Biological correlates of altered circadian rhythms, autonomic functions and sleep problems in autism spectrum disorder

The center of circadian rhythms regulation can be identified in the hypothalamus. It consists of several neuronal populations which produce neuropeptides and neurotransmitters with a crucial role for temperature, metabolic rate, thirst, hunger, sexual behavior, reproduction and emotional responses. The hypothalamus exerts its functions mainly through the productions of neuropeptides, such as the transcription factor orthopedia (OTP), crucial for proper development of diencephalic dopaminergic neurons or the steroidogenic factor 1 (SF-1), which is important for neuronal and structural connectivity. Another transcriptional factor, Sim-1, is implicated in hypothalamus differentiation [1]. Deregulation of autonomic system and of daily activity rhythms often precede underlying pathophysiological changes in the brain of subjects with neurodegenerative diseases, such as Alzheimer disease. Similarly, dysautonomias have been seen in metabolic syndromes, cardiovascular diseases, diabetes and cancer. In the last decades, several studies highlighted the importance of altered circadian rhythms, autonomic functions and sleep quality in ASD, while some authors reported associations between these alterations and specific autistic-like symptoms. Intriguingly, an abnormal connectivity between hypothalamus and amygdala has been associated with inappropriate responses to socially relevant stimuli in autism spectrum disorder (ASD) [2]. In particular, a dysfunction of the endogenous circadian system was hypothesized to be part of the neurodevelopmental alteration at the basis of the autistic condition [3]. Sleep problems are frequent among ASD individuals. Several hypotheses have been proposed for explaining this association, including a deficit in melatonin production, anxiety and arousal at the time of going to bed or specific gene mutations. Sleep disturbances have been related to the severity of ASD symptoms (restrictive and repetitive behaviors, difficulties in reciprocal and social interactions, irritability, etc.). It was recently reported an association between actigraphy-derived sleep efficiency, number of awakenings, anxiety and hyperactivity in ASD children. Other studies showed instead a positive correlation between sleep latency, affective problems and aggressive behavior in this population [4]. Some of the genes responsible of regulating circadian rhythms seem to be mutated or down-regulated in ASD, thus providing a molecular explanation for sleep problems: in this framework, deepening the knowledge about genes implicated in circadian rhythms deregulation could be crucial for the development of eventual biomarkers of sleep functions [5]. Circadian hormones such as melatonin and cortisol were also found to be altered in ASD. ASD individuals seem to show an altered production of melatonin, which, in turn, has been related to sleep difficulties, altered gastrointestinal motility, deficits in behavioral and emotional regulation, sensory protein dysfunction. As in the case of gene deregulation, alterations of melatonin/serotonin pathways could be promising biomarkers for ASD [6]. The increased cortisol levels of ASD people were related to hyperarousal and wake after sleep onset (WASO) duration in subjects with insomnia and in controls. On the other hand, a lower cortisol response was linked to poorer self-reported sleep-in healthy population [7]. Autonomic variables such as cardio-vagal activity and pupil size were also investigated in ASD, highlighting in ASD children a lower cardio-vagal activity as measured by heart rate variability (HRV) and increased sympathetic activity measured by means of urinary vanillylmandelic acid (VMA) [8]. It was pointed out that dysautonomias in ASD may be linked to the presence of repetitive behaviors [9]. Starting from these considerations, two important therapeutic strategies, such as melatonin supplementation and Early Start Denver Model (ESDM), a developmental and behavioral intervention, have been studied in ASD not only for sleep problems but also for improvements in social communications, stereotyped behaviors, rigidity, and anxiety [10]. In this review, we aimed to summarize previous reports about sleep problems in ASD children and adults, and about possible biological factors implicated in the alteration of circadian rhythms and autonomic function in this population, together with possible therapeutic approaches. Eventual shared biological underpinnings between ASD symptoms and altered circadian rhythms/autonomic functions will also be discussed.

To identify the studies focused on the relationship between ASD and sleep problems, a comprehensive literature review was performed using multiple databases (Pubmed, Web of Science, Scopus). Key words used were “autism spectrum disorder”, “autism”, “neurodevelopmental disorder”, “sleep problem”, “sleep disorder”, “circadian rhythm”, “autonomic nervous system”, “pharmacological therapy”, “pharmacological treatment”, “biological correlates”, “hormone”. Articles were considered eligible if written in English from 1980 to 2022.

Several studies highlighted a link between ASD and sleep problems, reporting specific correlations with some symptoms. Besides questionnaires, methods used for evaluating sleep in children and adolescents with ASD include polysomnography and actigraphy. The first one measures multiple physiological sleep parameters, such as brain waves, eye movements, etc. However, in ASD children, due to sensory sensitivity related problem, systematic desensitization is often required. The actigraph is a portable accelerometer device that can be worn on the wrist or ankle and records different sleep variables such as time of sleep onset, total sleep time, morning waking time and frequency of night-time waking [11]. Actigraphy enables to distinguish ASD poor sleepers from good sleepers, because the first ones would show prolonged initial sleep latency, decreased sleep efficiency and increased number of night awakenings [12].

Noticeably, some studies also highlighted an association between sleep problems and other symptoms in this population. Considering children, Schreck et al. [13] reported a relationship between reduced sleep time and deficits in social skills in this population. Moreover, Sikora et al. [14] found that sleep problems were often associated with both externalizing (such as impulsivity and aggression) and internalizing symptoms (such as anxiety and depression). In a study conducted on hospitalized ASD children, higher scores on the Aberrant Behavior Checklist-Community scale (ABC-C) at admission were associated with fewer minutes slept during the last five nights of hospitalization. Meanwhile, sleep problems seem not to be influenced by the subtype of ASD, or by the degree of cognitive impairment. Goldman et al. [15] compared sleep problems of older children and adolescents with ASD with those of control toddlers and younger children. Participants were divided in four groups: less than 5 years, from 5 to 7, from 7 to 11, more than 11 years. While all groups reported sleep related problems, differences in the kind of problem were reported depending on age: younger children showed greater levels of sleep anxiety, bedtime resistance, parasomnias and night walking, while older children and adolescents more frequently reported short sleep duration, delays in sleep onset and daytime sleepiness [12, 15]. A recent study evaluated the association between sleep problems and autistic traits in toddlers (n = 426). The classification was based on the total Modified Checklist for Autism in toddlers (M-CHAT-JV) scores. The autistic group (n = 26) was composed mainly by males. In terms of sleep problems, three items were more frequently endorsed in the autistic group: “bedtime resistance” (time attended before going to bed), “abnormality in circadian rhythm” (the total sleep time varies from one night to another), “sleepiness outside the naptime”. It has emerged that children aged 3–5 years with more difficult temperaments showed substantially more bedtime resistance than children with less difficult temperaments. Bedtime resistance in children at around 18 months of age may be related to subclinical autistic traits. It has been suggested that children with ASD have a deregulation of sleep homeostasis, resulting in a reduction of sleep pressure [16]. Stressful events, such as home confinement during COVID-19 pandemic, were recently shown to exert an impact on sleep in ASD children. In particular, a study was conducted among Turkish ASD children (n = 46). Children were given different questionnaires, such as the children’s chronotype questionnaire (CCQ), autism behavior checklist (AuBC) and children’s sleep habits questionnaire (CSHQ). Results reported that ASD children had greater sleep problems and altered chronotype during home confinement period than during normal state. The severity of ASD symptoms increased together with the chronotype score and sleep score during home confinement. Interestingly, the level of sleep problems was considered a mediating factor in the relationship between CCQ score and autism symptom severity. Moreover, a significant correlation was found between increased CSHQ and CCQ score and ASD symptoms during the home confinement period [17]. Other authors also hypothesized that neonatal-irritable sleep–wake rhythm may be involved in ASD pathogenesis. Miike et al. [18] conducted an analysis among 177 ASD children and 203 controls. In the ASD group irritable/over-reactive types of sleep–wake rhythms were observed more frequently than in control group. These alterations may cause difficulties in raising the child, due to frequent waking up, difficulties in falling asleep, short sleep hours, continuous crying and grumpiness. Noticeably, the number of mothers who went to bed after midnight during pregnancy was higher in ASD group than in controls. Yavuz-Kodat et al. [19] investigated 52 ASD children aged 3–10 years. Sleep and circadian rest-activity rhythms were evaluated objectively with actigraphy and subjectively with the CSHQ, while the ABC-C was used for assessing behavioral difficulties. A shorter continuous sleep period was reported among children with high irritability compared to those with lower irritability. A similar trend emerged in children with high stereotyped behaviors compared to those with less stereotypies.

Other studies focused instead on ASD adolescents and adults, finding that insomnia, parasomnias and circadian rhythms disorders were common also in this population. Ballester et al. [20] led a study on 41 adults with ASD and intellectual disability (ID), and 51 typically developed adults. They used an ambulatory circadian monitoring recording wrist temperature, motor activity, body position, sleep and light intensity. ASD individuals showed several sleep difficulties, such as low sleep efficiency, prolonged sleep latency, increased number and length of night awakenings, with daily sedentary behavior and increased nocturnal activity. ASD adults had an advanced sleep–wake phase disorder. Sleep and markers of the circadian system were significantly different between adults with both ASD and ID and the control group. Sleep disturbances described for adults with ASD and ID are similar to those for ASD adults without ID. Even elderly ASD individuals seem to report more difficulties to fall asleep than other people. Jovevska et al. [21] enrolled in their studies subjects aged from 15 to 80 years (297 ASD participants and 233 controls). Sleep quality, sleep onset latency, total night sleep and sleep efficiency, as measured by Pittsburgh Sleep Quality Index, were analyzed. Moreover, the authors investigated autistic traits, mental health condition, medication, employment and sex as predictors of sleep quality. Results highlighted lower sleep quality and increased sleep onset latency in the whole ASD group rather than in the control one. ASD subjects also showed poorer sleep quality and longer sleep onset latency than controls of similar age, especially in adulthood and middle age, while no significant difference was found in adolescent and elderly ASD groups with respect to age-matched controls. Predictors seemed to account for the 21% of sleep quality variance in the ASD group, with sex as the strongest predictor (female sex more related to sleep problems). In the comparison group, the strongest predictor was mental health condition, with predictors accounting for 25% of sleep quality variance [21].

Some authors evaluated in adult samples the relationship between sleep problems and other features. Baker et al. [22] investigated sleep–wake rhythms in 36 ASD adults. Participants were assessed through a 14-day sleep–wake diary and 14-day actigraphy. ASD individuals seemed to show sleep patterns associated with circadian rhythm disturbances, although authors highlighted that the role of employment status, comorbid anxiety and depression should be taken into account. Another study from the same group evaluated employment status in ASD adults with sleep problems. A total of 72 individuals were enrolled in the study (36 ASD people, 36 controls). Subjects were assessed by questionnaires and 14-day actigraphy. 20 ASD individuals versus 4 controls met criteria for insomnia and/or circadian rhythm sleep–wake disorders. Adults with ASD who met the criteria for sleep disorders according to the International classification of Sleep-Disorders, Third edition, reported higher scores on the Pittsburgh Sleep Quality Index (indicating worse sleep quality) and were more frequently unemployed when compared with ASD adults without sleep problems. This result highlighted an association between sleep problems and unemployment [22]. Deserno et al. [23] also evaluated the impact of sleep on quality of life among 598 ASD adults. Autistic features, comorbid complaints, daily functioning and socio-demographic characteristics were evaluated by means of questionnaires. Results reported that sleep problems were an important predictor of subjective quality of life, while subjective experience of individual contribution in the society seemed to statistically predict the level of daily activities. The authors stressed how sleep problems may be considered an important treatment target for improving quality of life in ASD individuals.

In conclusion, while the literature seems to have reached a global agreement on the presence of sleep alterations in ASD subjects, during both childhood and adult life, results about the possible relationship between sleep problems and ASD symptoms are more limited and controversial, also due to high heterogeneity in the methodology and in the specific symptoms measured (see Table 1).

Several studies highlighted altered ANS responses in ASD, which may be related to altered circadian rhythms. Alteration of autonomic variables in ASD was reported since early infancy [57]. Noticeably, alterations of ANS have also been implicated in metabolic, sleep, gastrointestinal disorders, as well as in seizures and hormonal dysfunction [58]. On the other hand, the association between autonomic instability and psychiatric symptoms like depression, anxiety and sleep disorders was previously stressed in the literature [57, 58].

In the field of ASD, differences in heart rate (HR) response between ASD children and controls were frequently highlighted, and both HR and respiratory sinus arrhythmia (RSA), which may be a good index of CNS regulation of HR, have been used as indicators of the integrity of the ANS [59]. In this framework, a recent study, evaluating HR and RSA among subjects from 1 to 72 months of age at 8 different time points, found a progressive reduction in HR and increases in RSA. However, the RSA increase was of smaller entity among those subjects who later received a diagnosis of ASD [59]. Bharath et al. [8] evaluated in a sample of ASD children (n = 40) and controls (n = 40) HRV together with urinary levels of VMA. They found that low frequency HRV was more represented in the ASD group, while high frequency HRV was less represented. In addition, urinary VMA concentrations were higher among ASD children. The authors hypothesized that ASD children may have a reduced cardio-vagal activity as measured by HRV, and increased sympathetic activity as measured by urinary VMA. Another study in a smaller sample further highlighted how HR variability (HRV) might be a promising parameter for differentiating ASD children from controls also in a very early stage. The authors enrolled 20 children (6 children with ASD, 14 controls), recording photo plethysmography (PPG) for 4 min in each child during resting state and color stimulus test condition. PPG was used for deriving HRV. Different responses between ASD and controls during resting state were highlighted for specific features of HRV [60]. Thapa et al. [61] compared ASD adults (n = 55) and controls from the general population (n = 55), and reported a significantly higher resting state HR among ASD patients, together with lower square root of mean squared differences of successive R–R intervals (RMSSD) and lower high frequency HRV. No significant difference was reported for low frequency HRV. These authors hypothesized that their results may suggest a general deregulation in resting autonomic activity among ASD adults, characterized by lower parasympathetic activity.

Among other factors linked to ANS alterations, pupil size was also considered as a potential biomarker for ASD. Anderson et al. [62] compared a group of ASD children with age matched controls, reporting that ASD individuals seemed to show a larger tonic pupil size and lower afternoon levels of sAA. The authors also highlighted a little diurnal variation of sAA among ASD children, while among controls a linear increase throughout the day was observed. Globally, they hypothesized that these features may be associated with an over-production of Norepinefrine (NE) [62]. In a recent work, Arora et al. [63] reviewed previous studies that evaluated autonomic arousal by measuring HR, pupillometry and electrodermal activity (EDA) in ASD individuals compared with controls. A general hyperarousal in ASD individuals with respect to controls was reported during resting states, eventually linked to the difficulties showed by ASD people in regulating their arousal.

Considering the issue of sleep problems, some studies specifically focused on the relationship between ANS and sleep alterations in ASD. In particular, Pace et al. [64] investigated HRV during sleep in high-functioning ASD children (n = 19) and controls (n = 19). Two indices of cardiac activity, the high frequency (HF) spectrum and RMSSD, were significantly higher in ASD individuals when compared with other ones. In both groups, lower mean HR values were found during sleep with respect to those registered during wakefulness. However, the ASD group showed a lower decrease in HR during deep sleep despite the presence of a higher parasympathetic tone. Another study analyzed HR and HRV during sleep in ASD. A group of ASD children (n = 21) and controls (n = 23) was evaluated by means of overnight polysomnography. HR was found higher in ASD subjects than in controls during stages N2 and N3 of non-REM sleep and during REM sleep. High frequency oscillations of HRV, which reflect vagal modulation, resulted lower in N3 stage of non-REM sleep and in REM sleep among ASD children. Low frequency to high frequency ratio (LF/HF) was instead higher during REM sleep. These data, together with the higher HR, lead to hypothesize a sympathetic dominance during sleep among ASD children, which would be associated with a decreased vagal influence [65]. Tessier et al. [66] investigated HRV in ASD subjects (17 adults and 13 children) and typically developed ones (16 adults and 13 children) before and after sleep. They divided the sample into four groups: ASD and control adults and ASD and control children. Each group was evaluated for two nights in a sleep laboratory. Normalized values of low frequency, normalized values of high frequency and LF/HF ratio were evaluated during sleep and vigilance state. Morning high frequency values resulted to be lower among ASD adults when compared with controls. LF/HF ratio during REM sleep was instead higher in both ASD adults and controls than in children. These results are in line with a lower parasympathetic activity in ASD individuals in the morning. However, it should be noted that while this study provided a good insight on possible differences in the investigated parameters depending on age, the presence of multiple group comparison analyses in a limited sample size limited the extensibility of their results. More recently, Chong et al. [67] investigated the relationship between EDA, another parameter used as a marker of sympathetic nervous system activation, and sleep deregulation in ASD children (n = 13). Two EDA indices, non-specific skin conductance responses (NSSCR) and tonic skin conductance levels (SCL) were measured. The group of children with deregulated sleep showed fewer NSSCRs and lower SCL in the afternoon.

Finally, other confirmations of the presence of altered ANS response in ASD may came from pharmacological studies: some authors proposed propranolol, a beta-blocker drug, as a possible therapeutic option for dysautonomias and emotional behaviors in ASD. According to the available literature, this drug seems to significantly improve cognitive performances, such as verbal problem solving, social skills, mouth fixation and conversation reciprocity. Moreover, it seems to improve anxiety, behavioral and autonomic deregulation, aggressive, self-injurious, and hypersexual behaviors, including among subjects with acquired brain injury. However, the specific benefits with respect to sleep still need to be further investigated [68].

The above described literature seems to suggest the presence of hyperarousal states in ASD individuals, possibly with a hyper-sympathetic state not properly counterbalanced by the vagal parasympathetic influences [62‐67]. However, also in this case the available studies are still scant in number and highly heterogeneous in methodology. Moreover, studies specifically focused on the possible relationship between altered ANS function and sleep alteration are even fewer and did not allow reaching a conclusive agreement on this topic (see Table 5).

To date, the first line treatment for ASD sleep problems, especially in children, are sleep hygiene practices and behavioral interventions [34, 69]. There are some extinction measures which parents of ASD children with sleep problems may follow, such as decreasing afternoon naps, or placing the child in his/her own bed without extraneous stimuli [70]. Associated behavioral manifestations, such as aggressive behaviors, may also be controlled by atypical antipsychotic agents such as risperidone or aripiprazole [71].

Among sleep related pharmacological interventions, administration of melatonin and its agonist drug agomelatine was reported to reduce latency of sleep and the number of awakenings in ASD children. As reported in the previous chapters, in ASD populations a reduction or an alteration in melatonin secretion was highlighted, as well as a reduction in gamma-Aminobutyric acid (GABA)_B receptors in the anterior and posterior cingulate cortex and in the fusiform gyrus cortex [12].

In this framework, Wasdell et al. [72] led a study on 51 children with neurodevelopmental disabilities (age range 2–18 years), who were administered melatonin 20–30 min before the desired bedtime and were invited not to eat during the 2–3 h before the moment of drug assumption. Melatonin resulted to be effective in 47/51 children in improving sleep quality and reducing family stress. Malow et al. [73] performed instead a dose escalation study on 24 children with ASD, who were free of psychotropic medications. The dose response, tolerability and safety were studied during a 14-week open label design. The supplementation of melatonin (1–3 g) showed to improve sleep latency and was proved to be tolerated and safe. A work by Maras et al. [74] has proved 3-month efficacy and safety of a novel pediatric-appropriate prolonged-release melatonin (PedPRM), an easily swallowed formulation shown to be efficacious versus placebo, for long-term treatment (that is up to 52 weeks) of children with ASD who suffered from insomnia. The study also reported an improvement in caregivers’ quality of life. While several studies were conducted in this field, also review and meta-analyses are available: in particular, a meta-synthesis by Cuomo et al. [75], collected data from eight previous systematic reviews based on a total of 38 studies about sleep intervention in ASD. Interventions were categorized in four groups: melatonin supplementation, other pharmacological therapies (such as risperidone, clonidine, benzodiazepines), behavioral interventions, parent education/education programs, alternative therapies. Melatonin supplementation was showed to be the most successful intervention for sleep initiation and maintenance of sleep, while for other parameters, such as night wakings or self-settlings, a strong effect of behavioral intervention and education was reported. Despite the general agreement on the efficacy of melatonin for improving sleep parameters and daytime behaviors in ASD [36], it should be noted that studies which evaluated melatonin efficacy are affected by several limitations, such as small sample sizes, comorbid neurodevelopmental disabilities, heterogeneity in methodologies and in dosages, and that not all the studies clearly confirmed melatonin beneficial effects on all the outcomes [76]. Adverse effects linked to melatonin use were also reported, although generally mild, such as morning tiredness, headache, irritability, diarrhea or night-time awakenings [35, 36].

More limited evidence is available for other pharmacological interventions. Ramelteon, a melatonin agonist and a sleep-cycle regulator used for people who have difficulties in falling asleep, was also hypothesized to have a preventive effect on the onset of sleep disorder after general anesthesia in patients with ASD [77]. A case series is also available, which reported, during a trial on 3 ASD children with Ramalteon, a decrease in sleep problems such as insomnia as well as an improvement of ASD behavioral symptoms [78, 79]. The use of agomelatine was also evaluated in this population. Ballester et al. [80] highlighted among ASD subjects (n = 23) with sleep problems a significant increase of night total sleep time, together with phase correction and improvement of sleep stability, after 3-months of treatment with agomelatine. Other studies reported that donepezil, a reversible inhibitor of acetyl cholinesterase, may have beneficial effects on ASD individuals for improving sleep, communication, eye contact, hyperactivity, expressive and receptive speech [81, 82]. Donepezil combined with choline supplement also showed a sustainable effect on language skills in children with ASD for 6 months after treatment, particularly in subject aged below 10 years [83]. Finally, benzodiazepines, which are commonly used in other neuropsychiatric conditions, were also reported to improve some sleep problems in ASD, even if paradoxical reactions with agitation and hyperactivity were reported [84, 85]. The use of benzodiazepines in ASD might be supported by evidences of a GABA-A alpha 5 receptor deficit in this population [85], while an impairment in GABAergic functions were also highlighted in ASD murine models [86‐88]. In a case series of 11 ASD children, the administration of 0.5–1 mg of clonazepam showed efficacy in 75% of the subjects [84]. However, the possibility that children with neurodisabilities could show an increased risk of paradoxical reaction prevents many clinicians from prescribing benzodiazepines in ASD children [88]. Another drug which has been object of investigation in this field is clonidine. In a case series of six children with neurodevelopmental disorders, clonidine was proven effective for reducing sleep problems and sleep initiation latency but showed a small impact on recurrent night time and early morning awakenings, mood instability and aggressiveness [89]. Among antidepressants, trazodone was suggested to be useful for treating the advanced sleep phase in ASD individuals [90]. Mirtazapine was reported to be effective on insomnia and symptoms related to sleep deprivation such as irritability and anxiety. In a study led by Posey et al. [91], 25 ASD participants were investigated to test the efficacy and tolerability of mirtazapine. An improvement of several symptoms, including aggressive behaviors, self-injury and irritability, was observed in nine subjects. A further research analyzed the effects of gabapentin administration to ASD children (n = 23). Gabapentin administered 30–45 min before the bedtime resulted to be effective in improving sleep in 18 participants [92, 93]. On the basis of the lower ferritin levels frequently reported among ASD children, an 8-week open-label treatment with oral iron supplementation was also conducted in 33 ASD children with restless sleep. A significant improvement in sleep quality was reported [11, 94].

Globally, it should be noted that while melatonin studies should be regarded in light of the previously mentioned limitations, it appears clear how studies on pharmacological interventions besides melatonin are still very limited in number, featuring also small sample sizes or, for some drugs, being limited to case series, thus increasing the risk of biases and placebo effects. As a result, the above reported research should be considered cautiously until confirmed by further investigations. In addition, another issue in this field lies in the fact that in most of the studies, including those on melatonin, the pharmacological treatment was reported to improve also other symptoms besides sleep problems. Although these data may further support the beneficial effect of the treatment object of investigation, it may be also considered a confounding factor: further studies should clarify if these drugs would exert a specific effect on sleep, or the sleep improvement should be considered a consequence of the improvement of other symptoms (see Table 6).

Studies on sleep showed how ASD subjects typically report more problems regarding bedtime resistance and reduced sleep pressure [16]. A link between sleep difficulties and irritability, deficits in social skills and behavioral problems was also highlighted in ASD children [13, 14, 19]. Generally, an insufficient sleep time seems to affect the quality of life of ASD individuals [23], including an important impact on employment status [22]. As reported in the previous chapters, several mechanisms may be responsible of this feature. The molecular explanation of circadian problems in ASD individuals may be found in mutation, variants or different expression of specific genes which regulate circadian rhythms, such as CLOCK [5, 26, 29], PER1 [27] and CACNA1C [30]. On the other hand, these problems may also be related to an alteration of melatonin levels (and, eventually, of PGV) as reported by several studies [3, 6, 32, 44, 45]. In addition, an impaired circadian regulation of cortisol [48, 56], as well as increased urinary corticosteroids levels were reported in ASD, possibly due to an altered HPA axis function [49]. Besides the development of anxiety and depressive symptoms, the chronic activation of HPA axis seems to exert a negative impact on ASD individuals’ daily life [48, 95] and to inversely correlate with gravity of repetitive behaviors [51]. Studies focused on autonomic functions in ASD globally pointed out the presence of hyperarousal state in ASD individuals [63]. ASD was hypothesized to be associated with the presence of a hyper-sympathetic state, which would be not adequately compensated by vagal parasympathetic influences. The analysis of autonomic indices such as pupil size [62], electrodermal activity [67] and HR highlighted a tendency towards a lower parasympathetic activity during daytime and a sympathetic dominance during sleep [64‐66].

It should be taken into account that some of the above reported alterations in ASD were not associated only with sleep problems but also with more ASD-specific clusters of symptoms, such as communication impairment or repetitive behaviors [51, 96]. These data suggest that the altered sleep patterns in ASD should be regarded in a broader perspective, considering the possible bidirectional interactions between ASD core symptoms and sleep problems in promoting and maintaining each other. Noticeably, the importance of rhythmicity and synchrony of motor, emotional and interpersonal rhythms for the development of social communication was recently stressed [97]. In addition, ASD children were reported to have difficulties in adapting their changes to the internal or external environment [98]. It would be worth mentioning that epileptic seizures, which are often comorbid with ASD, may be influenced by rapid rhythms of sensory stimuli in the external environment [99]. It was also hypothesized that melatonin may be implicated in the synchronization of the circadian clock network. As a result, interventions which combine melatonin treatment with behavioral measures may be useful for investigating the alterations of this clock [98]. The contribution of melatonin in the ontogenetic establishment of circadian rhythms and in the synchronization of circadian clocks suggests that melatonin may also be implicated in motor, emotional and interpersonal rhythms. An altered melatonin excretion or activity has been related to both timing problems in biological clocks and to the severity of social communication impairment [96, 100, 101]. Maternal variations in cardiac rhythm and in hormone levels were reported to influence the child’s ability to adjust to the environment, while parent–infant synchrony and the construction of shared timing were supposed to be important for social communication development [102]. Some authors also hypothesized that repetitive behaviors may be interpreted as a coping strategy for controlling anxiety and stress response as well as for experiencing more continuity and stabilizing circadian rhythms [51, 103]. In this framework, it should be remembered that melatonin seems to exert an effect also on the improvement of stereotyped behaviors [73]. Both melatonin supplementation and Early Start Denver Model (ESDM) have been studied in ASD not only for sleep problems but also for improving social communication, stereotyped behavior, rigidity, and anxiety. A significant improvement in global communication (including verbal and non-verbal scores) was reported [10], while improvements in rigidity were highlighted on the basis of parents’ and teachers’ comments [104]. Wasdell et al. [72] led an investigation to test the efficacy of melatonin in treatment of delayed sleep phase syndrome and sleep maintenance problems in children with neurodevelopmental disabilities and ASD, highlighting an anxiety reduction based on caregivers’ comment. ANS dysfunctions were also hypothesized to be involved in different ASD symptoms. Condy et al. [9] reported that repetitive behavior severity seems to be predicted by base line RSA and by its reactivity. Low baseline cardiac vagal control predicted less adaptive behaviors in children with and without ASD. ANS studies which analyzed HRV, pupil size and sAA, highlighted that ASD subjects differ from typically developed ones, showing a general reduction in parasympathetic function, while a tendency towards hypoarousal was reported in ASD by studies which measured electrodermal responsiveness [63]. Kaartinen et al. [105] investigated the relationship between autonomic arousal to direct gaze (measured by means of skin conductance responses) and social impairment in ASD children (n = 15) and controls (n = 16). Among ASD children, but not among controls, a positive association was found between impairment in social skills and increased arousal enhancement to direct gaze. In another study, RSA and HR were instead correlated with the accuracy and latency of recognition of facial emotions. Children with ASD were slower in recognizing facial emotions and made more errors in recognizing anger. While ASD subjects showed lower amplitude RSA and higher HR than controls, ASD children with more amplitude of RSA were faster at recognizing emotions. It was proposed that ASD children may live in a hyper-sympathetic state with a reduced ability to calm down. This asset may contribute to their impaired ability to engage in social interactions and to the increased levels of anxiety [106]. Bazelmans et al. [107] highlighted significant correlations between language skills and HR in non-ASD children and between language skills and HRV in ASD ones. These authors also stressed the possible use of autonomic control for detecting individual differences among ASD subjects. On the other hand, another study did not find specific associations between social functioning and autonomic/cortisol response in ASD and non-ASD subjects [108]. Recently, the role of gut microbiota in ASD physiopathology is also acquiring more attention. In particular, gut microbiota and CNS were supposed to influence each other through the “gut brain axis”, which would also involve mucosal immune system, enteric and autonomic nervous system (in particular the vagal branch) [109, 110]. In this framework, a recent investigation further broadened the perspective, stressing the importance of evaluating both microbiota and ANS alterations to properly differentiate specific ASD presentations. The authors evaluated autonomic functions and microbiota composition among ASD individuals comparing them with their non-affected first-degree relatives. A lower interbeat interval was reported in ASD patients, which might be eventually associated with higher sympathetic arousal and lower vagal tone. ASD individuals with sleep disturbances showed more gastrointestinal dysfunctions as well as altered interbeat interval and blood volume pulse, suggesting a possible role of these indices for detecting occult sleep disturbances among ASD subjects. Moreover, ASD subjects with greater microbial richness (increased alpha diversity of the gut flora) reported milder sensory/cognitive and language/communication impairment [110]. However, it should be noted that the use of first-degree relatives as controls should be considered as a major limitation for this study, considering the increasing interest about sub-threshold forms of autism [111‐113]. Increasing literature is pointing out that autistic traits would be distributed in a continuum in the general population, being highly represented in clinical populations of psychiatric patients with other disorders as well as in first-degree relatives of ASD probands [114‐118]. In this latter population, the presence of sub-threshold autistic traits also contributed in shaping the concept of “broad autism phenotype”. In this framework, biological studies in ASD field should include the category of patients’ relatives as an intermediate group and not as a control group [113, 118].

Globally, literature about the issue of sleep disturbance in ASD did not reach a sufficient agreement, especially when considering possible biochemical correlates. Further studies should investigate the link between ASD and sleep disturbances in a broader perspective, which would allow evaluating the shared biological underpinnings between ASD symptomatology and altered circadian rhythms. Moreover, an integrative model would allow clarifying the possible bidirectional interaction between sleep problems and other ASD symptoms: while disrupted sleep may worsen ASD clinical picture, exerting a detrimental effect on cognitive and behavioral features, on the other hand the typical core of ASD symptoms may facilitate the presence and maintenance of sleep problems in this population. In addition, pharmacological studies for sleep problems in ASD need to follow more standardized protocols to reach more repeatable and reliable results. As circadian rhythm alterations are known to exert a detrimental effect on psychopathological conditions, including ASD, improving sleep patterns in ASD subjects may lead to a significant improvement of the clinical outcome and of the psychopathological trajectory, eventually reducing also the risk of developing further disorders in comorbidity [47, 114]. Improve our knowledge in this field may lead to detect potential pathognomonic alterations or even biological markers for ASD, with a consequent improvement of diagnostic strategies. On the other hand, a better understanding of biological bases of ASD symptoms may pave the way for the investigation of further therapeutic targets, eventually improving actual treatment strategies for a population which is still poorly responsive to the currently approved treatments.