LAUSR

Abstract Genetic and other studies over the past decades have made it clear that it may be productive for both the diagnosis and treatment of depression if disease manifestation were to be described in multiple dimensions rather than by classical categories. As global smartphone usage has skyrocketed, one natural approach to improved phenotyping that can be deployed at scale is using smartphone-based apps to capture behavioral information from consented users. In particular, using such digital phenotyping techniques with a focus on passively-collected data can potentially inform on an individual's mood, sleep quality, physical activity, cognitive function, and both digital and in-person social activities. Tracking these over time can look at, e.g., the circadian rhythms of these clinically relevant behaviors, and more generally give insights into the longitudinal trajectories preceding and during depressive episodes. These technologies also have the potential to assist researchers in collecting large-scale behavior-rich phenotype datasets while keeping patients engaged over the course of the study process, for example, in the pursuit of recruiting 1 million patients with depression for genotype-phenotype association.