Although depression commonly precedes late-life suicidal behavior, clinicians still cannot confidently identify depressed elderly who are most likely to attempt or die by suicide. Thus, there is a great need for better predictive models regarding suicidal behavior in the elderly. This revised R01 (MH085651) application is to investigate specific cognitive vulnerabilities to late-life suicidal behavior. We focus on features that may cause accumulation of stressors, undermine deterrents, and facilitate the final decision to take one's life. Our preliminary data indicate that deficits in (1) specific aspects of cognitive control that involve reward/punishment processing, and in (2) social cognition distinguish depressed elderly suicide attempters from depressed non- suicidal elderly, while the two groups show similar global cognition, working memory, and forward planning. Building on this preliminary evidence, this new-investigator R01 will include key cognitive probes in a large- enough sample to test hypotheses that impairments in decision-making, affective processing, reversal learning, and social cognition are specifically associated with suicide attempts in depressed elders. We propose to assess 100 suicide attempters, 80 non-suicidal depressed individuals, and 60 non-psychiatric control subjects, aged 60 and older, using theory-driven computerized assessments as well as traditional tests of cognitive performance. Participants will undergo extensive clinical characterization of their suicidal behavior, psychopathology, psychosocial stressors, physical health, possible brain injury from suicide attempts, and medication exposure. The three groups will be similar in demographic characteristics and medical illness burden, and the two depressed groups will have similar severity of depression. To determine whether the identified impairments persist over time despite changes in mood state, we will repeat cognitive assessments four months after baseline (when substantial clinical improvement can reasonably be anticipated based on our pilot data). We will also prospectively explore the effect of cognitive status on suicide-related outcomes during this follow-up period. In collaboration with the biostatistical team of our late-life depression center and our external statistical consultant, we propose to use multivariate analyses of covariance to compare cognitive functions across groups, as well as discriminant function analysis to create a compact cognitive battery and to test its utility for correctly identifying suicide attempters beyond known risk factors. We will use mixed effects models to examine stability of cognitive impairments across mood states. Statistical analysis will account for factors that may affect cognition: severity of depression, medical illness burden, serum anticholinergicity, and other relevant factors identified by preliminary analyses. This project builds upon an ongoing K23, where the PI has shown the feasibility of recruiting, assessing, and longitudinally following suicidal elders with a high rate of suicidal behavior during follow-up. The research project will be conducted at the University of Pittsburgh, in collaboration with the Experimental Psychology Department, University of Cambridge.
Understanding cognitive deficits associated with late-life suicidal behavior and their relationship to other risk factors may help to advance translational neuroscience in geriatric mental health, identify elderly people at risk for suicide, and help to develop individualized treatment strategies in the service of preventing suicide in older people, who have the highest suicide rate in the US. The compact cognitive battery for assessing suicide risk derived from this research can be used in future prospective studies and in clinical settings.
Showing the most recent 10 out of 32 publications
