We are requesting a One-Year Administrative Supplement to the NIMH R01 MH085651-10 Decision Processes in Late-Life Suicide. The parent grant is a 5-year case-control longitudinal study of decision- making in older adults who have attempted suicide, with a special focus on those who have carried out medically serious (high-lethality) suicide attempts. Our overarching aim is to understand decision processes that may contribute to suicidal behavior in old age. Neurocognitive and decision-related factors may play a particularly salient role in late-life suicide due to age-related cognitive decline. Older adults have a higher risk of suicide, and that risk is even further elevated by those with a diagnosis of dementia, especially in the early stages of dementia. We have recently reported that there is a subgroup of older people characterized by high-lethality suicidal behavior, late-onset depression, marked cognitive decline, and higher risk of future suicide attempts, possibly representing a subtype of people with prodromal dementia that have a very high risk of suicide. In this supplement, we propose to investigate how neurodegeneration in the locus coeruleus, entorhinal cortex, and hippocampus, representative of neurodegeneration that occurs in early Alzheimer’s, relates to deficits in decision-making and to suicide attempts in a sample of 135 older adults. We propose to perform complex behavioral tasks that mimic real- life decisions. As indicators of Alzheimer’s-relevant neurodegeneration, we will analyze archival and newly- obtained structural MRI data to assess hippocampal and entorhinal cortex atrophy and will acquire (in a subgroup) neuromelanin-sensitive turbo spin echo scans to measure locus coeruleus signal intensity. We hypothesize that entorhinal cortex-hippocampus atrophy and reduced locus coeruleus function predict impaired decision making due to poorer ability to use information about task structure and reduced exploration of available options. We further hypothesize that Alzheimer’s-relevant dysfunction in exploration and learning task structure, as measured from task-derived computational parameters, will be greater in depressed elderly who are at high risk for suicide (those with a history of suicide attempt and current severe suicidal ideation) relative to depressed elders with low suicide risk. The Supplement will take advantage of a well-characterized sample of depressed older adults participating in the parent grant and will enrich it with individuals with prodromal Alzheimer’s disease, taking advantage of the Pittsburgh Alzheimer’s Disease Research Center’s registry. The investigators will contribute their complementary expertise in suicide risk factors, decision making, reinforcement learning, computational modeling, neuroimaging, and dementia.