Figure 6: MRCPs show differences as a function the preceding sensory cue. (A) SCPs contrasting ASD and TD participants separately for each cue type (AV, A, V). (B-D) Topographic visualizations of MRCPs plotted separately for TD (left) and ASD participants (right) averaged across a 50 ms time bins. (B) -250 to -200 ms before the button press response. (C) -200 to -150 ms before the button press response. (D) -100 to -50 before the button press. (E-F) Grand average waveforms plotted separately for TD (left) and ASD participants (right), stratified by the sensory cue type preceding the motor response (AV, A, V). (E) Mid-frontal electrode Fz. (F) Left parietal electrode P3.
Discussion
In this study, neural activity related to the generation of a cued motor movement was investigated in children with ASD and matched typically developing children using a simple speeded response task. Strengths of the study included the simplicity of the task which allowed for the investigation of basic sensorimotor processes, and the large sample size (n=84 per group) of children with similar age, sex, and IQ, allowing for the investigation of population differences with small and medium effect sizes. The inclusion of children and adolescents across a wide age range allowed for a cross-sectional investigation of sensorimotor processing through different phases of child development.
Evidence of earlier ramp-up of sensory-cued motor preparation in TD participants relative to ASD participants was shown over frontal and parietal scalp regions (Figure 3 ), and slower reaction times in ASD participants relative to TD were seen across all age ranges and sensory cue types (Figure 2 ). Across both participant groups, modulation in motor-related cortical potential (MRCP) morphology by reaction time was observed (Figure 4 ). Between-group differences in the topography and morphology of motor-related potentials were most prominent in the youngest group of participants (Figure 5 ). Differences in sensorimotor processes were also evident as a function of sensory cue type, with multisensory audio-visual stimuli evoking the highest-amplitude MRCPs and the strongest between-group differences between ASD and TD participants (Figure 6 ).
Participants with ASD showed poorer task performance than TD participants. When considering all trials, slower response time was observed in ASD participants. However, when analysis was constrained to remove trials with excessively fast or excessively slow RTs (TD: 4.48% of trials, ASD: 7.74%), significant between-group differences in mean RT were observed only in the slowest quartile of trials, although ASD participants continued to show greater RT variability. The current data, showing lower response rates in ASD (ASD: 86.3% hit rate; TD: 92.7% hit rate) coupled with greater RT variability support the idea that lower task performance in ASD may be driven by a greater tendency toward distractibility (Murphy et al., 2014). These data further suggest that ASD-related motor slowing—a common finding across many behavioral tasks (An et al., 2018; Sokhadze et al., 2016)—may be driven by a minority (< 10%) of extremely slow responses.
Evaluation of group differences between ASD and TD participants for three well-characterized MRCP components—the late berescheftspotential potential (BP), motor potential (MP), and reafferent potential (RAP)—revealed effects of diagnosis on sensorimotor processing across all components in various parcellations of the MRCP data, but consistent differences across all parcellations were observed in the late BP period. Previous research has shown reductions in amplitude of the late BP in patients with motor impairments, although previous work has largely focused on individuals with cerebellar lesions (Kitamura et al., 1999; Shibasaki et al., 1986) and focal hand dystonia (Deuschl et al., 1995; Toro et al., 2000; Yazawa et al., 1999). In patients with focal dystonia, BP amplitude reductions are suspected to arise due to abnormalities in primary motor cortex function, as suggested by early animal work. Evidence from early studies monitoring motor cortex activation in cats during voluntary movement indicate that pyramidal neurons located in the lateral motor cortex alter their firing rate within 500 ms of a voluntary movement: a time-window consistent with the late BP (Neafsey et al., 1978). In adults with cerebellar lesions, it is hypothesized that late BP dysfunction is caused by a deficit of facilitatory input from the cerebellum to these BP-generators in the lateral premotor cortex (Kitamura et al., 1999; Shibasaki and Hallett, 2006). In the Autism literature, the cerebellum has consistently emerged as a region showing both functional (Unruh et al., 2019) and morphological alterations (Mapelli et al., 2022). For example, post mortem examinations have identified ASD-related reductions in Purkinje cell size (Fatemi et al., 2002) and number (Kemper and Bauman, 1993). While cerebellar contributions to altered sensorimotor activity cannot be assessed under the current design and methodological approach, data from functional neuroimaging show hyperactivation of the cerebellum in ASD participants during a precision grip task, suggesting functional reorganization of motor systems which emphasizes subcortical processes (Unruh et al., 2019). Disrupted processing in this network could conceivably contribute to the more variable reaction times and altered sensorimotor processing uncovered by the current work.
For both participant groups, modulation of MRCPs by response speed was observed, with fast trials showing higher-amplitude activity during the late BP period occurring prior to the button-press response. However, the strongest differences between ASD and TD participants were observed in later phases of the MRCP surrounding the MP (0-25ms post-response). While modulation of motor potentials as a function of numerous task-related features such as task complexity (Mussini et al., 2021) and attention (Aliakbaryhosseinabadi et al., 2017) have been shown previously, the current results implicate response promptness as an additional modulatory factor. Although modest statistical differences were observed between ASD and TD participants, primarily in the fastest 50% of trials, largely similar topographic maps indicate recruitment of similar generators.
Differences in MRCP amplitude were also observed as a function of the preceding sensory cue type. Although event-related time-locking to the button-press response allows for examination of motor-related effects, movement execution in response to a sensory cue necessitates processing of that sensory cue. The current data, showing the largest between-group differences in the multisensory condition, accords with findings from a previous analysis of a subset of the current data showing higher-amplitude sensory-evoked responses to multisensory stimuli (AV) compared to unisensory stimuli (A or V) (Brandwein et al., 2013), underscoring the overlap of sensory processing and motor preparation in the context of a cued-motor paradigm. While between-group differences in MRCPs were present for all three cue types, differences between ASD and TD waveforms were greatest in the multisensory condition. This latter finding is consistent with neurophysiological findings that multisensory (here, audio-visual) inputs are processed differently in ASD compared to TD (Beker et al., 2018; Brandwein et al., 2013; Foxe et al., 2015; Russo et al., 2010) and may also reflect that ASD/TD differences in multisensory integration extend beyond the sensory processing stage into stages more classically associated with motor planning (Mercier et al., 2015).
Clear age-related differences in MRCP morphology were observed. In TD participants, these age-related differences were largely restricted to the post-response period while in ASD participants, additional differences were evident in the pre-response period beginning approximately 250 ms prior. Literature on the maturation of sensorimotor preparatory and post-movement processing is sparse. However, the current data would suggest that, in TD individuals, maturation effects in motor processes are most prominent in the RAP period (90-130 ms after movement onset) associated with somatosensory feedback processing. For ASD participants, maturation effects extend across the MRCP, spanning both the late BP and post-response periods. Between-group differences in MRCP maturation were most pronounced in younger participants. These data suggest ongoing maturation of the late BP in children with Autism and that early sensorimotor processes remain largely stable in TD children after age 6. These data are consistent with clinical reports of delayed fine motor skill acquisition in ASD relative to TD individuals (Fournier et al., 2010).
Finally, it should be noted that, while statistical differences between ASD and TD participants were visible across most parcellations of the current neural and behavioral data–such as when responses were binned by age, response promptness, and the preceding sensory cue type–these basic sensorimotor differences are unlikely to represent a sole explanation of the extensive and often severe motor impairment commonly reported in ASD (Green et al., 2009). Indeed, in the current work, the general topographic similarities of MRCPs between groups (Figure 3-6 ) suggest that processes upstream and downstream of sensorimotor preparation which were not examined in the current work—such as action-selection, motor sequence learning, or social communication— very likely also contribute to overall motor-related difficulties. Previous work has consistently demonstrated that individuals with ASD show neural and behavioral atypicalities when engaging in the high-level processes of movement imitation and pantomiming, indicating a breakdown of sensory integration, motor processing, social communication, or an interaction between all three variables. In neurotypical adults, for example, mu-band neuroscillitory suppression was shown to occur during both observed movements and self-initiated movements, whereas adults with ASD only showed such suppression during self-movement (Bernier et al., 2007; Oberman et al., 2005). While differences in sensorimotor processing were shown in the current work using a non-social task, it seems plausible that these differences may be exacerbated by tasks which additionally require the recruitment of higher-level cognitive processes. For example, in a motor response task to assess action-consequence learning in children with ASD, researchers observed differential modulation of the late BP component between groups (Migó et al., 2021). Children with ASD showed an enhanced late BP component during task segments in which a button press response was associated with an auditory consequence whereas, in contrast, TD participants showed an enhanced BP when the press did not result in a consequence. Although caution must be taken in generalizing from these relatively small sample studies, these data may reflect that basic differences in motor processing and preparation are mediated by cognitive factors such as the perceived meaningfulness of the generated movement or its social context.
This study is not without limitations. Although the quick and variable interstimulus interval used in the current investigation allowed for the presentation of a large number of trials (972 ± 58 trials per participant), the rapidity with which stimuli were presented precluded investigation of group differences in the early BP component onsetting between 1 and 3 seconds prior to the response. Further, although response-locked averaging is a common and well-documented approach for MRCP analysis, electromyographic (EMG) data monitoring skeletal muscle contraction prior to the response were not recorded in the current study, and therefore EMG-locked analyses could not be performed. Further, for the majority of participants, clinical assessments of attention deficit hyperactivity disorder (ADHD) and associated symptoms were not conducted. An investigation of the relationship between observed sensorimotor differences and symptoms of ADHD—a common comorbidity in ASD—is therefore outside the scope of the current work. While the broad age range in the current work included children as young as 6 years old, more work is needed to understand the neural trajectory of early sensorimotor development (e.g. in children 6 months to 5 years).
In summary, the current investigation presented data on the development of the MRCP in healthy children and adolescents, and evidence of moderate alteration in sensorimotor processes in children with ASD relative to typically developing peers. These data suggest that subtle sensorimotor atypicalities are present in non-social contexts and that these differences are most prominent in the youngest group of children (age 6-9). Future studies focused on younger children are still needed to understand the early development of sensorimotor processes in autism. Additionally, work examining movement imitation in ASD may consider the role of basic motor responses when interpreting between-group differences using more complex tasks.
Author Contributions: SM and JJF designed the original experiment. All authors conceived of the current study. KMW analyzed the data in consultation with SM and JJF. KMW produced the figures for this study. KMW wrote the first draft of the manuscript, and all authors provided editorial input on subsequent drafts. All authors read and approved the final version of this manuscript.
Acknowledgements : The authors thank Douwe Horsthuis for assistance with data collection and the Human Clinical Phenotyping Core (HCP) for their careful clinical and cognitive phenotyping of our ASD cohort.
Conflict of Interest : The authors report no conflict of interest, financial or otherwise.
Ethics and Consent : The institutional review board at Albert Einstein College of Medicine and the City College of New York approved all experimental procedures. Each participant provided written informed consent in accordance with the tenets laid out in the declaration of Helsinki.
Funding Information : This work was supported by the National Institute of Mental Health of the National Institutes of Health (NIH) under award number R01MH085322 (S.M. and J.J.F.), and the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the NIH under award number P50 HD105352 (Rose F. Kennedy Intellectual and Developmental Disabilities Research Center). Funding for KMW was provided through the Rose F. Kennedy Intellectual and Developmental Disabilities Research Center (T32 HD0980676). Ongoing work on Autism Spectrum Disorder (ASD) at The University of Rochester (UR) collaborating site is supported by the UR Intellectual and Developmental Disabilities Research Center (UR-IDDRC), through a center grant from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD P50 HD103536 to JJF).