the study.

Cognitive Abilities.

Cognitive development was assessed with the Mullen Scales of Early Learning (MSEL) [Mullen, 1995]. The MSEL is a standardized measure of cognitive ability for children ages 3-60 months that yields an Early Learning Composite (ELC) score, which is considered equivalent to a developmental quotient score, as well as five subscale scores for gross motor, visual reception, fine motor, expressive language, and receptive language.

Adaptive Behavior.

Adaptive behavior was assessed with the Vineland Adaptive Behavior Scale-II (VABS) [Sparrow et al., 1984]. The VABS is a semi-structured interview conducted with the parent that evaluates adaptive function in four domains: communication, daily living skills, socialization, and motor skills.

Statistical Analysis

We provide descriptive statistics of age, gender, height, and weight for the ASD and TD groups. We provide average age of onset of walking for the ASD group. Because of the potential for strong confounding by both chronological age (CA) and mental age (MA), comparisons between ASD and TD groups were assessed within age specific strata of 12-23 months and 24-36 months. The Mullen visual reception age equivalent score was used to determine the mental age for the ASD group [Staples et al., 2012]. As an example, a child with ASD with a chronological age of 24 months and a mental age of 15 months was grouped for comparison to TD toddlers in the 24-36 month chronological age (CA) category and then the 12-23 month mental age (MA) category, respectively. These age ranges were used given the variability in walking onset and cognitive development amongst typically developing children. Bivariate comparisons of gait variables between the two groups within the CA and MA groups of 12-23 months and 24-36 months as well as within the entire sample (12-36 months) were examined using Welch’s t-test and chi-square analysis. We also examined the age-adjusted mean difference between the gait variables for the ASD and TD groups using linear regression. Among the ASD group, linear regression models were used to assess the relationship of the gait variables to the cognitive and behavioral measures described above with age covariate adjustment. All analyses were conducted in Stata version 16.1 (StataCorp LP, College Station, Texas).

Results

Participant Characteristics

Table 1 describes the demographics of our sample. The ASD group was older, on average, by 2.4 months compared to the TD group (p=0.048), but there was a non-significant higher proportion of participants in the 24-36 month age range in the ASD group (71%) compared to the TD group (53%). There was no difference in the gender distribution between the ASD and TD cohorts, with both having a higher percentage of male participants (71% in ASD group, 67% in TD group). The TD cohort was significantly lower in weight (p = 0.02) and shorter (p <0.001) than the ASD cohort. We found that the average age of onset of walking for the ASD group was 12.8 months (SD=2.8).

Table 1.

Participant Characteristics

	ASD Group n = 51	TD Group n = 45	p-value
Age (months): M (SD), range	26.8 (5.5), 15-36	24.4 (6.1), 14-36	0.048*
% in 12-23 month age group	29%	47%	0.08
% in 24-36 month age group	71%	53%	0.08
% Male	71%	67%	0.68
Weight (lb): M (SD), range	28.8 (4.4), 20.2-44.4	26.7 (4.2), 20.0-38.0	0.02*
Height (in): M (SD), range	35.6 (2.6), 30.7-41.3	33.3 (2.6), 28.4-39.3	<0.001***
MSEL Early Learning Composite: M (SD), range	68 (21), 49-122
MSEL Gross Motor Domain: M (SD), range	21 (5), 14-33
MSEL Visual Reception Domain: M (SD), range	20 (7), 6-39
MSEL Fine Motor Domain: M (SD), range	21 (6), 11-34
MSEL Expressive Language Domain: M (SD), range	17 (8), 8-37
MSEL Receptive Language Domain: M (SD), range	16 (9), 4-34
VABS Composite: M (SD), range	67 (12), 47-97
VABS Communication Domain: M (SD), range	65 (21), 26-103
VABS Daily Living Skills Domain: M (SD), range	70 (14), 47-101
VABS Socialization Domain: M (SD), range	67 (13), 35-104
VABS Motor Skills Domain: M (SD), range	87 (13), 60-113

^*p < 0.05.

^***p < 0.001.

Abbreviations: M, mean; SD, standard deviation; lb, pound; in, inch; MSEL, Mullen Scales of Early Learning; VABS, Vineland Adaptive Behavior Scale-II. MSEL and VABS data are not available for the TD group.

Spatiotemporal Gait Variables

Tables 2–5 describe the differences in spatiotemporal gait variables between the ASD and TD cohorts within the domains of pace and postural control. We also include measures of variability for the ASD cohort in Table 6. Gait variables were normalized to height prior to comparison [Dusing & Thorpe, 2007].

Table 2.

Comparison of Spatiotemporal Gait Variables by Chronological Age in ASD and TD Cohorts

	ASD Group	TD Group	p-value
Velocity (cm/second)
12-36 months	0.89 (0.20)	1.13 (0.41)	<0.001***
12-23 months	0.90 (0.29)	1.16 (0.48)	0.046*
24-36 months	0.88 (0.16)	1.10 (0.36)	0.009**
Cadence (steps/minute)
12-36 months	1.78 (0.31)	2.17 (0.55)	<0.001***
12-23 months	1.98 (0.40)	2.43 (0.58)	0.008**
24-36 months	1.70 (0.23)	1.94 (0.43)	0.013*
Step Length (cm)
12-36 months	0.33 (0.05)	0.37 (0.06)	0.002**
12-23 months	0.32 (0.06)	0.35 (0.06)	0.102
24-36 months	0.34 (0.04)	0.38 (0.06)	0.004**
Stride Width (cm)
12-36 months	0.10 (0.04)	0.10 (0.04)	0.99
12-23 months	0.12 (0.05)	0.12 (0.04)	0.99
24-36 months	0.10 (0.02)	0.09 (0.04)	0.35

Mean and standard deviation are presented for each gait variable. All gait variables are normalized to height.

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

Abbreviations: cm, centimeter.

Table 5.

Mental Age-Adjusted Comparison of Spatiotemporal Gait Variables in ASD and TD Cohorts

	Mean Difference (TD - ASD)	95% Confidence Interval	p-value
Velocity (cm/second)	0.23	[0.09, 0.37]	0.001**
Cadence (steps/minute)	0.48	[0.30, 0.66]	<0.001***
Step Length (cm)	0.03	[0.002, 0.05]	0.03*
Stride Width (cm)	0.01	[−0.005, 0.025]	0.20

Mental age of the ASD group was determined using the Mullen Scales of Early Learning Visual Reception Age Equivalent Score. All gait variables are normalized to height.

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

Abbreviations: cm, centimeter.

Table 6.

Measures of Gait Variability in ASD Cohort

	12-36 months	12-23 months	24-36 months
Gait Variability Index	146 (13)	148 (7)	146 (6)
Step Length Coefficient of Variation (%)	13 (5)	15 (6)	13 (4)
Step Time Coefficient of Variation (%)	15 (5)	16 (6)	15 (5)
Stride Width Standard Deviation	3 (0.72)	3 (0.74)	3 (0.71)

Mean and standard deviation are presented for each measure of gait variability.

Compared to the TD cohort, both CA and MA matched autistic toddlers demonstrated significantly slower pace characterized by lower velocity, cadence, and step length. The differences in velocity and cadence were present within the 12-23 month age group, the 24-36 month age group, and the entire sample (age 12-36 months) as well as when examining the age-adjusted mean difference. The difference in step length was present within the 24-36 month age group and the entire sample (age 12-36 months) as well as when examining the age-adjusted mean difference, but was not significant within the 12-23 month age group. No statistical difference in postural control as characterized by stride width was demonstrated between the ASD and TD cohorts.

Gait and Developmental Measures

Table 7 describes the relationship of the spatiotemporal gait variables to developmental outcomes in the ASD cohort.

Table 7.

Age-Adjusted Linear Regression Evaluating Relationship of Spatiotemporal Gait Variables to Developmental Outcomes in ASD Cohort

	Velocity (cm/second)	Cadence (steps/minute)	Step Length (cm)	Stride Width (cm)
MSEL
Early Learning Composite	26.3 [−3.4, 55.9]	8.73 [−13.49, 30.95]	142.1 [7.0, 277.2] *	−118.0 [−307.2, 71.3]
Gross Motor	5.67 [0.17, 11.16] *	2.73 [−1.25, 6.70]	26.0 [−0.2, 52.1]	−25.8 [−60.8, 9.3]
Visual Reception	7.41 [−1.55, 16.38]	2.76 [−3.92, 9.45]	40.9 [0.1, 81.7]*	−62.1 [−117.1, −7.2] *
Fine Motor	5.82 [−0.36, 12.00]	2.72 [−1.89, 7.33]	27.0 [−1.6, 55.5]	−28.7 [−68.1, 10.6]
Expressive Language	9.55 [−0.16, 19.25]	3.42 [−3.88, 10.73]	38.9 [−6.3, 84.1]	−10.1 [−73.5, 53.2]
Receptive Language	9.77 [−1.55, 21.1]	4.19 [−4.25, 12.64]	48.9 [−3.1, 100.8]	−29.7 [−102.5, 43.2]
VABS
Composite	21.9 [4.7, 39.2] *	10.3 [−2.7, 23.4]	97.7 [19.2, 176.2] *	−63.3 [−176.1, 49.4]
Communication	33.8 [3.9, 63.6] *	13.0 [−9.5, 35.6]	165.3 [31.0, 299.5] *	−106.3 [−298.9, 86.3]
Daily Living Skills	23.3 [4.4, 42.3] *	10.8 [−3.5, 25.1]	106.5 [20.6, 192.5] *	−104.8 [−225.9, 16.3]
Socialization	18.6 [0.6, 36.7] *	11.0 [−2.3, 24.4]	73.5 [−9.6, 156.5]	−20.2 [−137.3, 96.9]
Motor Skills	19.2 [0.8, 37.6] *	2.65 [−11.32, 16.61]	130.5 [52.1, 208.9] **	−88.4 [−204.8, 28.1]

Coefficients and 95% confidence intervals are presented.

^*p < 0.05.

^**p < 0.01.

All gait variables are normalized to height. Abbreviations: MSEL, Mullen Scales of Early Learning; VABS, Vineland Adaptive Behavior Scale-II; cm, centimeter.

Within the ASD sample, velocity was significantly positively associated with the MSEL gross motor domain. Velocity was also associated with all four domains of the VABS (communication, daily living skills, socialization, and motor skills) and the VABS Adaptive Behavior Composite score (ABC). Velocity was not associated with the MSEL expressive or receptive language, visual reception, and fine motor domains or the MSEL Early Learning Composite score (ELC), although there was a trend toward a positive association between velocity and these variables. Step length was significantly positively associated with the MSEL visual reception domain and ELC score as well as the VABS communication, daily living skills, and motor skills domains and ABC. Step length was not associated with the MSEL expressive or receptive language and gross or fine motor domains or the VABS socialization domain, although there was a trend toward a positive association between step length and these variables. There was no significant relationship between cadence and the measured developmental outcomes. Stride width, a measure of postural control, was significantly negatively associated with the MSEL visual reception domain. There was a trend toward negative association between stride width and the remaining developmental outcomes, but none of these relationships were statistically significant.

Discussion

Early Gait Development in ASD vs. TD Toddlers

In this analysis we evaluated early gait development in toddlers with ASD and compared it to that of TD toddlers. In our cohort, the average age of onset of walking in the ASD group was found to be 12.8 months of age (SD = 2.8), based on parent report. Despite this being within the normal range of walking onset, our quantitative analysis of early gait development elucidated differences between ASD and TD toddlers that may negatively influence motor and behavioral development [Størvold et al., 2013]. Importantly, we found that differences between groups were present when matched on both chronological age and mental age. These findings indicate that the differences in gait are not solely attributed to differences in cognitive ability or general developmental delay between the ASD and TD groups, but rather may be indicative of atypical motor development more specific to ASD. We found that toddlers 1-3 years of age with ASD have gait features characterized by lower pace compared to TD children. Increasing pace (measured by velocity, step length, and cadence) is a marker of normal gait maturation [Bril & Breniere, 1992; Liu et al., 2022]. These results could indicate that even after the onset of locomotion at an expected age, gait development may be atypical in ASD. These findings have pertinent implications on development, clinical monitoring, and implementation of interventions. First, a child with slower pace would take a longer amount of time and greater energy to navigate and explore their environment, potentially reducing the time the child can simultaneously engage with others in social activities. This can negatively impact opportunities for environmental and social exploration. There are many developmental gains that come with the acquisition of gait, including visual access to the world as well as ability to manipulate objects with hands and show these objects to caregivers and peers [Adolph & Tamis-LeMonda, 2014]. Studies have shown that caregiver language and gestures are more likely to occur with bouts of infant walking [Schneider & Iverson, 2021]. Toddlers who are slow to engage with their environment may miss a larger range of meaningful opportunities for social bids and important verbal interactions with caregivers such as labeling of objects and actions. Studies of adults have shown that slow gait speed is associated with lower odds of social participation [Warren et al., 2016]. Extending this evaluation to children would improve our understanding of whether slower pace negatively impacts social engagement and physical activity in childhood. Second, gait is thought to stabilize in the first two years of walking, yet we found persistent gait differences at three years of age in ASD toddlers that had achieved typical age of walking onset. These data support the hypothesis that toddlers with ASD may not follow typical neuromotor maturation of gait even when developmental milestones are met on time. It is possible that differences in executive function (EF) may be associated with the gait differences observed in the autism group. There are known to be EF deficits in individuals with autism and gait is dependent on both motor and EF processes [Christoforou et al., 2023; Yogev-Seligmann et al., 2008]. Motor studies in adults and older children have found altered spatiotemporal variables of gait, such as slowing of gait speed, when measured while participants conduct concurrent EF tasks [Mohring et al., 2020; Rabaglietti et al., 2019]. Given that many domains of EF are forming at this age, future studies evaluating the influence of these cognitive functions during motor assessments would be beneficial in better understanding this relationship in autistic toddlers.

Third, there are important considerations for clinical monitoring and interventions given the findings of our study. We did not find a delay in the onset of walking for the autism group and clinicians should be aware that delayed walking onset is not a consistent finding in autistic individuals. Our findings indicate that non-delayed walking does not preclude autistic individuals from having differences in gait compared to TD children of similar age. Most clinic based developmental screening assessments consist of motor milestone achievement rather than quality of motor skills [Wilson et al., 2018a,b]. Similarly, large scale databases of developmental data on individuals with non-genetic and genetic forms of autism often focus on motor milestone achievement rather than quality of motor skills [Wickstrom et al., 2021]. From a clinical perspective, it is important for clinicians to be aware that screening for motor difficulties such as slower gait speed and referral to interventions focused on these specific gait abnormalities may improve long-term locomotive ability. These findings are also important for clinical screening in the educational system. Children with ASD may receive occupational or physical therapy intervention in or through public school to address educational needs. Currently, the IDEA act states that schools are required to provide children with services necessary for them to be involved in and make progress in the general education curriculum (US Department of Education). For gross motor skills, this often means physical therapy is deemed unnecessary unless there are clearly visible motor difficulties. More subtle motor impairments such as slower pace warrant monitoring as they may impact accessibility and engagement in school based physical activity programs.

Contrary to our initial hypothesis, we did not find a significant difference in postural control between the ASD and TD cohorts. Studies of older individuals with ASD have shown poor postural control and wider stride width compared to TD groups, and a recent meta-analysis using spatiotemporal variables of gait found that stride width is significantly wider in ASD compared to controls [Travers et al., 2013; Lum et al., 2020]. However, these studies do not include participants under four years of age. It is possible that differences in stride width were not found here because this cohort was evaluated in the window of walking onset or the first two years of independent ambulation. Previous studies have shown that base of support begins to narrow within the first 6 months of independent walking and stabilizes by the end of the second year of independent walking [Bril & Breniere, 1992]. A large portion of our cohort were assessed in the first two years of life, indicating that stride width may not have stabilized. Once stride width is stabilized, we may expect to see differences in stride width similar to those seen in studies in older participants with ASD and TD participants. The TD sample used in this study was also reported to have larger stride width values compared to other gait studies in toddlers [Dusing & Thorpe, 2007]. This may also have contributed to our findings. Using a larger normative sample in the future may elucidate if there are differences in stride width between ASD and TD toddlers.

Gait, Language, Social Communication, and Adaptive Function in ASD

We also evaluated the relationship of gait development to developmental outcomes of language, communication, motor skills, and adaptive function in toddlers with ASD. Prior literature supports that motor abnormalities can have a negative cascading effect on development in other domains [Iverson, 2021]. Here we found that lower measures of pace were associated with lower developmental scores of communication, motor skills, and adaptive function in children with ASD, demonstrating a potential modulating effect of the severity of gait abnormality on other developmental outcomes. Interestingly, we found that lower measures of pace were associated with a parent report of communication but not with a direct assessment of expressive language. It is possible that in early childhood, parental report, via a semi-structured interview format, may provide a better representation of the toddler’s language ability. It is also possible that with a larger sample size we may see these relationships emerge as significant given that there was a positive trend between measures of pace and MSEL expressive language. We did not find that cadence had any relationship to the measured developmental outcomes, which could be a reflection that cadence is more variable in this period of walking onset compared to the other spatiotemporal measures. We found that poorer postural control, characterized by larger stride width, was associated with lower scores of visual reception. It is possible that this relationship indicates a shared relationship in spatial and visual-motor difficulties in toddlers with autism. Our findings are in line with a previous study of infants at high likelihood of autism that found that infants who have faster control of postural adaptations have higher MSEL visual receptive scores [Kyvelidou et al., 2021]. Despite there being a trend toward a negative relationship between greater stride width and lower developmental scores on the other domains of interest, these relationships were not significant. The lack of stability of stride width at this age as described above could play a role in these findings.

Overall, these data highlight that aspects of atypical gait, such as lower pace, can negatively impact communication, motor skills, and adaptive function. Importantly, we also found that differences in spatiotemporal gait variables are not driven by cognitive ability. Thus, monitoring of atypical motor development beyond the onset of walking onset and implementation of appropriate intervention may have positive impacts on broad developmental outcomes in children with ASD.

Limitations

While our study is an important first step in characterizing quality of gait in toddlers with ASD, it has limitations that should be acknowledged. The data were cross sectional in nature and future studies should include longitudinal measurement of spatiotemporal gait variables to better understand the trajectory of gait development in the first three years of life. Although we created a mental age group for comparison, the availability of developmental assessments in the TD cohort would strengthen our ability to interpret the relationship of spatiotemporal gait variables to other areas of early development and cognitive ability and compare these findings to the ASD cohort.

Future Directions

Moving forward, we will evaluate the longitudinal development of gait and the relationship to other developmental domains in this cohort of children with ASD. These analyses will allow us to understand the trajectory of early gait development and further query findings such as differences in stride width between ASD and TD toddlers as gait begins to stabilize. Future studies should examine characteristics of locomotion such as locomotion type (e.g., crawling, cruising, and independent walking), number of steps, duration of locomotion bouts, number of falls, and coordination of gait on different surfaces in the home environment. This method will allow for a more detailed evaluation of factors that impact early gait development and stability. Longitudinal analysis will also allow for better interpretation of the cascading effects of motor development on the development of language, social communication, motor skills, adaptive function and symptoms of ASD. Furthermore, we hope to combine these gait measures with measures of brain function such as electroencephalogram (EEG) to inform our understanding of which brain circuits underly the motor impairments and motor variability that manifest in individuals with ASD. Evaluating differences in brain function during the emergence of motor abilities may better inform our knowledge of why motor differences occur in ASD.

Conclusions

Motor impairments have been well described in individuals with ASD across a lifespan. Quantitative measures of motor function that help elucidate differences in motor development early in life are needed and can improve our knowledge of the nature of motor impairments that may be more specific to ASD. These measures may also shed light on how persistent and often unrecognized motor differences in toddlers with ASD may perpetuate delays in the development of language, social communication, motor skills, and adaptive function. We utilized an objective and quantitative motor method to phenotype gait function in toddlers with ASD in the first three years of life. This method of quantitative gait analysis could be used for longitudinal evaluation of motor impairments that may be difficult to identify on standardized questionnaires or assessments of motor function that are predominantly focused on motor milestone achievement.

Table 3.

Comparison of Spatiotemporal Gait Variables by Mental Age in ASD and TD Cohorts

	ASD Group	TD Group	p-value
Velocity (cm/second)
12-36 months	0.89 (0.20)	1.13 (0.41)	<0.001***
12-23 months	0.86 (0.23)	1.16 (0.48)	0.014*
24-36 months	0.92 (0.16)	1.10 (0.36)	0.034*
Cadence (steps/minute)
12-36 months	1.78 (0.31)	2.17 (0.55)	<0.001***
12-23 months	1.82 (0.36)	2.43 (0.58)	<0.001***
24-36 months	1.71 (0.27)	1.94 (0.43)	0.037*
Step Length (cm)
12-36 months	0.33 (0.05)	0.37 (0.06)	0.002**
12-23 months	0.32 (0.05)	0.35 (0.06)	0.055
24-36 months	0.35 (0.03)	0.38 (0.06)	0.020*
Stride Width (cm)
12-36 months	0.10 (0.04)	0.10 (0.04)	0.99
12-23 months	0.11 (0.04)	0.12 (0.04)	0.18
24-36 months	0.09 (0.02)	0.09 (0.04)	0.62

Mental age of the ASD group was determined using the Mullen Scales of Early Learning Visual Reception Age Equivalent Score. Mean and standard deviation are presented for each gait variable. All gait variables are normalized to height.

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

Abbreviations: cm, centimeter.

Table 4.

Chronological Age-Adjusted Comparison of Spatiotemporal Gait Variables in ASD and TD Cohorts

	Mean Difference (TD - ASD)	95% Confidence Interval	p-value
Velocity (cm/second)	0.24	[0.10, 0.37]	<0.001***
Cadence (steps/minute)	0.31	[0.14, 0.47]	<0.001***
Step Length (cm)	0.04	[0.02, 0.06]	<0.001***
Stride Width (cm)	−0.006	[−0.02, 0.01]	0.43

All gait variables are normalized to height.

^***p < 0.001.

Abbreviations: cm, centimeter.

Data Availability Statement

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