Circadian activity and adhd: integrating genomic data with wearable technology

Atypical patterns of physical activity and circadian rhythmicity often co-occur with Attention-Deficit/Hyperactivity Disorder (ADHD), which contributes to worsened mental and physical health conditions. The utilization of large-scale digital phenotyping allows for objective quantification of the relationship between diseases and activity, offering improved temporal resolution and reliability compared to self-reported measures. In this study, we employed fine-grained wearable-derived measures to explore the intricate connections between ADHD genetic liability and various physical activity and sleep metrics in a time- and context-specific manner. Genomic and wearable-derived measures were collected from two community-dwelling cohorts: the UK Biobank (N = 103,712) and Adolescent Brain Cognitive Development Study (ABCD, N = 5,686). In the UK Biobank cohort, we examined genetic correlations between ADHD and physical movements based on time-of-day and day-of-week stratifications. Additionally, we assessed ADHD polygenic score (PGS) associations with activity duration across various intensity levels, daily and hourly. Precise sleep metrics were characterized using Fitbit devices with heart rate monitors in the ABCD cohort, where we explored the associations between ADHD-PGS and multi-stage sleep duration, quality, and fluctuations, alongside physical activity. While there was little to no genetic correlation between ADHD and daily average acceleration and sleep duration, robust patterns of correlations were observed across hourly activity within the day. Specifically, ADHD and overall acceleration exhibited reverse genetic correlation patterns across 24 hours, with positive correlations during late-night hours (12 am – 5 am, rg = 0.17 ∼ 0.36) and negative correlations during the day (rg = -0.18 ∼ -0.23). Further investigation into activity subtypes revealed that individuals with high ADHD-PGS had significantly reduced sleep time between 12 am and 6 am (β = -0.019∼ -0.036), but higher levels of sleep from 8 am to 7 pm (β = 0.016 ∼ 0.036). Throughout the day, high ADHD-PGS was associated with low sedentary behaviors and low moderate to vigorous activity, but high low-intensity activity. These effects were more prominent on weekdays than on weekends. All p values < 6e-4. Incorporating data from the ABCD study, the trends of associations for overall activities were largely consistent between adults and preadolescent children, with additional findings from the refined sleep measures yielding a positive correlation between ADHD-PGS and increased day-to-day variability of sleep onset (β = 0.05, p = 8e-3), but not the daily sleep onset time per se. The present study utilized large-scale genotyping data and wearable technology to investigate the relationships between ADHD, activity patterns, and sleep in two community-based cohorts. ADHD genetic risk is associated with delayed sleep-wake cycle, and increased daytime sleeping. The observed differences in energy expenditure in different activity intensities may reflect underlying neurobiological mechanisms related to ADHD symptomatology, such as attention problems and deficits in executive functions. These findings highlight the importance of using refined objective measures to understand the mechanisms underlying ADHD-related disturbances and emphasize the need to address sleep and circadian disruptions as part of a comprehensive approach to managing ADHD symptoms.