Anxiety Sat at the Centre of the Sleep Map

What the research found

Poor sleep and emotional difficulty travel together in autistic children. That much has been known for years. The harder question is how they connect: does bad sleep drive anxiety, does anxiety wreck sleep, or are both symptoms of something else again? A 2024 study in the Journal of Autism and Developmental Disorders takes a careful look at that tangle using a method called network analysis.

The researchers studied 240 autistic children aged between five and thirteen, with a mean age of 8.8 years. All were recruited from paediatric clinics in Victoria, Australia, and all had moderate to severe sleep problems that had lasted at least four weeks. Around two-thirds were boys. Children with an intellectual disability, and those from non-English-speaking backgrounds, were not included. Parents completed three questionnaires: one on sleep habits designed specifically for autism, one on emotional and behavioural difficulties, and one on autism-related traits.

Network analysis is worth explaining in plain terms. Instead of asking whether sleep and anxiety are linked overall, it maps every symptom as a point and draws a line between two points only when they are related after accounting for everything else in the picture. The result is a web. Some points sit at the busy centre, connected to many others; some sit at the quiet edge. The method cannot tell you which symptom came first, but it can show you which ones hold the whole structure together.

The central points were not the sleep symptoms. They were the emotional ones. Depression, anxiety and behavioural difficulties emerged as the most central variables in the network of eleven nodes. Anxiety was directly connected to seven other variables, more than almost any other symptom. Its strongest links ran to depression, to night waking and parasomnias, and to a cluster the questionnaire calls sleep anxiety and co-sleeping. Depression and anxiety also formed the strongest bridges between the emotional side of the network and the sleep side.

The authors also checked whether their map was stable, which matters because network models estimate a great many relationships at once and can be fragile in smaller samples. Stability is reported as a coefficient between zero and one, where higher is sturdier. The connections themselves and the bridge measures scored 0.75, which is comfortably robust. The measure of which individual nodes were most central scored 0.52, which is adequate but weaker, so claims about the exact ranking of central symptoms should be held a little more loosely than claims about the connections.

What it means for you and your child

The practical message is quietly useful. If your autistic child struggles with sleep, the instinct is often to treat sleep as a standalone plumbing problem: a better bedtime routine, a darker room, perhaps melatonin. Those things matter. But this study suggests anxiety sits close to the heart of the difficulty, wired into both the emotional distress and the specific sleep symptoms of night waking and fear around bedtime. Addressing the worry may do more than addressing the pillow.

That reframing fits what many parents already sense. A child who cannot fall asleep alone, who wakes in the night, who needs a parent in the room, is often a child who is anxious. Treating the anxiety, whether through the child’s clinical team, through anxiety-focused strategies, or through support at bedtime that lowers fear rather than simply enforcing routine, is a reasonable thing to raise with a professional. The study does not prescribe a treatment, but it points to where the pressure in the system seems to concentrate.

Now the honest limits, because they are real and the authors state them plainly. This is a snapshot in time. The data were gathered at a single moment, so the study cannot say whether anxiety causes the sleep problems, the sleep problems cause the anxiety, or a third factor drives both. A central position in the network is a strong hint about where to intervene, not proof of a cause. Anyone who tells you this study shows anxiety causes sleep problems has overstated it.

Several other cautions apply. Everything was reported by parents, which introduces the ordinary biases of a tired adult rating their own child. The depression questionnaire happened to include two items about sleep, which can artificially inflate the apparent link between low mood and poor sleep, a problem the authors flag as possible circularity. And the sample was deliberately narrow: autistic children without intellectual disability, all with existing sleep problems, all from English-speaking families in one Australian region. Whether the same map holds for children with intellectual disability, or for those whose sleep is only mildly disrupted, is simply unknown from this work.

So hold this study for what it is. It is one well-conducted analysis of a specific group, and its finding is a direction rather than a verdict: in autistic children with entrenched sleep problems, anxiety appears to be doing a lot of the connective work. That is a good reason to take a child’s night-time fear seriously as its own target, and a good reason to be sceptical of any single tidy fix. The value of a map is that it shows you where the roads meet. It does not tell you which way the traffic is flowing.