The study was conducted in accordance with the Declaration of Helsinki, and approved by the Human Experimental Ethics Inspection of Guangzhou Sport University (ethics number: 2022LCLL-21, date of approval: 7 July 2022). At the outset, the study recruited 56 ASD patients aged between 6 and 11 years, all hailing from the Guangzhou Association of Persons with Intellectual Disabilities and Their Relatives (GAPIDR). Owing to the inherent heterogeneity of ASD (Lai et al., 2014), and to ensure experimental data remained within pertinent observational parameters, participants underwent preliminary screenings based on their IQ, the severity of their autism, among other factors (Figure 1).

We screened children for inclusion in the study based on the following criteria:

Initially, we identified 20 potential participants whose mother tongue was Chinese. Guardians of these participants were asked to read and sign informed consent forms, detailing specifics of the experiment, prior to participation.

Over the last thirty years, Ayres developed a collection of clinical assessments focused on sensory integration. For each subtype of sensory integration dysfunction, checklists were established (Ayres, 2005), to be completed by parents. These evaluations were then used to assess the severity of sensory integration issues in children. Reflecting the cultural context of China, Jung (2018) of Taiwan amalgamated various symptom checklists to create a consolidated sensory integration checklist (Jung, 2018). Researchers from the Mental Health Institute of Beijing Medical University, such as Ren et al. (1994), imported the Child Sensory Integration Scale (Ren et al., 1994) from Taiwan’s Child Brain Development Alliance. Initially, they conducted trials in the Beijing region, and subsequently expanded their testing to 14 provinces and cities nationwide, establishing a standard model. For this scale, age-adjusted standard scores were chosen as the scoring metric, ensuring both accuracy and objectivity in the assessment results. The scale’s scoring is straightforward, allowing examiners to understand and use it effortlessly. Moreover, since the scale’s questions relate to scenarios seen in children’s everyday life, parents find them easy to answer. Such features highlight the scale’s strong applicability and acceptability. The sensory integration scale is validated for its objectivity and utility. It’s effective in evaluating the developmental progress of children’s sensory integration capabilities, gauging the severity of sensory integration disorders, and serving as a benchmark to compare the efficacy of sensory integration treatments. With a test-retest reliability between 0.48 and 0.74, and structural validity from 0.48 to 0.93, the scale exhibits robust reliability and validity. This attests to its strong usability and acceptability in mainland China. In our research, the scale’s Cronbach’s α is 0.802.

The “Child Sensory Integration Scale” is designed to evaluate sensory integration development in school-aged children aged 2–11 years. This scale encompasses 5 factors and 58 items in total, addressing vestibular sensation, tactile perception, proprioception, learning abilities, and issues specific to children aged 10 and older. Given that our study primarily involves children under 10, the last factor was excluded. The scores from each item on the scale, when combined with the age of the participants, are converted into standard scores. A score below 40 indicates a mild sensory integration disorder, while a score below 30 suggests a severe disorder. The scale utilizes a 5-point Likert scoring system, ranging from “never” (highest score) to “always” (lowest score). Parents or knowledgeable individuals should diligently complete the scale based on the child’s behavior in the most recent month. The four items of the scale are as follows:

Upon application of a sensory integration rating scale, we found 18 of the 20 eligible ASD children showed symptoms of Sensory Processing Disorders (SPD) (Table 1). This prevalence aligns with Professor Yi Shin Chang’s (University of California, San Francisco) research indicating over 90% of children with ASD demonstrate SPD symptoms (Chang et al., 2014). The remaining 18 participants were randomly divided into equal experimental and control groups (n = 9 per group).

We utilized the Footscan plantar pressure and balance analyzer, a widely employed Center of Pressure (COP) analyzer and balance analysis system in motor biomechanics. The Footscan system consists of a 50 cm × 40 cm pressure plate system and balance ability analysis software. The pressure plate contains 4 pressure sensors per square centimeter, totaling approximately 4,000 sensors, which can effectively depict pressure distribution under both stationary and dynamic conditions. The software generates a detailed plantar pressure distribution map with multi-dimensional coloring, which offers a comprehensive overview and quantitative measurements of pressure distribution (Yu et al., 2022).

Participants were instructed to stand on the Footscan and perform the Sharpened Romberg Test (SRT) to gather data on their balance capabilities. The Sharpened or Tandem Romberg test is a variation of the original test (Black et al., 1982; Johnson et al., 2005). By analyzing participants’ responses during the SRT and the biomechanical feedback from Footscan, we offered biomechanical insights into the effects of SIT on the balance function in children with ASD. The strength of the SRT lies in its role as an enhanced static balance assessment task. It can heighten the participants’ bodily control and focus during the task (Fournier et al., 2010b), and the increased challenge in posture control can partially showcase the participants’ executive functions (Smith et al., 2021). The Romberg Test is a tool designed to evaluate individuals’ sense of balance, particularly assessing the function of the spinal cord’s dorsal column. It serves as a diagnostic instrument for sensory ataxia (Nguyen et al., 2018), encompassing conditions such as subacute combined degeneration of the spinal cord, posterior cord syndrome, and spinal cord hemisection (Brown-Séquard syndrome). The test has been recognized as a sensitive and precise measure for evaluating imbalances due to central vertigo, peripheral vertigo, and head injuries, and its clinical application spans nearly 150 years (Lanska and Goetz, 2000; Forbes et al., 2023). Nevertheless, our objective in employing the SRT wasn’t to clinically diagnose participants. Instead, we aimed to have them adopt a scientifically sound standing posture for data collection on Footscan, thereby bolstering the scientific rigor of our experimental data.

The SRT balance assessment comprises three stances: feet together, semi-tandem, and tandem. In this study, we utilized the tandem stance of the SRT to capture participants’ balance metrics on Footscan. During the SRT balance assessment, participants were instructed to remove their shoes and stand on the Footscan device. They were required to align their feet precisely from heel to toe, cross their arms over the chest, rest their open palms on the respective shoulders, and maintain their chin parallel to the ground. Once the participant achieved a stable posture, the official task commenced. The experiment integrated two conditions: Visual Deprivation (VD) and Normal Visual (NV), with a duration threshold set at 30 s for each test (Figure 2).

All participants underwent three testing phases. In the first phase, guardians provided the participants’ basic information and signed the experiment’s informed consent forms. In the second phase, participants completed a Footscan pre-test. Here, participants were familiarised with the testing procedure, which was formally executed after a practice run. In the final phase, within 2 days of concluding the intervention, participants completed post-test data collection with Footscan. All testing sessions were scheduled between 3:00 and 5:00 PM.

The SPD screening process dictated the intervention scheme for the participants, based on group classifications. The experimental group received sensory integration therapy (SIT) (Schaaf and Mailloux, 2015) as an intervention for 8 weeks, while the control group was assigned to the usual treatment. During the formal intervention process, all participants wore heart rate monitors to maintain a heart rate within 50–69% HRmax, achieving moderate to low-intensity aerobic exercise. However, these data weren’t analyzed. As a rule, participants were not permitted to partake in regular exercise outside of the experiment, or their data would be discarded. This requirement was discussed and agreed upon with the guardians before the experiment, with detailed explanations provided. We checked each participant’s daily condition and physical feelings via oral inquiries before each intervention session. The sensory integration therapist for the experimental group hailed from the GZSU Sports Medicine Rehabilitation Center. This therapist’s senior certification in sensory integration was co-endorsed by the National Autism Rehabilitation Research Center under the China Rehabilitation Research Center (CRRC) and the Beijing Association for Rehabilitation of Autistic Children. The therapist was not briefed on the specific details or hypotheses of the study and solely performed professional sensory integration therapies. The therapeutic sessions occurred in a sensory integration therapy room, outfitted with equipment essential for SIT. The intervention strategies adhered to the SIT principles set by Ayres (2005), which have since been substantiated by a plethora of scholars (Parham et al., 2007, 2011; Schaaf et al., 2014). The SIT intervention specifically encompasses the subsequent seven sensory systems:

Implementation Feedback. Administer sensory integration therapy three times weekly, with each session spanning 60 min. Allocate 10–15 min to each sensory system. The therapy regimen extends over 8 weeks. Document participants’ responses and advancements for every activity. Adjust activity challenges and intensities as dictated by observed outcomes. Maintain regular communication with parents and teachers, gaining insights into participants’ behavior at home and school. Offer guidance and support based on these insights.

Experimental data were analyzed using SPSS Statistics 22. Repeated measure ANOVA was used to test the main effects and interactions of the Go/No-Go cognitive neural tasks. For reaction time (RT), accuracy ratio (AR) of Go/No-Go effects, blood oxygen signal differences, and the SRT, we combined visual display (VD) task conditions to collect data on COPx, COPy, TTW, and EA. A t-test was performed pre- and post-experiment to analyze data differences, using repeated measure ANOVA as a basis.

Out of 18 eligible ASD children screened for SPD using the Children’s Sensory Integration Rating Scale, 10 showed moderate SPD symptoms, six had mild SPD symptoms, and two were normal. The 18 SPD participants were randomly divided into control and experimental groups, nine in each. There were no statistical differences between the two groups in terms of age, sex, BMI, and skeletal muscle mass (P > 0.05, Table 1), suggesting that all participants met the intervention conditions for this experiment.

The centre of pressure (COP), the instantaneous locus of forces applied and counteracted during the contact between the human body’s sole and the ground, primarily navigates the heel-to-toe range when the body is standing. The displacement remains relatively minimal (Cornwall and McPoil, 2000). The projection of the COP path, which consists of the entire instantaneous locus, onto the Footscan force-measuring platform, forms a gait line (Fuller, 1999). The Footscan employs a resistive pressure-sensitive sensor in an X-Y matrix, capable of recording real-time pressure data from participants. This data is further processed and visualized via Footscan 9 to extract measurement characteristics.

Data related to the SRT was pre-processed for both control and experimental groups, collected before and after the experiment (Table 2). Following an 8-week intervention, very significant differences appeared in COPx, COPy, TTW, and EA across the NV and VD tasks within the experimental group. Lower TTW and EA values indicate superior static balance ability. Consequently, these findings demonstrate that sensory integration training (SIT) can significantly enhance ASD children’s balance function.

The data gathered for COPx, COPy, TTW, and EA during NV and VD tests before and after the intervention was dynamically compared for both control and experimental groups (Figure 4). The t-test results showed that under NV conditions, all four indices exhibited significant differences in the control group (P < 0.05), and even more significant ones in the experimental group (P < 0.01).

Building upon the descriptive statistics and the t-test, a repetitive measure ANOVA was carried out on the experimental results. The independent variables were Time, Group, and SRT condition (NV, VD), while the dependent variables were COPx, COPy, TTW, and EA results. A 2 × 2 × 2 three-way repetitive measure ANOVA was constructed.

The findings were as follows:

The repetitive measure ANOVA results highlighted that the main effects of Time and SRT condition were significant across all four dependent variables. The experimental results affirmed that the experimental design in this study achieved the expected value. Specifically, when VD was treated as an independent variable, the results revealed an improvement in the balance ability of participants in both groups under the NV task. However, the performance of the experimental group, under the SIT intervention during the VD task, was more resilient to interference, thereby showcasing a more stable performance.

The acquired plantar balance results were visualized through the 3D scan view of Footscan 9^® (Figure 5). The balance change curve was in alignment with the results of this study, further attesting to the effectiveness of SIT intervention in enhancing the balance function of children with ASD.

The behavioral indices for the Go/No-Go tasks encompassed reaction time (RT) and accuracy ratio (AR). Preliminary data processing results are depicted in the descriptive statistics found in Table 3.

The main and interaction effects among all indices underwent analysis through repeated measure ANOVA using SPSS. Results revealed significant main effects for both RT, F_(1,16) = 193.85, P < 0.001, η2 = 0.93, and AR, F_(1,16) = 449.42, P < 0.001, η_p² = 0.95, when considered as independent variables. This finding implies an interaction between RT and AR in Go/No-Go tasks, further validating the suitability of the selected Go/No-Go cognitive neural task for this experiment.

Pre- and post-experiment data from both groups underwent analysis through a t-test to compare the RT differences between experimental and control groups. The findings demonstrated that the RT of the experimental group was significantly lower than that of the control group (P < 0.01), further suggesting that SIT intervention can enhance the performance level of cognitive neural tasks in children with ASD (Figure 6).

Data pre-processing occurred before data analysis: linear polynomial fitting and band-pass filtering were performed on the oxy-Hb signals from each channel. Signal wavebands demonstrating substantial fluctuations in the baseline region and during the Go/No-Go tasks were removed, and a waveband frequency of 0.01–0.8 Hz was selected to minimize data bias caused by cardiac pulse, natural breathing, and head movement noise.

Post-processing, the data from Go/No-Go tasks were categorized into GO and NO-GO task data. Although fNIRS can record data on three blood oxygen signals – oxy-Hb, deoxy-Hb, and total Hb, the analysis focused on oxy-Hb signal changes due to their higher sensitivity to cerebral blood flow changes, neural activation levels in the prefrontal lobe, signal-to-noise ratio, and test-retest reliability (Sato et al., 2004; Uludag et al., 2004; Zhang et al., 2011).

During the Go/No-Go tasks, high activation was noted in the right inferior frontal gyrus (R-IFG) and rostral middle frontal gyrus (R-MFG). Therefore, these regions were selected as the study’s regions of interest (ROI) (Figure 7).

A repeated measure ANOVA tested the main effect of all independent variables and interactions between all dependent variables. Tests were performed on oxy-Hb concentration, and the SIT and Go/No-Go effects in the two ROIs. The results revealed significant interactions between SIT and Go/No-Go effects when considering the oxy-Hb concentration in the R-IFG, F_(1,16) = 18.02, P < 0.001, η² = 0.55, and R-MFG regions, F_(1,16) = 16.50, P < 0.001, η² = 0.52, as independent variables. This outcome further affirms that SIT can effectively elevate oxy-Hb blood oxygen concentration in the aforementioned brain regions.

A paired sample t-test further compared the differences in oxy-Hb signals between experimental and control groups before and after the Go/No-Go effects test. Results showed that in both the R-IFG, t_(1,16) = 4.25, P < 0.001, r = 0.55, and R-MFG regions, t_(1,16) = 4.06, P < 0.001, r = 0.52, oxy-Hb signals in the experimental group were significantly higher than in the control group. This suggests that SIT significantly enhances neural activation in the R-IFG and R-MFG regions (Figure 8).

Firstly, our findings indicate that, compared to traditional therapies, Sensory Integration Therapy (SIT) significantly enhances the balance capabilities of children with autism spectrum disorder (ASD). More specifically, biomechanical evidence derived from Footscan reveals that while traditional therapy showed some improvement in participants’ balance under the SRT’s NV task conditions, it failed to demonstrate efficacy in VD task scenarios. For SIT, however, improvements in participants’ balance were observed under both NV and VD conditions, with the enhancements being particularly pronounced under VD tasks. Under typical visual circumstances, children with ASD predominantly rely on visual signals transmitted to the cerebral cortex. They integrate visual and proprioceptive systems with various bodily components to execute a range of intricate movements. Throughout this process, vision stands out as the most vital information conduit (Little, 2018; Lim et al., 2020). When deficits in sensory integration led to visual impairments, the stability and balance of posture become correspondingly impacted (Cornilleau-Pérès et al., 2005; Stins et al., 2009). Molloy et al. (2003) discovered that children with ASD encounter more significant challenges in balance tasks deprived of visual cues compared to typically developing (TD) children (Molloy et al., 2003). This underscores an excessive dependence on visual cues among children with ASD. Two subsequent studies further corroborated the visual reliance observed in children with ASD. These studies indicated that, when subjected to visual deprivation, children with ASD demonstrated greater postural sway during a visual searching task (Stins et al., 2015) compared to an auditory digit span task (Memari et al., 2014). Such postural swaying is attributed to the inadequate integration of vestibular, proprioceptive, and visual systems (Molloy et al., 2003; Minshew et al., 2004). Utilizing SIT for intervention, we integrated the seven sensory systems of children with ASD, facilitating coordinated completion of numerous physical activities. The results highlight that the enhanced balance capabilities weren’t solely a result of visual dependence. Intriguingly, SIT seemed to induce a compensatory effect on the sensory input deficiencies in children with ASD, particularly in vision, vestibular perception, and proprioception. This emergent compensatory effect warrants further in-depth exploration in future research.

On another note, the heterogeneity and frequent comorbidities associated with ASD contribute to the intricacy of its etiology. Leo Kanner, renowned as one of the most influential psychiatrists of the 20^th century and often dubbed the father of child psychiatry, established as early as 1943 through clinical observations that children with ASD typically show motor skill impairments (Kanner, 1943). In 2014, Professor Lai Mengquan of the University of Toronto’s Psychiatry Department published an article titled “Autism” (Lai et al., 2014) in The Lancet. In this piece, he reaffirmed the prevalent motor skill challenges among individuals with ASD. Following research has empirically indicated (Gowen and Hamilton, 2013) that children with ASD face delays in both fine and gross motor development, often demonstrating uncharacteristic and aberrant motor patterns. Whyatt and Craig (2012) undertook a study evaluating the factors influencing the motor capabilities of children with ASD (Whyatt and Craig, 2012). Their results suggest that balance abilities are the most prominently affected skills among the various motor challenges faced by children with ASD. The effectiveness of balance ability heavily relies on the quality of posture control. Relevant studies indicate that compared to typically developing children, children with ASD exhibit specific deficits in static postural control (Wang et al., 2016). This, to some extent, highlights the challenges faced by children with ASD in various control mechanism movements, potentially impacting their daily life and physical activities. The issues stem from abnormalities in the cortical cerebellar brain structure network impacting the balance function of postural control. The cerebellum holds a key role in orchestrating movement, encompassing guided feedback and reactive behavioral correction during movement (Baev et al., 2002), and synchronizing the timing and spatial coordination of different body joints (Whitney et al., 2008). Our findings suggest that SIT can enhance the balance abilities of children with ASD by mediating corresponding postural control mechanisms via various brain regions (Palliyath et al., 1998). However, SIT is not only associated with the balance abilities of children with ASD but also holds a significant correlation with visuomotor integration, cognitive abilities, motor skills, and postural control. The intrinsic mechanisms of its action require extensive and repeated experimental validation in subsequent studies.

Furthermore, evidence from the Go/No-Go task highlights that children with ASD undergoing SIT demonstrate a notable reduction in RT, pointing to enhanced response efficiency and faster cognitive processing. Conversely, the rise in AR emphasizes the favorable effects of SIT on response accuracy and precision, signifying improvements in participants’ executive functions. In tandem with this, fNIRS neurology findings corroborate the underlying neural pathways through which SIT optimizes the executive functions in children with ASD.

These findings suggest that SIT may have beneficial effects on executive function in ASD children by modulating neural activation in the PFC. One possible mechanism is that SIT may improve sensory processing and integration in ASD children, which may in turn facilitate cognitive processing and regulation. Sensory processing refers to the ability to receive, organize, and interpret sensory information from the environment and the body (Ayres, 1972). Sensory integration refers to the ability to use sensory information to plan and execute adaptive responses (Ayres et al., 2005). ASD children often exhibit sensory processing and integration difficulties, such as hypersensitivity or hyposensitivity to sensory stimuli, sensory seeking or avoiding behaviors, and poor sensory discrimination or modulation (Tomchek and Dunn, 2007). These difficulties may impair their attention, perception, memory, learning, and emotion regulation (Baranek, 2002).

Addressed concretely, we focused on two subregions of the PFC: the right inferior frontal gyrus (R-IFG) and the rostral middle frontal gyrus (R-MFG). The R-IFG is implicated in inhibitory control, as it inhibits prepotent responses and resolves response conflicts (Aron et al., 2014). The R-MFG is involved in working memory, cognitive flexibility, and attentional control, as it maintains task-relevant information and switches between different task rules or strategies (Derrfuss et al., 2005). Our results showed that SIT significantly increased oxy-Hb signals in both R-IFG and R-MFG regions during Go/No-Go tasks, indicating enhanced neural activation and blood flow in these regions. Moreover, our behavioral results showed that SIT significantly reduced reaction time and increased accuracy ratio during Go/No-Go tasks, indicating improved performance and efficiency in executive function. By enhancing neural activation in the R-IFG and R-MFG regions during Go/No-Go tasks, SIT may improve inhibitory control and cognitive flexibility among ASD children through these pathways. Inhibitory control is crucial for suppressing prepotent responses or distractors during Go/No-Go tasks, such as: they were instructed to quickly press the space bar when they saw an elephant, whereas for other animals, they needed to suppress their impulses. Cognitive flexibility is essential for switching between different task rules or stimuli during Go/No-Go tasks, such as: the different animals in this study. By improving performance and efficiency in Go/No-Go tasks, SIT may improve behavioral and functional outcomes among ASD children.

SIT is based on the theory that providing appropriate and individualized sensory stimulation can enhance neural plasticity and promote optimal sensory processing and integration (Schaaf and Miller, 2005). SIT involves various activities that challenge the vestibular, proprioceptive, tactile, auditory, and visual systems, such as swinging, bouncing, spinning, jumping, crawling, balancing, touching, listening, and looking (Bundy and Lane, 2019). SIT aims to improve sensory modulation, discrimination, praxis, and organization skills in ASD children, which may lead to better behavioral and functional outcomes (Parham et al., 2011).

By improving sensory processing and integration in ASD children, SIT may also improve executive function through several pathways. First, SIT may enhance motor coordination and balance function in ASD children, which may in turn improve their inhibitory control and cognitive flexibility. Previous studies have shown that motor skills and executive function are closely related in both typical and atypical development (Diamond, 2000; Hill, 2001; Piek et al., 2008). Motor skills require inhibitory control to suppress unwanted movements and cognitive flexibility to adapt to changing environments or task demands (Best and Miller, 2010). Conversely, executive function requires motor skills to execute planned actions and monitor feedback (Diamond, 2013). Therefore, SIT may improve executive function by improving motor skills in ASD children.

On the other hand, SIT may enhance attentional control and working memory in ASD children, which may in turn improve their inhibitory control and cognitive flexibility. Attentional control refers to the ability to selectively focus on relevant stimuli or tasks while ignoring irrelevant distractors (Posner and Petersen, 1990). Working memory refers to the ability to temporarily store and manipulate information for complex cognitive tasks (Baddeley, 2003). Both attentional control and working memory are essential for executive function, as they enable goal-directed behavior and problem-solving (Miyake et al., 2000). SIT may improve attentional control and working memory in ASD children by providing optimal levels of arousal and stimulation that match their individual sensory preferences and needs (Schaaf et al., 2014). SIT may also improve attentional control and working memory by enhancing interoception, which is the sense of the internal state of the body (Craig, 2002). Interoception is important for attentional control and working memory, as it helps regulate emotional states, arousal levels, motivation, and self-awareness (Critchley and Garfinkel, 2017).

This study has several implications for the theory and practice of SIT for ASD children. First, this study provides empirical evidence for the effectiveness and potential benefits of SIT on balance function and executive function in ASD children, using objective and quantitative measures. Previous studies on SIT have been mostly qualitative or based on subjective measures, which have limited validity and reliability. This study uses Footscan and fNIRS technologies to measure biomechanical and neurophysiological outcomes of SIT intervention, which enhances the credibility and rigor of the research. Second, this study demonstrates a novel and multidisciplinary approach to assess the effects of SIT on ASD children, by integrating biomechanics, hemodynamics, and cognitive neuroscience. This approach allows us to explore the underlying mechanisms of SIT intervention from multiple perspectives and levels of analysis, which may reveal new insights and discoveries that could not be obtained from a single discipline or method. Third, this study contributes to the existing literature on SIT for ASD by providing a comprehensive and critical discussion of the findings, highlighting the implications, limitations, and future directions of the research. This discussion may help researchers and practitioners to better understand the strengths and weaknesses of SIT intervention, as well as the gaps and challenges that need to be addressed in future studies.

This study also has several limitations that need to be acknowledged and addressed in future research. First, the sample size is relatively small (n = 18) and lacks diversity (all Chinese, mostly male, right-handed). A larger and more representative sample would increase the statistical power and external validity of the study. Future studies should recruit more participants from different backgrounds, genders, handedness, and severity levels of ASD. Second, the control group received a different intervention (treatment as usual) than the experimental group (SIT), which may introduce confounding factors and bias. A better control condition would be a sham or placebo intervention that mimics the SIT without its active ingredients. Future studies should use a randomized controlled trial design with a proper control group to eliminate potential confounds and establish causal inference. Third, the study did not provide sufficient details on how the SIT intervention was implemented, such as the frequency, duration, intensity, content, and individualization of the sessions. A clear description of the intervention protocol would enhance the replicability and transparency of the study. Future studies should provide more information on how SIT intervention was delivered and tailored to each participant’s needs and preferences. Fourth, the study did not report any measures of adherence, satisfaction, or adverse effects of the intervention, which are important indicators of feasibility and acceptability of SIT for ASD children and their families. Future studies should collect and analyze data on these aspects to evaluate the practicality and safety of SIT intervention.

In addition to addressing these limitations, future studies should also explore other aspects of SIT intervention that were not covered in this study. For example, future studies should measure other aspects of cognitive function and quality of life in ASD children that may be affected by SIT intervention, such as attention span, memory capacity, language skills, social skills, and self-esteem.