Sleep appears to be essential for most living organisms’ survival and integrity, especially for typical synaptic development and brain plasticity (Fogel et al., 2012; Hartsock and Spencer, 2020). Over the past decades, the role of sleep in learning and memory has been probed by many studies at behavioral, systemic, cellular, and molecular levels (Walker and Stickgold, 2006; Rawashdeh et al., 2007; Sawangjit et al., 2018; Kim et al., 2019). The American Academy of Sleep Medicine (AASM) has recently released the third edition of the International Classification of Sleep Disorders (ICSD-3) in 2014. This guideline grouped sleep disorders into seven basic types: insomnia disorders, central disorders of hypersomnolence, circadian rhythm sleep-wake disorders, sleep-disordered breathing, movement disorders, parasomnias, and other sleep disorders (Sateia, 2014; Ito and Inoue, 2015). There are universal physiologic changes during sleep, and some biologic, environmental, psychological, social as well as genetic factors can affect change in the sleep pattern (Jenni and O’Connor, 2005; Bathory and Tomopoulos, 2017). The sleep-wake circadian rhythm is regulated through both circadian and homeostatic processes. Arousal and sleep are active and involved in neurophysiologic processes, including both activation and suppression of neural pathways. Sleep disorders can be an early symptom of the disease, and the presence of rapid eye movement (REM) sleep behavior disorder (RBD) can be used as an early diagnostic indicator for neurodegenerative diseases (Galland et al., 2012; Kotterba, 2015). Besides, sleep disorders have observable effects on physical and mental health of children with autism spectrum disorder (ASD) and their parents (Zhang et al., 2021).

Autism spectrum disorder is a neurodevelopmental disorder and the prevalence of ASD is increasing, with 1 in 59 children in the United States. diagnosed with ASD (Orefice, 2019). ASD is approximately four times more prevalent among males than females (Christensen et al., 2016; Satterstrom et al., 2020). According to the DSM-V, the previous categories of pervasive developmental disorders, pervasive developmental disorder-not otherwise specified (PDD-NOS) and Asperger disorder were combined into ASD. The new diagnostic criteria from DSM-V defined ASD as a heterogeneous spectrum disorder with deficits in social interaction and communication, restricted and repetitive interests, and stereotyped behaviors (American Psychiatric Association, 2013; Devnani and Hegde, 2015).

It is not surprising when considering the numerous health and behavioral issues that sleep disturbance are commonly observed in the clinical progression of ASD. Children with ASD experience more sleep problems compared with the general population, particularly insomnia. Sleep-wake cycle abnormalities are associated with communication deficits, stereotyped behaviors, and autism severity (Tudor et al., 2012). Disrupted sleep may exacerbate the daily dysfunction of ASD children, such as social and communication skills, behavioral performance, stereotypical behaviors, and motor output on non-verbal performance tasks (Schreck et al., 2004; Limoges et al., 2013).

As we gain deeper knowledge of the neural mechanisms of ASD and sleep, more contributions from sleep-related biomarkers to the study of neurophysiology in ASD. Prospectively, the emergence of digital technologies and devices is making studies of sleep physiology more flexible and convenient. The sleep study provided new insights for research on the children with ASD when compared with the other behavioral tests currently used in human subjects and animal experimental models. This review’s main objective is to explore animal models’ role, especially non-human primate (NHP) models, as a useful tool to investigate sleep disorders in ASD children. Firstly, we present data on sleep disorders in autistic children, emphasizing their prevalence, clinical characteristics, phenotypic analyses, and pathophysiological mechanisms. Next, we highlight the current state of animal models for ASD and explore their implications and future prospects in translational research. We suggest that using NHP animal models may provide insights into sleep disorders in ASD.

In most mammalian species, sleep amounts are highest during the neonatal period (Weber and Dan, 2016). Sleep loss can significantly affect a child’s health-related quality and activities of daily living (Maski and Kothare, 2013). The brain is one of the organs most impacted by sleep or the lack thereof while adolescence is a critical period for brain reorganization. It is beyond doubt that sleep disorders during this period exert irreversible effects on children’s brain development (Roffwarg et al., 1966; Klein et al., 2000; Weber et al., 2018). REM sleep can prune newly formed postsynaptic dendritic spines during motor learning (Li et al., 2017), and the balance of newly formed and original dendritic spines is crucial for neuronal circuit development and behavioral improvement in children. Two studies found that sleep enhance cortical plasticity in the visual cortex during the developmental critical period (Frank et al., 2001; Aton et al., 2009). In conclusion, sleep seems to be important for brain development, learning, and memory consolidation by selectively eliminating and maintaining newly formed synapses (Li et al., 2017).

Sleep deprivation may cause physical diseases and developmental problems. During early life, sleep deprivation has been shown to have long-term implications for social behaviors in adulthood (Hudson et al., 2020). Neural substrates can be affected by sleep deprivation, including the prefrontal cortex, basal ganglia, and amygdala. Furthermore, sleep deprivation may cause difficulties in executive functioning, reward learning as well as emotional reactivity. Such issues may contribute to difficulties in judgment, resolution of problems, challenging behaviors, emotional control, and public health concerns, such as depression, suicide, and risk-taking behavior (Maski and Kothare, 2013). These findings indicate that insufficient sleep during early life has persistent effects on brain development and later behavioral performance.

It has been assumed that sleep can clear out brain’s toxins, such as beta-amyloid which was associated with Alzheimer’s disease (Xie et al., 2013). Sleep is essential for maintaining the body’s physical health and is associated with neurodegeneration, metabolic diseases, cancer, and aging. The processes of growth and development are related to sleep quality. The abnormal sleep and circadian also affect hormones and metabolism (Li et al., 2018a). Getting adequate sleep can help the immune system to better react against infection (Grigg-Damberger, 2009; Herculano-Houzel, 2013; Welberg, 2013).

Many neurodevelopmental processes have been reported in the children with ASD, such as synaptic plasticity, neurogenesis and migration of neuron (Gilbert and Man, 2017). About 40–80% of children with ASD exhibit at least one sleep-related problems (Verhoeff et al., 2018), including irregular sleeping and waking patterns, decreased sleep efficiency, reductions in total sleep time and REM sleep time, sleep onset delays, decreased sleep efficiency, increased wakening after sleep onset, bedtime resistance, and daytime sleepiness (Humphreys et al., 2014). Studies utilizing Actigraphy (ACT) and Polysomnogram (PSG) have found that increased sleep latency, and decreased sleep duration and sleep efficiency in ASD children (Elrod and Hood, 2015). A comprehensive review in children with ASD reported that insomnia is one of the most common sleep problems (Souders et al., 2009). Another study also documented that the predominant sleep disorder included insomnia, difficulty falling, and staying asleep (Malow et al., 2006). Mutluer et al. (2016) found that the most common symptoms reported were troubles falling asleep, sleep after waking up and tired after sleeping.

Childhood sleep disorders which are mostly reported by parents are associated with emotional, cognitive, and behavioral disturbances. Sleep disturbances occur in approximately 20–30% of preschool children, including bedtime resistance, sleep onset delays, night terrors or nightmares, and repetitive rhythmic behaviors (Lozoff et al., 1985; Krakowiak et al., 2008; Knappe et al., 2020). The abnormalities of ASD may predispose children to various threaten of sleep and make them especially susceptible to sleep problems (Morgenthaler et al., 2007; Maxwell-Horn and Malow, 2017). Sleep problems have become one of the most common symptoms among ASD children (Richdale, 1999; Wiggs, 2001; Liu et al., 2006; Uren et al., 2019). Two studies compared sleep behaviors of ASD with typically developing (TD) children, they found that 66% of ASD children exhibited moderate sleep disturbances (Souders et al., 2009) and 71% in another study (Malow et al., 2016). A parent-reported study found that 35% of ASD children had at least one sleep dysfunction (Krakowiak et al., 2008). The risk of sleep disturbance is 2.8-fold higher in children with ASD (Köse et al., 2017). A recent study repeated sleep measures at different age in 5,151 children, and found that ASD children have an increase in sleep problems with age, whereas TD children decrease (Verhoeff et al., 2018).

Over the past years, many different sleep analysis methods have been reported (Lomeli et al., 2008; Kelly et al., 2012; Ibáñez et al., 2018; de Zambotti et al., 2019). Infection, pain as well as trauma can disrupt sleep and activity (Vandekerckhove and Cluydts, 2010; Doufas et al., 2012), even some issues that might seem minor to us are often very significant to a child. As children progress from infancy to adolescence, sleep structure, sleep behavior and sleep duration will also change (Williams et al., 2013), it is crucial to take into account the specificity of different ages of children when investigating sleep states. Some sleep studies require an intimate contact of the electrode with the skin and even require surgical implantation of electrodes, which are difficult to apply in freely moving animals and humans, particularly in children related to the lively side of their nature. Even though several new technological developments have been brought to reduce inconvenience, pain, and further damage of these methods, expensive and burdensome must also be considered, especially for long-term studies that include large samples. Some assessments were developed to monitor sleep through observation of body motion and posture. These methods could obviate the need for direct contact and even avoid surgery or electrode implantation. It is non-invasive and low cost. Nevertheless, the behavioral observation does not provide sufficient information compared to those provided by electroencephalogram (EEG) and electromyogram (EMG). In general, both humans’ and animals’ sleep analyses include sleep patterns, locomotor activity, temperature, and food intake. The current study summarized phenotype analyses for sleep disorders obtained from sleep diaries, parent-reported questionnaires, electroencephalography, actigraphy, and videosomnography. We summarized the main types of experimental approaches applicable to assessment methods of sleep studies and all of these methods have advantages and disadvantages (Figure 1). The selection of clinical sleep assessment should be tailored to children’s unique characteristics, and safety and feasibility must also be taken into consideration.

Numerous animal models of ASD have been generated in the last decade (Peñagarikano et al., 2011; Li et al., 2015; Kazdoba et al., 2016; Sacai et al., 2020). Many ASD-associated genes such as Neuroligins play a crucial role in regulation of synaptic adhesion and keeping imbalance between excitatory and inhibitory control in brain circuits (Wintler et al., 2020). The gene editing tools have rapidly been adopted by scientists to parse the role of genetic abnormalities in the etiology and symptomology of ASD. Because the more established gene editing technologies were used in the mice, mice have become the primary animal model of genetic diseases (Crawley, 2012). Growing studies of NHP models have been generated because their close phylogenetic relatedness to humans (Gadad et al., 2013). Moreover, mounting evidence suggests that environmental factors during early development is important. Animal models of maternal exposure to valproic acid and maternal immune activation appear to be the most commonly used. Frequent blood draws and PSG recordings, which are difficult procedures for children with ASD, also make the ASD models becoming ideal candidates. Here, we summarize some rodent (Table 1) and NHP (Table 2) models of ASD, which may have potential value to investigate the causes and effects of ASD, as well as their effects on brain development and sleep disorders.

An ideal animal model of human disease should show tight junctions with clinical characteristics of the disease. The statistics from United States government in 2010 indicated that almost 90% of the laboratory animals used in science research are mice, rats, and other rodents. NHP only represents 0.28% among all animals (Phillips et al., 2014). However, rodents diverged from humans by more than 70 million years of evolution. There are significantly evolutionary differences in brain anatomy, cognitive capacity, and social behavior between humans and rodents (Kumar and Hedges, 1998; Gibbs et al., 2004). Compared with rodents, rhesus macaque (Macaca mulatta), most common NHP used in study, are separated from humans approximately 25 million years ago and are more similar to humans in genetics, neurobiology, and behavior. Thus, NHP have reasonable behavioral correlates to the characteristics of patients in ASD, such as repetitive behaviors, communication deficits, and stereotyped behavior (Watson and Platt, 2012; Parker et al., 2018). As mentioned previously, prenatal environment and gestational timing may impact neurodevelopment of offspring. The gestational period of rhesus monkeys (165 days) and humans (280 days) is much longer than mouse (18–23 days) (Clancy et al., 2001). Besides, the prenatal immune challenge and neuron development of primates occur mostly during the third trimester of prenatal and during early postnatal period (Careaga et al., 2017). The mouse is becoming increasingly popular for genetic studies. However, the mouse’s brain weighs a few grams, and ours weighs one and a half kilos. Can we use the mouse to learn something about our brain? The region of the neocortex is almost 80% in the human brain, which is just 28% in the rat (Roberts and Clarke, 2019). Human prefrontal cortex includes granular and agranular cortex, while rat prefrontal cortex only contains agranular cortex (Ongür and Price, 2000; Uylings et al., 2003). It has been proposed that the prefrontal cortex has a substantial role in social processing, and its potential dysfunction may cause ASD (Sugranyes et al., 2011). The temporal lobe is a morphological brain region which is unique to primates (Colombo et al., 2000; Bryant and Preuss, 2018). The major areas of the human brain classified by Brodmann have also been identified in NHP. Structure and function of the amygdala are nearly the same in the human and non-human primate (Gowen et al., 2007; Rutishauser et al., 2015; Schumann et al., 2016), but remarkably different from the rodent brain (Chareyron et al., 2011). The close relationship of development and evolution between NHP and human show that great prospects to mimic clinical realities by designing NHP animal models.

As mentioned previously, sleep problems in children with ASD are caused by multi-factorial risks such as abnormal neurodevelopment and environmental factors (Bourgeron, 2007; Owens and Mindell, 2011). Modeling clinical disorders in animals provide an opportunity to improve translational research although the human disorder’s clinical phenotype is complex and heterogeneous and lacks objective homologous endpoints across species (Missig et al., 2020). Many previous studies of ASD animal models exhibited several hallmark features which have been documented in humans.

Sleeping studies in humans must be done in accessible samples, predominantly saliva or blood, and confounded by environmental factors. Species-specific differences including light and biological rhythm, as well as sleep features have been noted in studies of sleep (Campbell and Tobler, 1984). For example, rodents are commonly thought to awake during the dark phase and asleep during the light phase. However, researchers found that mice are not explicitly nocturnal, and they have diurnal feeding activity. Researchers also reported that seasonal influences were demonstrated to be more potent on activity than specific genes which was generally considered to control sleep (Daan et al., 2011). Effective use of animals to study normal sleep and sleep disorders must consider known similarities and differences between human and animals. Likewise, sleep is important to keep health and can significantly influence daily activity schedules in NHP (Fruth et al., 2018; Qiu et al., 2019). Sleep structure and EEG patterns of NHP are closely related to the consolidated and monophasic organization observed in humans (Reite et al., 1965; Hsieh et al., 2008), which contrasts with the more fragmented sleep patterns in rodents (Fifel and Cooper, 2014) (Table 3). NHP is also a diurnal animal to better recapitulate clinical conditions with behavioral and metabolic properties closer to humans. Humans pass through 4–6 cycles of NREM and REM within a night’s sleep, which are much shorter in rats and mice (Fuller et al., 2006; Toth and Bhargava, 2013). Nunn and Samson compared sleep patterns in 30 different species of primates, including humans. Most species generally sleep between 9 and 15 h, while humans averaged just 7 h (Nunn and Samson, 2018). In summary, measuring behavioral and sleep states in NHP may provide a better understanding of sleep disorders in children with ASD compared with rodents.

Autism spectrum disorder is a neurodevelopmental disorder and the origins of ASD remain unresolved. The potential estimates including genetic, maternal, and environmental effects (Bai et al., 2019). At first, there are various types of genetic variation, such as single-nucleotide polymorphism (SNP) or rare genetic mutations (Weiner et al., 2017). Animal model studies have shown that the impact of genetic on behavior is complex and not completely correspond to specific behavior. Brain development can be influenced by not only the expression of genes, but also modified by environmental factors during the pregnancy and postnatal period. Therefore, the application of gene editing technology in animal models of disease may not completely mimic clinical phenotypes of humans. Secondly, although the NHP animal model has been used to study the impact of environmental modifications on the brain development. Genetic influences may also affect individual responses to different situations and different types of environmental challenges (Machado and Bachevalier, 2003). In this area, rodent models may be more appropriate which have more identical genetic backgrounds compared with NHPs. Besides, NHP exhibit significant and stable individual differences in social commination (Capitanio, 1999; Capitanio and Widaman, 2005).

Non-human primates have much longer reproductive cycle and lower reproduction efficiency compared with rodents, which may bring the difficulties to prepare an adequate quantity of experimental animals. Furthermore, the giant body size of NHP may cause a significant challenge for experimental design. Limitations associated with the gene-editing technique, including editing efficiency, chimeras, and off-target effects, should also be brought to attention (Niu et al., 2014; Chen et al., 2016). Lastly, NHP, rather than other animals, require more significant ethical consideration because its significant cognitive capacity and complex social behavior. Researchers have moral responsibility to ensure that experimental animals receive reduced negative effects and suffering (Bentham, 1996; Beauchamp and Frey, 2011). Animal experiments should follow the principles of the 3Rs, including replacement, reduction, and refinement (Russell and Burch, 1959; Jennings et al., 2009).

Autism spectrum disorder is a neurodevelopmental disorder and with the increasing incidence of ASD, it is essential to understand what has changed in our genes and environments that may contribute to these disorders. It has been showed that ASD is not a single disease, but rather several conditions including genetic, maternal, and environmental effects that ultimately cause similar behavioral impairments. The abnormalities of ASD may predispose children to various threaten of sleep and make them especially susceptible to sleep problems. Sleep disorders have been reported as one of the most common symptoms and in up to 80% of children with ASD have sleep problems which may even contribute to the altered brain structure and activity (Blakemore et al., 2006). Thus, understanding how sleep affected children with ASD by specific mechanisms such as brain development and synaptic plasticity will enable a broader understanding of the disorders’ causes and provide insights into specific treatments. Over the years, many different sleep analysis methods have been reported. The selection of sleep assessment method should be tailored to specific subjects and taken into consideration of their unique characteristics.

Animal models hold great potential values to investigate the causes and treatments for sleep problems in children with ASD. Numerous animal models of ASD have been generated in the last decade. An ideal animal model should show tight junctions with clinical characteristics of the disease. Rodents are the most common experimental animals and growing studies of NHP models have been generated because their close phylogenetic relatedness to humans.

Because of the complexity and heterogeneity of the ASD, it is still inadequate to understand how genes control and influence complex behavior. The animal models of ASD are currently oversimplified and have many issues. Recently, Liu et al. (2019) demonstrated that an approach to generate cloned monkeys by somatic cell nuclear transfer (SCNT), which can creatively solve the problem of generating NHP models with uniform genetic backgrounds. This study is profoundly improving the overall reproducibility of the model. Continued research used this technique to generate five BMAL1 knockout monkeys for sleeping study, and these monkeys exhibited more activities and reduced sleep during night (Qiu et al., 2019). The development of genome editing technologies (such as CRISPR/Cas9 and base editing, etc.) has opened up the revolutionary ways to directly target and modify genomic sequences in animals (Kang et al., 2019; Zhao J. et al., 2019; Li et al., 2020). We anticipate greater numbers of applications will materialize shortly, such as genome-edited NHP combined with SCNT. Although some issues still need to be solved, studying the sleep disorder across multiple biological scales can offer the hope in the field of translational medicine for ASD and other human diseases. The role of NHP animal model in this process is irreplaceable and must be recognized.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.