Nightmares in Children with Foetal Alcohol Spectrum Disorders, Autism Spectrum Disorders, and Their Typically Developing Peers

Children with Foetal Alcohol Spectrum Disorders (FASD) and Autism Spectrum Disorders (ASD) experience significantly higher rates of sleep disturbances than their typically developing (TD) peers. Pre-sleep anxiety and waking emotional content is known to affect the content and frequency of nightmares, which can be distressing to children and caregivers. This is the first study to analyse nightmare frequency and content in FASD, and to assess its association with psychometric outcomes. Using online caregiver questionnaires, we assessed reports from 277 caregivers of children with ASD (n = 61), FASD (n = 112), and TD children (n = 104) using the Children’s Sleep Habits Questionnaire (CSHQ), the Child Behaviour Checklist (CBCL), the Spence Children’s Anxiety Scale (SCAS), and the Behaviour Rating Inventory for Executive Functioning (BRIEF). Within the ASD group, 40.3% of caregivers reported their children had nightmares. Within the FASD group, 73.62% of caregivers reported their children had nightmares, and within the TD group, 21.36% of caregivers reported their children had nightmares. Correlation analysis revealed significant associations between anxiety and nightmares, maladaptive behaviour and nightmares, and executive functioning and nightmares in the TD and FASD groups, but not ASD group. This paper adds to the emerging body of work supporting the need for sleep interventions as part of clinical practice with regard to children with ASD and FASD. As a relatively niche but important area of study, this warrants much needed further research.


Introduction
During rapid eye movement (REM) sleep, it is thought that brainstem generators physically activate motor, perceptual, affective, cognitive, and amnestic circuits whose information is then synthesised into what is known as dreaming [1]. The sequential operation of non-REM (NREM) and REM sleep throughout the night is crucial for a fully functioning cognitive system [2,3], and is fundamental to the cognitive and emotional development of the child [4].
Nightmares are defined as disturbing mental experiences that awaken the dreamer and generally occur in REM sleep [5][6][7][8][9], alongside typical negative emotions of terror, fear, and anxiety [9][10][11][12][13]. First nightmare experiences tend to be at around three years old, peaking between six and ten years old [14][15][16]. The nightmare frequency of children is also found to be associated with higher levels of anxiety, as rated by caregivers [17]. Furthermore, nightmare frequency is associated with the mood regulation function of dreaming [18], suggesting nightmares occur as a result of intense pre-sleep negative emotions. Moreover, nightmares may prompt bedtime worries and anxiety and can therefore result in bedtime resistance behavioural problems [17], such as leaving the bedroom, stalling bedtimes, or the need to fall asleep next to the parent [4]. It is also reasonable to note the bidirectional relation between nightmare frequency and child anxiety [18,19]. Indeed, night time fears, the two neurodevelopmental conditions, ASD and FASD, were more likely to experience nightmares, and whether there is a correlational trend between nightmare content and psychometric outcomes of affect, behaviour, or executive control.

Ethical Approval
This study was approved by the UCL Institute of Education Research Ethics Committee (Approval number 16683/001). All caregiver participants gave written informed consent.

Questionnaires
This study used several questionnaires to assess parental reported sleep and daytime functioning, and relevant background information of the child.
Children's Sleep Habits Questionnaire (CSHQ; [46]): A 33-item caregiver report which screens for common sleep problems in school aged children. It is rated on a 3-point scale ('rarely' for an event that occurs 0-1 times per week; 'sometimes' for an event that occurs 2-4 times per week; 'usually' for an event that occurs between 5-7 times per week). The CSHQ is a widely used assessment tool in paediatric sleep, with high internal validity and Cronbach alpha score of 0.83 [47]. Items are grouped into eight subscales: bedtime resistance, sleep onset delay, sleep duration, sleep anxiety, night waking, parasomnia, sleep disordered breathing, daytime sleepiness. Clinical scores are determined as 41 or above. Nightmare frequency and content were taken from Subscale Item 6: Parasomnias ("Does your child have nightmares? If so, describe the frequency and content"). Spence Children's Anxiety Scale (SCAS; [48]): A 39-item caregiver report which screens for clinical anxiety levels in children aged between 6 and 16. It is rated on a 4-point scale (0 = never to 3 = always). The SCAS is a widely used instrument with high internal validity and a Cronbach's alpha score of 0.92. The SCAS assesses various anxiety symptoms specified in the DSM-V, including subscales for separation anxiety, social phobia, obsessivecompulsive disorder, panic, physical injury fears, and generalised anxiety disorder. Clinical scores are determined as 31 and above [49].
Behaviour Rating Inventory of Executive Function (BRIEF; [50]) An 83-item caregiver report which screens for atypical executive functioning in children aged between 5 and 16 years. It is rated on a 3-point scale (1 = never to 3 = often). The BRIEF is a widely used measurement of executive functioning with a Cronbach alpha score of 0.8-0.98. Eight aspects of executive functioning are assessed. Subscales are working memory, inhibition, shifting, emotional control, and planning and organisation. Clinical scores are determined as 65 and above [50].
Child Behaviour Checklist (CBCL; [51]). A 118-item caregiver report which measures maladaptive behaviours in children aged between 5 and 16. It is rated on a 3-point scale (0 = not true to 2 = very true). The CBCL is widely used in research settings with mental health workers, hospital staff, foster parents, clinicians, and teachers. It has high internal validity, with a 0.94 Cronbach's alpha score. Subscales are divided into 'Internalising' and 'Externalising' behaviours, as well as the subsets of: withdrawn, somatic complaints, anxious/depressed, social problems, thought problems, attention problems, delinquency, and aggression. Clinical scores are determined as a total of 64 and above [52].
The Childhood Autism Rating Scale, Parents Version (CARS; [53]). A 15-item caregiver questionnaire that screens for the severity of autistic symptoms, using a seven-point (including midpoint) Likert Scale, ranging from typical to atypical behaviour. Categories are: relating to people, imitation, emotional responsiveness, body use, object use, adaptation to change, visual responses, listening responses, taste, smell, touch responses, fear or nervousness, verbal communication, nonverbal communication, activity levels, intellectual responsiveness, and general observations. The CARS demonstrates moderate to good sensitivity and specificity (81.4% and 78.6%, respectively) and good internal consistency (Cronbach's alpha = 0.79) however cannot be used in place of a diagnostic assessment. A CARS score of > = 33 indicates possible ASD [54].
Neurobehavioural Screening Tool (NST; [55]). A ten-item binary checklist that screens for possible FASD in children. Questions examine whether children meet the more common neurobehavioural characteristics of FASD, however these are not always accurate or representative of all children with FASD. Categories are: acting young, lying and cheating, lacking guilt after misbehaving, difficulty concentrating, impulsivity, hyperactivity, displays of cruelty, stealing at home, and stealing outside of home. The NST has low sensitivity but high specificity (62% and 100%, respectively) and in the absence of a more accurate measurement tool, is a widely used screening mechanism for FASD in children. Scores above 8, plus confirmed prenatal alcohol exposure indicate that a FASD diagnostic evaluation should be conducted [55].

Participants
Participants were caregivers of children aged between 6 and 15 years old (between the 6th and 16th birthday). This age group was chosen since it represents a large age range from which a cross sectional analysis could be made.
All children had received either a diagnosis of FASD or ASD, and no diagnoses of any physical or mental health conditions in the TD group. Screening tools for FASD and ASD were additionally included in the battery of questionnaires that caregivers completed. A number of children also had secondary diagnoses, which are outlined in Table 1 below. This study was initially advertised as an online questionnaire assessing sleep and daytime functioning in children with FASD or ASD. Caregivers of TD participants were recruited through three schools in West London, South London, and Milton Keynes. Caregivers of children with FASD were recruited through the UK FASD Network mailing list, while caregivers of children with ASD were recruited through online ASD forums. Caregivers were directed to an online questionnaire, at the end of which there was an option to receive feedback on the child's score as well as sleep hygiene intervention ideas.
A total of 322 caregivers completed the online questionnaire. Nine responses were excluded as large sections of the questionnaire were not completed. Twenty-four responses were excluded as they did not have a diagnosis of FASD or ASD, and were not TD. A further 12 were excluded as they did not meet the age criteria. One-way between group ANOVAs indicated no age (F(1,235) = 1.06, p = 0.43, η p 2 = 0.85) or SES differences (F(1,3) = 1.06, p = 0.49, η p 2 = 0.01), but there were significant differences between sex, with significantly more boys than girls (F(1,2) = 6.58, p = 0.01, η p 2 = 0.02). The final sample consisted of 277 participants, outlined in Table 1.

Statistical Analysis
We performed quantitative methodology using SPSS Statistics for Windows, Version 23.0 (IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY, USA: IBM Corp). The questionnaire had a 92% response rate with 146 missing values. Where there were missing data, these were imputed in SPSS. One-way ANOVA with Gabriel post-hoc tests were conducted to examine group differences of nightmare frequency, child sleep habits, child anxiety, executive function, and behaviour. Pearson's correlation coefficient was performed to test if the nightmare frequency was associated with clinical outcome measures. Qualitative content analysis was conducted using NVivo V.12 on novel items, which could not be categorised into the Garfield [56] and Shredl and Pallmer [21] categorisation of 'being chased', 'injury or death', or 'sensing scary'. Table 2 shows the percentage of children whose nightmare content fell into the categories of 'being chased', 'injury or death', and 'sensing scary', as reported by Garfield [56] and Schredl and Pallmer [21].  Table 3 shows the characters each group experienced within the three themes. Children in the TD group reported being chased by a 'dog' whilst children in the clinical groups reported being chased by more fantastic characters such as 'aliens', 'monsters', or 'large insects'. Similarly, whilst TD children reported dreaming of the death of 'people' and 'puppies', the children from the ASD and FASD groups had dreamed of the death of unusual characters like 'toys' and 'aliens', or significant family members. Children in the FASD group were the only ones to dream of concrete characters (e.g., witches) in the nightmare theme of 'sensing scary'.  Table 4 presents the results of the thematic analysis based on the nightmare themes of Gakenbach [57] and Revonsuo and Salmivalli [58]. Nightmare content from the FASD group was classified into the theme of fantasy (30%), persecuted/being chased (18%), and robbery/theft/crime (9%), but less for the theme of real life (9%), while the pattern was somewhat similar between the TD and ASD groups. A full list of nightmare content can be found in the Appendix A.

Group Comparison on Nightmare Frequency and All Clinical Outcome Measures
One-way ANOVA with the Gabriel post-hoc test was conducted to examine if nightmare frequency, child sleep habit, child anxiety, behaviour, and executive function varied between the three groups. Under the timeframe from 'less than monthly (1)' to 'nightly (5)', nightmare frequency was significantly varied among the three groups (F(2, 58) = 7.142, p = 0.002, η 2 = 0.203), with the children from the TD group (M = 2.25, SD = 1.06) scoring significantly lower in nightmare frequency than the ASD group (M = 3.57, SD = 1.09, p = 0.009) and the FASD group (M = 3.58, SD = 1.09, p = 0.001), while there were no differences of nightmare frequency between the ASD and FASD groups.

The Relation between Nightmare Frequency and Clinical Outcome Measures
Pearson's correlation coefficient with bootstrapping was conducted to examine whether nightmare frequency was associated with any of the clinical outcome measures in each group of children. As Table 6 shows, the nightmare frequency of TD children was positively significantly associated with maladaptive behaviour and executive function, while the nightmare frequency of the children from the FASD group was positively significantly related to child maladaptive behaviour and anxiety. No correlations were found withing the ASD group.

Discussion
This is the first study to explore the associations between nightmares and psychometric outcomes in children with ASD, FASD, and in TD children. We aimed to explore nightmare frequency and content among the three groups, as well as correlational associations between nightmare frequency, anxiety, maladaptive behaviour, and executive functioning in children with ASD, children with FASD, and TD children. Children in both clinical groups scored significantly higher on anxiety, maladaptive behaviour, and executive functioning measures than their TD peers. Children in both clinical groups experienced significantly more sleep problems than the TD group. Within our sample, children with ASD and FASD had significantly more nightmares than TD children: seventy-four percent of children in the FASD group were reported to have nightmares; 40% of the ASD group were reported to have nightmares, which was in comparison to 21% reported in the TD group. Nine percent of the TD group reported to have weekly nightmares, which is comparable to previous data which reported weekly nightmares in 5.2% of a sample of 6359 TD children [26]. In addition, we report that within our sample nightmare frequency was significantly associated with maladaptive behaviour and anxiety in children with FASD. Nightmare frequency was also significantly associated with maladaptive behaviour and executive functioning difficulties in TD children. These results demonstrate much needed further investigation into this area of research.

Nightmare Content
Children in the TD group were more likely to report nightmares involving insects and dogs. This is in line with previous studies in TD populations, reporting frequent appearances of animals in children's dreams [22]. Children in the ASD group reported more death, injury, and violence related nightmares than the TD group, but also similar amounts of real life, being chased, fantasy, and separation related nightmares. Children in the ASD group were more likely to report nightmares of themselves being injured, whereas children in the TD and FASD groups were more likely to report nightmares of significant others being injured. In one comparison study [38] between children with ASD, siblings, and a control group (n = 345), it was noted that the experience being raised, i.e., having more than one caregiver and changing caregivers frequently, was positively associated with nightmare frequency. This could perhaps be an explanation for the higher number of appearances of attachment figures within the FASD nightmares, as well as higher frequency of nightmares in the FASD group. Children in the FASD group reported the most robbery, theft and crime related nightmares, persecutory nightmares such as being chased (usually by wolves, monsters, or strangers), separation nightmares, and nightmares involving people, toys or animals with significance being injured, dying, or taken away. According to the Threat Simulation Theory, threatening experiences during wake are rehearsed during sleep, possibly due to dream consciousness working as an evolutionary defence mechanism, in which the brain repeatedly simulates threatening events [58]. Children with FASD are significantly more likely to experience attachment issues as a result of upbringing differences, including the frequent changing of caregivers. This may explain why threatening situations, such as being chased or family members dying were rehearsed by children with FASD. Meanwhile, injury to self was more prominent in children in the ASD group which could perhaps be due to the 'self-other' perception within ASD [59].

Nightmare Frequency and Psychometric Associations
In the TD group, nightmare frequency was significantly associated with composite CBCL scores, as well as the subscales of anxious and depressed behaviour, attentional problems, and aggression. Nightmare frequency was significantly associated with composite BRIEF scores, as well as the subscales of working memory, shifting, planning and organising, initiation, and emotional control. In the FASD group, nightmare frequency was significantly correlated with composite CBCL scores, as well as the subscales of withdrawn, somatic, anxious/depressed, and thought problems. Nightmare frequency was also significantly associated with total SCAS as well as the subscale of panic symptoms. No associations were found within the ASD group. One narrative review [37] has indicated the positive relation between nightmare frequency and psychiatric outcomes such as posttraumatic stress, depressive, bipolar, anxiety, and obsessive compulsive disorders, but not in ASD where they report the association between nightmares and psychometric outcomes is less clear. In this narrative review, adults with ASD were reported to have significantly fewer nightmares than controls, and there appeared to be no relationship between ASD and nightmares in children and adolescents [37]. Inconsistent to previously reported data on nightmares in children with ASD, our study reports that children with ASD experienced a significantly higher number of nightmares than their TD peers. The frequency of nightmares has previously been positively associated with ASD symptomology and IQ [31], but not to psychometric outcomes, whilst parasomnias in general have previously been reported to have various associations with psychometric outcomes, some positive, some non-conclusive [29]. This is consistent with our findings. This nonlinear relationship may be due to lower scores on the CBCL, SCAS, and BRIEF in contrast to the ceiling effect noted below.
This lack of association in the ASD group may also be due to the processes involved in emotional regulation and hypersensation. Higher emotional regulation difficulty is associated with higher nightmare frequency in light of the influence of waking intense emotions [60], and the theory of the adaptive emotion regulation function in dreaming [22]. Based upon these findings, emotion regulation difficulty could be the underlying factor for the nightmare frequency in the ASD group, in which difficulties in emotion regulation are well reported [28]. In addition, hypersensitisation in ASD may be a contributory factor. Early adverse experiences during sensitive windows can have a long-lasting effect on nightmare frequency, that is, a trait variable maybe associated with sensory processing sensitivity and a trait marker encompassing increased emotional reactivity, greater depth of processing, and subtle awareness of environment stimuli. Experiences resulting from hypersensation can be distressful and contribute to the occurrence of nightmares, according to the Continuity Hypothesis in which negative experiences in waking life transfer to dreaming, alongside intense negative emotions before sleep [60]. In addition, according to the nightmare theory of the selective mood regulatory function of dreaming [61], the occurrence of nightmares are due to the emotional surge in REM sleep that dreaming fails to adaptively regulate. Based on the findings demonstrating the disruption of REM sleep in ASD [62], it might be the case that REM disruption contributes to the failure of dreaming to adapt waking emotions, which thereby leads to nightmares.

CSHQ/Nightmare Frequency
Although nightmare frequency positively correlated with CBCL, BRIEF, and SCAS in the TD and FASD groups, nightmare frequency was not associated with sleep related measurements from the CSHQ. These measure bedtime resistance, sleep onset delay, sleep anxiety, night waking, parasomnia, sleep disordered breathing, and daytime sleepiness. This may be because CSHQ scores did not contribute to nightmare frequency as sleep quality is generally not a leading factor of nightmare frequency [22]. However, within some subscales of the CSHQ, particularly sleep anxiety, night waking, and parasomnias, caregivers were asked whether during pre-sleep children experienced negative affect, troubling thoughts, or whether children woke up at night screaming or were inconsolable. It would have followed, then, for these items within the CSHQ subscales to correlate with nightmare frequency. Perhaps due to the small number of participants who reported nightmares such data were not found, or this may be due to the ceiling effect of high nightmare frequency and high parasomnia scores. Within previous analyses of the larger dataset [43,44], parasomnias were significantly associated with executive functioning and behavioural outcomes in the TD and FASD group, but less so in the ASD group. Night wakings however were associated with somatic complaints and attention in the TD group, which is comparable to the present analysis. However, this result warrants further investigation.

CBCL/Nightmare Frequency
In the TD group, nightmare frequency was significantly correlated with several CBCL subscores, particularly those related to affect. These were the subscales of anxious/depressed, attention, and aggression. In the FASD group, nightmare frequency was significantly correlated with withdrawn, somatic, anxious/depressed, and thought problems. Influential factors of negative emotions before sleep [18], the susceptibility of dreaming to waking intense emotions [23], as well as the theoretical models of continuity hypothesis and pre-sleep states as the behavioural problems displayed by children could explain the relationship between the manifestation of intense emotional experiences and nightmare frequency and content.
The CBCL assesses children's withdrawn, somatic, anxious, social, thought, attention, delinquency, and aggressive behaviour, however the instrument itself is used differently between the two clinical groups. Due to the widespread neural damage incurred by prenatal alcohol exposure, children with FASD may have difficulties with a number of behaviours, which are compounded by the environmental stressors common to the psychosocial environment [38]. Children with FASD therefore tend to score within clinical levels in a pattern-like manner on the CBCL, due to developmental, regulatory, affect-based and executive functioning issues, and as a result the CBCL is used as a screening tool for FASD [38]. Behavioural manifestations in ASD are not as homogenous, and children with ASD present with wide range of internalised and externalised behaviours across the several subsections of the CBCL [63]. This heterogeneity may explain why CBCL scores were not associated with nightmare frequency in the ASD group. Within the TD group, the present results are comparable to previous data on nightmare frequency and aggression in children.
In the previously mentioned large scale study [26], nightmare frequency was significantly correlated with aggression, as well as mood disturbance and hyperactivity.

SCAS/Nightmare Frequency
TD children were more likely to experience nightmares if they also scored within the clinical range of 'anxious/depressed' on the CBCL. In the FASD group, SCAS total score and the subscale of panic symptoms were significantly correlated with nightmare frequency. Anxiety subscales measured panic symptoms, separation anxiety, physical injury fears, social anxiety, obsessive compulsive symptomology, and generalised anxiety. It is reasonable to assume a relationship between anxiety and nightmare frequency given the involvement of pre-sleep emotions on the occurrence of nightmares, as well as the role of emotional continuity and threat simulation [18,23,58]. Anxiety is consistently associated with sleep problems both in children with ASD [28] and FASD [44]. In addition, nightmares are conceptualised as an anxiety-based phenomenon [60]. Additionally, nightmares are connected to emotional regulation [61], fear [64], repressed fear [65], and the transformation of shame into fear [66]. As mentioned above, threatening experiences during wake may be rehearsed during sleep which may explain why anxiety was significantly associated with nightmare frequency within our sample.

BRIEF/Nightmare Frequency
Executive functioning related measurements that were measured related to working memory, emotional control, inhibition, monitoring, organisation of materials, planning and organising, and shifting. Simor et al. [67], for example, refer to the association between executive functioning and nightmare frequency due to impaired fronto-limbic functioning during REM sleep. It is interesting that this association was found only in the TD group and not the clinical groups, both of which are clinical conditions related to frontal lobe functioning, but this is likely due to a higher proportion of the clinical group already reaching maximum executive functioning scores and creating a ceiling effect which was not seen in the TD group.

Strengths and Limitations
This is the first study to report nightmare frequency and content in children with FASD. Sleep problems in ASD and FASD are currently understudied, particularly those pertaining to behavioural and affect based outcomes. Through further research such as this it may be possible to develop sleep interventions for children with neurodevelopmental conditions. This study additionally benefited from a large sample size from which nightmare data could be extracted, but further studies on nightmares in these clinical populations should take into consideration the smaller subsamples of children who experience nightmares. Some of the strengths of this study are within the large sample size, its novelty, and access to a relatively understudied population of children with FASD.
Several limitations are also of importance here. Firstly, parental reports are subjective views of children's sleep problems, and further analysis can benefit from objective sleep measures such as actigraphy or polysomnography (PSG). Such objective investigations may be used to correlate PSG or actigraphy data with nightmare frequency or content. Secondly, the results reported here may be attributed to a selection bias, in that parents of children with sleep problems were more likely to want to take part, and the sample may not have had the heterogeneity of a population-based sample. Around a third of participants had to be excluded because of the lack of a formal diagnosis of FASD or ASD, due to a general lack or delay in diagnostic services available in the UK. Thirdly, predictions within this study are statistical descriptions. Without an experimental design and with the limits of cross-sectional data, causation cannot be implied. However, this study has provided a platform for the need for further examination using objective sleep and psychometric measures in these clinical populations.

Summary
We report that, analogous with theory, nightmares may be associated with waking emotions in children with ASD or FASD, although the causative, bidirectional associations are less clear. The cross-sectional nature of the present data means that any inference about direction of the associations between nightmares and psychometric outcomes can only be hypothetical, and further analysis could possibly assess nightmare content and frequency in greater depth. We recommend the development of sleep intervention strategies to reduce general anxiety, as well as affect, aggression, hyperactive, and other maladaptive behaviours of the child during pre-sleep activities, thereby reducing the occurrence of nightmares. Data Availability Statement: All data pertaining to this paper can be found in Appendix A.

Conflicts of Interest:
The authors declare no conflict of interest. TD At least once a week, they tend to involve wolves, aliens or family members going away/being hurt.

TD
Rarely at the minute and usually he will get up out of bed but not fully awake, needs consoling but will go back to bed easily and fall back to sleep.

FASD
He tells me he has nightmares most nights. He dreams people die or he dies. He dreams he is alone and can't find anyone. He feels he can't come and get me or call out for help as he thinks there are people in the house who will come and kill him.

FASD
Usually twice a week shouting n muttering in sleep is noted but child will not recall a nightmare. Once a fortnight usually a nightmare about her safety or my safety. Yes, sometimes, but content almost never disclosed, just 'scary'.

AUTISM
Patches of nightmares, subject not always disclosed. A couple of times a man who followed son and friend in school and in wood and clearing and had fireballs for hands.
AUTISM I think he sometimes has nightmares, but he is unable to explain the content of his dreams. This occurs probably 2 or 3 times a week.

AUTISM
Sometimes he does; he rarely remembers but they seem to involve him being chased. He went through a rough patch when someone at school told him about 'Five Nights at Freddie's', he was anxious and had nightmares about that for a while.
AUTISM Nightmares 2-3 times a week usually of some form of harm happening to family member.
AUTISM Sometimes-only a few times a month. He does talk in his sleep though and sleep walk more frequently.
AUTISM Over past few months since starting high school has been happening about 10 times a month.

AUTISM
Yes, probably 2 to 3 times a week. Sometimes night terrors but this is lessening with age. AUTISM Apparently most nights but he can't describe them.

AUTISM
Occasional nightmare about being lost and unable to find people or being chased by wolves. AUTISM Wakes crying saying had bad dream a hand full of times a year.