Project Summary/Abstract
Suicidal ideation and behavior are growing public health problems in the United States. Unfortunately, our current
ability to predict suicide is only slightly above chance, which may be attributable to an overreliance on
distal/cross-sectional risk factors that are weak proximal predictors of suicide risk. Modeling the complex process
by which atypical sleep impacts daily functioning in conjunction with established proximal risk factors can aid in
identifying contexts and time periods of greatest suicidal risk, modeled at the individual level. The proposed study
builds upon our team’s extensive expertise in sleep/wake cycles, psychophysiology, deep phenotyping, and
multi-method, multivariate, ecologically valid models of suicide vulnerability in high-risk psychiatric populations.
We will examine how a holistic model of atypical sleep relates to known trait (baseline neurocognitive
performance; e.g., greater impulsive tendencies, higher loss sensitivity, reduced ability to regulate emotions) and
state (time-varying, occurring hours to days before SI/SB; e.g., momentary fluctuations in emotional reactivity,
impulsivity; greater emotional lability; greater isolative tendencies), risk factors for suicide, and examine how
these factors together proximally influence suicidal ideation and confer risk for future suicidal behavior. We will
recruit 200 psychiatric inpatients at high risk for suicide and conduct a baseline assessment of sleep/wake
functioning and trait risk factors and use laboratory-based tasks coupled with psychophysiology (i.e., event-
related potentials, heart rate variability, and electrodermal activity) to phenotype risk processes linked to arousal
and cognitive systems. This baseline assessment will be followed by four weeks of EMA and digital phenotyping
coupled with SAFTE-derived actigraphy to characterize key state risk factors. We will conduct follow-up
assessments at 1-, 3-, and 6-months post hospital discharge to determine how our proximal model of risk
prospectively predicts SI and SB. The proposed study aims to characterize proximal risk for suicide using
intensive longitudinal methods and to identify “windows” of greatest risk for suicide, which may vary from person
to person, that serve as markers for intensive intervention. Finally, we will leverage this extensive dataset to
develop a model of the sleep-suicide relationship emphasizing the contribution of trait and state factors. The
results of this study have the potential to greatly enhance our understanding of the phenomenology of suicide
risk as it exists in the real world, with the potential to improve our ability to predict, prevent, and intervene using
both traditional and technology-enhanced psychotherapies.
