SEE MAIN PROJECT WIKI FOR GENERAL INFORMATION ON STUDY BACKGROUND AND DATA COLLECTION **Primary Hypotheses** *Specific Aim 1:* Investigate age as a predictor of risk factors for cognitive impairment in this sample of AI/AN. Hypothesis 1a: Older AI/AN will have higher levels of biomarkers of allostatic load including inflammation (hs-CRP, TNF-alpha, IL1 beta, IL6, IL10), metabolism (cholesterol, HDL, LDL, triglycerides, hemoglobin A1c), steroids (cortisol, DHEA), anthropomorphic measures (BMI, waist-to-hip ratio), and autonomic nervous system activity (resting epinephrine & norepinephrine, heart rate variability, & BP). Hypothesis 1b: Older AI/AN will show worse cardiovascular functioning (e.g., higher ambulatory BP and less nocturnal BP dipping). We will also investigate age as a predictor of depression and rumination. We do not have a clear hypothesis in these cases because depression (and possibly rumination) is sometimes lower in older adults, which may relate to better coping skills or social support in those who surpass the average lifespan, or the involvement of other processes. *Specific Aim 2:* Assess whether exposure to discrimination and historical trauma drive the effect of age on risk factors for cognitive impairment (i.e., depression, rumination, and allostatic load). Hypothesis 2: Aging AI/AN will exhibit more risk factors for cognitive impairment but this will be especially true of those who indicate high perceived discrimination or historical trauma. *Specific Aim 3:* Assess whether exposure to discrimination and historical trauma partially explains the effects of age on risk factors for cognitive impairment. Hypothesis 3a: The association between age and allostatic load will be partially mediated by exposure to discrimination and historical trauma. Hypothesis 3b: The association between age and cardiovascular functioning will be partially mediated by exposure to discrimination and historical trauma. Again, the associations among age, depression or rumination, and exposure to colonial harms will be investigated, but may be more complex. **MEASURES FOR INCLUSION IN SECONDARY ANALYSES** **Independent Variable (Primary Predictor)** Age was self-reported by participants, who gave their age, in years, on the day they participated in the study. **Primary Outcome** *Depression.* The Center for Epidemiologic Studies Depression Scale (CESD-R) assessed the extent of depressive symptoms in the prior two weeks (e.g., “My appetite was poor”, “I couldn’t shake off the blues”, and “I felt depressed”). Response options ranged from 0 (0 or less than 1 day last week) to 4 (nearly every day for two weeks). Two items asking about suicidal ideation were not included in our assessment because the research staff did not have the expertise to follow-up with people who admit to suicidal ideation. This measure will be scored for each participant as the average of responses to the 18 questions. *Rumination.* Participants completed a face-valid measure of rumination on four occasions, between 1PM and 9:30PM after their lab session, using an electronic diary. Participants were asked, “How much were you thinking about problems at work or in your personal life?” and “How much were you thinking about times you were treated unfairly because of your race/ethnicity?”. Responses were made on a 0-100 sliding scale. *Allostatic Load.* An index of allostatic load will be calculated by dividing biomarkers of allostatic load into quartiles and then assigning a value of 1 for each biomarker for which participants appear in the top quartile (bottom quartile for DHEA and HDL-cholesterol). These values will then be summed into a single index of allostatic load. Biomarkers of allostatic load measured in this study include glucocorticoids (complete blood count glucocorticoid sensitivity [CBC GC sensitivity]; hair cortisol and DHEA), measures of inflammation (high sensitivity C-reactive protein [hs CRP], tumor necrosis factor alpha [TNF-alpha], interleukin-1 beta [IL-1 beta] and IL-6 and IL-10 by multiplex), glucose regulation and lipid metabolism markers (high-density lipoprotein [HDL], low-density lipoprotein [LDL], triglycerides, total cholesterol, and hemoglobin A1c [HA1c]), catecholamines such as epinephrine and norepinephrine, anthropomorphic measures (body mass index [BMI], waist-to-hip ratio, and cardiovascular function (resting BP and heart rate variability). Particularly important for our purposes are measures that reflect activity over a period of time (e.g., hemoglobin A1c9 and hair cortisol). *Ambulatory Blood Pressure.* Participants were fitted with an electronic, ambulatory BP monitor, that automatically recorded their BP for 24 hrs after they left the lab – every 20 minutes during the day and every 60 minutes at night. Participants were asked to complete a separate electronic diary entry at every recording (when the BP cuff inflated) during the day. This diary measured activities, events, and posture, that would affect the BP reading. For the purposes of this project, we will investigate means and slopes (trends) in daytime and nighttime BP. BP typically drops at night.18 Chronic stress exposure is associated with reduced nocturnal dipping. Reduced nocturnal BP dipping predicts risk for cognitive impairment. **Primary Mediator/Moderator** *Discrimination.* The Brief PEDQ-CV33, with additional subscales on media discrimination, past-week discrimination, and friends & family discrimination, were used to assess levels of perceived discrimination. Items included, “Because you are Native or American Indian, how often have you been treated unfairly by teachers, principals, or other staff at school?”, “…have others threatened to hurt you (ex: said they would hit you)?”, and “…have policemen or security officers been unfair to you?”. Items were rated on a scale from 0 (never) to 4 (very often). We will examine overall perceived discrimination (average across items), as well as differences among subscales. Scores will be treated as continuous measures. *Historical Trauma.* The Historical Loss Scale19 measured frequency of thoughts (exposure to) historical trauma experienced by AI/AN. Items included, “How often do you think about these losses to American Indians?”; “Loss of our land.”, “Loss of our language.”, and “Loss of our culture.” Items were rated on a scale from 0 (never) to 4 (very often). We will examine overall exposure to historical trauma (average across items) as well as explore differences among items. **DATA ANALYTIC PLAN** We will use each of the above measures to explore whether colonial harms (i.e., discrimination, historical trauma) interact with age to predict risk factors for cognitive impairment and AD. We will also use these measures to examine whether colonial harms are possible mechanisms through which age affects these risk factors. **All models will control for other factors typically associated with our primary outcomes (e.g., cynicism, hostility, and anxiety as well as indicators of participants’ diet, alcohol, tobacco, and drug use).** *Statistical Analysis.* Descriptive statistics will be computed to describe baseline patient characteristics. Allostatic load, will be calculated following prior research and analyzed as a count with a Poisson distribution. Structural equation modeling will also be used to confirm that allostatic load is a unified construct with multiple latent factors (e.g. inflammation, metabolism, autonomic system, etc.). Hypotheses 1-2 will be tested using multiple regression models, run using the stats package for the statistical package R (version 3.4.2). The analytic approach for hypothesis 1 will include unadjusted regressions and adjusted regressions that control for known covariates of our dependent measures (outlined in the proposal) and hypothesis 2 will use similar regression models that include perceived discrimination and historical trauma as moderators. Hypotheses 3a-3b will be tested using structural equation modeling, run using the lavaan.survey package. Each structural equation model will have a total of nine free parameters. According to prior recommendations, a “gold standard” for considering sample size suggests a minimum of 20 observations per free parameter (Tanaka, 1987). According to these recommendations, with nine free parameters, we would require a minimum of 180 observations. By the end of data collection we will have close to 300 observations. All tests will be two-sided (α = 0.05). We will employ methods that utilize all available data and impute missing data when justifiable, using the mice package.