Autism spectrum disorder (ASD) involves difficulties in social communication and restricted, repetitive behaviors (1). Over 90% of children with ASD exhibit at least one type of problem behavior, such as self-injury or tantrums (2, 3). Furthermore, children with ASD are more likely to have comorbid physical and mental health conditions, including epilepsy and depression (4). Among these, sleep disturbances are some of the most frequently reported comorbidities, with an estimated 50%–80% of children with ASD experiencing sleep-related issues (5, 48). Compared to typically developing (TD) children, those with ASD exhibit a higher prevalence of difficulties such as sleep-onset delay, nighttime awakenings, and suboptimal sleep hygiene, and these symptoms often persist chronically (6, 49).

Research has also demonstrated associations between sleep problems and core ASD symptoms. For instance, children with ASD who obtain less total sleep time per night tend to exhibit greater symptom severity, particularly in the domain of social communication deficits (7). In addition, shorter sleep duration and frequent nighttime awakenings have been shown to predict the presence of stereotyped behaviors (8). Heightened sensory hypersensitivity is also linked to more severe sleep difficulties in children with ASD. Specifically, greater sensory sensitivity has been associated with increased sleep-onset delays, more frequent nighttime awakenings, and reduced total sleep duration (9, 10).

Numerous studies have reported atypical autonomic nervous system (ANS) functioning in individuals with ASD (11, 12). A commonly used index of ANS activity is heart rate variability (HRV), which refers to the variation in intervals between successive R-waves (RR intervals) on an electrocardiogram. HRV can be measured noninvasively, making it a widely adopted parameter in recent psychophysiological research [e.g., (13, 14)]. Studies have shown that children with ASD exhibit lower resting-state HRV than their TD peers, and this reduction in HRV has been associated with atypical attentional responses, difficulties in emotion regulation, and deficits in social communication (15, 16). However, recent meta-analytic and theoretical work suggests that autonomic differences in ASD and other neurodevelopmental conditions may be more heterogeneous than previously assumed, and not all studies report reduced HRV in ASD (17, 18). This variability underscores the importance of transparent and replicable protocols that can support nuanced interpretation of HRV findings.

Given that ANS activity fluctuates cyclically during sleep, HRV is also widely used as a physiological indicator of sleep state and quality (19, 20). Under typical conditions, sympathetic nervous system activity increases during rapid eye movement (REM) sleep—resulting in elevated heart rate—while parasympathetic activity dominates during non-REM (NREM) sleep—resulting in a lower heart rate (20). In contrast, children with ASD tend to show elevated sympathetic activity and higher heart rates during REM sleep, as well as reduced parasympathetic activity and persistently elevated heart rates during NREM sleep, leading to lower HRV across the night compared to TD children (21). However, to date, no studies have investigated HRV during sleep specifically in preschool-aged children with ASD.

Sleep plays an important role in emotion regulation and daytime functioning, and in autistic children poorer sleep quality has been associated with more behavior problems and social difficulties (7). From this perspective, nocturnal HRV may index the extent to which autonomic functioning “resets” during sleep, which may subsequently influence behavioral and emotional regulation the following day. Examining night-to-day carryover effects therefore provides a mechanistic rationale for focusing on nocturnal HRV rather than solely on daytime physiological indices.

Although some studies have examined the relationship between sleep problems and daytime behavioral problems, or assessed daytime resting-state HRV in children with ASD, the association between nocturnal HRV and subsequent problem behaviors remains largely unexplored. Although several studies have measured HRV during sleep, most have relied on polysomnography [e.g.m (13, 19, 21, 22)]. Polysomnography offers the advantage of accurately capturing a wide range of physiological signals, including electrocardiography (ECG), electroencephalography (EEG), and eye movements during sleep (23, 24). However, it requires the attachment of multiple electrodes to the body, which cannot be easily removed and may be particularly distressing for infants and young children—especially those with sensory hypersensitivity, as is common in ASD (24). Moreover, polysomnography is expensive and requires specialized expertise, often necessitating visits to laboratories or hospitals. This makes it less suitable for naturalistic, home-based monitoring (25). In response to these limitations, recent attention has shifted to wearable garment-type devices (e.g., 25–27). These devices use conductive textiles to measure HRV with minimal tactile stimulation, enabling comfortable and unobtrusive data collection during both active and resting states (26). They are better tolerated by individuals with sensory sensitivities, including children with ASD, and have been found suitable for use in this population (28). Additionally, unlike traditional polysomnography, these wireless, portable devices can be used at home, making them ideal for continuous or long-term monitoring in natural environments.

Building on this perspective, the present study aims to longitudinally and quantitatively evaluate the everyday sleep patterns of a child with ASD by measuring electrocardiographic data over a one-month period using a wearable garment-type device. Specifically, we investigate whether variations in HRV during sleep predict subsequent daytime problem behaviors. We hypothesize that lower HRV and increased sympathetic dominance during sleep will be associated with a higher likelihood of problem behaviors occurring the following day.

The primary aim of this study was to evaluate the feasibility of one-month home-based nocturnal HRV monitoring using smart clothing in a preschooler with ASD. A secondary exploratory aim was to examine whether variations in sleep-related HRV might be associated with next-day problem behaviors. Based on the hypothesis that increased sympathetic activity during sleep would predict the occurrence of problem behaviors on the following day, we compared HRV during nights preceding days with and without reported problem behaviors. However, no statistically significant differences were observed between the two conditions. Recent meta-analytic and theoretical work has emphasized that autonomic alterations in ASD are heterogeneous and not consistently observed across studies (17, 18). Thus, the null findings in this single-case dataset should not be interpreted as evidence of no relationship, but rather as reflecting this heterogeneity and the limited statistical power of the present design.

These results suggest that it may be difficult to directly predict next-day problem behaviors based solely on total sleep duration and HRV indices. HRV indices also showed considerable night-to-night fluctuation even on nights with similar caregiver-reported sleep duration, further limiting the ability to detect stable associations within this dataset. One possible explanation lies in the multifactorial nature of problem behaviors in children with ASD, which are known to serve four primary functions: access to tangibles, attention-seeking, escape from aversive stimuli, and automatic reinforcement (37). Given the strong interaction between these behaviors and environmental contingencies, HRV alone may not be sufficient as a predictive variable. To improve prediction accuracy, it may be necessary to obtain more detailed information about each day's problem behaviors. Although caregivers in this study reported whether problem behaviors occurred and described their nature, future studies should also consider collecting information on the intensity of such behaviors. This could help reveal patterns indicating, for example, that problem behaviors are more intense following poor sleep and less intense following better sleep. Goodwin et al. (38) demonstrated the use of actigraphy—a wrist-worn accelerometer—for quantifying problem behaviors in individuals with ASD. Actigraphy allows for the measurement of behavior that is difficult to capture through human observation, including frequency, duration, and temporal patterns. The application of such objective behavioral quantification methods in future studies may lead to more accurate and consistent prediction of problem behaviors. Future work may integrate garment-based HRV monitoring with wrist-worn actigraphy and standardized caregiver reports, aligning these modalities within shared temporal windows to better model the dynamic interactions between sleep-related autonomic activity, movement patterns, and observed behaviors.

To date, most studies on sleep in children with ASD have employed short-term polysomnographic assessments, while even longer-term studies have largely relied on caregiver-reported questionnaires or visual observation. Few, if any, longitudinal studies have examined sleep-related HRV specifically. Given that sleep patterns fluctuate from day to day—and even more so in children with ASD compared to their typically developing peers (39)—short-term measurements are unlikely to capture stable sleep profiles. To determine which types of sleep changes are associated with behavioral outcomes, long-term monitoring is essential. This study is, to our knowledge, the first to measure HRV during sleep over a one-month period in a child with ASD and to examine its relationship with next-day problem behaviors. Although the findings did not support the predictive utility of HRV in this context, this study represents an important first step in exploring physiological-behavioral associations in naturalistic settings. Longitudinal designs not only enable between-subject analyses but also facilitate the modeling of within-subject variability, which is particularly relevant for predicting individualized behavioral responses. In the future, sleep-related HRV may also be useful for predicting other socially significant responses, such as joint attention (40) and emotional regulation (41).

The present study also demonstrated the utility of garment-type wearable devices for real-time measurement of everyday sleep. While not as precise as polysomnography and limited in the range of physiological signals they can capture (42), wearable garments are well-suited for long-term sleep monitoring. From a feasibility standpoint, one of the two initially enrolled children was unable to tolerate the garment due to pronounced sensory hypersensitivity. This highlights the need for design considerations such as fabric texture, seam placement, and graded familiarization protocols in future wearable systems for ASD. Across the 32 planned nights, 25 (78%) yielded usable paired physiological and behavioral data, suggesting that this level of adherence may represent a realistic expectation for month-long home-based monitoring. In particular, the reduced noise during sleep enhances the reliability of HRV data collection. With the continued accumulation and analysis of high-quality physiological data during sleep, such measures may eventually be used to predict a variety of psychological and behavioral outcomes. Although garment-type wearables have been reported to be tolerable for children with ASD, including those with sensory sensitivities (28), the present study showed that in cases of extreme sensory hypersensitivity, sustained use of the device may be difficult. Thus, future studies should consider using sensory profiling tools [e.g., (43, 44)] to assess the likelihood that children will tolerate wearable devices for physiological monitoring.

A key limitation of the present study is that it involved only a single participant. To further examine the predictive value of HRV during sleep for problem behaviors, future studies should include participants who exhibit a broader range of behaviors in terms of intensity, frequency, and topography. It may also be beneficial to collect additional profile variables, such as cognitive and language development, adaptive behavior, and ASD symptom severity, to examine how the relationship between sleep-related HRV and behavior may vary across subgroups.

If behavioral prediction based on physiological and other quantitative indices becomes feasible, it may eventually enable caregivers and professionals to anticipate episodes of challenging behavior and implement preventive interventions (45–47). Such predictive systems may be especially beneficial for children who engage in frequent and severe behaviors, or for those with limited verbal repertoires who struggle to communicate internal states. In these cases, physiological signals may offer a critical early warning system to reduce the risk of problem behavior and enhance proactive support.