Precision sleep signatures to predict mental health outcomes in youth.
Article
Soehner, Adriane M, McMakin, Dana L, Jalbrzikowski, Maria et al. (2026). Precision sleep signatures to predict mental health outcomes in youth.
. BIOLOGICAL PSYCHIATRY-COGNITIVE NEUROSCIENCE AND NEUROIMAGING, S2451-9022(26)00231-4. 10.1016/j.bpsc.2026.07.017
Soehner, Adriane M, McMakin, Dana L, Jalbrzikowski, Maria et al. (2026). Precision sleep signatures to predict mental health outcomes in youth.
. BIOLOGICAL PSYCHIATRY-COGNITIVE NEUROSCIENCE AND NEUROIMAGING, S2451-9022(26)00231-4. 10.1016/j.bpsc.2026.07.017
The transition from childhood to adolescence heralds a marked escalation in pediatric mental health risk, as well as major developmental shifts in sleep. Poor sleep health is a common, causal, and modifiable transdiagnostic mental health symptom and risk factor in youth. Yet, unraveling the sleep-mental health risk relationship over adolescence has proven to be deceptively challenging. Sleep health arises from complex biopsychosocial processes and can be measured across multiple methods and time scales. The interplay between profound developmental shifts in sleep over adolescence and the high-dimensional nature of sleep measurement often leads to significant data heterogeneity. As a result, computational approaches are necessary to parse out typical variation from at-risk patterns that may reflect warning signs of emerging mental illness. In this review, we propose sleep signatures as a strategy to characterize sleep health and accurately predict psychiatric outcomes in adolescence. Sleep signatures are within-person combinations of multiple sleep features that more holistically characterize individual-level patterns of sleep health. We propose the multidimensional sleep health framework as a basis for sleep signature development and discuss unique complexities in sleep measurement for adolescents, highlighting classic sleep measurement methods and new opportunities provided by modern wearable and smartphone-based sleep monitoring. Next, we review computational techniques to derive multidimensional, multimodal sleep signatures in a developmental context, focusing on variable-centered (factor analysis) and person-centered (clustering) approaches. Finally, we offer a roadmap for leveraging these approaches to identify sleep signatures salient to adolescent mental health through the ongoing Pediatric Precision Sleep Network project.
