Susan Resnick - US grants

Summary of work: The neuroanatomic and neurophysiologic underpinnings of age-associated cognitive and memory change remain unclear, as there are a limited number of studies of longitudinal brain changes in individuals without dementia. We are performing serial magnetic resonance imaging (MRI), positron emission tomography (PET), and neuropsychological assessments in participants from the Baltimore Longitudinal Study of Aging (BLSA) to investigate the neurobiological basis of memory change and cognitive impairment. These evaluations allow us to examine changes in brain structure and function which may be early predictors of cognitive change and impairment, including Alzheimer's Disease (AD). An understanding of these associations and early detection of brain changes will be critical in identifying individuals likely to benefit from new interventions. Since 2005, we also have acquired 11-C-PIB PET imaging studies of amyloid distribution in the brain to enhance the identification of preclinical AD. In addition, we are using neuroimaging tools to investigate modulators of cognitive and brain changes, including sex differences in brain aging, genetic risk factors and the effects of sex steroid and other hormones. We continue to perform serial MRI and PET scans for neuroimaging study participants of the Baltimore Longitudinal Study of Aging. Our initial longitudinal MRI investigations demonstrated significant longitudinal gray and white matter tissue loss over a four-year interval even in healthy older adults (Resnick et al., J Neuroscience 2003). Recently, we confirmed and extended these findings over a longer (up to 9-year) follow-up interval (Driscoll et al., Neurology 2009). Furthermore, we demonstrated that a number of brain regions, including the hippocampus and orbital frontal region in addition to whole brain volume generally, showed accelerated tissue loss in individuals with mild cognitive impairment (MCI). We use support vector machine algorithms to develop classifiers for prediction of diagnostic status on an individual person basis. We identified a network of abnormalities that contributed to maximal discrimination between MRI scans of normal and MCI individuals in a sample of individuals matched for important demographic characteristics (Fan et al., Neuroimage 2008). We then applied this approach to the AD patients and controls from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and used the classifier based on that sample to calculate the Spatial Pattern of Abnormality for Recognition of Early AD (SPARE-AD) scores for all MRI scans of each individual BLSA participant. We demonstrated that higher SPARE-AD abnormality scores were associated with lower verbal memory performance (Davatzikos et al., Brain 2009) and that abnormality scores increased with age and cognitive impairment. Using the 11-C-PIB PET scan data, we confirmed the experience at other centers that 20-30% of cognitively normal older adults have PIB positive scans, i.e., deposition of brain amyloid. In addition, we investigated whether PIB retention was associated with longitudinal rCBF changes in the preceding years (Sojkova et al., 2008). PIB distribution volume ratios (DVR) of regions of interest were estimated by fitting a reference tissue model to the measured time activity curves (Zhou et al, Neuroimage, 2007). The mean cortical DVR was used to divide participants into high and low PIB retention groups. Both regions of greater longitudinal decrease and greater longitudinal increase in rCBF were observed in association with high PIB. While longitudinal declines may reflect greater decrements in neuronal function in the high PIB group, greater longitudinal increases in rCBF are also observed in those with higher amyloid load and may represent a compensatory attempt to preserve neuronal function in these regions. Approximately half of the neuroimaging study participants are enrolled in the BLSA autopsy program, and the integration of autopsy and imaging findings is an active area of investigation (e.g. Iacono et al 2008). BLSA studies have shown that individuals with AD pathology at autopsy but without antemortem cognitive impairment (Asymptomatic AD) have larger nuclei and nucleoli of the anterior and posterior cingulate and the hippocampus, suggesting either a compensatory or inflammatory process. Ongoing studies investigate the cellular basis of neuroimaging findings, including regional volume loss, to try to better define characteristics that accelerate cognitive impairment as well as those that support the maintenance of cognitive health.

The Baltimore Longitudinal Study of Aging (BLSA) was established in 1958 and is one the oldest prospective studies of aging in the USA and the world. The mission of the BLSA is to learn what happens to people as they get old and how to sort out changes due to aging from those due to disease or other causes. Over the last year, we have published a number of papers investigating the effects of possible modifiers of cognitive aging, risk for dementia, and the presence of Alzheimer's pathology at autopsy. These potential early markers and modifiers include glucose intolerance and insulin resistance, personality characteristics, changes in regional cerebral blood flow, and evidence of vascular/demyelinating disease on antemortem and postmortem MRI scans. Based on the established link between diabetes and increased risk of dementia and AD, we expected to find an association between glucose intolerance and AD-type neuropathology at autopsy or on in vivo PET amyloid scans. However, in this prospective cohort with multiple assessments of glucose intolerance and insulin resistance over time, measures of glucose and insulin homeostasis were not associated with either amyloid plaques or neurofibrillary tangles. In contrast, we found significant differences in regional cerebral blood flow changes over time between participants with normal glucose tolerance (NGT) and those with impaired glucose tolerance (IGT). This study included 15 participants with IGT and 49 with NGT at mid-life who subsequently received PET studies of regional cerebral blood flow (rCBF) over an 8-year follow-up period. Voxel-based analysis of differences between groups of rCBF changes showed greater rCBF decline over time in the frontal, parietal, and temporal cortices. Some brain regions in the frontal and temporal cortices also showed greater longitudinal increments in rCBF in the IGT group. Our findings suggest that IGT in midlife may be associated with changes in brain function and are consistent with a role of vascular factors in the diabetes-dementia relationship. Although glucose intolerance and insulin resistance were not associated with AD pathology, we did find a relationship between higher amounts of white matter disease on MRI scans and amyloid plaques and neurofibrillary tangles at autopsy. White matter disease was quantified in pre- and postmortem MRI scans for 50 BLSA autopsy study participants using the Cardiovascular Health Study (CHS) criteria in a blinded manner. We also examined amyloid angiopathy and found that amyloid angiopathy was not independently associated with CHS score. Our finding that AD pathology at autopsy is associated with MRI-detected cerebral white matter disease may explain, in part, the association between cerebral white matter disease and cognitive decline in the elderly individuals. In addition to investigation of the associations between metabolic and vascular risk factors and AD pathology at autopsy, we have also investigated associations between behavioral characteristics and later risk for clinical dementia and AD pathology. We found that early personality characteristics were associated with later risk and expression of clinical dementia. Broad factors and specific facets of personality were assessed up to 28 years (mean 11 years) before onset of dementia and up to 30 years (mean 15 years) before death in a cohort (n = 111) evaluated for AD neuropathology at autopsy. Individuals with higher baseline Neuroticism scores (vulnerability to stress, anxiety, and depression) or lower scores on Conscientiousness (order and competence) were less likely to remain asymptomatic in the presence of AD neuropathology. These findings suggest that a resilient personality profile is associated with lower risk or delay of clinical dementia even in persons with AD neuropathology. In a second study, including a meta-analysis of the association between personality domains and risk for AD, we found that high Neuroticism and low Conscientiousness were associated with increased risk for incident diagnosis of clinical AD.

The Womens Health Initiative (WHI) randomized, placebo-controlled clinical trials of hormone therapy (HT) were designed to test the hypothesis that conjugated equine estrogens alone (CEE-Alone) or in combination with medroxyprogesterone acetate (CEE+MPA) protected postmenopausal women against the development of heart disease. The WHI Memory Study (WHIMS)was an ancillary study to the WHI trials, which consisted of parallel placebo-controlled randomized clinical trials of 0.625 mg/day CEE therapy with and without 2.5 mg/day MPA in women with a uterus or post-hysterectomy, respectively. WHIMS investigated the effect of CEE-Alone and CEE+MPA on risk for probable dementia and mild cognitive impairment in women age 65 and older, as well as the effects of these treatments on global cognitive function. WHIMS followed 7479 women at 38 of the 40 WHI clinical sites participating in the HT trial. The WHI Study of Cognitive Aging (WHISCA), an ancillary study to WHIMS, was developed to investigate the effects of HT on domain-specific cognitive function in women without dementia. Based on observational studies from our group and others and randomized trials of younger women following surgical menopause, we hypothesized that hormone therapy (HT) in postmenopausal women would benefit memory and possibly other cognitive functions. WHISCA enrolled 2305 women at 14 of the WHIMS sites, distributed across the two parallel trials. WHISCA was initiated on average 3 years after WHI randomization and the primary outcome was the effect of HT on rates of cognitive change, adjusted for time since randomization. Through the main WHI study and WHIMS, incident diagnoses of stroke, probable dementia, and mild cognitive impairment (MCI) are available for WHISCA participants. The WHIMS CEE+MPA trial terminated earlier than planned (July, 2002) due to an adverse risk-to-benefit profile in the main WHI trial. Subsequently, the WHI CEE-Alone Trial also was terminated early (February, 2004). Results from the WHIMS trials showed that CEE-Alone or CEE+MPA increase the risk of dementia and have adverse effects on global cognition in women aged 65 years or older. HT also has been shown to increase the risk of clinical stroke in women 65 years and older. The initial report of WHISCA findings (Resnick et al., Journal of Clinical Endocrinology and Metabolism, 2006) included 1416 women free of probable dementia who were randomized to CEE+MPA or placebo and followed for a 1.35 yr prior to the early termination of the main WHI trial due to an increase in risk for breast cancer in HT-treated women and an adverse risk-benefit profile. Analysis of annual rates of change in specific cognitive functions and affect, adjusted for time since randomization, showed that CEE + MPA had a negative impact on verbal memory (p <0.01) and a trend to a positive impact on figural memory (p = 0.012) over time compared with placebo with no effect on other cognitive domains. In addition, these effects were evident only after long-term therapy. CEE + MPA did not significantly influence positive affect, negative affect, or depressive symptoms. These findings suggest that HT may have different effects across different cognitive domains. The results of the CEE-Alone Trial in women with prior hysterectomy who were randomized to CEE or placebo are currently in press in Journal of Clinical Endocrinology and Metabolism. More recently, through the related WHIMS-MRI study, we have reported with our WHIMS collaborators that HT was associated with decreased hippocampal and frontal regional brain volumes in more than 1400 women evaluated post-trial with brain MRI scans. This adverse effect of HT on hippocampal volume was most pronounced in women who had the lowest scores on a measure of global cognitive function at WHI baseline.

In health-related research, we have examined the association between personality traits and several biomarkers of mental and physical health. Brain derived neurotrophic factor (BDNF) is among the most promising biomarkers of mood disorders. BDNF regulates synaptic plasticity and neurogenesis, and BDNF plasma and serum levels have been associated with depression and other psychiatric and neurodegenerative disorders. In clinical samples, circulating levels of BDNF are reported to be lower in depressed patients compared to controls and are reported to increase significantly with antidepressant treatment. In large community-based samples, we are examining whether BDNF concentration in plasma and serum are associated with the personality trait of Neuroticism, a major risk-factor for depression. Among the markers of physical health, we have examined the association between personality traits and obesity, dyslipedemia, metabolic syndrome, arterial thickening, and inflammation. In particular, Interleukin-6 (IL-6) is a peripheral marker of chronic inflammation that increases with age and is implicated in a wide range of health outcomes. IL-6 is an inflammatory cytokine that promotes inflammation in response to infection or injury. Although beneficial in response to acute injuries, chronic production of IL-6 leads to increased morbidity and mortality. Higher levels of IL-6 are associated with frailty and disability among the elderly, and elevated IL-6 has been linked to numerous chronic conditions, such as diabetes, osteoporosis, and cardiovascular disease. High Neuroticism and low Conscientiousness are frequently implicated in health-risk behaviors, such as smoking and overeating, as well as health outcomes, including mortality. Their associations with physiological markers of morbidity and mortality, such as inflammation, are less well documented. The present research examines the association between the five major dimensions of personality and interleukin-6 (IL-6), a pro-inflammatory cytokine often elevated in patients with chronic morbidity and frailty. A population-based sample (N=4,923) from four towns in Sardinia, Italy, had their levels of IL-6 measured and completed a comprehensive personality questionnaire, the NEO-PI-R. Analyses controlled for factors known to have an effect on IL-6: age, sex, smoking, weight, aspirin use, and disease burden. High Neuroticism and low Conscientiousness were both associated with higher levels of IL-6. The findings remained significant after controlling for the relevant covariates. Similar results were found for C-reactive protein, a related marker of chronic inflammation. Further, smoking and weight partially mediated the association between impulsivity-related traits and higher IL-6 levels. Finally, logistic regressions revealed that participants either in the top 10% of the distribution of Neuroticism or the bottom 10% of Conscientiousness had an approximately 40% greater risk of exceeding clinically-relevant thresholds of IL-6. Consistent with the literature on personality and self-reported health, individuals high on Neuroticism or low on Conscientiousness show elevated levels of this inflammatory cytokine. Identifying critical medical biomarkers associated with personality may help to elucidate the physiological mechanisms responsible for the observed connections between personality traits and physical health.

The Womens Health Initiative (WHI) randomized, placebo-controlled clinical trials of hormone therapy (HT) were designed to test the hypothesis that conjugated equine estrogens alone (CEE-Alone) or in combination with medroxyprogesterone acetate (CEE+MPA) protected postmenopausal women against the development of heart disease. The WHI Memory Study (WHIMS) was an ancillary study to the WHI trials, which consisted of parallel placebo-controlled randomized clinical trials of 0.625 mg/day CEE therapy with and without 2.5 mg/day MPA in women with a uterus or post-hysterectomy, respectively. WHIMS investigated the effect of CEE-Alone and CEE+MPA on risk for probable dementia and mild cognitive impairment in women age 65 and older, as well as the effects of these treatments on global cognitive function. The WHI Study of Cognitive Aging (WHISCA), an ancillary study to WHIMS, was developed to investigate the effects of HT on domain-specific cognitive function in women without dementia. WHISCA enrolled 2305 women at 14 of the WHIMS sites, distributed across the two parallel trials. WHISCA was initiated on average 3 years after WHI randomization and the primary outcome was the effect of HT on rates of cognitive change, adjusted for time since randomization. The WHIMS CEE+MPA trial terminated earlier than planned (July, 2002) due to an adverse risk-to-benefit profile in the main WHI trial. Subsequently, the WHI CEE-Alone Trial also was terminated early (February, 2004). Results from the WHIMS trials showed that CEE-Alone or CEE+MPA increase the risk of dementia and have adverse effects on global cognition in women aged 65 years or older. HT also has been shown to increase the risk of clinical stroke in women 65 years and older. The initial report of WHISCA findings showed that CEE + MPA had a negative impact on verbal memory (p <0.01) and a trend to a positive impact on figural memory (p = 0.012) over time compared with placebo with no effect on other cognitive domains. In addition, these effects were evident only after long-term therapy. CEE + MPA did not significantly influence positive affect, negative affect, or depressive symptoms. These findings suggest that HT may have different effects across different cognitive domains. The findings from the CEE-Alone Trial in women with prior hysterectomy who were randomized to CEE or placebo show that CEE alone did not affect domain-specific cognitive function over time. Participants in the WHISCA and WHIMS studies continue to be followed through telephone cognitive assessments as they pass through the risk period for cognitive decline. Over the next 5-year follow-up the NIA will assume the primary funding role for the WHIMS Suite of Studies through a Research and Development Contract. This R and D contract also includes continued cognitive follow-up of women in the WHIMS-Younger (WHIMS-Y) study, who were randomized to hormone therapy through the WHI when aged 50-54 years. The WHIMS-Y study will test the hypothesis that hormone therapy around the time of the menopause may benefit cognitive function later in life.

As researchers seek to understand individual differences in peoples psychological resilience to life stressors, both personality traits and spirituality/religiousness (S/R) have emerged as important predictors of sustained well-being in the face of adversity, though it is not clear how these variables are jointly related to emotional well-being. Further, most research on the relation between personality traits and religiousness has been conducted in healthy, well-adjusted samples and little is known about their association in vulnerable populations. To address these gaps in the research record, in a recent study LPC/PSCS investigators examined the relative association of personality traits and S/R with mental health in a sample of people living with HIV. Research on the psychology of religion relies on a wide variety of measures to capture peoples experience of the sacred, yet a basic distinction can be drawn between religiousness and spirituality. Religiousness captures adherence to traditional religious creeds, often centered on a specific community of faith and encompasses a set of behaviors (e.g., prayer, church attendance) associated with a specific belief system. Spirituality typically refers to subjective experiences of the transcendent which imbue everyday life with a sense of deeper and may also encompass a sense of communion with humanity and compassion for others. This differentiation is particularly relevant when studying populations affected by HIV/AIDS who often face stigmatization and may be more likely to report being spiritual but not religious. We focus on the Five-Factor Model of personality traits, a comprehensive and widely replicated trait taxonomy that describes personality along the following five dimensions: Neuroticism (N), Extraversion (E), Openness to Experience (O), Agreeableness (A), and Conscientiousness (C). In previous studies, high A appears to show somewhat stronger associations to religiousness than to spirituality, high O shows the opposite pattern, high C is related to both concepts, whereas low N and high E show weaker associations to either construct. Method Participants (n=112) were paid volunteers drawn from a larger longitudinal sample and were recruited into the parent sample if they were HIV positive and had CD4 counts ranging from 150 to 500. Participants disease status was assessed via CD4 lymphocite count and viral load at baseline. Personality traits were assessed with the Revised NEO Personality Inventory (NEO-PI-R, Costa &McCrae, 1992). Spirituality and religiousness were assessed with the short form of the Ironson-Wood Spirituality/Religiousness Index (IWSRI;Ironson et al., 2002). Mental health was captured using a test battery including the Beck Depression Inventory (BDI, Beck, Steer, &Brown, 1996), the Perceived Stress Scale (Cohen, Kamarck, &Mermelstein, 1983), and others. Results Our findings suggest that the general pattern of associations between personality traits and S/R holds true across diverse samples. C showed the most consistent pattern of associations to S/R with significant partial correlations to all of the IWSRI subscales, followed by A and O. For N, significant associations were limited to the Religious Behavior scale and E was the only factor that showed no associations to the IWSRI. Evidence from facet-level analyses illustrate that religiousness is associated with low scores on Impulsiveness (N) and high scores on Altruism and Compliance (A). Also, consistently across studies, aspects of spirituality were associated with high scores on Fantasy, Aesthetics, and Feelings (O). Further, high scores on Tender-Mindedness (A) and Competence (C) were associated with both spirituality and religiousness. Personality factors and IWSRI scores were significantly associated with mental health and the general pattern suggested that N was associated with lower levels of mental health (i.e., higher levels of depression, hopelessness, anxiety and stress as well as lower levels of optimism). The remaining personality factors as well as the IWSRI subscales were positively associated with mental health. Discussion Religious and spiritual individuals feel in charge of their lives and show high levels of sympathy and concern for others. In addition, primarily religious individuals are uniquely characterized by the ability to resist temptations coupled with the tendency to de-escalate social conflicts and assist others in need. Primarily spiritual individuals, in turn, show vivid imagination, appreciation for beauty, and receptiveness to their inner feelings. Our findings suggest that when the NEO-PI-R domains are considered in combination, underlying personality traits fully account for the association of S/R with mental health, consistent with the theoretical claim that religiousness and spirituality are characteristic adaptations that develop as basic personality traits are channeled by cultural and environmental influences (source). Furthermore, a set of mediation analyses provide good initial evidence that certain aspects of S/R mediate the relationship of both O and A with mental health. Our findings also have important practical implications because the mediation analyses suggest that the benefits of S/R for mental health are largely dependent on a persons underlying personality traits. This suggests that any interventions aimed at inducing a turn towards religion in order to improve mental health may be misguided. Instead, medical professionals may want to assess their clients personality profile to determine who may benefit most from access to spiritual and/or religious support and guidance.

The notion of stability is central to the definition of personality traits, which are generally thought of as enduring tendencies or habitual patterns of behavior, thoughts, and emotions (McCrae & Costa, 2003), but stability does not imply immutability. Under normal circumstances, adult traits are largely stable, as indicated by high correlation coefficients computed for a group assessed twice on the same trait. These coefficients represent the average stability for a sample, but individuals vary in terms of their intra-individual stability. Despite the relative stability of individual differences, we have extended the studies of mean-level change in personality by examining longitudinal trajectories in more diverse samples and cross-sectional trends across cultures. Furthermore, we aim to delineate the neural and cognitive correlates of personality dimensions across longitudinal assessments. Most recently, we have begun investigating the association between personality traits and decision making. We recently investigated the personality correlates of risky decision making in healthy older adults. We also continue to collect fMRI studies of brain activation during decision making and will relate personality variables to patterns of brain activation during gains and losses, as well as in relation to individual differences in risk preference. We continue to collaborate with Dr. Antonio Terracciano on studies of personality characteristics that may predict subsequent cognitive impairment and personality change during the preclinical and prodromal phases of Alzheimer's Disease (AD). In previous work, we have shown that traits of high Neuroticism and low Conscientiousness are associated with increased risk for Alzheimer's disease later in life. A more recent paper investigated whether personality is stable or changes during preclinical and prodromal AD. We examined 2046 participants in the Baltimore Longitudinal Study of Aging. During 24,569 person-years, mild cognitive impairment was diagnosed in 104 (5.1%) individuals, and all-cause dementia was diagnosed in 255 (12.5%) participants, including 194 (9.5%) with Alzheimer disease. Individuals who developed dementia scored higher on Neuroticism (=2.83; 95% CI, 1.44 to 4.22; P<.001) and lower on Conscientiousness (=-3.34; 95% CI, -4.93 to -1.75; P<.001) and Extraversion (=-1.74; 95% CI, -3.23 to -0.25; P=.02). However, the unimpaired group did not differ significantly in personality change over time from the Alzheimer disease and mild cognitive impairment groups. Thus, we found no evidence for preclinical change in personality before the onset of mild cognitive impairment or dementia. These findings provide evidence against the reverse causality hypothesis and strengthen evidence for personality traits as a risk factor for dementia.

Personality traits are summarized by five broad dimensions with pervasive influences on major life outcomes, strong links to psychiatric disorders and clear heritable components. To identify genetic variants associated with personality traits we perform genome-wide association studies(GWAS) on a genetically isolated population within Sardinia, Italy. In the first study of the five major dimensions of personality the analyses of 362129 single nucleotide polymorphisms (SNPs) revelead several strong signals within or near genes previously implicated in psychiatric disorders. They include the association of neuroticism with SNAP25, extraversion with BDNF and two cadherin genes (CDH13 and CDH23), openness with CNTNAP2, agreeableness with CLOCK, and conscientiousness with DYRK1A. Effect sizes were small (less than 1% of variance), and most failed to replicate in the follow-up independent samples (N up to 3903), though the association between agreeableness and CLOCK was supported in two of three replication samples. We infer that a large number of loci may influence personality traits and disorders, requiring larger sample sizes for the GWAS approach to confidently identify associated genetic variants.

Previous research in this Laboratory has examined personality profiles of cultures obtained by averaging traits assessed in samples of college students and adults. We have extended this line of research by examining ratings of 12 to 17 years old in 24 cultures. Observer ratings of adolescents were obtained using the NEO Personality Inventory-3 (NEO-PI-3), a more readable version of the Revised NEO Personality Inventory (NEO-PI-R), which showed good psychometric properties across cultures. Aggregate scores were generalizable across age and gender and showed convergence with culture-level scores from previous studies of self-reports and observer ratings of adults. Non-Western cultures tend to score slightly lower on a dimension related to Extraversion. Further, aggregate scores of adolescent ratings were unrelated to national character stereotypes, consistent with previous research that challenge the accuracy of these widely shared beliefs. The multinational study includes systematic assessments of personality stereotypes of age groups and genders as well as nations. Personality stereotypes of age and gender groups were assessed in each of the 26 cultures (N = 3,323) using the National Character Stereotypes measure, which includes items for each of the 30 facets included in the NEO-PI-3. Participants rated either males or females in the age groups adolescent, adult, and old in their own country. We compared assessed personality stereotypes of different age groups with previously published self-reported and observer-rated personality at each of the age groups. Across cultures, adolescents were seen as impulsive, rebellious, undisciplined, preferring excitement and novelty, whereas old people were consistently considered lower on impulsivity, activity, antagonism, and Openness. Results showed that stereotypes of age groups were remarkably accurate, with only minor variability, though personal stereotypes of any particular rater were less accurate than the aggregate stereotype of each age group. This pattern of findings suggests that the formation of age stereotypes may have culture-neutral origins. In the assessment of National Character Stereotypes, participants were asked to rate the personality of the typical member of their culture, as well as the typical personality of Americans. Across cultures, the rating of Americans warmth and competence was systematically related to the rater cultures GDP, such that wealthier countries rated Americans as more warm and less competent. Results were similar when raters own family income is used as the predictor. Findings were consistent with theories on the dimensions of social perception.

Summary of work: The neuroanatomic and neurophysiologic underpinnings of age-associated cognitive and memory change remain unclear, as there are a limited number of studies of longitudinal brain changes in individuals without dementia. We are performing serial magnetic resonance imaging (MRI), including measures of vascular changes, positron emission tomography (PET), and neuropsychological assessments in participants from the Baltimore Longitudinal Study of Aging (BLSA) to investigate the neurobiological basis of memory change and cognitive impairment. These evaluations allow us to examine changes in brain structure and function which may be early predictors of cognitive change and impairment, including Alzheimer's disease (AD). We are continuing longitudinal testing of older participants and evaluating new participants, including MRI and concurrent neuropsychological assessments of participants less than 55 years old. For individuals aged 55 and older, we also currently perform a single PET measurement of CBF, followed by a PET scan using 11-C-Pittsburgh Compound B (PiB) to measure in vivo amyloid distribution. Our progress over the last year includes continued acquisition of new neuroimaging assessments as well as continued analysis of existing data and methods development. Approximately half of the neuroimaging study participants are enrolled in the BLSA autopsy program, and the integration of autopsy and imaging findings is an active area of investigation to gain a better understanding of factors that promote cognitive resilience in individuals who have amyloid pathology but do not show memory impairment. In addition, we are using neuroimaging tools to investigate modulators of cognitive and brain changes, including sex differences in brain aging, genetic risk factors, and the effects of sex steroid and other hormones. An understanding of these brain-behavior associations and early detection of accelerated brain changes that predict cognitive decline and impairment will be critical in identifying individuals likely to benefit from new interventions. Over the last year, we have published a number of papers from this study. Consistent with imaging findings at other centers and autopsy studies, we find that about 30 percent of cognitively normal older adults have detectable levels of amyloid in the brain. Our PiB studies have demonstrated that higher PiB levels in cognitively normal individuals are associated with greater decline over time in mental status and memory (Resnick et al, 2010) but PiB was not significantly associated with regional tissue loss in normal individuals (Driscoll et al, 2011). These studies also have revealed longitudinal increases in PiB retention in individuals with higher PiB retention at initial PiB assessment (Sojkova et al, 2011) and the concordance and discordance between in vivo amyloid imaging patterns and pathological ratings of amyloid plaques according to the CERAD classification for pathological diagnosis of AD (Sojkova et al, 2010). Our imaging-neuropathology analyses have highlighted difficulties in using standard neuropathological diagnosis for autopsy validation of PiB due to differences in regions examined under the standard CERAD assessment and the brain regions showing the earliest amyloid deposition on PiB imaging. We are using the spatial patterns of PiB binding (and MRI tissue loss and lesions) to guide more detailed autopsy analyses. In addition, we have investigated clusterin and other plasma protein concentrations and genetic risk in relation to PiB levels and patterns. Higher clusterin concentration in plasma at baseline neuroimaging assessment was associated with higher medial temporal PiB retention more than 10 years later (Thambisetty et al, 2010). In addition, we combined proteomics with in in vivo amyloid imaging to identify a panel of 18 2DGE plasma protein spots that discriminated between individuals with high and low brain Aβ(Thambisetty et al, 2010). Mass spectrometry identified these proteins, many of which have established roles in Aβclearance, including a strong signal from apolipoprotein-E (ApoE). Plasma ApoE concentration was associated with increased Aβburden in the medial temporal lobe, most pronounced in the hippocampus and entorhinal cortex. APOE ε4 carriers also showed greater Aβlevels in several brain regions relative to ε4 non-carriers. These results suggest that both peripheral concentration of ApoE protein and APOE genotype are related to early neuropathological changes in brain regions vulnerable to AD pathology even in the non-demented elderly. The data from this project also continue to be used for important methodological developments to enhance analysis of longitudinal neuroimaging data, including papers describing new approaches for skull-stripping on MRI (Carass et al., 2011), cluster analysis of imaging data for detection of a cluster-based measure of pathology that reflects the deviation of a subject's MR image from a normal (i.e. cognitively stable) state (Filipovych et al, 2011), and an extension of Biological Parametric Mapping to include robust regression and robust inference in the neuroimaging context of application of the general linear model (Xue et al, 2011).

The Womens Health Initiative (WHI) randomized, placebo-controlled clinical trials of hormone therapy (HT) were designed to test the hypothesis that conjugated equine estrogens alone (CEE-Alone) or in combination with medroxyprogesterone acetate (CEE+MPA) protected postmenopausal women against the development of heart disease. The WHI Memory Study (WHIMS) was an ancillary study to the WHI trials, which consisted of parallel placebo-controlled randomized clinical trials of 0.625 mg/day CEE therapy with and without 2.5 mg/day MPA in women with a uterus or post-hysterectomy, respectively. WHIMS investigated the effect of CEE-Alone and CEE+MPA on risk for probable dementia and mild cognitive impairment in women age 65 and older, as well as the effects of these treatments on global cognitive function. The WHI Study of Cognitive Aging (WHISCA), an ancillary study to WHIMS, was developed to investigate the effects of HT on domain-specific cognitive function in women without dementia. WHISCA enrolled 2305 women at 14 of the WHIMS sites, distributed across the two parallel trials. WHISCA was initiated on average 3 years after WHI randomization and the primary outcome was the effect of HT on rates of cognitive change, adjusted for time since randomization. The WHIMS CEE+MPA trial terminated earlier than planned (July, 2002) due to an adverse risk-to-benefit profile in the main WHI trial. Subsequently, the WHI CEE-Alone Trial also was terminated early (February, 2004). Results from the WHIMS trials showed that CEE-Alone or CEE+MPA increase the risk of dementia and have adverse effects on global cognition in women aged 65 years or older. HT also has been shown to increase the risk of clinical stroke in women 65 years and older. The initial report of WHISCA findings showed that CEE + MPA had a negative impact on verbal memory (p < 0.01) and a trend to a positive impact on figural memory (p = 0.012) over time compared with placebo with no effect on other cognitive domains. In addition, these effects were evident only after long-term therapy. CEE + MPA did not significantly influence positive affect, negative affect, or depressive symptoms. These findings suggest that HT may have different effects across different cognitive domains. The findings from the CEE-Alone Trial in women with prior hysterectomy who were randomized to CEE or placebo show that CEE alone did not affect domain-specific cognitive function over time. Participants in the WHISCA and WHIMS studies continue to be followed through telephone cognitive assessments as they pass through the risk period for cognitive decline. The WHIMS Suite of Studies also includes cognitive follow-up of women in the WHIMS-Younger (WHIMS-Y) study, who were randomized to hormone therapy through the WHI when aged 50-54 years. The WHIMS-Y study tested the hypothesis that hormone therapy around the time of the menopause may benefit cognitive function later in life. Initial results from the WHIMS-Y study were based on 1326 postmenopausal women studied with a validated telephone cognitive assessment battery an average of 7.2 years after the trials ended, when women had a mean age of 67.2 years. The initial report included the first two administrations of the cognitive battery and showed neither harm nor benefit of earlyHT on later cognitive function. The WHIMS Suite of Studies is conducted by Wake Forest University, which is also the site for the Southeast Regional Center for WHI and leads the Aging, Cognition and Functional Status interest group for the WHI. Over the last year, we have continued to perform cognitive follow-up evaluations through telephone assessments in the original WHIMS cohort (WHIMS extension study) and have closed out the WHIMS-Y study as sufficient data are available in the latter cohort to test the study hypotheses. Over the last year, we have continued to investigate factors that may modulate cognitive and brain outcomes in older women. Following on a previous paper that showed that presence of diabetes influence the effect of postmenopausal hormone therapy on brain volumes and ischemic lesion volumes, we examined interactions between prior hormone assignment and diabetes on cognitive outcomes. In older women, higher levels of estrogen may exacerbate the increased risk for cognitive impairment conveyed by diabetes. We examined whether the effect of postmenopausal HT on cognitive impairment incidence differed depending on type 2 diabetes. WHIMS participants (N = 7,233), aged 65-80 years, were classified according to type 2 diabetes status and followed for probable dementia and cognitive impairment (mild cognitive impairment or dementia). Through a maximum of 18 years of follow-up, women with diabetes had increased risk of dementia (hazard ratio HR 1.54 95% CI 1.16-2.06) and cognitive impairment (HR 1.83 1.50-2.23). The combination of diabetes and random assignment to HT increased their risk of dementia (HR 2.12 1.47-3.06) and cognitive impairment (HR 2.20 1.70-2.87) compared with women without these conditions, interaction P = 0.09 and P = 0.08. These interactions appeared to be limited to women assigned to unopposed conjugated equine estrogens. Our findings are consistent with a prior report that higher endogenous estrogen may exacerbate risks that type 2 diabetes poses for cognitive function in older women. We also investigated whether obesity influenced regional brain volumes and white matter lesion loads in WHIMS participants. While midlife obesity has been linked to age-related brain atrophy and risk of dementia, relationships are less clear for older individuals. We examined the associations that obesity (body mass index, BMI) and change in BMI over an average of 6.6 (1.0-9.1) years had with global and regional brain and white matter lesion volumes in a sample of 1,366 women aged 65-80. Both global obesity and increase in BMI were associated with lower cerebrospinal fluid and higher region-specific brain volumes, after controlling for diabetes and other cerebrovascular disease risk factors. Obesity, but not change in BMI, predicted lower lesion loads in some brain areas. Thus, contrary to predictions based on studies of midlife obesity, we found that higher BMI in late life was associated with less brain atrophy and lower ischemic lesion loads, perhaps reflecting weight loss during prodromal phases of neurodegenerative disease in older women. In a recent paper published as part of a series characterizing WHI participants who were 80 years or older, we investigated factors that predicted optimal or resilient cognitive aging in WHIMS participants. Two thousand two hundred twenty-eight women with a mean age of 85 years who participated in the WHIMS were classified as cognitively normal (85.5%), mild cognitive impairment (3.9%), dementia (5.4%) or other cognitive impairment (5.1%) by central adjudication. Global cognitive functioning was assessed using the TICS-M (telephone interview for cognitive status-modified). Backward stepwise logistic regression was used to select factors that were independently associated with cognitive status. Factors associated with preserved cognitive functioning were younger age, higher education and family incomes, being non-Hispanic white, better emotional wellbeing, fewer depressive symptoms, more insomnia complaints, being free of diabetes, and not carrying the APOE 4 Alzheimers disease risk allele. Cognitively normal women who demonstrated sustained high levels of cognition were younger, more educated, and endorsed better self-reported general health, emotional wellbeing, and higher physical functioning. These findings provide a broad overview of factors associated with maintenance of cognitive health at older ages and are motivating the next phase of the WHIMS follow-up.

In previous work, we found larger nuclear and nucleolar volumes in hippocampus and cingulate regions in asymptomatic individuals with Alzheimer's type pathology (ASYMAD) compared to both older controls and cognitively impaired individuals. More recently, we compared antemortem changes in regional cerebral blood flow in BLSA participants subsequently diagnosed at autopsy as ASYMAD (n=6) with changes in cognitively impaired (CI=6) and normal (CN=7) older individuals. At autopsy, the ASYMAD and CI groups had similar CERAD and Braak scores, similar amounts of beta-amyloid and tau determined with immunohistochemistry in middle frontal (MFG), middle temporal (MTG), and inferior parietal (IP) regions, and more beta-amyloid than CN in precuneus, MFG and IP areas. PET rCBF scans beginning 10 years before death and while all participants were cognitively intact identified similarities and differences in longitudinal rCBF change among groups across a mean 7.2 year interval. Both ASYMAD and CI groups showed similar longitudinal rCBF declines in precuneus, lingual and MTG regions relative to CN. The CI also showed greater rCBF decreases in anterior and posterior cingulate, cuneus and brainstem regions relative to ASYMAD and CN, whereas ASYMAD showed greater relative rCBF increases over time in medial temporal and thalamic regions relative to CI and CN. Our findings provide evidence of early functional alterations that may contribute to cognitive resilience in those who accumulate AD pathology but maintain normal cognition. In a second study, we applied machine learning classifiers to discriminate between cognitively normal individuals who show cognitive decline over time versus those who remain cognitively stable. This study incorporated both MRI and PET resting cerebral blood flow data into the classifier. We showed that functional and structural changes were detected well before (2.3-2.9 years) changes in verbal memory performance. This type of approach will be extended to incorporate additional factors that may protect against cognitive decline and impairment.

Summary of work: As part of our program of research on early markers of Alzheimers disease (EMAD), we are performing serial magnetic resonance imaging (MRI), including measures of vascular changes, positron emission tomography (PET), and neuropsychological assessments in participants from the Baltimore Longitudinal Study of Aging (BLSA) to investigate the neurobiological basis of memory change and cognitive impairment. These evaluations allow us to examine changes in brain structure and function which may be early predictors of cognitive change and impairment, including Alzheimer's disease (AD). We are continuing longitudinal testing of older participants and evaluating new participants, including MRI and concurrent neuropsychological assessments of participants less than 55 years old. For a subsample of individuals aged 55 and older, we also perform a single PET measurement of CBF, followed by a PET scan using 11-C-Pittsburgh Compound B (PiB) to measure in vivo amyloid distribution. Our progress over the last year includes continued acquisition of new neuroimaging assessments as well as continued analysis of existing data and methods development. Approximately half of the neuroimaging study participants are enrolled in the BLSA autopsy program, and we continue to use the 3-dimensional imaging findings to guide more detailed neuropathological investigations. In addition, we are using neuroimaging tools to investigate modulators of cognitive and brain changes, including sex differences in brain aging, genetic, metabolic, and inflammatory risk factors, and the effects of sex steroid and other hormones. An understanding of these brain-behavior associations and early detection of accelerated brain changes that predict cognitive decline and impairment will be critical in identifying individuals likely to benefit from interventions if a successful treatment for prevention or delaying onset of disease is identified. Over the last year, we have published a number of papers from this study. In a prior study, we found both concordance and discordance between in vivo amyloid imaging patterns and pathological ratings of amyloid plaques according to the CERAD classification for pathological diagnosis of AD (Sojkova et al, 2010). Our imaging-neuropathology analyses highlighted difficulties in using standard neuropathological diagnosis for autopsy validation of PiB due to differences in regions examined under the standard CERAD assessment and the brain regions showing the earliest amyloid deposition on PiB imaging. In a follow-up study (Driscoll et al, 2012), we examined the correspondence among regional (11)C-PiB load, region-matched quantitative immunohistological assessments of amyloid-beta and neurofibrillary tangles (NFT), and brain atrophy (MRI) in the same six older Baltimore Longitudinal Study of Aging participants who came to autopsy. We used unbiased stereological principles to quantify amyloid-beta plaques and NFTs in hippocampus, orbito-frontal cortex, anterior and posterior cingulate gyrus, precuneus and cerebellum. We found a general agreement between the regional measures of amyloid obtained via stereological assessment and imaging, with significant relationships evident for the anterior and posterior cingulate gyri, and the precuneus. No associations were observed between (11)C-PiB load and NFT count for any of the regions examined, or between regional Aβ or NFT counts and corresponding brain volumes. These findings support the validity of the PiB imaging for quantification of amyloid plaque burden in vivo. In a follow-up to our earlier work investigating plasma clusterin, we found that higher plasma clusterin was associated with a slower rate of brain atrophy in whole brain, ventricular CSF, temporal gray matter as well as parahippocampal, superior temporal and cingulate gyri. However, this relationship was not observed in individuals who remained cognitively normal (Thambisetty et al, 2012). In a second study, we examined the risk variant rs11136000 single nucleotide polymorphism in the clusterin gene (CLU) in relation to longitudinal changes in resting state regional cerebral blood flow (rCBF) during normal aging, as well as its influence on cognitive decline in presymptomatic stages of disease progression. We examined the effect of the novel Alzheimer's disease (AD) risk variant rs11136000 single nucleotide polymorphism in the clusterin gene (CLU) on longitudinal changes in resting state regional cerebral blood flow (rCBF) during normal aging and investigated its influence on cognitive decline in presymptomatic stages of disease progression. A subset of 88 cognitively normal older individuals had longitudinal (15)O-water positron emission tomography measurements of rCBF at baseline and up to eight annual follow-up visits. We also analyzed trajectories of cognitive decline among CLU risk carriers and noncarriers in individuals who remained cognitively normal (n = 599), as well as in those who subsequently converted to mild cognitive impairment or AD (n = 95). Cognitively normal carriers of the CLU risk allele showed significant and dose-dependent longitudinal increases in resting state rCBF in brain regions intrinsic to memory processes. There were no differences in trajectories of memory performance between CLU risk carriers and noncarriers who remained cognitively normal. However, in cognitively normal individuals who eventually converted to mild cognitive impairment or AD, CLU risk carriers showed faster rates of decline in memory performance relative to noncarriers in the presymptomatic stages of disease progression. Thus, the CLU influences longitudinal changes in brain function in asymptomatic individuals and is associated with faster cognitive decline in presymptomatic stages of disease progression. In addition to the association between CLU and patterns of regional cerebral blood flow change, we found associations between baseline cardiovascular risk, measures with the Framingham Cardiovascular Risk Profile (FCRP), and subsequent change in resting rCBF. Images quantifying voxel-wise longitudinal rates of CBF change were calculated and used to examine the relationship between baseline FCRP score and changes over time in regional CBF in 97 individuals. Higher baseline FCRP scores were associated with accelerated longitudinal decline in CBF in orbitofrontal, medial frontal/anterior cingulate, insular, precuneus, and brain stem regions. Of the components that comprise the FCRP score, higher diastolic blood pressure and diabetes were associated independently with greater decline in the medial frontal/anterior cingulate and insular regions, respectively. The association between baseline risk profile and later rCBF decline emphasizes the importance of treating potentially modifiable risk factors for cognitive and brain changes. The data from this project also continue to be used for important methodological developments to enhance analysis of longitudinal neuroimaging data, including papers which aim to determine the temporal relationships among brain changes and cognitive or affective changes (Chen et al, 2012; Dotson et al, 2012; Clark et al, epub).

PET imaging with the Tau radioligand AV-1451 was introduced into the BLSA study in 2016. In combination with PET amyloid imaging and MRI measures, these studies will provide information on the preclinical stages of Alzheimer's Disease. As of August 22, 2018, 79 BLSA participants have been enrolled in the study and more than 30 have had at least one repeated study. A customized analysis pipeline has been developed with refined MRI-based regions using the MUSE analysis for structural MRI region of interest definition. Images are acquired on a High Resolution Research Tomograph for PET brain imaging and transferred to the Laboratory of Behavioral Neuroscience for analysis. Preliminary analysis of the first 60 images show the expected finding of greater Tau deposition in individuals with positive amyloid PET scans. In addition, associations of extracerebral off-target binding with age have been demonstrated. Analysis of additional images is underway and longitudinal repeats, as well as new enrollment, continue.

Summary of work: As part of our program of research on early markers of Alzheimers disease, we are performing serial magnetic resonance imaging (MRI), including measures of vascular changes, positron emission tomography (PET), and neuropsychological assessments in participants from the Baltimore Longitudinal Study of Aging (BLSA) to investigate the neurobiological basis of memory change and cognitive impairment. These evaluations allow us to examine changes in brain structure and function which may be early preclinical predictors of cognitive change and impairment, including Alzheimer's disease (AD). Neuroimaging evaluations of BLSA participants began in 1994 with approximately 160 individuals aged 55-85 at enrollment. We continue longitudinal testing of older participants and evaluation of new participants, including MRI and neuropsychological assessments of participants younger than 55 years old. Since 2009, we have expanded MRI assessments to a larger portion of the BLSA with more than 1000 individuals receiving MRI scans and concurrent cognitive assessments. For a subsample aged 55 and older, we performed PET measurements of cerebral blood flow, followed by a PET scan using 11-C-Pittsburgh Compound B (PiB) to measure in vivo amyloid deposition. Over the last 2 years we expanded our amyloid imaging studies from 50 scans per year to 70 scans per year through Alzheimer's disease funding initiatives. In addition, we initiated Tau PET (AV-1451) studies of BLSA participants receiving PET amyloid scans (separate annual report). Our progress includes continued acquisition of new neuroimaging assessments and continued analysis of existing data and methods development. We use neuroimaging tools to investigate modulators of cognitive and brain changes, including sex differences in cognitive and brain aging, genetic, metabolic, and inflammatory risk factors, and the effects of sex steroid and other hormones. An understanding of these brain-behavior associations and early detection of accelerated brain changes during the preclinical or asymptomatic stage of disease will be critical in identifying individuals likely to benefit from interventions if a successful treatment for prevention or delaying onset of disease is available. Over the last year we have published a number of papers, including demonstrations that elevated markers of inflammation are associated with longitudinal changes in regional cerebral blood flow measured by PET (Warren et al, 2018) and that rates of amyloid-beta accumulation are greater in amyloid positive individuals who also have metabolic syndrome (Gomez et al, 2018). We have used multimodal imaging with measures of brain structure and resting state fMRI connectivity to characterize the heterogeneity of structural and functional MR imaging patterns in five groups of individuals demonstrating patterns of advanced aging versus a group of resilient agers (Eavani et al, 2018). We have also investigated associations between regional white matter lesion volumes, representing 4 clusters of distinct anatomic distribution, and clinical and cognitive characteristics in BLSA participants (Habes et al, 2018). We have used the rich data available from participants receiving amyloid PET and structural MRI scans in combination with serial cognitive assessments to investigate the independent and joint effects of amyloid pathology, a hallmark Alzheimers neuropathology, and neurodegeneration, measured by hippocampal atrophy, on changes in memory and other cognitive domains (Bilgel et al, 2018). We categorized 171 cognitively normal BLSA participants (562 cognitive assessments, 3.7 years follow-up) based on dichotomous amyloid pathology (A) and hippocampal neurodegeneration (N) status at baseline: A-N-, A+N-, A-N+, A+N+. We conducted linear mixed effects analyses to assess cross-sectional and longitudinal trends in cognitive test z-scores by amyloid and neurodegeneration group. Individuals with amyloidosis or hippocampal atrophy had steeper longitudinal declines in verbal episodic memory and learning compared to those with neither condition. Among individuals with hippocampal atrophy, amyloid positivity was associated with steeper declines in verbal memory, visual memory, language, and mental status. Similarly, among individuals with amyloidosis, hippocampal atrophy was associated with steeper declines in verbal memory, visual memory, language, and mental status. Presence of both amyloidosis and hippocampal atrophy was associated with greater declines than would be expected by their additive contributions in visual memory, language, and mental status. Neither amyloidosis nor hippocampal atrophy was predictive of declines in executive function, processing speed, or visuospatial ability. Our results suggest that both amyloid pathology and neurodegeneration have disadvantageous, in part synergistic, effects on prospective cognition. These cognitive effects are detectable early among cognitively normal individuals with amyloidosis, who are in preclinical stages of Alzheimer's disease and highlight the importance of early intervention to target both amyloidosis and atrophy to preserve cognitive function before further damage occurs. Methodological Developments. We have continued to optimize our approach for MRI volumetric analysis. We have used the MUSE analysis pipeline to harmonize MRI volumetric data over more than 20 years, allowing analysis of predictors of neurodegeneration. This approach was validated for MRI data acquired across different field strengths and vendors (Erus et al, 2018). In addition to these selected highlights, we have a number of manuscripts at various stages of review and development, including papers describing accelerated MRI-assessed brain volume loss in individuals who subsequently develop cognitive impairment, sex differences in volume loss, and brain volume changes in participants with hearing loss.

The Womens Health Initiative (WHI) randomized, placebo-controlled clinical trials of hormone therapy (HT) were designed to test the hypothesis that conjugated equine estrogens alone (CEE-Alone) or in combination with medroxyprogesterone acetate (CEE+MPA) protected postmenopausal women against the development of heart disease. The WHI Memory Study (WHIMS) was an ancillary study to the WHI trials, which consisted of parallel placebo-controlled randomized clinical trials of 0.625 mg/day CEE therapy with and without 2.5 mg/day MPA in women with a uterus or post-hysterectomy, respectively. WHIMS investigated the effect of CEE-Alone and CEE+MPA on risk for probable dementia and mild cognitive impairment in women age 65 and older, as well as the effects of these treatments on global cognitive function. The WHI Study of Cognitive Aging (WHISCA), an ancillary study to WHIMS, was developed to investigate the effects of HT on domain-specific cognitive function in women without dementia. WHISCA enrolled 2305 women at 14 of the WHIMS sites, distributed across the two parallel trials. WHISCA was initiated on average 3 years after WHI randomization and the primary outcome was the effect of HT on rates of cognitive change, adjusted for time since randomization. The WHIMS CEE+MPA trial terminated earlier than planned (July, 2002) due to an adverse risk-to-benefit profile in the main WHI trial. Subsequently, the WHI CEE-Alone Trial also was terminated early (February, 2004). Results from the WHIMS trials showed that CEE-Alone or CEE+MPA increase the risk of dementia and have adverse effects on global cognition in women aged 65 years or older. HT also has been shown to increase the risk of clinical stroke in women 65 years and older. The initial report of WHISCA findings showed that CEE + MPA had a negative impact on verbal memory (p < 0.01) and a trend to a positive impact on figural memory (p = 0.012) over time compared with placebo with no effect on other cognitive domains. In addition, these effects were evident only after long-term therapy. CEE + MPA did not significantly influence positive affect, negative affect, or depressive symptoms. These findings suggest that HT may have different effects across different cognitive domains. The findings from the CEE-Alone Trial in women with prior hysterectomy who were randomized to CEE or placebo show that CEE alone did not affect domain-specific cognitive function over time. Participants in the WHISCA and WHIMS studies continue to be followed through telephone cognitive assessments as they pass through the risk period for cognitive decline. The WHIMS Suite of Studies also included cognitive follow-up of women in the WHIMS-Younger (WHIMS-Y) study, who were randomized to hormone therapy through the WHI when aged 50-54 years. The WHIMS-Y study tested the hypothesis that hormone therapy around the time of the menopause may benefit cognitive function later in life. Initial results from the WHIMS-Y study were based on 1326 postmenopausal women studied with a validated telephone cognitive assessment battery an average of 7.2 years after the trials ended, when women had a mean age of 67.2 years. The initial report included the first two administrations of the cognitive battery and showed neither harm nor benefit of early post-menopausal HT on later cognitive function. The WHIMS Suite of Studies is conducted by Wake Forest University, which is also the site for the Southeast Regional Center for WHI. Wake Forest Investigators lead the Aging, Cognition and Function interest group for the WHI. Follow-up cognitive evaluations continue through telephone assessments in the original WHIMS cohort (WHIMS extension study). The WHIMS-Y study has concluded as sufficient data were obtained in this cohort to test the study hypotheses. In addition to ongoing evaluations in the current WHIMS extension cohort, analysis of existing data continues. These analyses include genetic associations with risk for cognitive impairment and with cognitive trajectories, effects of particulate air pollution on cognitive, affective and brain outcomes, and determination of factors that promote cognitive resilience even in the presence of the APOE e4 genetic risk allele. In one recent paper, annual cognitive assessments from 2561 women, aged 75 to 92, were used to identify distinct trajectories, ranking women according to age-, race- and ethnicity-adjusted performance levels. Participants assigned to the resulting trajectories were compared for selected risk factor profiles. Using this approach, women were grouped into five trajectories according to relative cognitive performance over time. These groups differed significantly according to 3 known risk factors for cognitive decline - education level, apolipoprotein E-e4 genotype, and type 2 diabetes mellitus -and a biomarker based on brain structure that has been linked to cognitive decline and Alzheimer's disease. Participants with consistently low relative levels of cognitive function over time and those whose relative performance over time declined to these levels tended to have poorer risk factor profiles. Thus, longitudinal measures of an individual's relative performance on different assessment protocols for global cognitive function can be used to identify trajectories of change over time that appear to have internal validity with respect to known risk factors. In a second paper, we used WHISCA data on positive affect (PA) and negative affect (NA) to determine whether affective state at baseline influenced risk of cognitive impairment (mild cognitive impairment (MCI) and probable dementia) in women without depression. Baseline PA and NA were assessed in 2137 postmenopausal women (mean age 73.8 years) who had participated in WHISCA and had received annual assessments with the Positive and Negative Affect Schedule (PANAS). Women were followed annually for an average of 11.3 years; those with elevated depressive symptoms at baseline were excluded. We found that higher NA was associated with a higher risk of MCI and probable dementia, even after adjusting for covariates including age, education, sociodemographic, lifestyle, and cardiovascular risk factors, global cognition, and HT assignment. PA was not significantly associated with either outcome. These results show that NA, even in the absence of elevated depressive symptoms, is associated with higher risk of MCI and dementia. The next phase of the WHIMS analyses are focusing on women who survive to old age (80+) and remain cognitively healthy, cognitive resilience. In addition, a new study is underway to evaluate the biology of cognitive resilience in women who carry the APOE e4 genetic risk allele for Alzheimer's disease but remain cognitively healthy in late life. Biospecimens obtained at WHI baseline will be used to conduct studies of insulin resistance, proteomics, metabolomics and auto-antibody profiles in APOE e4 escapees in comparison to those who develop cognitive impairment and in comparison to women with the APOE e3/e3 genotype.