Using wearable accelerometers over 3 weeks, 52 actimetry measures were identified as associated with autism, with long inactivity periods and lower physical activity levels linked to less sleep, later sleep onset, and greater social impairment in autistic adults.
Key Findings
Results
52 actimetry measures derived from wearable accelerometers were associated with autism, achieving good discriminative performance.
Study recruited 318 autistic individuals and 130 non-autistic family members, collecting accelerometer data over 3 weeks (8,249 days total)
308 traits were examined using elastic net algorithm and linear mixed effects regressions
The model achieved an area under the curve (AUC) of 0.812 (CI: 0.761–0.862) in the training set
Validated in a test set with AUC of 0.756 (CI: 0.700–0.813)
154 distinct physical activity, sleep, and behavioral traits were identified from the recordings
Results
Both mean levels and day-to-day variability in several actimetry measures, including time spent sedentary and total light physical activity, were associated with autism.
Autistic individuals were more likely to be less active for longer periods and have lower overall physical activity levels
Long inactivity periods and lower physical activity levels were associated with autism
These patterns correlated with less sleep and later sleep onset in autistic individuals
Day-to-day variability in these measures was specifically identified as a feature associated with autism diagnosis
Results
Reduced physical activity during wake was more strongly associated with shorter sleep time in autistic individuals than in non-autistic relatives.
The association between reduced physical activity and shorter sleep was significantly stronger in autistic individuals compared to non-autistic relatives (likelihood ratio: 41.6; p = 1.13e-10)
For every hour of inactivity, autistic participants had on average approximately 23 minutes less sleep compared to approximately 17 minutes less in family members
This differential association suggests a unique relationship between activity and sleep in autism
Results
Reduced physical activity in autistic individuals was linked to increased social impairment as measured by the Social Responsiveness Scale.
The Social Responsiveness Scale (SRS) was used to quantify social impairment
Autistic individuals with lower levels of physical activity showed higher social impairment scores on the SRS
This association suggests a potential relationship between activity levels and core features of autism
Methods
The study enrolled autistic adults without intellectual disability and their non-autistic relatives using a family-based design with wearable devices.
318 autistic individuals and 130 family members were recruited
Participants wore accelerometer-based wrist-worn devices over 3 consecutive weeks
A data-driven approach was used to identify actimetry-derived features associated with autism
Study population was limited to autistic adults without intellectual disability
The family-based design allowed comparison between autistic individuals and non-autistic relatives, controlling for shared genetic and environmental factors
What This Means
This research suggests that autistic adults, compared to their non-autistic family members, show distinct patterns of physical activity and sleep that can be detected by inexpensive wrist-worn activity trackers worn over three weeks. By analyzing data from 318 autistic adults and 130 of their relatives — totaling over 8,000 days of recordings — the researchers identified 52 measurable features that together could reliably distinguish autistic individuals from their non-autistic relatives. The most prominent differences involved physical activity: autistic adults tended to have longer periods of inactivity, lower overall activity levels, and more day-to-day variability in these patterns. These activity differences were also linked to sleeping less and going to sleep later.
A particularly notable finding was that inactivity was more strongly connected to reduced sleep in autistic individuals than in their relatives. For every additional hour of inactivity, autistic participants slept about 23 minutes less on average, compared to about 17 minutes less for family members. Additionally, autistic individuals who were less physically active also tended to score higher on a measure of social difficulties, suggesting that low activity levels may be connected to core features of autism, not just general health.
This research suggests that low physical activity may contribute to both poorer sleep and greater social difficulties in autistic adults. Because this was an observational study, it cannot prove that increasing activity would improve sleep or social functioning, but the authors call for intervention studies to test whether programs aimed at increasing physical activity could improve quality of life for autistic individuals. The use of affordable consumer wearables also shows promise as a low-burden tool for studying behavioral patterns in autism research.
Weissenkampen J, Ghorai A, Carneiro T, Fasolino M, Gehringer B, Rajan M, et al.. (2026). Sleep and activity patterns in autism.. Autism : the international journal of research and practice. https://doi.org/10.1177/13623613251413538