Simons Sleep Project (SSP): An open science resource for accelerating scalable digital health research in autism and other psychiatric conditions

Abstract

Wearable and nearable devices offer a novel opportunity to measure extensive behavioral and neurophysiological data directly from participants in their home environment. The Simons Sleep Project (SSP) was designed to accelerate research into sleep and daily behaviors in individuals with autism using such techniques. This open-science resource contains raw and processed data from Dreem3 EEG headbands, multi-sensor EmbracePlus smartwatches, and Withings Sleep mats, as well as parent questionnaires and daily sleep diaries. Data were collected successfully for >3600 days/nights from 102 adolescents (10-17 years old) with idiopathic autism and 98 of their non-autistic siblings. Whole-exome sequencing data is also available for all participants and their parents. To demonstrate the utility of this extensive dataset, we first present the breadth of synchronized high-resolution data available across multiple sensors/devices. We then demonstrate that objective sleep measures (e.g., total sleep time) from the three devices are more accurate and reliable than parent reported measures and reveal that sleep onset latency (SOL) was the only objectively defined sleep measure that differed significantly between autistic children and their siblings (of those examined in the study). Moreover, SOL was reliably associated with the severity of multiple behavioral difficulties in all children, regardless of autism diagnosis. These results highlight the importance of measuring sleep directly from participants using objective measures and demonstrate the extensive opportunities afforded by the SSP to further study autism and develop new digital phenotyping techniques for multiple research domains.