PROJECT SUMMARY/ABSTRACT
This proposal responds to RFA-MH-22-135. Non-disclosure of suicide ideation prior to attempt is normative
and even more common in later life. Similarly, our preliminary data indicates lower inclusion of social contacts
in suicide safety plans among at-risk older adults. Unfortunately, interventions to increase informal help
seeking are ineffective to date. Therefore, major roadblocks exist in the interpersonal processes seen as
essential to late-life suicide prevention, and new targets are needed for increasing help-seeking behavior. The
purpose of this longitudinal observational project is to understand why older adults do not disclose suicide
ideation or seek help from others amidst suicide crises. We focus on social cognitive abilities and biases,
which influence how individuals perceive (or misperceive) others and have previously been associated with
loneliness, suicidal ideation and behavior. The premise of this research is that social cognitive impairments
impact the positive and negative valence systems involved in affiliation and the social structures that determine
help seeking in suicide. Facilitating this research, our group has developed and validated a dyadic social
affiliation task that objectively quantifies RDoC positive and negative valence systems involved in support
seeking. This task enables evaluation of valence systems across multiple units of analysis, including behavior,
self-report, facial affect and natural language processing. Measurement of these intrinsic processes will be
integrated with quantification of extrinsic help-seeking opportunity, via structural social network assessment. A
translational aspect of our proposed social network analyses is to augment network measurement with the
social elements of suicide safety planning. In a longitudinal study, we will recruit a sample of diverse older
adults from a variety of urgent, primary, and mental health care settings. Recruitment will be stratified by
groups defined by current active suicide ideation, depressive symptoms without ideation or attempt histories,
and healthy comparators. In Aim 1 of the study, we will compare groups on social cognition, positive and
negative valence indicators from our social affiliation task, along with social network structure, including past
help seeking and disclosure in the subgroup with current suicide ideation. In Aim 2, we will administer mobile
emotion recognition tasks and social affiliation in predicting in real-time suicide disclosure through ecological
momentary assessment over 30 days. We will then evaluate whether social cognition and affiliation markers
predict trajectories of help seeking and suicide ideation through longitudinal data gathered over one year.
Exploratory analyses will leverage the data derived by applying time series network analyses and natural
language processing to model processes underlying non-disclosure. Our long-term goal is to translate this
research to novel personalized interventions that improve interpersonal processes embedded in both current
and novel suicide prevention approaches for older adults. This proposal responds directly to NIMH Strategic
Plan, Aim 2.2 and 3.2, and the NIMH’s Priority Area in Digital Mental Health.