Skip to main content
Download PDF
Download ePub

Visual feedback and motor memory contributions to sustained motor control deficits in autism spectrum disorder across childhood and into adulthood

Journal of Neurodevelopmental Disorders volume 17, Article number: 26 (2025) Cite this article

Abstract

Background

Autistic individuals show deficits in sustained fine motor control which are associated with an over-reliance on visual feedback. Motor memory deficits also have been reported during sustained fine motor control in autism spectrum disorders (ASD). The development of motor memory and visuomotor feedback processes contributing to sustained motor control issues in ASD are not known. The present study aimed to characterize age-related changes in visual feedback and motor memory processes contributing to sustained fine motor control issues in ASD.

Methods

Fifty-four autistic participants and 31 neurotypical (NT) controls ages 10–25 years completed visually guided and memory guided sustained precision gripping tests by pressing on force sensors with their dominant hand index finger and thumb. For visually guided trials, participants viewed a stationary target bar and a force bar that moved upwards with increased force for 15s. During memory guided trials, the force bar was visible for 3s, after which participants attempted to maintain their force output without visual feedback for another 12s. To assess visual feedback processing, force accuracy, variability (standard deviation), and regularity (sample entropy) were examined. To assess motor memory, force decay latency, slope, and magnitude were examined during epochs without visual feedback.

Results

Relative to NT controls, autistic individuals showed a greater magnitude and a trend for a steeper slope of force decay during memory guided trials. Across conditions, the ASD group showed reduced force accuracy (β = 0.41, R2 = 0.043, t79.3=2.36, p = .021) and greater force variability (β=-2.16, R2 = 0.143, t77.1=-4.04, p = .0001) and regularity (β=-0.52, R2 = 0.021, t77.4=-2.21, p = .030) relative to NT controls at younger ages, but these differences normalized by adolescence (age x group interactions). Lower force accuracy and greater force variability during visually guided trials and steeper decay slope during memory guided trials were associated with overall autism severity.

Conclusions

Our findings that autistic individuals show a greater magnitude and tendency for a greater rate of force decay than NT individuals following the removal of visual feedback indicate that motor memory deficits contribute to fine motor control issues in ASD. Findings that sensorimotor differences in ASD were specific to younger ages suggest delayed development across multiple motor control processes.

Introduction

Sensorimotor impairments are highly prevalent in autism spectrum disorder (ASD) [1]. They are associated with the severity of core social, communication, and repetitive behaviors, as well as cognitive outcomes [2,3,4,5,6], and they are some of the earliest signs of atypical development in children who later receive a diagnosis of ASD [2, 7, 8]. Sensorimotor differences in autistic persons have been observed across multiple behaviors and effector systems [9,10,11,12,13,14,15], and they involve multiple motor control processes including motor planning [11, 14], online motor control [11, 14], and motor learning [4, 16, 17]. Additionally, structural and functional differences in cerebellar-cortical brain networks involved in sensorimotor control repeatedly have been observed in ASD [17,18,19,20,21,22,23]. These findings highlight an important role of sensorimotor differences in ASD and support the need to identify motor control and neurodevelopmental mechanisms of sensorimotor impairments in autistic individuals.

Deficits of sustained, or online sensorimotor control, including the ability to reactively adjust motor output in response to sensory feedback or sustaining motor control from memory, have been repeatedly shown in ASD. Challenges with sensory feedback processing and motor memory during sustained motor control (e.g., precision grip control) likely relate to challenges with daily living skills that have been documented in autistic individuals including holding and using writing utensils [24], grasping and controlling eating utensils during feeding [25], and unintentionally dropping objects (e.g., cups and dishes) [26].

Visuomotor feedback and motor memory processes involved in sustained motor control have been well characterized in studies of precision gripping in NT development [27,28,29,30,31]. Precision gripping relies on the integration of sensory feedback (e.g., visual and somatosensory/cutaneous) to provide information on precision grip force error that is used to make corrective adjustments to the ongoing motor command and update internal models (motor learning) of grip control [27, 29,30,31]. During visually guided precision gripping, the visuomotor feedback loop lasts approximately one second, and includes integration of visual feedback error information since the last motor command and the execution of a corrective movement based the accumulated visual feedback information [27, 29]. In healthy adults, removing visual feedback during precision gripping leads to a decay in force output beginning 0.5–1.5s after the removal of visual feedback [29]. The maintenance of force output for 0.5–1.5s is consistent with the duration of the visuomotor feedback loop and reflects the capacity for visual feedback error information to be stored in short-term memory. In the absence of visual feedback, the motor system relies on cutaneous feedback (e.g., whether the gripped object slipping or the exerted grip force is appropriate for the object’s weight) and the implicit knowledge of force exertion from prior grip experiences (e.g., sensorimotor memory) [30, 31]. When visual feedback is occluded during sustained precision grip tasks where the gripped object is squeezed but not lifted (i.e., the object cannot be dropped and its weight is irrelevant), force variability and regularity increase [27, 32, 33], and grip force decreases [29] at latencies that exceed the short-term visuomotor memory (> 1.5s) and with greater effects over time [27], indicating that corrective motor commands become less accurate and less dynamic as this longer-term sensorimotor memory fades. These findings reflect motor memory processes that are distinct from, and more persistent than, short-term visuomotor memory (e.g., somato-motor memory [29, 34]).

During tests of visually guided precision gripping, our lab has demonstrated that autistic individuals have increased variability and regularity of sustained grip force relative to NT controls, suggesting that the ability to use sensory feedback to make precise and accurate corrective adjustments to ongoing movements is impaired [11, 19, 35, 36], though a separate study of 22 autistic individuals suggested that elevations in force variability were specific to, or at least more severe, during tasks with a dynamic (moving) target rather than a static (fixed) target [37]. These deficits are exacerbated when the spatial resolution of visual feedback is enhanced or degraded, indicating that autistic persons are over-reliant on visual feedback to correct error in grip force [11, 19, 36]. Autistic individuals also show reduced effects of somatosensory feedback manipulations (tendon vibration) on grip force control relative to NT controls, regardless of the resolution of visual feedback, suggesting that autistic persons have reduced reliance on secondary sensory inputs for issuing corrective updates to ongoing motor commands during sensory feedback guided motor behaviors [35].

In addition to atypical visuomotor feedback processing, findings from studies evaluating motor learning from sensory error suggest that motor memory may be atypical in autistic individuals, which may contribute to motor control deficits. Autistic individuals show a stronger adaptation to proprioceptive errors than NT controls during reaching [4, 16, 17]. These findings suggest that autistic persons may be more biased toward updating internal models based on proprioceptive feedback than NT controls but may be deficient in using visual feedback to update internal models. As a result, while autistic individuals may be over-reliant on visual feedback error information to update an ongoing motor command, they may not be as effective or accurate at using visual feedback to update internal models for future motor plans. Though to date, only one prior study has assessed the role of motor memory during sustained motor control in ASD.

During a precision gripping test in which visual feedback was removed and individuals were instructed to continue gripping at a constant force level, we previously documented a faster rate of force decay among autistic individuals relative to age-, IQ-, and sex-matched NT controls implicating deficient motor memory [38]. That study did not assess short-term visuomotor memory or analyze age-related effects on motor memory or visuomotor processes, though there are known developmental changes in visuomotor and memory guided motor control throughout childhood and into adulthood [32, 39, 40]. NT individuals (6–22 years old) show age-associated reductions in visually guided precision grip force regularity that are absent or significantly attenuated for memory guided grip force control [32, 39, 40]. Conversely, they show significant age-associated decreases in memory-guided grip force variability that are not observed in visually guided grip force control [32, 39]. The development of motor memory and visuomotor feedback processes and their contributions to sustained motor control issues in ASD are still not known.

Here, we aimed to assess differences in short-term visuomotor memory as well as longer-term sensorimotor memory and visual feedback processes during sustained fine motor control in ASD as a function of age. We examined age-associated differences in motor precision, variability, and regularity during visually guided and memory guided precision gripping to test the hypothesis that visual feedback and motor memory processes are different in ASD, especially during early childhood, suggesting delayed development. We expected autistic individuals to show stronger age-related reductions in force variability and regularity relative to NT controls during visually guided precision gripping, such that group differences are most pronounced in younger individuals, consistent with our prior findings [35]. To further assess motor memory, we examined the latency, slope, and magnitude of force decay following the removal of visual feedback (memory guided precision gripping). We expected that autistic individuals would show reduced duration of the short-term visuomotor memory, as well as faster force decay compared to NT controls, that are more pronounced at younger ages, consistent with the hypothesis that early issues with motor memory processes contribute to differences in sustained motor control in autism. To examine whether motor behaviors vary as a function of the severity of clinical traits, we also examined associations between motor control and clinical ratings of ASD severity, motor behavior, and IQ.

Methods

Participants

Fifty-four autistic participants (16 females) and 31 neurotypical (NT) controls (18 females) completed tests of precision gripping with their dominant hand (Table 1). The autistic and NT groups did not differ on age (t = -1.16, p = .25, range: 10–25 years). Autistic participants were recruited through our research registries comprised of individuals evaluated through the University of Kansas Health System who have consented to be contacted for research purposes, and through community advertisements. NT controls were recruited through community advertisements. ASD diagnoses were confirmed through diagnostic consensus by our study team based on Diagnostic and Statistical Manual of Mental Disorders, Edition 5 (DSM-5) [41] criteria and classification criteria from the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) [42] and Autism Diagnostic Interview – Revised (ADI-R) [43]. The ADOS-2 and ADI-R were administered by study clinicians and clinical trainees on the study team who are trained to research reliability. Autistic participants were excluded if they had a known genetic or metabolic disorder associated with ASD (e.g., Fragile X syndrome) or a full scale IQ (FSIQ) below 60 as measured using the Wechsler Abbreviated Scales of Intelligence, Second Edition (WASI-II) [44]. Since cognitive ability ranges widely in autism, this FSIQ cutoff was used to maximize representation of autistic individuals across the autism spectrum while ensuring that participants would be able to understand instructions and complete experimental tasks. As a conservative cutoff to ensure that NT participants did not show elevated autistic traits or have a confounding medical history, they were excluded if they scored ≥ 8 on the Social Communication Questionnaire [45], reported a history of psychiatric or neurologic disorders, had a family history of ASD in first- or second-degree relatives, had a family history of a developmental or learning disorder, psychosis, or obsessive compulsive disorder in first-degree relatives, or had a FSIQ below 85 as measured using the WASI-II. The FSIQ cutoff was used to limit the possibility that individuals included in the NT controls had underlying and possibly undiagnosed or unreported neurodevelopmental or neurologic conditions. Participants also were excluded if they had a history of head injury with neurological sequelae, birth injury, or seizure disorder. No participants were taking medications known to affect sensorimotor behavior, including antipsychotics, stimulants, or anticonvulsants at the time of testing [46]. All participants had corrected or uncorrected visual acuity of at least 20/40. Adult participants provided written informed consent after a complete description of the study, in accordance with the Declaration of Helsinki and the approved Institutional Review Board study protocol (IRB#: STUDY00140269). For participants under the age of 18 and adults who were under legal guardianship, a parent or legal guardian provided written informed consent, and the participant provided written assent. All study procedures were approved by the local Institutional Review Board.

Table 1 Demographic and clinical characteristics of autistic individuals (ASD) and neurotypical controls (NT)
Full size table

Clinical assessments

Participants completed the Wechsler Abbreviated Scales of Intelligence, Second Edition (WASI-II) to assess verbal IQ (VIQ), perceptual IQ (PIQ), and full-scale IQ. The WASI-II is validated for individuals aged 6–89 years. For this study we report scores for the full-scale IQ value that is calculated from all four of the administered subscales, with the exception of two autistic participants who did not complete all four subtests. For these two participants, the two subtest scores are used.

Autistic participants completed the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) [42] to confirm diagnosis and to quantify severity of autism for analysis. The ADOS-2 is a semi-structured play-based assessment of autistic traits that is the gold-standard diagnostic assessment for autism. It was conducted by a trained research reliable study clinician or clinical trainee. Participants in our study were administered module 2 (phrase speech), 3 (verbally fluent children), or 4 (verbally fluent adolescents and adults) according to age and language abilities. The composite severity score (ADOS-CSS) is a standardized score indicating severity of autism that can be compared across modules (higher scores indicate greater severity). The ADOS-CSS are reported and used for analyses.

Autistic participants completed the Repetitive Behavior Scale – Revised (RBS-R) [47, 48], to assess restricted and repetitive behaviors associated with autism. The RBS-R is a questionnaire that asks individuals to rate items from five categories of repetitive behavior (motor stereotypy, self-injurious behavior, compulsions, routines/sameness, and restricted interests). Parents or caregivers completed the RBS-R for participants under 18 years of age, and adult participants completed it as a self-report questionnaire. The overall summed severity scores were used for analyses. Higher scores indicate more severe repetitive behavior.

To determine handedness, participants completed the Annett Hand Preference Questionnaire (Annett) [49]. The Annett is a 12-item questionnaire that asks the participant which hand they prefer to use for various daily activities (e.g., writing, throwing, using a hammer, etc.), with endorsement of left hand, right hand, or either hand. If left hand is endorsed more than right hand, the participant is considered left-handed, and vice-versa for a classification of right-handed. If left and right hand are equally endorsed, the participant is classified as mixed-handed. Two NT controls scored as mixed-handed and two autistic participants did not complete the Annett, so their dominant hand for precision grip testing was determined based on which hand they used for writing. One NT control scored as left-handed on the Annett but self-reported as right-handed and used right hand for writing, so they completed precision grip testing with right hand. Handedness counts in Table 1 are based on the hand used to complete precision grip testing.

Participants completed the Bruininks-Oseretsky Test of Motor Proficiency, Second Edition (BOT-2) to assess motor abilities. The BOT-2 is a structured skill-based motor assessment. Participants completed a series of structured motor tasks from three areas: Fine Manual Control, Manual Coordination, and Body Coordination. Standard composite scores (t-scores) from the Fine Manual Control tests are reported and analyzed for the present study. Higher scores on the BOT-2 reflect better motor performance.

Precision grip testing

Participants completed tests of precision gripping while seated 52 cm from a 67 cm (27in) Samsung liquid crystal display monitor with a resolution of 1920 × 1080 and a 120 Hz refresh rate (Fig. 1). Participants sat with the elbow of their dominant hand comfortably positioned at 90° and their forearm resting in a custom arm brace fixed to the table to provide stability during testing. The participants used their thumb and index finger of their dominant hand to press against two opposing precision load cells that were secured to a custom grip device attached to the arm brace. A Coulbourn (V72-25) resistive bridge strain amplifier received analog signals from the load cells, which were converted to digital signals sampled at 100 Hz with a 16-bit analog-to-digital converter (NI USB-6341; National Instruments Corporation). During the first part of the study, ELFF load cells (ELFF-B4-100 N; Entran) 1.27 cm in diameter were used. Due to normal wear, the ELFF load cells were replaced with Honeywell load cells (Model 53, Honeywell International, Inc.) 1.5 cm in diameter during the course of the study. ELFF load cells were used for 35.3% of participants (35.5% of NT controls, 35.2% of autistic participants), and Honeywell load cells were used for 64.7% of participants (64.5% of NT controls, 64.8% of autistic participants. The voltage-to-Newton calibration was different for each type of load cell, so a correction calculated from known weights was applied to the force trace after data collection to correct for calibration errors. Additionally, load cell type was included as a covariate in our analyses to account for differences in load cell design and calibration effects, including the visual angles of feedback during the task (described below).

Fig. 1
figure 1

Task design. (A) During visually guided trials, participants see a target bar that turns from yellow to green to indicate that they should start pressing. Participants also view feedback of their force output (white bar) for the entire trial. (B) During memory guided trials, participants see visual feedback of their force output (white bar) and the green target bar for the first 3s of the trial, after which the white force bar disappears, and they are instructed to keep pressing at the same force level until the target turns red (12s later). (C) Example force output (dark blue) in Newtons (N) for a neurotypical participant during a visually guided trial. The grey line represents target force. (D) Example force output (dark blue) in Newtons (N) for a neurotypical participant during a memory guided trial with target force indicated by the grey line. The participants’ force usually begins to decay (black arrow) after the visual feedback disappears. (E) Example force output (dark blue) in Newtons (N) for an autistic participant during a visually guided trial. (F) Example force output (dark blue) in Newtons (N) for an autistic participant during a memory guided trial

Full size image

Prior to precision grip testing, participants completed an assessment of their maximum grip force, or maximum voluntary contraction (MVC) using their dominant hand. Participants completed three trials in which they were asked to press as hard as they could for three seconds. The average of the participant’s maximum force output across these trials comprised their MVC. Since the precision gripping tasks rely on sustained grip force, MVC trial averages were used to capture the maximum sustained force participants could maintain. To ensure that the measurement of MVC was valid, the three trials had to be within 15 N of each other to be included in the calculation, otherwise the participant was provided a break and the MVC test was readministered. For the precision gripping tasks, the target force was set at 45% of the participants’ MVC to account for differences in strength across participants.

During the precision gripping task, participants viewed two horizontal bars on the screen (Fig. 1A). A horizontal white force bar moved upward with increased force and downward with decreased force, and a static bar representing the target force was red during periods of rest. The target bar turned yellow to cue the participant to get ready for the start of the trial, and it turned green to cue the participant to begin pressing at the beginning of each trial. Participants were instructed to press the load cells as quickly as possible when the yellow target bar turned green and to keep pressing so that the white force bar stayed as steady as possible at the level of the green target bar until the target bar turned red, marking the end of the trial.

To test the impact of visual feedback and motor memory processes on grip force behavior, participants completed precision grip testing with and without visual feedback. During visual feedback trials, visual feedback was presented continuously throughout the 15s trial. Due to calibration differences for the two types of load cells that were used during the study and variance in the distance between the screen and the participants’ eyes during naturalistic viewing, the visual angles ranged from 0.74 to 1.15 degrees per 1 N increase in force output. Visual angles between 0.623 and 2.023 degrees result in small changes in the spatial amplitude of visual feedback and small changes in force error (Coombes et al. 2010). This range is also associated with stable and optimal variability and regularity of grip force in autistic individuals and NT controls [11].

For the trials without visual feedback (“memory guided” trials), the initial part of the trial was the same as for the visually guided trials – the target bar turned from yellow to green, and the participant pressed on the force transducers to match the white bar to the level of the green target bar (Fig. 1B). After three seconds, the white force output bar disappeared, and the participant was instructed to continue pressing at the same level until the target bar turned red (12s after the visual feedback was removed). Participants completed blocks of five trials of each condition using their dominant hand (5 trials x 2 conditions = 10 trials). Trials were 15s in duration and alternated with 15s rest periods. Each block was separated by 30s of rest. The target force was set to 45% of the participant’s MVC for all trials. The order of the blocks was pseudo-randomized and counterbalanced across participants.

Data processing

Grip force data were processed using custom applications developed by our lab in R and MATLAB (MathWorks, Inc., Natick, Massachusetts). Trials were excluded if the load cells were not properly re-zeroed between trials or if there were indications that the participant was not following instructions (e.g., the mean force exceeded twice the target force, there was evidence that the participants used fingers other than dominant hand index finger and thumb to press, participants stopped pressing during the trial). For the memory guided condition, trials were excluded if the participant did not reach a stable level of force output within ± 2 Newtons of the target force before visual feedback was removed. Based on these criteria, 19.8% of trials were excluded for the ASD group (12.3% of visually guided and 26.5% of memory guided feedback trials) and 1.3% of trials were excluded for the NT control group (1.9% of visually guided and 0.7% of memory guided trials). Trial-level data were averaged for each participant within each condition. Participants needed to have at least two useable trials of a condition for their data to be included. Four autistic participants were excluded from analyses due to insufficient data. An additional five autistic participants had insufficient data for only the memory guided condition, and one had insufficient data for only the visual feedback condition. Final analyses included 50 autistic participants and 31 NT controls with valid data for at least one condition.

To compare force output across conditions, the sustained force output for each trial was analyzed. To account for the differences in trial structure between conditions, only the last 12s of each trial were used for analysis of sustained force output. This 12s phase corresponds to the segment of the memory guided trials where visual feedback was not available and the analogous segment of the trials with visual feedback.

The force traces for each trial were low-pass filtered via a double-pass fourth-order Butterworth filter at a low-pass cutoff of 15 Hz following previous studies from our lab [20, 35]. Force accuracy was calculated as a proportion of the target force (mean of the sustained force data divided by the target force), such that values close to 1 reflect greater accuracy. For the other dependent variables, sustained force data were linearly detrended to account for drift in participants’ force output over the duration of the trial that is due to a change in force accuracy across the trial rather than moment-to-moment fluctuations sustained grip force that reflect variability in grip force control. Force variability was calculated as the standard deviation (SD) across the force time series, and it reflects how well participants are able to maintain a stable force level. To test the time dependent regularity of the force time series, sample entropy (SampEn) was calculated for each trial [50, 51]. SampEn is defined as the natural logarithm of the conditional probability that two similar sequences of m data points in a timeseries of a given length (N) remain similar within a tolerance level (r) at the next data point in the series. SampEn returns a value between 0 and 2. Lower values of SampEn indicate greater regularity of the timeseries. For example, a sine wave, with its predictable oscillating pattern, would have a SampEn value near 0. Whereas, an unpredictable timeseries – like feedback guided grip force, where the timing of changes in force output is determined, in part, by the timing of sensory feedback error information – show high SampEn (low regularity). Parameter settings for SampEn calculations were m = 2 and r = .2 × SD of the timeseries. The timeseries length was 1200 data points (12s sampled at 100 Hz). The sampenc.m function (for MATLAB) from the PhysioNet Toolbox [52, 53] was used to calculate SampEn values for each trial.

To characterize the trajectory of force output after visual feedback was removed during the memory guided trials, models were fit to the force traces for each trial. The model consisted of three segments: (1) a horizontal line fit to the stable force output at the beginning of the trial, starting before visual feedback was removed, (2) a logarithmic function fit to the decay in force output after visual feedback was removed, (3) for trials where participants reached a stable force output after their force decayed and before the end of the trial, a horizontal line was fit to the data to model this secondary stable force output. The cut points between the horizontal and logarithmic models were determined using an automated fitting process. The nls function in R was used to fit a piecewise model to the data, incorporating two cut points. Initial estimates for the cut points were provided based on visual inspection of the force trace, and the final values were optimized to minimize the sum of squared residuals across the time course from the end of the initial stable force output (first horizontal segment) to the end of the trial.

Latency, slope, and magnitude of the force decay were analyzed. Decay latency was measured as the difference between the removal of visual feedback and the beginning of the logarithmic decay model segment. The decay slope was calculated as the log slope of the logarithmic decay model segment. For 2.4% of trials (ASD: 2.5% and NT 2.0%), the force showed little to no decay (visual inspection) and was best modeled using a linear fit rather than a logarithmic fit. The determination to use a linear fit was made based on at least one of the following criteria: (a) the program could not fit a logarithmic function to the force timeseries for the trial, (b) the logarithmic fit was < 500ms in duration and < 20% of the target force in magnitude, and/or (c) the three-segment model using a logarithmic fit over- or under-estimated the force at the start or end of the logarithmic segment, resulting in greater residuals across the trial than the linear model fit. Descriptive statistics were calculated and reported for the trials with linear decay slopes, but these trials were not factored into the participants’ trial averages or the linear regression models, as they are not directly comparable to the trials with logarithmic slopes. The magnitude of decay was calculated by taking the difference between the force output at the onset of the logarithmic (or linear) decay model segment and the end of the decay model segment or the end of the trial, if force did not stabilize before the end of the trial. This value was then converted to percentage of target force by dividing the raw difference by the participant’s target force to account for the differences in initial force.

Statistical analysis

Force accuracy, SD, and SampEn were analyzed using separate linear multilevel mixed effects models (MLM) [54, 55] with the lme4 package in R version 4.0.0 [54]. MLM allows for the analysis of within- and between-subjects fixed effects while allowing within-subjects effects to vary randomly and is robust to missing data. Task condition (visually guided, memory guided) was included as a level 1 predictor. Group (ASD, NT) and age were included as level 2 predictors. For all dependent variables, the models also included two two-factor covariates to account for different load cells used during testing (“load cell type”) and participant sex (male, female). Random intercepts of participant also were included in the models.

Initial models for force accuracy, SD, and SampEn included the three-way interaction of group x task condition (visually guided, memory guided) x age, all relevant two-way interactions and main effects terms, as well as the covariates for load cell type and sex. To maintain the most parsimonious models possible, other 3-way interactions were not included. Models were fit using the maximum likelihood approach to allow for model comparisons. Terms were removed systematically, and model fit was compared between the previous model and the model with the removed term using likelihood ratio tests. Terms that did not significantly improve model fit (p < .05), based on the model comparisons, were not included in the final models. Satterthwaite’s method was used to calculate degrees of freedom for the final model and post hoc comparisons [56]. Due to the inherent challenge in determining denominator degrees-of-freedom and calculating p-values for MLMs, we treated the t-value as a z-value and used a z > 1.96 threshold as an additional guideline for determining whether terms explained significant variance in the model [56].

The latency, slope, and magnitude of force decay following the removal of visual feedback were analyzed using separate linear regression models with the lm (linear model) function in R. Group (ASD, NT), age, and the group x age interaction were included as predictors. For all dependent variables, the models also included covariates for load cell type and sex.

Simple coding was used for group (NT = -0.5, ASD = 0.5), task condition (memory guided = -0.5, visually guided = 0.5), and sex (male = -0.5, female = 0.5). Age was log10 transformed. SD, SampEn, and decay magnitude were log10 transformed and decay slope and decay latency were square root transformed to correct for skewed distributions. Based on this coding system, the intercept for each model represented the grand mean of the sample.

Pearson correlations were used to assess the relation between experimental variables and ASD symptom severity measured using the ADOS Composite Severity Score (ADOS-CSS) as well as repetitive behaviors measured using the RBS-R. Pearson correlations also were used to assess the relation between visuomotor and motor memory behaviors and IQ for each group.

To determine whether visual feedback guided motor control and motor memory during precision gripping relate to clinically relevant fine motor skills, Pearson correlations were run between the motor variables and BOT-2 Fine Manual Control subscale scores. For each set of correlations, p-values were adjusted using false discovery rate (FDR) to limit Type I error.

Results

Force accuracy

Figure 2 shows results for force accuracy, and the model summary is reported in Table 2. There was a significant group x age interaction (β = 0.40, R2 = 0.042, t79.9 = 2.32, p = .023). Follow-up comparisons revealed that autistic individuals showed a greater increase in accuracy with age than the NT group (slopeASD = 0.56 ± 0.12, slopeNT = 0.15 ± 0.14). Across ages and groups, force accuracy was greater in the visually guided condition relative to the memory guided condition (β = − 0.22, R2 = 0.610, t76.7 = -15.74, p < .0001).

Fig. 2
figure 2

Force accuracy. Age (years; log10 scale) associations with force accuracy (proportion of mean force to target force) for (A) the autism (ASD; red circles) and (B) the neurotypical (NT; blue triangles) groups. The regression lines represent the significant group x age effects (across conditions), though data for visually guided (Vis; solid points) and memory guided (Mem; empty points) are also shown. (C) Force accuracy depicted as a function of group (NT: blue triangles; ASD: red circles) during visually guided (Vis; filled points) and memory guided (Mem; empty points) precision gripping. The large points represent group x condition marginal means adjusted for random intercepts of subject in the LMER models. The * represents the significant condition main effect. Error bands (A, B) and bars (C) represent the 95% confidence intervals from the MLM models after accounting for random intercepts of participant

Full size image
Table 2 Linear mixed effects model summary for force accuracy
Full size table

Force variability

Results of the model for force SD are summarized in Table 3; Fig. 3. There was a significant group x age interaction (β = -1.99, R2 = 0.127, t77.4 = -3.82, p = .0003). Follow-up comparisons revealed that the ASD group showed a stronger age-related decrease in force SD than the NT group (slopeASD = -1.11 ± 0.36, slopeNT = 0.88 ± 0.42). Overall, the ASD group showed higher force SD than the NT group (β = 2.41, R2 = 0.131, t77.5 = 3.89, p = .0002). Sex was a significant covariate (β = -0.14, R2 = 0.51, t76.6 = -2.32, p = .023), such that males showed greater force variability than females (meanM = 0.20 ± 0.04; meanF = 0.06 ± 0.04). No effects of task condition were observed.

Table 3 Linear mixed effects model summary for force variability (SD)
Full size table
Fig. 3
figure 3

Force Variability in Newtons (N). Age (years; log10 scale) associations with force standard deviation in Newtons (N; log10 scale) for (A) the autism (ASD; red circles) and (B) neurotypical control (NT; blue triangles) groups. The regression lines represent the significant group x age effects (across conditions), though data for visually guided (Vis; solid points) and memory guided (Mem; empty points) are also shown. (C) Force variability depicted as a function of group (NT: blue triangles; ASD: red circles) during visually guided (Vis; filled points) and memory guided (Mem; empty points) precision gripping. The large points represent group x condition marginal means adjusted for random intercepts of subject in the LMER models. The * represents the significant group main effect. Error bands (A, B) and bars (C) represent the 95% confidence intervals from the MLM models after accounting for random intercepts of participant

Full size image

Force regularity

Results of the model for force SampEn are summarized in Table 4; Fig. 4. There were significant group x task condition (β = 0.11, R2 = 0.021, t74.9 = 2.23, p = .029) and group x age interactions (β = 0.95, R2 = 0.063, t80.2 = 2.73, p = .008). Follow-up comparisons revealed that the ASD group showed a greater age-related increase in force SampEn than the NT group (slopeASD = 1.17 ± 0.24, slopeNT = 0.22 ± 0.28), and the ASD group showed lower SampEn than the NT group only in the visually guided condition (meanASD = -0.61 ± 0.03, meanNT = -0.50 ± 0.03). There was a significant task condition x age interaction (β = − 0.52, R2 = 0.021, t77.3 = -2.21, p = .030). Follow-up comparisons revealed that age related increases in SampEn are stronger in the visually guided condition relative to the memory guided condition (slopeVisual = 0.95 ± 0.23, slopeMemory = 0.43 ± 0.22).

Table 4 Linear mixed effects model summary for force regularity (SampEn)
Full size table
Fig. 4
figure 4

Force regularity. Age (years; log10 scale) associations with force SampEn (unitless; log10 scale) for (A) the autism (ASD; red circles) and (B) neurotypical control (NT; blue triangles) groups. Higher SampEn corresponds to lower regularity. The regression lines represent the significant group x age effects (across conditions), though data for visually guided (Vis; solid points) and memory guided (Mem; empty points) are also shown. (C) Force regularity depicted as a function of group (NT: blue triangles; ASD: red circles) during the visually guided feedback (Vis; solid points) and memory guided (Mem; empty points) conditions. The large points represent group x condition marginal means adjusted for random intercepts of subject in the LMER models. The * represents the significant effect of group in the visually guided condition (group x condition interaction). Error bands (A, B) and bars (C) represent the 95% confidence intervals from the MLM models after accounting for random intercepts of subject

Full size image

Slope of force decay

The results of the linear model for the slope of force decay in the memory guided condition are summarized in Table 5; Fig. 5. In the model for the slope of force decay there was a trend for the ASD group to have a steeper (more negative) decay slope than the NT group (β = − 0.07, R2 = 0.052, t70 = -1.96, p = .054) (meanASD = -0.51 ± 02, meanNT = -0.45 ± 03). In the ASD group, seven of 275 (2.5%) trials were fit with a linear slope (mean = − 0.00006 ± 0.00011), and in the NT control group, three of 150 (2.0%) trials were fit with a linear slope (mean = − 0.00023 ± 0.00025), indicating that each group had a comparable proportion of trials that showed little to no decay.

Table 5 Linear model summary for decay slope
Full size table
Fig. 5
figure 5

Decay slope and magnitude. (A) Slope of the force decay (square root scale) following the removal of visual feedback (memory guided condition) for the autism (ASD; red circles) and neurotypical control (NT; blue triangles) groups. (B) Magnitude of the force decay (log10 scale) following the removal of visual feedback (memory guided condition) for the ASD (red circles) and NT control (blue triangles) groups. Large points represent group means adjusted for random intercepts of subject in the MLM models. The * represents a significant effect of group. Error bars represent the 95% confidence intervals from the MLM models after accounting for random intercepts of subject

Full size image

Decay latency

In the model for the latency of force decay onset, no terms were significant. For reference, Table 6 shows the model summary for the model containing nonsignificant main effects of group and age, as well as the covariate for the different load cells used over the course of the study. To interpret findings of the decay onset latency in the context of short-term visuomotor memory processes, we calculated the group medians of the raw (untransformed) latency values. Median decay onset latencies were.842s for ASD and.669s for NT. Median values are reported due to a rightward skew in the distribution of the untransformed latency data (skewness = 1.35).

Table 6 Linear model summary for decay latency
Full size table

Decay magnitude

Results of the model for the magnitude of the force decay in the memory guided condition are summarized in Table 7; Fig. 5. The magnitude of decay was calculated as a proportion of the participant’s target force with larger values representing greater decay. The values were log10 transformed to correct for skewed distributions. The ASD group had a greater magnitude of force decay than the NT group (β = − 0.12, R2 = 0.09, t70 = 2.55, p = .013) (meanASD = -0.29 ± 0.03, meanNT = -0.41 ± 0.04). Additionally, the magnitude of force decay decreased significantly with age (β = -0.89, R2 = 0.153, t70 = -3.56, p = .0007).

Table 7 Linear model summary for decay magnitude
Full size table

Relation to clinical features

Autism severity

More severe clinical ratings of autism on the ADOS-CSS were positively correlated with SD (r = .49, pFDR = 0.009) and negatively correlated with force accuracy (r = −.34, pFDR = 023) in the visually guided condition (Fig. 6). ADOS-CSS was negatively correlated with decay slope in the memory guided condition (r = −.41, pFDR = 0.029). No other dependent variables correlated with ADOS-CSS for either condition. Scores on the RBS-R were not significantly correlated with any measure of grip force control. To ensure that these results are not a consequence of sampling bias (e.g., autistic individuals with less intense autistic traits were not more likely to participate than those with more intense traits), we ran an exploratory correlational analysis of ADOS-CSS with age. ADOS-CSS and age were not significantly correlated (r = −.012, p = .93).

Fig. 6
figure 6

Relation of ASD Symptomatology to grip control. Association between Autism Diagnostic Observation Schedule – Composite Severity Scores (ADOS-CSS) for the autism group (ASD; red circles) and (A) force accuracy (proportion of mean force to target force) and (B) force variability (standard deviation; log10 scale) in Newtons (N) during visually guided (Vis; solid points) precision gripping. (C) Association between ADOS-CSS scores for the ASD group (red circles) and the slope of logarithmic force decay (square root scale) during memory guided precision gripping (Mem; empty points). Error bands represent standard error

Full size image

IQ

FSIQ, VIQ, and PIQ were not significantly correlated with any of the dependent variables for either group or feedback condition.

Fine motor behavior

For the ASD group only, BOT Fine Manual Control scores were negatively correlated with force SD in the visually guided condition (r = −.50, pFDR = 0.014). The ASD group also showed trending correlations of BOT Fine Manual Control scores with force accuracy (r = .37, pFDR = 0.095) in the visually guided condition after FDR corrections. No other dependent variables were significantly correlated with BOT Fine Manual Control scores in either group or condition.

Discussion

The present study assessed the unique contributions and age-dependent patterns of visual feedback processing and motor memory to well-established differences in sustained precision motor control in autistic individuals. We observed a trend for autistic individuals to show a faster rate of decay in their grip force than NT controls following the removal of visual feedback, consistent with our prior findings [38]. These results suggest that for some autistic individuals, deficient motor memory may contribute to sensorimotor impairments. Additionally, we observed greater age-associated improvements in sustained motor control in autistic individuals relative to NT individuals reflecting reduced abilities during early childhood followed by normalization of sensorimotor control during early adolescence and adulthood. These results suggest sensory feedback processes that contribute to sensorimotor precision follow a protracted course of maturation in ASD. Age-related improvements were not specific to visually guided or memory guided motor control, suggesting that developmental delays may impact multiple motor control mechanisms in ASD.

Motor memory and visuomotor feedback differences in ASD

During memory-guided grip control, force decayed at a greater rate (trend) and to a greater extent in autistic individuals compared to NT controls. These findings are consistent with our prior study [38]. However, we did not find group differences in the latency of force decay after the removal of visual feedback. The visuomotor memory lasts 0.5–1.5s following the removal of feedback, so latencies that exceed 0.5–1.5s would reflect the involvement of motor memory processes for maintaining force output that are distinct from short-term visuomotor memory processes [27, 29]. For both groups, decay onset latencies were within the range of visuomotor memory (medianASD =.842s; medianNT =.669s), supporting the interpretation that short-term visuomotor memory is not impaired in ASD in the context of visuomotor control. Together, these findings indicate that short-term visuomotor motor memory processes are not impacted in ASD. Instead, results support the hypothesis that autistic individuals show deficiencies in longer-term sensorimotor memory processes. More severe force decay and the trend for more rapid force decay of the longer-term motor memory in autistic individuals may result from reduced reliance on somatosensory feedback during visually guided precision gripping relative to NT controls [35]. In NT individuals, cutaneous inputs are critical for maintaining grip force control in the absence of visual feedback and establishing longer-term sensorimotor memories [30, 31]. Autistic individuals are less reliant on somatosensory feedback than NT individuals during visually guided gripping [35], which may impair their ability to update internal models based on somatosensory input and, therefore, limit the formation of the sensorimotor memory. This difference may result in autistic individuals’ reduced ability to maintain force output at a consistent level after visual feedback is removed.

Autistic individuals showed greater regularity of grip force than NT controls, specifically during visually guided precision gripping, consistent with previous studies [11, 35]. Regularity represents the degrees of freedom – including physical (muscles and joints) and neurophysiological (motor control brain networks and processes) – of motor control. Lower regularity (higher entropy) indicates greater processing and integration of sensory feedback information for updating ongoing motor behaviors. Our finding of reduced entropy of sustained visuomotor control in autism suggests that autistic individuals do not integrate visual feedback as effectively as NT individuals to optimize control of precision movements [35]. Force regularity was the only variable for which condition effects varied as a function of group. This could indicate that force regularity is more sensitive to group differences in visual feedback and motor memory processes than motor variability or accuracy, as regularity has been shown to be more sensitive to slight changes in visual feedback during visually guided gripping than variability [27]. Alternatively, these findings could indicate that autistic and NT individuals may use different mechanisms to decrease force regularity when visual feedback is present, though, future studies are needed to test these hypotheses.

Age-related differences in visuomotor control among autistic individuals

Autistic individuals show stronger age associations than NT individuals for all sustained force variables. In the ASD group, force variability and regularity decreased with age and force accuracy increased with age, while these metrics were stable across ages in the NT group. Specifically, younger autistic children showed poorer force control relative to NT children, while adolescents and young adults were comparable across groups. These findings are consistent with our prior study of sustained visually guided precision gripping showing that sensorimotor impairments are more robust in ASD at younger ages [35]. However, in a separate study that included a broader age range (5–35 years), we found stronger age-related improvements in grip force regularity in NT individuals relative to autistic individuals and comparable age-associations in force variability across groups at the visual angle and target force level used in the present study [36]. These findings are likely driven by the rapid maturation of motor processes in NT development that occur at younger ages than were included in the present study, as well as variation across autistic individuals in the extent to which sensorimotor processes are disrupted. Together, these findings suggest that autistic individuals have delayed development of sustained, precision motor control processes relative to NT individuals. Age associations of sustained force control outcomes did not differ across task conditions, indicating that delayed motor development impacts multiple motor processes and is not specific to visuomotor control.

While literature tracking developmental trajectories of motor function in ASD from childhood to adulthood is sparse, some studies suggest a normalization or improvement in motor skills with age in autistic individuals. Young autistic children showed elevated variability in stride velocity, stride time, and stride length relative to NT children during gait, but these differences were not present during adolescence [57]. Autistic children show deficits in reach-to-grasp behaviors, including larger normalized jerk and more motor units than NT children [58], but a separate study using similar methods did not find differences between autistic and NT adults [59]. While these studies are consistent with our findings of age-related normalization of motor skills in autistic individuals, longitudinal studies are needed to understand the developmental trajectories of discrete sensorimotor processes in autism.

Neurophysiology of visuomotor and motor memory processes

Our findings that autistic individuals show deficits in visually guided and memory guided precision gripping are consistent with neuroimaging studies showing atypical activation and connectivity in cerebellar-posterior parietal cortical circuits that translate sensory information into reactive motor adjustments, as well as frontal-cerebellar networks involved in cognitive processing [23, 60]. In functional MRI studies of precision gripping, we found increased activation of cortical sensory and motor control regions (e.g., supplementary motor area, superior parietal lobule, middle frontal gyrus, inferior frontal gyrus) in ASD that was more pronounced when visual feedback was amplified as well as reduced parietal-cerebellar functional connectivity at high force levels, and reduced parietal-premotor and parietal-putamen functional connectivity across visual feedback conditions (i.e., feedback presented at different visual angles). These results indicate autistic individuals show reduced modulation of circuits involved in sensory processing and precision motor control as well as reduced integration of multimodal sensory feedback with motor control systems for making dynamic adjustments to motor output and reducing motor error [19, 61]. Further, these studies also found differential age-associated changes in cerebellar-cortical networks in autistic compared to NT individuals (ages 8–33 years). These studies further found evidence of delayed or protracted development of these cerebellar-cortical networks in autism. Autistic individuals had stronger age-associated increases in functional connectivity of cerebellar-cortical networks involved in sensory processing (visual cortex, inferior parietal cortex) and motor control (primary motor and premotor cortices) than NT individuals, showing greater differences in functional connectivity at younger ages. For autistic individuals, increases in cerebellar-cortical connectivity with increased age were associated with reduced force variability and regularity, consistent with the present finding of age-associated increases in precision motor control in ASD.

Neuroimaging studies of ASD have found atypical activation and functional connectivity of brain systems involved in motor memory. In NT individuals, prefrontal cortex is selectively responsive to a change from visually guided to memory guided precision gripping (compared to a change from high to low resolution visual feedback) and is associated with short-term visuomotor memory [62]. Anterior cingulate cortex also shows selective responses to a change from visually guided to memory guided precision gripping, but it responds after the temporal capacity of short-term visuomotor memory, suggesting its involvement in longer-term motor memory processes. At rest, autistic individuals show reduced functional connectivity between cerebellum and prefrontal cortex [23, 60], indicating atypical function of networks involved in short-term visuomotor memory. Additionally, we found reduced functional connectivity between primary motor cortex and anterior cingulate cortex in ASD during visuomotor behavior, specifically under more challenging (higher target force) conditions [61], indicating that under conditions that place greater demand on the motor system, autistic individuals show atypical activation of networks involved in longer-term motor memory. While these studies demonstrate that autistic individuals show atypical function of brain circuits that have been implicated in motor memory, future neuroimaging studies during memory guided motor behavior in ASD are needed to determine whether behavioral metrics of motor memory impairments in ASD are associated with atypical activation or connectivity within these brain networks.

Clinical associations

Clinically rated ASD severity was associated with force regularity and accuracy during visually guided precision gripping and rate of force decay during memory guided precision gripping consistent with the hypothesis that differences in visuomotor integration and motor memory may both contribute to the development of autism. These clinical associations were not observed with the RBS-R or IQ, indicating sensorimotor and motor memory differences may be selectively associated with social-communication differences in ASD. ASD trait severity was derived from the ADOS-CSS. While the ADOS-CSS is an indicator of overall autistic trait severity, it is derived predominantly from items relating to social-communication with few restricted, repetitive behavior items, and therefore, it largely reflects the severity of social-communication traits. Sensorimotor integration is fundamental to the development of social and communication behaviors [63, 64]. Early development of sensorimotor processes including the integration of sensory feedback and updating internal models based on the sensory consequences of the individual’s movements are necessary for mapping others’ behavior (e.g., social gestures, facial expressions, speech) onto one’s own sensorimotor representations [63, 64]. This process provides information on the timing and intent of the other person’s movements and facilitates the learning of social-communication behaviors through imitation, interpreting others’ social-communication behaviors, and understanding the dynamics of reciprocal social interaction. Given that sensorimotor deficits are some of the earliest indicators of atypical development in ASD, it is possible that they contribute to later deficits in social interaction and communication [65,66,67,68]. Future studies that specifically characterize social-communication traits and their relation to visual motor and motor memory development in autism will be important for elucidating these mechanisms.

Fine manual motor skills as measured by the BOT-2 were selectively associated with visually guided precision grip control. The fine motor items on the BOT-2 are heavily reliant on visual feedback processes (e.g., coloring inside the lines, cutting around a circle, tracing a path, drawing a shape from a visual reference), so the finding that visually guided precision grip control was correlated with the BOT-2 Fine Manual Control scores is expected. Memory guided precision gripping abilities were not associated with fine motor abilities as measured on the BOT-2. This result is consistent with hypotheses; specifically, the BOT-2 items are not timed; they do not require continuous motor output that would rely on short-term visuomotor memory processes, and they do not require individuals to generate movements from memory, suggesting that they do not depend on longer-term motor memory processes.

Limitations

Very few studies have characterized age-related patterns of sensorimotor behavior and processing across broad ranges in autism. The present study used a cross-sectional design to assess age-associated differences in visuomotor feedback and motor memory processes during sustained precision gripping in autism. However, to better understand developmental trajectories of sensorimotor function in autism, longitudinal studies of multiple distinct sensorimotor processes and effector systems are needed. Additionally, while we observed differences in memory guided motor control in autistic individuals relative to NT controls, we were not able to determine how other sensorimotor control processes – including reliance on somatosensory feedback, attention, or motor learning – may be impacting behavior in this condition. Our prior findings suggested that autistic individuals have reduced reliance on somatosensory feedback for sustained precision gripping than NT controls, which may contribute to the greater and more rapid decay of force we observed during memory guided control in autistic individuals in the present study. Additionally, a separate study of memory guided motor control found that adults with attention deficit hyperactivity disorder (ADHD) showed a faster rate of force decay (steeper negative slope) following the removal of visual feedback relative to NT adults, and the faster rate of force decay in ADHD was associated with ADHD trait severity [69]. While we did not assess ADHD traits in our sample, we aimed to reduce attention related effects during task administration by monitoring participants’ attention to the stimulus display, as well as during data processing by excluding trials for which participants displayed pressing behavior that indicated that they were not performing the task as instructed (e.g., stopped pressing at any point during the trial). These procedures minimize but do not preclude attention related effects on force control during the memory guided condition, though our findings that force variability was not influenced by condition suggest that participant attention was comparable across conditions. However, future studies are needed to determine the effects of reduced somatosensory feedback processes, attention, or learning on memory guided sustained precision motor control in ASD. Additionally, it is possible that autistic adults with less intense autistic traits are more likely to volunteer for research than autistic adults with more intense autistic traits, though age was not associated with ADOS-CSS, suggesting that this scenario is unlikely in the present sample.

Conclusion

The present study demonstrates that differences in visual feedback and motor memory processing contribute to sustained precision motor control deficits in autistic individuals, though short-term visuomotor memory is unaffected. Precision motor control deficits in autism are most pronounced in childhood and normalize in adolescence and early adulthood suggesting that autistic individuals have protracted development of precision motor control. Our findings provide novel insights into neurodevelopmental processes underlying precision sensorimotor behavior in ASD.

Data availability

Data is provided within the manuscript or supplementary information files. Raw data will be made available to researchers upon reasonable request to the corresponding author(s).

Abbreviations

ASD:

Autism spectrum disorder

NT:

Neurotypical

IQ:

Intelligence quotient

ADOS-2:

Autism Diagnostic Observation Schedule, Second Edition

ADI-R:

Autism Diagnostic Interview – Revised

FSIQ:

Full-Scale IQ

WASI-II:

Wechsler Abbreviated Scales of Intelligence – Second Edition

OR:

Odds ratio from Fisher’s Exact test

M:

Male

F:

Female

L:

Left-handed

R:

Right-handed

SD:

Standard deviation

ADOS-CSS:

Autism Diagnostic Observation Scale – Composite Severity Score

RBS-R:

Repetitive Behavior Scale – Revised

VIQ:

Verbal intelligence quotient

PIQ:

Perceptual intelligence quotient

BOT-2:

Bruininks-Oseretsky Test of Motor Proficiency, 2nd Edition

MVC:

Maximum voluntary contraction

SampEn:

Sample entropy

MLM:

Muli-level mixed effects model

FDR:

False discovery rate

SE:

Standard error

Sqrt:

Square root

References

  1. Jong MD, Punt M, Groot ED, Minderaa RB, Hadders-Algra M. Minor neurological dysfunction in children with autism spectrum disorder. Dev Med Child Neurol. 2011;53(7):641–6.

    Article  PubMed  Google Scholar 

  2. Estes A, Zwaigenbaum L, Gu H, St. John T, Paterson S, Elison JT, et al. Behavioral, cognitive, and adaptive development in infants with autism spectrum disorder in the first 2 years of life. J Neurodev Disord. 2015;7(1):24.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Hannant P, Cassidy S, Tavassoli T, Mann F. Sensorimotor Difficulties Are Associated with the Severity of Autism Spectrum Conditions. Front Integr Neurosci. 2016;10. Available from: http://journal.frontiersin.org/Article/https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fnint.2016.00028/abstract

  4. Haswell CC, Izawa J, Dowell LR, Mostofsky SH, Shadmehr R. Representation of internal models of action in the autistic brain. Nat Neurosci. 2009;12(8):970–2.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. LeBarton ES, Iverson JM. Fine motor skill predicts expressive Language in infant siblings of children with autism. Dev Sci. 2013;16(6):815–27.

    Article  PubMed  Google Scholar 

  6. Ravizza SM, Solomon M, Ivry RB, Carter CS. Restricted and repetitive behaviors in autism spectrum disorders: the relationship of attention and motor deficits. Dev Psychopathol. 2013;25(03):773–84.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Sacrey LAR, Bennett JA, Zwaigenbaum L. Early infant development and intervention for autism spectrum disorder. J Child Neurol. 2015;30(14):1921–9.

    Article  PubMed  Google Scholar 

  8. Wilson RB, Vangala S, Reetzke R, Piergies A, Ozonoff S, Miller M. Objective measurement of movement variability using wearable sensors predicts ASD outcomes in infants at high likelihood for ASD and ADHD. Autism Res. 2024;17(6):1094–105.

    Article  PubMed  Google Scholar 

  9. Miller HL, Templin TN, Fears NE, Sherrod GM, Patterson RM, Bugnariu NL. Movement smoothness during dynamic postural control to a static target differs between autistic and neurotypical children. Gait Posture. 2023;99:76–82.

    Article  PubMed  Google Scholar 

  10. Gong L, Liu Y, Yi L, Fang J, Yang Y, Wei K. Abnormal gait patterns in autism spectrum disorder and their correlations with social impairments. Autism Res. 2020;13(7):1215–26.

    Article  PubMed  Google Scholar 

  11. Mosconi MW, Mohanty S, Greene RK, Cook EH, Vaillancourt DE, Sweeney JA. Feedforward and feedback motor control abnormalities implicate cerebellar dysfunctions in autism spectrum disorder. J Neurosci. 2015;35(5):2015–25.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Schmitt LM, Cook EH, Sweeney JA, Mosconi MW. Saccadic eye movement abnormalities in autism spectrum disorder indicate dysfunctions in cerebellum and brainstem. Mol Autism. 2014;5(1):47.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Abrams G, Jayashankar A, Kilroy E, Butera C, Harrison L, Ring P, et al. Differences in praxis errors in autism spectrum disorder compared to developmental coordination disorder. J Autism Dev Disord. 2024;54(3):1113–29.

    Article  PubMed  Google Scholar 

  14. Wang Z, Magnon GC, White SP, Greene RK, Vaillancourt DE, Mosconi MW. Individuals with autism spectrum disorder show abnormalities during initial and subsequent phases of precision gripping. J Neurophysiol. 2015;113:1989–2001.

    Article  PubMed  Google Scholar 

  15. Takarae Y, Minshew NJ, Luna B, Krisky CM, Sweeney JA. Pursuit eye movement deficits in autism. Brain. 2004;127(12):2584–94.

    Article  PubMed  Google Scholar 

  16. Izawa J, Pekny SE, Marko MK, Haswell CC, Shadmehr R, Mostofsky SH. Motor learning relies on integrated sensory inputs in ADHD, but Over-Selectively on proprioception in autism spectrum conditions: distinct patterns of motor memory in autism. Autism Res. 2012;5(2):124–36.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Marko MK, Crocetti D, Hulst T, Donchin O, Shadmehr R, Mostofsky SH. Behavioural and neural basis of anomalous motor learning in children with autism. Brain. 2015;138(3):784–97.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Cerliani L, Mennes M, Thomas RM, Martino AD, Thioux M, Keysers C. Increased functional connectivity between subcortical and cortical Resting-State networks in autism spectrum disorder. JAMA Psychiatry. 2015;72(8):767–77.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Lepping RJ, McKinney WS, Magnon GC, Keedy SK, Wang Z, Coombes SA et al. Visuomotor brain network activation and functional connectivity among individuals with autism spectrum disorder. Hum Brain Mapp. 2021.

  20. McKinney WS, Kelly SE, Unruh KE, Shafer RL, Sweeney JA, Styner M, et al. Cerebellar volumes and sensorimotor behavior in autism spectrum disorder. Front Integr Neurosci. 2022;16:821109.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Oldehinkel M, Mennes M, Marquand A, Charman T, Tillmann J, Ecker C, et al. Altered connectivity between cerebellum, visual, and Sensory-Motor networks in autism spectrum disorder: results from the EU-AIMS longitudinal European autism project. Biol Psychiatry Cogn Neurosci Neuroimaging. 2018;4(3):260–70.

    PubMed  Google Scholar 

  22. Unruh KE, Martin LE, Magnon G, Vaillancourt DE, Sweeney JA, Mosconi MW. Cortical and subcortical alterations associated with precision visuomotor behavior in individuals with autism spectrum disorder. J Neurophysiol. 2019;122(4):1330–41.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Wang Z, Wang Y, Sweeney JA, Gong Q, Lui S, Mosconi MW. Resting-State Brain Network Dysfunctions Associated With Visuomotor Impairments in Autism Spectrum Disorder. Front Integr Neurosci. 2019 [cited 2019 Sep 16];13. Available from: https://www.frontiersin.org/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fnint.2019.00017/full

  24. van den Bos N, Houwen S, Schoemaker M, Rozenblum S. Using structural equation modeling to analyze handwriting of children and youth with autism spectrum disorder. J Autism Dev Disord. 2024;54(1):155–67.

    Article  PubMed  Google Scholar 

  25. Sparaci L, Northrup JB, Capirci O, Iverson JM. From using tools to using Language in infant siblings of children with autism. J Autism Dev Disord. 2018;48(7):2319–34.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Gowen E, Earley L, Waheed A, Poliakoff E. From one big clumsy mess to a fundamental part of my character. Autistic adults’ experiences of motor coordination. PLoS ONE. 2023;18(6):e0286753.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Slifkin AB, Vaillancourt DE, Newell KM. Intermittency in the control of continuous force production. J Neurophysiol. 2000;84(4):1708–18.

    Article  CAS  PubMed  Google Scholar 

  28. Vaillancourt DE, Thulborn KR, Corcos DM. Neural basis for the processes that underlie visually guided and internally guided force control in humans. J Neurophysiol. 2003;90(5):3330–40.

    Article  PubMed  Google Scholar 

  29. Vaillancourt DE, Russell DM. Temporal capacity of short-term visuomotor memory in continuous force production. Exp Brain Res. 2002;145(3):275–85.

    Article  PubMed  Google Scholar 

  30. van Polanen V, Davare M. Sensorimotor Memory Biases Weight Perception During Object Lifting. Front Hum Neurosci. 2015 Dec 23 [cited 2025 Feb 13];9. Available from: https://www.frontiersin.org/journals/human-neuroscience/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fnhum.2015.00700/full

  31. Augurelle AS, Smith AM, Lejeune T, Thonnard JL. Importance of cutaneous feedback in maintaining a secure grip during manipulation of hand-held objects. J Neurophysiol. 2003;89(2):665–71.

    Article  PubMed  Google Scholar 

  32. Deutsch KM, Newell KM. Age differences in noise and variability of isometric force production. J Exp Child Psychol. 2001;80(4):392–408.

    Article  CAS  PubMed  Google Scholar 

  33. Shafer RL, Solomon EM, Newell KM, Lewis MH, Bodfish JW. Visual feedback during motor performance is associated with increased complexity and adaptability of motor and neural output. Behav Brain Res. 2019;376:112214.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Johansson RS, Westling G. Roles of glabrous skin receptors and sensorimotor memory in automatic control of precision grip when lifting rougher or more slippery objects. Exp Brain Res. 1984;56(3):550–64.

    Article  CAS  PubMed  Google Scholar 

  35. Shafer RL, Wang Z, Bartolotti J, Mosconi MW. Visual and somatosensory feedback mechanisms of precision manual motor control in autism spectrum disorder. J Neurodev Disord. 2021;13(1):32.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Unruh KE, McKinney WS, Bojanek EK, Fleming KK, Sweeney JA, Mosconi MW. Initial action output and feedback-guided motor behaviors in autism spectrum disorder. Mol Autism. 2021;12(1):52.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Lidstone DE, Miah FZ, Poston B, Beasley JF, Mostofsky SH, Dufek JS. Children with autism spectrum disorder show impairments during dynamic versus static Grip-force tracking. Autism Res Off J Int Soc Autism Res. 2020;13(12):2177–89.

    Article  Google Scholar 

  38. Neely KA, Mohanty S, Schmitt LM, Wang Z, Sweeney JA, Mosconi MW. Motor memory deficits contribute to motor impairments in autism spectrum disorder. J Autism Dev Disord. 2019;49(7):2675–84.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Deutsch KM, Newell KM. Children’s coordination of force output in a pinch grip task. Dev Psychobiol. 2002;41(3):253–64.

    Article  PubMed  Google Scholar 

  40. Deutsch KM, Newell KM. Deterministic and stochastic processes in children’s isometric force variability. Dev Psychobiol. 2003;43(4):335–45.

    Article  PubMed  Google Scholar 

  41. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (DSM-5®). Arlington, VA: American Psychiatric Publishing; 2013.

    Book  Google Scholar 

  42. Lord C, Rutter M, DiLavore PC, Risi S, Gotham K, Bishop S. Autism diagnostic observation schedule: ADOS-2. Los Angeles, CA: Western Psychological Services; 2012.

    Google Scholar 

  43. Lord C, Rutter M, Le Couteur A. Autism diagnostic Interview-Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. J Autism Dev Disord. 1994;24(5):659–85.

    Article  CAS  PubMed  Google Scholar 

  44. Wechsler D, Zhou X, WASI-II. Wechsler abbreviated scale of intelligence. Second. San Antonio, TX: The Psychological Corporation; 2011.

    Google Scholar 

  45. Rutter M, Bailey A, Lord C. The social communication questionnaire: manual. Los Angeles, CA: Western Psychological Services; 2003.

    Google Scholar 

  46. Reilly JL, Lencer R, Bishop JR, Keedy S, Sweeney JA. Pharmacological treatment effects on eye movement control. Brain Cogn. 2008;68(3):415–35.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Bodfish JW, Symons FJ, Parker DE, Lewis MH. Varieties of repetitive behavior in autism: comparisons to mental retardation. J Autism Dev Disord. 2000;30(3):237–43.

    Article  CAS  PubMed  Google Scholar 

  48. Lam KSL, Aman MG. The repetitive behavior Scale-Revised: independent validation in individuals with autism spectrum disorders. J Autism Dev Disord. 2007;37(5):855–66.

    Article  PubMed  Google Scholar 

  49. Annett M. Hand preference observed in large healthy samples: classification,norms and interpretations of increased non-right-handedness by the right shift theory. Br J Psychol. 2004;95(3):339–53.

    Article  PubMed  Google Scholar 

  50. Richman JS, Moorman JR. Physiological time-series analysis using approximate entropy and sample entropy. Am J Physiol-Heart Circ Physiol. 2000;278(6):H2039–49.

    Article  CAS  PubMed  Google Scholar 

  51. Yentes JM, Hunt N, Schmid KK, Kaipust JP, McGrath D, Stergiou N. The appropriate use of approximate entropy and sample entropy with short data sets. Ann Biomed Eng. 2013;41(2):349–65.

    Article  PubMed  Google Scholar 

  52. Goldberger Ary L, Amaral Luis AN, Leon G, Hausdorff Jeffrey M, Ivanov Plamen Ch., Roger M et al. G.,. PhysioBank, PhysioToolkit, and PhysioNet. Circulation. 2000;101(23):e215–20.

  53. Lake DE, Richman JS, Griffin MP, Moorman JR. Sample entropy analysis of neonatal heart rate variability. Am J Physiol-Regul Integr Comp Physiol. 2002;283(3):R789–97.

    Article  CAS  PubMed  Google Scholar 

  54. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear Mixed-Effects Models Using lme4. J Stat Softw. 2015 [cited 2020 May 27];67(1). Available from: http://www.jstatsoft.org/v67/i01/

  55. Hox J. Multilevel Analysis: Techniques and Applications, Second Edition. 2nd ed. Routledge; 2010 [cited 2020 Sep 4]. Available from: https://www.taylorfrancis.com/books/9780203852279

  56. Luke SG. Evaluating significance in linear mixed-effects models in R. Behav Res Methods. 2017;49(4):1494–502.

    Article  PubMed  Google Scholar 

  57. Manicolo O, Brotzmann M, Hagmann-von Arx P, Grob A, Weber P. Gait in children with infantile/atypical autism: Age-dependent decrease in gait variability and associations with motor skills. Eur J Paediatr Neurol. 2019;23(1):117–25.

    Article  PubMed  Google Scholar 

  58. Yang HC, Lee IC, Lee IC. Visual feedback and target size effects on Reach-to-Grasp tasks in children with autism. J Autism Dev Disord. 2014;44(12):3129–39.

    Article  PubMed  Google Scholar 

  59. Fukui T, Sano M, Tanaka A, Suzuki M, Kim S, Agarie H et al. Older Adolescents and Young Adults With Autism Spectrum Disorder Have Difficulty Chaining Motor Acts When Performing Prehension Movements Compared to Typically Developing Peers. Front Hum Neurosci. 2018 Oct 23 [cited 2024 Jun 18];12. Available from: https://www.frontiersin.org/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.3389/fnhum.2018.00430

  60. Khan AJ, Nair A, Keown CL, Datko MC, Lincoln AJ, Müller RA. Cerebro-cerebellar Resting-State functional connectivity in children and adolescents with autism spectrum disorder. Biol Psychiatry. 2015;78(9):625–34.

    Article  PubMed  PubMed Central  Google Scholar 

  61. Unruh KE, Bartolotti JV, McKinney WS, Schmitt LM, Sweeney JA, Mosconi MW. Functional connectivity of cortical-cerebellar networks in relation to sensorimotor behavior and clinical features in autism spectrum disorder. Cereb Cortex. 2023;33(14):8990–9002.

    Article  PubMed  PubMed Central  Google Scholar 

  62. Poon C, Chin-Cottongim LG, Coombes SA, Corcos DM, Vaillancourt DE. Spatiotemporal dynamics of brain activity during the transition from visually guided to memory-guided force control. J Neurophysiol. 2012;108(5):1335–48.

    Article  PubMed  PubMed Central  Google Scholar 

  63. Miall RC. Connecting mirror neurons and forward models. NeuroReport. 2003;14(17):2135–7.

    Article  CAS  PubMed  Google Scholar 

  64. Rizzolatti G, Fogassi L, Gallese V. Neurophysiological mechanisms underlying the Understanding and imitation of action. Nat Rev Neurosci. 2001;2(9):661–70.

    Article  CAS  PubMed  Google Scholar 

  65. Cook JL, Blakemore SJ, Press C. Atypical basic movement kinematics in autism spectrum conditions. Brain. 2013;136(9):2816–24.

    Article  PubMed  PubMed Central  Google Scholar 

  66. Hellendoorn A, Wijnroks L, van Daalen E, Dietz C, Buitelaar JK, Leseman P. Motor functioning, exploration, visuospatial cognition and Language development in preschool children with autism. Res Dev Disabil. 2015;39:32–42.

    Article  PubMed  Google Scholar 

  67. LeBarton ES, Landa RJ. Infant motor skill predicts later expressive Language and autism spectrum disorder diagnosis. Infant Behav Dev. 2019;54:37–47.

    Article  PubMed  Google Scholar 

  68. Wilson RB, Burdekin ED, Jackson NJ, Hughart L, Anderson J, Dusing SC, et al. Slower Pace in early walking onset is related to communication, motor skills, and adaptive function in autistic toddlers. Autism Res Off J Int Soc Autism Res. 2024;17(1):27–36.

    Article  Google Scholar 

  69. Neely KA, Chennavasin AP, Yoder A, Williams GKR, Loken E, Huang-Pollock CL. Memory-guided force output is associated with self-reported ADHD symptoms in young adults. Exp Brain Res. 2016;234(11):3203–13.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

The authors would like to thank the participants and their families for their involvement in our study. We would also like to thank members of our research team EA and LH for providing technical support for the data processing programs used for the present manuscript.

Funding

This study was supported by K01 MH137518, TL1 TR002368, NIGMS P20 GM103418, R01 MH112734, U54 HD090216, R21 AG065621, UL1 TR001427, R01 NS121120, R01 AG086493, the University of Kansas – Life Span Institute Endowment, and the University of Florida APK Research Investment Grants Award.

Author information

Authors and Affiliations

  1. Life Span Institute, University of Kansas, Lawrence, KS, USA

    Robin L. Shafer, Abigail Driggers & Matthew W. Mosconi

  2. Kansas Center for Autism Research and Training (K-CART), University of Kansas, Lawrence, KS, USA

    Robin L. Shafer, Abigail Driggers & Matthew W. Mosconi

  3. Hoglund Biomedical Imaging Center, University of Kansas Medical Center, Kansas City, KS, USA

    James Bartolotti

  4. Emory Graduate Division of Biological and Biomedical Science, Emory University, Atlanta, GA, USA

    Zheng Wang

  5. The Frederick J. and Marion A. Schindler Cognitive Neurophysiology Laboratory, Department of Neuroscience, University of Rochester School of Medicine and Dentistry, Rochester, NY, USA

    Erin Bojanek

  6. Department of Applied Physiology and Kinesiology, University of Florida, Gainesville, FL, USA

    Zheng Wang

  7. Department of Clinical Child Psychology, University of Kansas, Lawrence, KS, USA

    Matthew W. Mosconi

Authors
  1. Robin L. Shafer
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. James Bartolotti
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Abigail Driggers
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Erin Bojanek
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Zheng Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Matthew W. Mosconi
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

EB, MWM, and ZW were involved in the conception and design of the study. AD and RLS were involved with data analyses for the study, and JB developed the data processing and analysis programs to model trial-level logarithmic decay. AD, MWM, and RLS were involved in the interpretation of the data. RLS drafted the manuscript and substantively revised it with assistance from MWM. All authors approved the submitted version of the work and have agreed both to be personally accountable for the author’s own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature.

Corresponding author

Correspondence to Matthew W. Mosconi.

Ethics declarations

Ethics approval and consent to participate

Adult participants provided written informed consent after a complete description of the study, in accordance with the Declaration of Helsinki. For participants under the age of 18 and adults who were under legal guardianship, a parent or legal guardian provided written informed consent on behalf of the participant, and the participant provided written assent. All study procedures were approved by the University of Kansas Medical Center Institutional Review Board (IRB#: STUDY00140269).

Consent for publication

Not applicable.

Competing interests

MWM is PI on an investigator initiated clinical trial of behavioral inflexibility in autism funded by Acadia Pharmaceuticals. MWM and ZW received funding from Novartis Pharmaceuticals Corporation for an investigator-initiated study of Phelan McDermid Syndrome. The other authors declare that they have no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shafer, R.L., Bartolotti, J., Driggers, A. et al. Visual feedback and motor memory contributions to sustained motor control deficits in autism spectrum disorder across childhood and into adulthood. J Neurodevelop Disord 17, 26 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s11689-025-09607-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s11689-025-09607-7

Keywords

Journal of Neurodevelopmental Disorders

ISSN: 1866-1955

Contact us