Increases in Neuroticism May Be an Early Indicator of Dementia: A Coordinated Analysis

Increases in Neuroticism May Be an Early Indicator of Dementia: A Coordinated Analysis Abstract Objectives Although personality change is typically considered a symptom of dementia, some studies suggest that personality change may be an early indication of dementia. One prospective study found increases in neuroticism preceding dementia diagnosis (Yoneda, T., Rush, J., Berg, A. I., Johansson, B., & Piccinin, A. M. (2017). Trajectories of personality traits preceding dementia diagnosis. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 72, 922–931. doi:10.1093/geronb/gbw006). This study extends this research by examining trajectories of personality traits in additional longitudinal studies of aging. Methods Three independent series of latent growth curve models were fitted to data from the Longitudinal Aging Study Amsterdam and Einstein Aging Study to estimate trajectories of personality traits in individuals with incident dementia diagnosis (total N = 210), in individuals with incident Mild Cognitive Impairment (N = 135), and in individuals who did not receive a diagnosis during follow-up periods (total N = 1740). Results Controlling for sex, age, education, depressive symptoms, and the interaction between age and education, growth curve analyses consistently revealed significant linear increases in neuroticism preceding dementia diagnosis in both datasets and in individuals with mild cognitive impairment. Analyses examining individuals without a diagnosis revealed nonsignificant change in neuroticism overtime. Discussion Replication of our previous work in 2 additional datasets provides compelling evidence that increases in neuroticism may be early indication of dementia, which can facilitate development of screening assessments. Longitudinal change, Mild cognitive impairment, Multi-study conceptual replication, Personality Although research examining personality in healthy older adults indicates minimal variation in personality traits over time (Allemand, Zimprich, & Hertzog, 2007; Berg & Johansson, 2013; Mroczek & Spiro, 2003; Small, Hertzog, Hultsch, & Dixon, 2003), several studies report substantial personality change in individuals who have been diagnosed with dementia. Studies using retrospective informant-reports of personality have found personality changes in individuals diagnosed with dementia, including increases in neuroticism (Dawson, Welsh-Bohmer, & Siegler, 2000; Torrente et al., 2014), decreases in extraversion (Dawson et al., 2000; Rankin, Kramer, Mychack, & Miller, 2003; Torrente et al., 2014) and decreases in conscientiousness (Dawson et al., 2000; Torrente et al., 2014) relative to premorbid personality traits. The connection between personality change and dementia at the behavioral level is consistent with neurological research utilizing sophisticated neuroimaging techniques; that is, previous work suggests a relation between the neurological underpinnings of personality and dementia-related neuropathology. Research by Mahoney and associates (2011) used volumetric brain MRI to examine the neuroanatomical profiles of personality change in individuals diagnosed with frontotemporal lobar degeneration (FTD) and healthy controls (Mahoney, Rohrer, Omar, Rossor, & Warren, 2011). Personality change was assessed by informant-report of current personality relative to estimated premorbid personality for the FTD group, and by informant-report of current personality relative to estimated personality 10 years previously for the control group. The individuals with FTD showed significant declines in extraversion, agreeableness, openness, and conscientiousness, and increases in neuroticism, compared with trait stability for the control group. MRI results revealed that trait changes were associated with cortical losses in regional gray matter in overlapping anterior and dorsal cortical areas, emphasizing that personality change in FTD is associated with brain damage. Mahoney and associates (2011) concluded that by examining individuals with FTD neuropathology, their research provided an opportunity to identify the neural substrates that are critical for maintenance of personality trait stability. Although Mahoney’s study focused on FTD, other types of dementia neuropathology may also affect the biological mechanisms and structures that support personality, particularly because similar patterns of personality change are seen in individuals with Alzheimer’s disease (AD), FTD, and temporal-variant FTD (Rankin et al., 2003), as well as generalized pattern of personality change in individuals with AD and behavioral-variant of FTD (Torrente et al., 2014). Given that dementia neuropathology and dementia-related cognitive decline develop many years prior to the clinical diagnosis of dementia (Caselli, Beach, Knopman, & Graff-Radford, 2017; Hall, Lipton, Sliwinski, & Stewart, 2000), dementia neuropathology may also affect aspects of personality in the years preceding dementia diagnosis. Indeed, a small number of studies suggest that personality changes may precede dementia diagnosis. Lykou and associates (2013) found that increases in neuroticism occurred in individuals with mild cognitive impairment (MCI) as well as those with AD. Although this research applied a cross-sectional retrospective research design, the results indicate that a similar pattern of personality change may occur as early as the MCI disease stage. Two studies have prospectively examined the timing of informant-rated personality change, operationalized as an affirmative answer to any item regarding specific change in personality (Balsis, Carpenter, & Storandt, 2005; Smith-Gamble et al., 2002). Both studies found that individuals with informant-rated personality change at baseline were significantly more likely to be diagnosed with dementia at follow-up occasions. More recently, a study by Yoneda, Rush, Berg, Johansson, and Piccinin (2017) prospectively investigated trajectories of self-reported personality traits using latent growth curve modeling, rather than general personality change using chi square tests. Based on intraindividual trajectories of personality traits in individuals who received a dementia diagnosis after the first measurement occasion, analyses revealed significant increases in neuroticism and no change in extraversion preceding dementia diagnosis. In contrast, personality trait stability over time was found in the individuals who were not diagnosed, suggesting that trajectories of personality are different for individuals who did and did not receive a dementia diagnosis during the observation period. Conceptual replication of this work in additional longitudinal studies of aging would offer additional evidence that personality change is possibly an early indication of dementia, which may be valuable in a clinical setting. Identification of the early signs of dementia can aid in implementing early treatment strategies, planning dementia care services, and facilitating development of screening assessments. Most importantly, although there is not yet a cure for dementia, mounting evidence suggests that progression to dementia may be slowed by adherence to a healthy lifestyle (Di Marco et al., 2014; Middleton & Kristine, 2010; Polidori, Nelles, & Pientka, 2010). Therefore, the sooner that dementia can be identified, the sooner individuals can be educated and, ideally, motivated to commit to a healthier lifestyle. The current work extends Yoneda and colleagues’ research by investigating personality change in two additional affiliates of the Integrative Analysis of Longitudinal Studies on Aging and Dementia (IALSA) network: The Longitudinal Aging Study of Amsterdam (LASA; Huisman et al., 2011) and Einstein Aging Study (EAS; Katz et al., 2012). IALSA was established to enable and encourage replication of research using comparable methods and measures across longitudinal studies of aging and dementia. Further, a coordinated analysis provides the opportunity for immediate conceptual replication and comparison of the results, which protects against Type I errors, allows for evaluation of consistencies and differences between samples, and is important for effective cumulative science (Graham et al., 2017; Hofer & Piccinin, 2009; Hofer & Piccinin, 2010). LASA and EAS were selected based on inclusion of both repeated assessment of personality traits and an indication of definite or probable dementia diagnosis. Our study includes individuals with all types of dementia to increase power within our analyses and because previous research indicates a generalized pattern of personality trait changes in individuals with dementia (Rankin et al., 2003; Torrente et al., 2014). The EAS also provided data for trajectories of personality traits in individuals who were diagnosed with MCI at some point during the study. To our knowledge, no research has examined trajectories of personality traits in individuals with MCI. The rationale for including these analyses is based on robust evidence suggesting that the cognitive impairment observed in MCI represents an early stage of dementia (Petersen et al., 2014). A systematic review including 19 longitudinal studies investigating conversion from MCI to dementia concluded that mild cognitive decline is not a normal part of aging and significantly predicts conversion to dementia (Bruscoli & Lovestone, 2004). Furthermore, Roberts and associates (2014) found that individuals with prevalent or incident MCI (N = 534) have a high risk of progressing to dementia, even when they reverted to normal cognition at some point during the longitudinal study. Therefore, our analyses examining trajectories of personality in individuals with MCI may provide further information regarding personality change in the earliest stages of the disease process, though it is important to note that not all individuals classified with MCI will convert to dementia. The research examining personality change preceding (Yoneda et al., 2017) and following (Dawson et al., 2000; Lykou et al., 2013; Mahoney et al., 2011; Rankin et al., 2003; Torrente et al., 2014) diagnosis of dementia informs our hypotheses: that (a) increases in neuroticism and (b) decreases in extraversion, conscientiousness, agreeableness, and openness will precede diagnosis of dementia. Furthermore, based on retrospective research examining individuals with incident MCI and AD (Lykou et al., 2013), we expect a similar pattern of personality change in individuals who are classified as having MCI. Method Participants and Procedures For both datasets, individuals were not included in the analyses if they were diagnosed with dementia at baseline. Three independent series of latent growth curve models were estimated to examine (a) individuals diagnosed with incident dementia, (b) individuals diagnosed with incident MCI (in EAS), and (c) individuals not diagnosed with dementia (nor MCI, in EAS). For the first set of analyses examining individuals diagnosed with dementia, personality data were excluded after the occasion in which participants were diagnosed based on the following rationales: our primary interest was personality changes preceding diagnosis; self-report following diagnosis has questionable accuracy; personality was generally no longer assessed following diagnosis. For the second set of analyses examining individuals with MCI in the EAS dataset, all available personality data from these individuals were included, resulting in examination of trajectories preceding and following MCI classification. In order to ensure that data were only used to represent one of either MCI or dementia, individuals who were eventually diagnosed with dementia during the study were excluded from the MCI analysis (and vice versa). The third set of analyses examined trajectories of personality traits for individuals who had repeated measurement of personality and did not receive a dementia diagnosis, and, in the case of the EAS, also did not receive diagnosis of MCI. Longitudinal Aging Study Amsterdam participants (LASA; Huisman et al., 2011; Amsterdam, Zwolle and Oss, the Netherlands). The LASA is a population-based cross-sequential longitudinal study consisting of three independent and geographically representative cohorts of older adults, aged 55–85 years. Participants were recruited in 1992–1993 and tested every 3 years following recruitment into the study. Participants were visited in their homes by trained interviewers, who completed a main interview and a medical interview at each occasion. Participants were asked to complete and mail back several questionnaires, such as the Neuroticism scale. Data from individuals with incident probable dementia classification (N = 162) and individuals not diagnosed with dementia during the study (N = 1138) were used in the current analysis. Einstein Aging Study participants (EAS; Katz et al., 2012; New York, United States). The primary objective of the EAS is to study the aging brain and healthy aging, and the risk factors associated with AD. The study began in 1993, and the sample consists of community dwelling adults aged 70 and older who were systematically recruited from Bronx County, New York. After study entry, participants contributed data annually by undergoing physical, cognitive, and psychosocial assessments at a clinical research center (Katz et al., 2012). Due to a limited subsample of individuals classified as having dementia (N = 48), only three waves of personality data were included in the analyses examining the years preceding dementia diagnosis. Additional analyses include five waves of data examining individuals diagnosed with MCI (N = 135) and individuals not diagnosed with either MCI or dementia during the study (N = 602). Measures Each study involved a large battery of measures including assessment of health, functional and mental capacity, well-being, personality, depressive symptoms, and social network. A subsample of these assessments was used in the current analyses. Due to heterogeneity across samples, personality measures and depressive symptoms were not standardized, as standardized scores may be sample-specific. Although the personality and depression scales are not equivalent across samples, unstandardized scores permit investigation of measures of change between different samples (Schumacker & Lomax, 2015). Personality measures Each of the studies assessed personality using different self-report measurements. Although the assessments are not identical, different measurement of the same trait are highly correlated (Luteijn, Starren, & Van Dijk, 2000; McCrae & Costa, 1985), allowing for comparison of the same construct between studies. The LASA used the Neuroticism scale from the Dutch Personality Questionnaire (DPQ; Luteijn et al., 2000), which contains 15 items with three response options: applicable, do not know, and not applicable. Scores on the Neuroticism scale can range from 0 to 30, with higher scores indicating a higher degree of neuroticism. Good internal consistency reliability is reported (α = .85). The EAS assessed the Big Five personality traits using items from the International Personality Item Pool (IPIP; Goldberg, 1999). Each trait scale contains 10 items with five response options ranging from 1 (very inaccurate) to 5 (very accurate). A mean score was computed for each trait scale; thus, scores can range from 1 to 5, with higher scores indicating higher endorsement of that trait. Good internal consistency reliability estimates are reported: neuroticism (α = .76), extraversion (α = .76), agreeableness (α = .72), contentiousness (α = .81), and openness (α = .72). Covariates Diagnosis of dementia and MCI Individuals were included in the analyses examining trajectories of personality traits aligned according to dementia and MCI diagnosis based on their corresponding diagnosis. The LASA did not include a formal dementia diagnosis, but rather a composed variable of probable dementia, determined by taking into account the Mini Mental State Examination (MMSE) score, a telephone administered MMSE, or an Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) (Van den Kommer et al., 2008). Specifically, individuals who scored two or more standard deviations below the mean decline on the MMSE since the last occasion of measurement or at least 28 points on the IQCODE were classified as having probable dementia. Further, continued decline at the next measurement occasion was required. Finally, several factors were considered to obtain information regarding probable diagnosis including reason for participant dropout or cause of death as well as pertinent data from General Practitioners, specialists, and nursing home admittance. The date of probable dementia classification corresponds to the measurement occasion in which participants met the above requirements, only if those participants also demonstrated continued decline at the next wave. Incident MCI diagnosis was not available in LASA. The EAS included a clinical conference in which dementia diagnosis was based on standardized clinical criteria from the DSM-IV (American Psychiatric Association, 1994). Criteria were deliberated according to a comprehensive review of cognitive, functional and neurological assessments. Furthermore, the EAS included criteria to classify amnestic MCI (aMCI) and non-amnestic MCI (naMCI) for participants who were not diagnosed with dementia, but did demonstrate cognitive impairment and had subjective complaints of decline. Specifically, individuals with aMCI demonstrated objective and subjective memory impairment, while individuals with naMCI did not meet memory criteria for aMCI, but scored at least 1.5 SDs below the mean in at least one of the following domains: executive function, language, visuospatial ability, or attention (Katz et al., 2012). The date of dementia or MCI diagnosis was set when the diagnostic team assembled shortly after the in-person examinations. Depressive symptoms Depressive symptoms measured at the first occasion were included as a covariate in this study. Assessment at baseline was used to minimize missingness. Controlling for depression is important because both dementia and neuroticism have been associated with higher levels of depressive symptoms (Nussbaum, 1997; and Kendler, Neale, Kessler, Heath, & Eaves, 1993, respectively). For LASA, depressive symptoms were measured using the self-report Center for Epidemiologic Studies Depressive Scale (CES-D; Radloff, 1977), translated into Dutch. The scale includes 20 items rated on a 4-point scale ranging from 0 (“rarely or none of the time”) to 3 (“most or all of the time”). Participants are asked about the frequency of certain experiences during the past week. Four items are reverse coded prior to summing scores. Total scores range from 0 to 60, with higher scores indicating more depressive symptoms. Good internal consistency reliabilities are reported (α = .87–.90). For EAS, depressive symptoms were measured using the self-report Geriatric Depression Scale (GDS; Yesavage & Sheikh, 1986). The scale includes 15 items with dichotomous response options (yes = 1; no = 0). Participants are asked about the frequency of experiences during the past week. Total scores range from 0 to 15, with higher values indicating more depressive symptoms. Good internal consistency reliabilities are reported (α = .87). Statistical Analysis The nature of a longitudinal research design makes the use of latent growth curve modeling (LGM) desirable. Each personality trait was examined longitudinally with latent growth curve models in Mplus version 7.3 (Muthén and Muthén (1998–2013)) estimating individual trajectories of change in each personality trait with and without covariates. The models were estimated independently in each study. The first set of analyses examined trajectories of personality traits and the preclinical onset of dementia. Individuals were aligned according to the occasion of diagnosis; thus, time was specified as “years-preceding-dementia,” with the occasion in which individuals were diagnosed with dementia specified as time zero (i.e., the intercept), and no measurement of personality after this point. The second set of analyses examined individuals diagnosed with MCI from the EAS. Individuals were aligned chronologically according to the occasion at which they received a classification of MCI (i.e., the intercept, or “0” on each individual’s timeline, is the wave when MCI classification was assigned). As noted, individuals who were diagnosed with incident MCI continued to complete the personality assessments and, for these analyses, the data following diagnosis were included; therefore, the “MCI timeline structured” analyses examined trajectories of personality preceding and after MCI classification. Any individuals who also received a dementia diagnosis during the study were excluded from these analyses, however. The third set of analyses examined individual trajectories of personality for individuals who did not receive a dementia diagnosis during the study, and, for the EAS, were also not classified with MCI. Time was specified as time-in-study and the intercept was specified as the baseline assessment, resulting in the intercept corresponding to a younger age compared with the analyses examining individuals diagnosed with dementia. Although the same number of occasions and set of covariates that were used in the first two series of analyses were used in these models, the analyses aim to provide context rather than a direct comparison due to the necessary difference in the structuring of time and the meaning of the intercept. For EAS, the same number of occasions (five) and set of covariates were used as the analyses examining individuals with MCI in order to maximize the number of waves of data included in these analyses, as structuring according to the same constraints as the analyses examining individuals with dementia would have resulted in three waves of data. For each dataset, sex was included as a dichotomous variable, with male as the reference group (see Table 1 for demographic information). Age and education were measured in years and centered at each sample mean. Centering age and education across samples at a common value may have facilitated interpretability; however, because the mean age is different for each sample, centering at a common value would have resulted in extrapolation. The interaction between age and education was also included to acknowledge expected age-related differences in educational attainment. Across datasets the modal response to depressive symptoms was zero, so depressive symptoms were entered into the models uncentered. Table 1. Demographic Information and Baseline Personality Descriptive Statistics of Participants Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Note. LASA = Longitudinal Aging Study Amsterdam; EAS = Einstein Aging Study; Dem Dx = Individuals diagnosed with dementia; MCI Dx = Individuals diagnosed with MCI; No Dem Dx = Individuals who did not receive a dementia diagnosis during the study; No Dem/MCI Dx = Individuals who did not receive an incident dementia or MCI diagnosis; Time to Dx = time in years to dementia or MCI diagnosis at baseline; Max waves = maximum waves included in analyses; Average waves = average waves of data included in analyses; Depression= depressive symptoms, CES-D (LASA) or GDS (EAS) at Time 1; Wave 1 personality = Means and standard deviations reported from the conditional models; — = not available. View Large Table 1. Demographic Information and Baseline Personality Descriptive Statistics of Participants Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Note. LASA = Longitudinal Aging Study Amsterdam; EAS = Einstein Aging Study; Dem Dx = Individuals diagnosed with dementia; MCI Dx = Individuals diagnosed with MCI; No Dem Dx = Individuals who did not receive a dementia diagnosis during the study; No Dem/MCI Dx = Individuals who did not receive an incident dementia or MCI diagnosis; Time to Dx = time in years to dementia or MCI diagnosis at baseline; Max waves = maximum waves included in analyses; Average waves = average waves of data included in analyses; Depression= depressive symptoms, CES-D (LASA) or GDS (EAS) at Time 1; Wave 1 personality = Means and standard deviations reported from the conditional models; — = not available. View Large For the EAS years preceding dementia analyses that included a maximum of three waves, only linear trajectories were considered. Quadratic trajectories were examined for all other analyses to determine the most appropriate model for trajectory shape. For some personality traits, the quadratic models would not converge; however, in the models that did, the quadratic trajectories revealed higher Bayesian information criteria (BIC), suggesting that the linear models afforded a more parsimonious fit to the data. In addition, the EAS years-preceding-dementia models had some model convergence issues, which may have been due to limited sample size (<100 individuals), or possibly due to the variation in dementia diagnosis between studies. Although standard estimates and errors were reported for the EAS models, a warning that the estimation had reached a saddle point was included in the output. The slope variance was nonsignificant for these models, so analyses were rerun with the slope variance constrained to zero, which resulted in model convergence with no warnings. Therefore, results are reported for the EAS models with the slope variance constrained to zero. Due to larger sample sizes and slope variances, it was not necessary to constrain the slope variances to zero in the analyses examining individuals who were classified with probable dementia in LASA, nor for the analyses examining individuals without a diagnosis in either dataset. An alpha level of .05 was used for all statistical tests. Results Demographic information and baseline personality descriptive statistics are presented in Table 1. Univariate Latent Growth Curve Modeling For each study, univariate growth curve models for each of the personality traits were fitted with and without covariates to examine rates of change in personality traits preceding dementia diagnosis, as well as preceding and following MCI diagnosis. Results are reported for the model with sex, age, education, depressive symptoms at the first occasion, and the interaction between age and education as covariates. All covariates were included in the model as covariates of the intercept (centered at time of dementia or MCI diagnosis) and the slope, allowing both trajectory characteristics to differ across individuals who varied on these characteristics. The following results focus on the main trajectories of individuals diagnosed with dementia and MCI (in the case of EAS), and of individuals who did not receive a diagnosis of dementia (nor MCI, in the case of EAS). Identification and interpretation of significant covariates within each model is included in the supplementary material. Results from the unconditional models are not reported, but this information is available by request. Personality Change in Individuals Diagnosed With Incident Dementia and MCI Neuroticism Both the LASA and EAS datasets included measurements of neuroticism. The conditional models for both studies consistently revealed significant mean linear increases in neuroticism. Figure 1 depicts the results for the average trajectories of neuroticism for each study; that is, for a male participant who entered the study at the mean age with the mean years of education and no depressive symptoms. The intercept (i.e., time = 0) represents the average expected score on the neuroticism scale at time of diagnosis of dementia or of MCI in the case of EAS, and baseline for individuals not diagnosed with MCI or dementia. The slope represents the average change per year between study entry and dementia diagnosis, MCI diagnosis, or time in study, respectively. The average trajectory of neuroticism for individuals with MCI spans the intercept because data were included leading up to and following diagnosis of MCI. The intercept and slope differ substantially between datasets because personality was measured using different scales, and the inter-occasion intervals differ between datasets. Parameter estimates and standard errors for conditional models for trajectories of neuroticism are presented in Table 2. Table 2. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models Personality Trait Neuroticism Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES-D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × Education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Table 2. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models Personality Trait Neuroticism Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES-D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × Education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Figure 1. View largeDownload slide Average trajectories of neuroticism. Note. The left plot shows trajectories of neuroticism based on years preceding dementia diagnosis for individuals with incident dementia from the Longitudinal Aging Study of Amsterdam (LASA) and Einstein Aging Study (EAS) datasets. The left plot also shows the average trajectory of neuroticism based on years preceding and following diagnosis of MCI for individuals from EAS. The right plot shows trajectories of neuroticism for individuals from LASA and EAS who did not receive a dementia diagnosis during the course of the study. Each line represents a male participant who entered the study at the mean age with the mean years of education for each study and no depressive symptoms. Figure 1. View largeDownload slide Average trajectories of neuroticism. Note. The left plot shows trajectories of neuroticism based on years preceding dementia diagnosis for individuals with incident dementia from the Longitudinal Aging Study of Amsterdam (LASA) and Einstein Aging Study (EAS) datasets. The left plot also shows the average trajectory of neuroticism based on years preceding and following diagnosis of MCI for individuals from EAS. The right plot shows trajectories of neuroticism for individuals from LASA and EAS who did not receive a dementia diagnosis during the course of the study. Each line represents a male participant who entered the study at the mean age with the mean years of education for each study and no depressive symptoms. Extraversion, conscientiousness, agreeableness, and openness The EAS protocols included assessment of the four additional personality traits. For conscientiousness, agreeableness, and openness, none of the conditional models revealed significant change, suggesting relative stability in these traits for individuals who were eventually diagnosed with dementia and MCI. The parameter estimates and standard errors for the conditional models of these personality traits are presented in Supplemental Table S1 for both time metrics. Likewise, the extraversion models revealed nonsignificant linear slope means and variances in the years-preceding-dementia models. However, for the MCI-timeline-structured models, both the conditional and unconditional models revealed significant negative linear slope means (p < .01). The results suggest that at the point of MCI classification, a man who entered the study at age 80 with 14 years of education scored an average of 3.3 on the extraversion scale and had been decreasing by .04 per year between study entry and MCI classification. The parameter estimates and standard errors for the conditional extraversion trajectory models are presented in Table 3. Table 3. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models for Personality Trait Extraversion Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES−D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Table 3. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models for Personality Trait Extraversion Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES−D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Personality Change in Individuals Not Diagnosed With Dementia Across both datasets, the time-in-study trajectory models for personality traits in individuals who were not diagnosed with dementia revealed limited longitudinal changes. Although slopes in the unconditional models were occasionally significant (two out of six models: neuroticism decreases in LASA and conscientiousness decreases in EAS), once covariates were added to the models, none of the personality trait trajectories identified significant change, suggesting relative stability of all personality traits in individuals who were not diagnosed with dementia (or MCI, in the case of EAS) during the study. The parameter estimates and standard errors for the conditional models are presented in the following tables: neuroticism in Table 2; extraversion in Table 3; and conscientiousness, openness and agreeableness in Supplemental Table S2. Discussion Our previous work (Yoneda et al., 2017) examined trajectories of personality traits using data from Origins of Variance in the Oldest-Old (OCTO-Twin; McClearn et al., 1997). Analyses revealed significant increases in neuroticism and stability in extraversion preceding dementia diagnosis. The primary aim of this study was to replicate these analyses to examine whether consistent changes in personality traits precede dementia diagnosis. Applying a coordinated analysis approach, two additional datasets were selected through the IALSA network based on availability of repeated assessment of personality and classification of dementia. Conceptual replication, utilizing identical statistical models to the extent possible, provides a powerful basis to evaluate replicability, protects against Type I errors, and allows for accelerated accumulation of knowledge (Hofer & Piccinin, 2009). Consistent with our previous work (Yoneda et al., 2017) and supporting our first hypothesis, we identified significant linear increases in neuroticism preceding dementia diagnosis in both EAS and LASA datasets. Furthermore, analyses revealed linear increases in neuroticism in the analyses examining individuals with an incident MCI diagnosis. Finally, consistent with prospective research finding that individuals who were not diagnosed with incident dementia were less likely to show personality change compared with individuals who eventually receive a diagnosis (Balsis et al., 2005; Smith-Gamble et al., 2002), the analyses examining individuals who did not receive a diagnosis during the study revealed stability in neuroticism over time. These findings provide reliable evidence of a consistent pattern of neuroticism increases preceding dementia diagnosis, and, further, suggest that change in neuroticism may occur early in the disease process. Additionally, these results indicate that individuals who remain undiagnosed have markedly different trajectories of neuroticism compared with individuals not diagnosed with incident dementia or MCI. In conducting the current analyses examining trajectories of personality traits in individuals diagnosed with incident dementia, we noted very small slope variance for the EAS years preceding dementia models. Our previous OCTO-Twin models demonstrated similarly low slope variance, and did not converge when all of the covariates were included for the slope; therefore, we had trimmed two of the covariates. However, by constraining the slope variance to zero in the EAS years preceding dementia models, we were able to include all of the covariates on the slope of each personality trait, with no convergence issues. For the purposes of a more meticulous replication (i.e., models in which all covariates of interest were included as predictors of the slope of each personality trait across all three datasets), we re-fitted the OCTO-Twin models to estimate trajectories of neuroticism and extraversion preceding dementia diagnosis with the slope variance constrained to zero. Consistent with the results from the current EAS analyses and our previous OCTO-Twin models, the updated OCTO-Twin models revealed significant increases in neuroticism and stability in extraversion preceding dementia diagnosis, while controlling for all covariates on the slope. Although the results from the updated OCTO-Twin analyses are not presented here, the information is available by request. Assessments of extraversion, conscientiousness, openness, and agreeableness were also available in EAS. Our analyses revealed significant decreases in extraversion only, and solely for individuals with MCI. These results may indicate that individuals with MCI might feel more cognitively challenged in the presence of others, possibly leading to avoidance of social activity. Although consistent with expectations, these results do not explicitly support our second hypothesis that a decrease in extraversion will precede dementia because the finding was seen only in individuals with MCI. The inability to detect significant decreases in extraversion in the individuals with dementia may be accounted for by a variety of reasons, including limited power due to small sample size, or operation of a different process for the individuals with MCI versus those eventually diagnosed with dementia. A further possibility is that individuals with MCI have more insight into the social changes that are occurring. Although self-report does not typically differ from informant-report for individuals with MCI compared with controls (Farias, Mungas, & Jagust, 2005; Ready, Ott, & Grace, 2004), future research including both assessment types may provide unique information about personality changes in individuals with MCI. Finally, our analyses did not support the hypothesis of a decrease in conscientiousness, openness, or agreeableness. The hypotheses were based on retrospective research; thus, several factors regarding methodology could be responsible for the discrepancy between our expectations and findings (see Yoneda et al., 2017 for a more detailed account). Our findings regarding neuroticism increases prior to diagnosis of dementia are consistent with expectations based on neurological research (Mahoney et al., 2011). Although Mahoney and associates (2011) found increases in neuroticism related to neuropathology, their results also indicated that neuropathology was related to decreases in the four personality traits. The discrepancy between our findings and Mahoney and associates’ findings may be due to more substantial personality changes occurring in individuals with FTD, recollection bias, or the timing of personality measurement. Namely, their research assessed changes after diagnosis of FTD while our focus was on change preceding diagnosis; thus, changes in these traits may not yet be measureable, or the brain regions responsible for extraversion, conscientiousness, openness, and agreeableness may not be implicated until the later stages of dementia. Our findings are also consistent with research examining neural correlates of personality in healthy individuals, which finds an association between higher levels of neuroticism and smaller brain volume or reduced cortical thickness (Jackson, Balota, & Head, 2011; Knutson et al., 2001; Wright et al., 2007; Xu & Potenza, 2012). These personality-related neurological findings, in combination with neurological characteristics of dementia such as heightened levels of brain lesions, decreased neuroreactivity, and diminished grey and white matter (Debette & Markus, 2010; Jagust et al., 2008), provide a theoretical context for why increases in neuroticism may be associated with neurodegeneration. However, the comparison is complicated, as our research examines within-person changes and does not include neurological assessments, while the majority of neurology research examines between-person differences. Future research investigating endorsement of neuroticism and intracranial fluid at repeated occasions, which allows evaluation of the amount of brain degeneration since maximum lifetime brain volume, could improve our understanding of personality change, neurology, and dementia. Attrition is a concern in any longitudinal study of aging; however, the majority of missing data across both datasets is due to mortality rather than refusal to participate, which reduces potential selection bias, though could be affected by factors related to mortality. Furthermore, our analyses were limited by the number of individuals who received a dementia or MCI diagnosis. For example, our analyses included individuals with aMCI and naMCI as one group to increase sample size, but trajectories of personality may differ depending on MCI subtype and, importantly, whether individuals eventually convert to dementia. Likewise, although previous research indicates similar personality changes at the domain level after diagnosis of dementia, trajectories of personality in the years leading up to diagnosis may differ depending on the type of dementia. Our analyses could not evaluate whether changes occurred in differing directions for certain personality traits depending on the type of dementia. Future research based on larger samples could extend this conceptual replication by examining trajectories of personality traits according to specific types of dementia or MCI also using LGM. Our syntax is available by request, or we would happily analyze any additional datasets that include assessment of the required constructs. Future research could also apply alternative models to explore a variety of more detailed research questions. Bivariate LGM could examine the association between change in both personality traits and domains of cognitive functioning. Growth mixture modeling (Nylund, Asparouhov, & Muthén, 2007) could examine the possible existence of subpopulations based on change in personality traits. Multistate modeling (Andersen, 1993) could examine change in personality traits corresponding to a series of states such as MCI or dementia diagnosis. Joint growth survival modeling (Tsiatsis & Davidian, 2004) could examine dementia diagnosis or survival as distal outcomes based on change in certain personality traits. A further limitation of this study is the relatively small increases in neuroticism in both datasets. Although these increases are statistically significant, one could question the clinical significance of neuroticism increase, and it may be difficult to detect at the level of the individual patient. However, the magnitude of quantifiable changes is constrained by the properties of the available personality assessments; that is, answers are limited by the nature of the response options. For both datasets, the items used to measure personality traits were relatively limited in quantity; therefore, any measureable change seems potentially meaningful, particularly given the strikingly similar results across datasets. Moreover, significant changes were not detected in the analyses examining individuals who did not receive a diagnosis, which, given the substantially larger samples, would have provided more power to detect significant change. Thus, assessments of personality traits that capture more detail and are more sensitive to variation would enhance this study. For example, The NEO-Personality Inventory Revised (NEO PI-R; Costa & McCrae, 1992) assesses each trait using 40 items answered on a 5-point scale. Future research using the neuroticism scale from the NEO PI-R to assess personality at more frequently administered repeated occasions could improve our understanding of personality change and progression to dementia. Many of this study’s limitations are characteristic of any comparison across independent studies because the study designs are not identical. Although each study included repeated measurement of neuroticism, the measures differed across datasets, and assessments for all five personality traits were available only in EAS. Assessment of depressive symptoms also differed between datasets, with LASA administering the CES-D (Radloff, 1997) and EAS administering the GDS (Yesavage & Sheikh, 1986). In addition, classification of dementia varied across datasets. For EAS, diagnoses were made based on the DSM-IV. Although research comparing different versions of the DSM criteria has found slight differences in rate of diagnosis (Erikinjuntti et al., 1997; Pohjasvaara et al., 1997), there is more similarity in classification using DSM indexes compared with other diagnostic tools (e.g., the Cambridge Mental Disorders in the Elderly), strengthening the replicability of the EAS analyses to those of the OCTO-Twin analyses. The LASA created a “probable dementia” category based on declines on the MMSE or IQCODE, rather than a formal diagnosis, which may have resulted in differential dementia diagnoses. The differences in diagnostic criteria could have affected our results because individuals who may have been classified with dementia in one sample may not have been classified in another sample, and if an individual was not classified with dementia, they were included in the analyses examining individuals who did not receive a dementia diagnosis. Indeed, the studies that utilized similar diagnosis criteria (EAS and OCTO-Twin) also had low slope variance, indicating that trajectories of personality traits may be more similar at the interindividual level for individuals classified according to more formal diagnosis criteria. Finally, the intervals between measurement occasions and available number of occasions also varied between datasets, so quadratic models could not be fitted without additional constraints. Yet, despite the differences in designs, assessments, and diagnostic criteria, considerable similarities were found in the results for both datasets, as well as our previous work using OCTO-Twin data. Heterogeneity in the key features of datasets reinforces significant findings in a coordinated analysis (Hofer & Piccinin, 2009; Lindwall et al., 2012); namely, evidence for generalizability of the findings, based on a consistent pattern of personality change across studies with such different characteristics, is more substantial. The primary strength of this study is the ability to investigate within-person change in personality using data from multiple longitudinal studies of aging. The included studies are heterogeneous in several features including age and cultural background, which offers conceptual rather than identical replication. Despite the differences across datasets, the results revealed a clear pattern of personality trait change, specifically increases in neuroticism, in individuals eventually diagnosed with dementia. The MCI analysis was also consistent with this pattern, and to our knowledge, this was the first study to examine trajectories of personality traits in individuals with incident MCI diagnosis. Overall, this study contributes consistent, replicated findings to the existing literature. A conceptual replication applying a coordinated analysis approach with consistent findings across diverse datasets offers compelling evidence that an increase in neuroticism may indeed be an early indicator of dementia. Supplementary Material Supplementary material are available at The Journals of Gerontology Series B: Psychological and Social Sciences online. Funding This work was supported by the National Institute on Aging (NIA) at the National Institutes of Health (NIH; grant number P01AG043362; 2013–2018) in support of the Integrative Analysis of Longitudinal Studies of Aging (IALSA) research network. Origins of Variance in the Oldest-Old (OCTO-Twin) data collection was funded by the National Institute on Aging at the National Institutes of Health (grant number AG08861); the Swedish Council for Working Life and Social Research; the Adlerbertska Foundation; the Wenner-Gren Foundations; and the Wilhelm and Martina Lundgrens Foundation. The Longitudinal Aging Study Amsterdam (LASA) is largely supported by the Netherlands Ministry of Health, Welfare and Sports, Directorate of Long-Term Care. Einstein Aging Study (EAS) data collection was supported by the National Institutes of Health NIA (National Institute on Aging) Grant AG03949; the Sylvia & Leonard Marx Foundation, and; the Czap Foundation. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official view of the NIH. Conflict of Interest None reported. References Allemand, M., Zimprich, D., & Hertzog, C. ( 2007). Cross-sectional age differences and longitudinal age changes of personality in middle adulthood and old age. Journal of Personality , 75, 323– 358. doi: 10.1111/j.1467-6494.2006.00441.x Google Scholar CrossRef Search ADS   American Psychiatric Association. ( 1987). Diagnostic and statistical manual of mental disorders  ( 3rd ed.). Washington, DC: Author. PubMed PubMed  American Psychiatric Association. ( 1994). Diagnostic and statistical manual of mental disorders  ( 4th ed.). Washington, DC: Author. PubMed PubMed  Andersen, P. K. ( 1993). Statistical models based on counting processes . New York: Springer-Verlag. Google Scholar CrossRef Search ADS   Balsis, S., Carpenter, B. D., & Storandt, M. ( 2005). Personality change precedes clinical diagnosis of dementia of the Alzheimer Type. Journal of Gerontology: Psychological Sciences , 60B, 98– 101. doi: 10.1093/geronb/60.2.P98 Google Scholar CrossRef Search ADS   Berg, A. I., & Johansson, B. ( 2013). Personality change in the oldest-old: Is it a matter of compromised health and functioning? Journal of Personality , 82, 25– 31. doi: 10.1111/jopy.12030 Google Scholar CrossRef Search ADS   Bruscoli, M., & Lovestone, S. ( 2004). Is MCI really just early dementia? A systematic review of conversion studies. International Psychogeriatrics , 16, 129– 140. doi: 10.1017/S1041610204000092 Google Scholar CrossRef Search ADS   Caselli, R., Beach, T., Knopman, D., & Graff-Radford, N. ( 2017). Alzheimer disease: Scientific breakthroughs and translational challenges. Mayo Clinic Proceedings , 92, 978– 994. doi: 10.1016/j.mayocp.2017.02.011 Google Scholar CrossRef Search ADS   Costa, P. T., & McCrae, R. R. ( 1992). Neo Pi-R. The revised NEO personality inventory . Odessa, FL: Psychological Assessment Resources. Dawson, D. V., Welsh-Bohmer, K. A., & Siegler, I. C. ( 2000). Premorbid personality predicts level of rated personality change in patients with Alzheimer disease. Alzheimer Disease and Associated Disorders , 14, 11– 19. doi: 10.1097/ 00002093-200001000-00002 Google Scholar CrossRef Search ADS   Debette, S., & Markus, H. S. ( 2010). The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: Systematic review and meta-analysis. British Medical Journal , 341, 341– 350. http://dx.doi.org/10.1136/bmj.c3666 Google Scholar CrossRef Search ADS   Di Marco, L. Y., Marzo, A., Muñoz-Ruiz, M., Ikram, M. A., Kivipelto, M., Ruefenacht, D.,… Frangi, A. F. ( 2014). Modifiable lifestyle factors in dementia: A systematic review of longitudinal observational cohort studies. Journal of Alzheimer’s Disease: JAD , 42, 119– 135. doi: 10.3233/JAD-132225 Erkinjuntti, T., Ostbye, T., Steenhuis, R., & Hachinski, V. ( 1997). The effect of different diagnostic criteria on the prevalence of dementia. The New England Journal of Medicine , 337, 1667– 1674. doi: 10.1056/NEJM199712043372306 Google Scholar CrossRef Search ADS   Eysenck, H. J., & Eysenck, S. G. G. ( 1964). Manual of the Eysenck personality inventory . London: University of London Press. Farias, S. T., Mungas, D., & Jagust, W. ( 2005). Degree of discrepancy between self and other-reported everyday functioning by cognitive status: Dementia, mild cognitive impairment, and healthy elders. International Journal of Geriatric Psychiatry , 20, 827– 834. doi: 10.1002/gps.1367 Google Scholar CrossRef Search ADS   Floderus-Myrhed, B., Pedersen, N., & Rasmuson, I. ( 1980). Assessment of heritability for personality, based on a short-form of the Eysenck Personality Inventory: A study of 12,898 twin pairs. Behavior Genetics , 10, 153– 162. doi: 10.1007/BF01066265 Google Scholar CrossRef Search ADS   Goldberg, L. R. ( 1999). A broad-bandwidth, public-domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe , Vol. 7 (pp. 7– 28). Tilburg, The Netherlands: Tilburg University Press. Graham, E. K., Rutsohn, J. P., Turiano, N. A., Bendayan, R., Batterham, P. J., Gerstorf, D.,… Mroczek, D. K. ( 2017). Personality predicts mortality risk: An integrative data analysis of 15 international longitudinal studies. Journal of Research in Personality , 70, 174– 186. doi: 10.1016/j.jrp.2017.07.005 Google Scholar CrossRef Search ADS   Hall, C. B., Lipton, R. B., Sliwinski, M., & Stewart, W. F. ( 2000). A change point model for estimating the onset of cognitive decline in preclinical Alzheimer’s disease. Statistics in Medicine , 19, 1555– 1566. doi: 10.1002/(SICI)1097-0258(20000615/30)19:11/12<1555::AID-SIM445>3.0.CO;2–3 Google Scholar CrossRef Search ADS   Hofer, S. M., & Piccinin, A. M. ( 2009). Integrative data analysis through coordination of measurement and analysis protocol across independent longitudinal studies. Psychological Methods , 14, 150– 164. doi: 10.1037/a0015566 Google Scholar CrossRef Search ADS   Hofer, S. M., & Piccinin, A. M. ( 2010). Toward an integrative science of life-span development and aging. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences , 65B, 269– 278. doi: 10.1093/geronb/gbq017 Google Scholar CrossRef Search ADS   Huisman, M., Poppelaars, J., van der Horst, M., Beekman, A. T., Brug, J., van Tilburg, T. G., & Deeg, D. J. ( 2011). Cohort profile: The longitudinal aging study Amsterdam. International Journal of Epidemiology , 40, 868– 876. doi: 10.1093/ije/dyq219 Google Scholar CrossRef Search ADS   Jackson, J., Balota, D. A., & Head, D. ( 2011). Exploring the relationship between personality and regional brain volume in healthy aging. Neurobiology of Aging , 32, 2162– 2171. doi: 10.1016/j.neurobiolaging.2009.12.009 Google Scholar CrossRef Search ADS   Jagust, W. J., Zheng, L., Harvey, D. J., Mack, W. J., Vinters, H. V., Weiner, M. W.,… Chui, H. C. ( 2008). Neuropathological basis of magnetic resonance images in aging and dementia. Annals of Neurology , 63, 72– 80. doi: 10.1002/ana.21296 Google Scholar CrossRef Search ADS   Katz, M. J., Lipton, R. B., Hall, C. B., Zimmerman, M. E., Sanders, A. E., Verghese, J.,… Derby, C. A. ( 2012). Age-specific and sex-specific prevalence and incidence of mild cognitive impairment, dementia, and Alzheimer dementia in blacks and whites: A report from the Einstein Aging Study. Alzheimer Disease and Associated Disorders , 26, 335– 343. doi: 10.1097/WAD.0b013e31823dbcfc Google Scholar CrossRef Search ADS   Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. ( 1993). A longitudinal twin study of personality and major depression in women. Archives of General Psychiatry , 50, 853– 862. doi: 10.1001/archpsyc.1993.01820230023002 Google Scholar CrossRef Search ADS   Knutson, B., Momenan, R., Rawlings, R. R., Fong, G. W., & Hommer, D. ( 2001). Negative association of neuroticism with brain volume ratio in healthy humans. Biological Psychiatry , 50, 685– 690. doi: 10.1016/S0006-3223(01)01220-3 Google Scholar CrossRef Search ADS   van den Kommer, T. N., Comijs, H. C., Dik, M. G., Jonker, C., & Deeg, D. J. ( 2008). Development of classification models for early identification of persons at risk for persistent cognitive decline. Journal of Neurology , 255, 1486– 1494. doi: 10.1007/s00415-008-0942-3 Google Scholar CrossRef Search ADS   Lindwall, M., Cimino, C. R., Gibbons, L. E., Mitchell, M. B., Benitez, A., Brown, C. L.,… Piccinin, A. M. ( 2012). Dynamic associations of change in physical activity and change in cognitive function: Coordinated analyses of four longitudinal studies. Journal of Aging Research , 2012, 493598. doi: 10.1155/2012/493598 Google Scholar CrossRef Search ADS   Luteijn, F., Starren, J., & Van Dijk, H. ( 2000). NPV; Nederlandse Persoonlijkheids Vragenlijst, handleiding (herziene uitgave)  . [DPQ; Dutch Personality Questionnaire, manual (revised edition)]. Lisse: Swets & Zeitlinger. Lykou, E., Rankin, K. P., Chatziantoniou, L., Boulas, C., Papatriantafyllou, O., Tsaousis, I.,… Papatriantafyllou, J. D. ( 2013). Big 5 personality changes in Greek bvFTD, AD, and MCI patients. Alzheimer Disease and Associated Disorders , 27, 258– 264. doi: 10.1097/WAD.0b013e31826e5504 Google Scholar CrossRef Search ADS   Mahoney, C. J., Rohrer, J. D., Omar, R., Rossor, M. N., & Warren, J. D. ( 2011). Neuroanatomical profiles of personality change in frontotemporal lobar degeneration. The British Journal of Psychiatry: The Journal of Mental Science , 198, 365– 372. doi: 10.1192/bjp.bp.110.082677 Google Scholar CrossRef Search ADS   McClearn, G. E., Johansson, B., Berg, S., Pedersen, N. L., Ahern, F., Petrill, S. A., & Plomin, R. ( 1997). Substantial genetic influence on cognitive abilities in twins 80 or more years old. Science (New York, N.Y.) , 276, 1560– 1563. doi: 10.1126/science.276.5318.1560 Google Scholar CrossRef Search ADS   McCrae, R. R., & Costa, P. T. ( 1985). Comparison of EPI and psychoticism scales with measures of the five-factor model of personality. Personality and Individual Differences , 6, 587– 597. doi: 10.1016/0191-8869(85)90008-X Google Scholar CrossRef Search ADS   Middleton, L. E., & Yaffe, K. ( 2010). Targets for the prevention of dementia. Journal of Alzheimer’s Disease: JAD , 20, 915– 924. doi: 10.3233/JAD-2010-091657 Google Scholar CrossRef Search ADS   Mroczek, D. K., & Spiro, A. ( 2003). Modeling intraindividual change in personality traits: Findings from the normative aging study. Journal of Gerontology , 58B, 153– 165. doi: 10.1093/geronb/58.3.P153 Google Scholar CrossRef Search ADS   Muthén, L. K., & Muthén, B. O. ( 1998–2013). Mplus user’s guide  ( 6th ed.). Los Angeles, CA: Muthén & Muthén. Nussbaum, P. D. ( 1997). Late-life depression. In P. D. Nussbaum (ed.), Handbook of Neuropsychology and Aging. Critical Issues in Neuropsychology . Boston, MA: Springer. Google Scholar CrossRef Search ADS   Nylund, K. L., Asparouhov, T., & Muthén, B. O. ( 2007 ). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal , 14, 535– 569. doi: 10.1080/10705510701575396 Google Scholar CrossRef Search ADS   Petersen, R. C., Caracciolo, B., Brayne, C., Gauthier, S., Jelic, V., & Fratiglioni, L. ( 2014). Mild cognitive impairment: A concept in evolution. Journal of Internal Medicine , 275, 214– 228. doi: 10.1111/joim.12190 Google Scholar CrossRef Search ADS   Pohjasvaara, T., Erkinjuntti, T., Vataja, R., & Kaste, M. ( 1997). Dementia three months after stroke. Baseline frequency and effect of different definitions of dementia in the Helsinki Stroke Aging Memory Study (SAM) cohort. Stroke , 28, 785– 792. doi: 10.1161/01.STR.28.4.785 Google Scholar CrossRef Search ADS   Polidori, M. C., Nelles, G., & Pientka, L. ( 2010). Prevention of dementia: Focus on lifestyle. International Journal of Alzheimer’s Disease , 2010, 1– 9. doi: 10.4061/2010/393579 Google Scholar CrossRef Search ADS   Radloff, L. S. 1977. The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement , 1, 385– 401. doi: 10.1177/014662167700100306 Google Scholar CrossRef Search ADS   Rankin, K. P., Kramer, J. H., Mychack, P., & Miller, B. L. ( 2003). Double dissociation of social functioning in frontotemporal dementia. Neurology , 60, 266– 271. doi: 10.1212/ 01.WNL.0000041497.07694.D2 Google Scholar CrossRef Search ADS   Ready, R. E., Ott, B. R., & Grace, J. ( 2004). Patient versus informant perspectives of Quality of Life in Mild Cognitive Impairment and Alzheimer’s disease. International Journal of Geriatric Psychiatry , 19, 256– 265. doi: 10.1002/gps.1075 Google Scholar CrossRef Search ADS   Roberts, R. O., Knopman, D. S., Mielke, M. M., Cha, R. H., Pankratz, V. S., Christianson, T. J. H., … Petersen, R. C. ( 2014). Higher risk of progression to dementia in mild cognitive impairment cases who revert to normal. Neurology , 82, 317– 325. doi: 10.1212/WNL.0000000000000055 Google Scholar CrossRef Search ADS   Schumacker, R. E., & Lomax, R. G. ( 2015). A beginner’s guide to structural equation modeling  ( 4th ed.). New York: Routledge. Small, B. J., Hertzog, C., Hultsch, D. F., & Dixon, R. A. ( 2003). Stability and change in adult personality over 6 years: Findings from the Victoria Longitudinal Study. Journal of Gerontology: Psychological Sciences , 58B, 166– 176. doi: 10.1093/geronb/58.3.P166 Google Scholar CrossRef Search ADS   Smith-Gamble, V., Baiyewu, O., Perkins, A. J., Gureje, O., Hall, K. S., Ogunniyi, A.,… Hendrie, H. C. ( 2002). Informant reports of changes in personality predict dementia in a population-based study of elderly African Americans and Yoruba. The American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry , 10, 724– 732. doi: 10.1097/00019442-200211000-00011 Google Scholar CrossRef Search ADS   Torrente, F., Pose, M., Gleichgerrcht, E., Torralva, T., López, P., Cetkovich-Bakmas, M., & Manes, F. ( 2014). Personality changes in dementia: Are they disease specific and universal? Alzheimer Disease and Associated Disorders , 28, 261– 268. doi: 10.1097/WAD.0000000000000030 Google Scholar CrossRef Search ADS   Tsiatis, A. A., & Davidian, M. ( 2004). Joint modeling of longitudinal and time-to-event data: An overview. Statistica Sinica , 14, 809– 834. Retrieved from http://www.jstor.org/stable/24307417 Wright, C. I., Feczko, E., Dickerson, B., & Williams, D. ( 2007). Neuroanatomical correlates of personality in the elderly. Neuroimage , 35, 263– 272. doi: 10.1016/ j.neuroimage.2006.11.039 Google Scholar CrossRef Search ADS   Xu, J., & Potenza, M. N. ( 2012). White matter integrity and five-factor personality measures in healthy adults. Neuroimage , 59, 800– 807. doi: 10.1016/j.neuroimage.2011.07.040 Google Scholar CrossRef Search ADS   Yesavage, J. A., & Sheikh, J. I. ( 1986). Geriatric depression scale (GDS): Recent evidence and development of a shorter version. Clinical Gerontologist , 5, 165– 173. doi: 10.1300/J018v05n01_09 Google Scholar CrossRef Search ADS   Yoneda, T., Rush, J., Berg, A. I., Johansson, B., & Piccinin, A. M. ( 2017). Trajectories of personality traits preceding dementia diagnosis. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences , 72, 922– 931. doi: 10.1093/geronb/gbw006 © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journals of Gerontology Series B: Psychological Sciences and Social Sciences Oxford University Press

Loading next page...
 
/lp/ou_press/increases-in-neuroticism-may-be-an-early-indicator-of-dementia-a-Ei4805ahWR
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1079-5014
eISSN
1758-5368
D.O.I.
10.1093/geronb/gby034
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives Although personality change is typically considered a symptom of dementia, some studies suggest that personality change may be an early indication of dementia. One prospective study found increases in neuroticism preceding dementia diagnosis (Yoneda, T., Rush, J., Berg, A. I., Johansson, B., & Piccinin, A. M. (2017). Trajectories of personality traits preceding dementia diagnosis. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 72, 922–931. doi:10.1093/geronb/gbw006). This study extends this research by examining trajectories of personality traits in additional longitudinal studies of aging. Methods Three independent series of latent growth curve models were fitted to data from the Longitudinal Aging Study Amsterdam and Einstein Aging Study to estimate trajectories of personality traits in individuals with incident dementia diagnosis (total N = 210), in individuals with incident Mild Cognitive Impairment (N = 135), and in individuals who did not receive a diagnosis during follow-up periods (total N = 1740). Results Controlling for sex, age, education, depressive symptoms, and the interaction between age and education, growth curve analyses consistently revealed significant linear increases in neuroticism preceding dementia diagnosis in both datasets and in individuals with mild cognitive impairment. Analyses examining individuals without a diagnosis revealed nonsignificant change in neuroticism overtime. Discussion Replication of our previous work in 2 additional datasets provides compelling evidence that increases in neuroticism may be early indication of dementia, which can facilitate development of screening assessments. Longitudinal change, Mild cognitive impairment, Multi-study conceptual replication, Personality Although research examining personality in healthy older adults indicates minimal variation in personality traits over time (Allemand, Zimprich, & Hertzog, 2007; Berg & Johansson, 2013; Mroczek & Spiro, 2003; Small, Hertzog, Hultsch, & Dixon, 2003), several studies report substantial personality change in individuals who have been diagnosed with dementia. Studies using retrospective informant-reports of personality have found personality changes in individuals diagnosed with dementia, including increases in neuroticism (Dawson, Welsh-Bohmer, & Siegler, 2000; Torrente et al., 2014), decreases in extraversion (Dawson et al., 2000; Rankin, Kramer, Mychack, & Miller, 2003; Torrente et al., 2014) and decreases in conscientiousness (Dawson et al., 2000; Torrente et al., 2014) relative to premorbid personality traits. The connection between personality change and dementia at the behavioral level is consistent with neurological research utilizing sophisticated neuroimaging techniques; that is, previous work suggests a relation between the neurological underpinnings of personality and dementia-related neuropathology. Research by Mahoney and associates (2011) used volumetric brain MRI to examine the neuroanatomical profiles of personality change in individuals diagnosed with frontotemporal lobar degeneration (FTD) and healthy controls (Mahoney, Rohrer, Omar, Rossor, & Warren, 2011). Personality change was assessed by informant-report of current personality relative to estimated premorbid personality for the FTD group, and by informant-report of current personality relative to estimated personality 10 years previously for the control group. The individuals with FTD showed significant declines in extraversion, agreeableness, openness, and conscientiousness, and increases in neuroticism, compared with trait stability for the control group. MRI results revealed that trait changes were associated with cortical losses in regional gray matter in overlapping anterior and dorsal cortical areas, emphasizing that personality change in FTD is associated with brain damage. Mahoney and associates (2011) concluded that by examining individuals with FTD neuropathology, their research provided an opportunity to identify the neural substrates that are critical for maintenance of personality trait stability. Although Mahoney’s study focused on FTD, other types of dementia neuropathology may also affect the biological mechanisms and structures that support personality, particularly because similar patterns of personality change are seen in individuals with Alzheimer’s disease (AD), FTD, and temporal-variant FTD (Rankin et al., 2003), as well as generalized pattern of personality change in individuals with AD and behavioral-variant of FTD (Torrente et al., 2014). Given that dementia neuropathology and dementia-related cognitive decline develop many years prior to the clinical diagnosis of dementia (Caselli, Beach, Knopman, & Graff-Radford, 2017; Hall, Lipton, Sliwinski, & Stewart, 2000), dementia neuropathology may also affect aspects of personality in the years preceding dementia diagnosis. Indeed, a small number of studies suggest that personality changes may precede dementia diagnosis. Lykou and associates (2013) found that increases in neuroticism occurred in individuals with mild cognitive impairment (MCI) as well as those with AD. Although this research applied a cross-sectional retrospective research design, the results indicate that a similar pattern of personality change may occur as early as the MCI disease stage. Two studies have prospectively examined the timing of informant-rated personality change, operationalized as an affirmative answer to any item regarding specific change in personality (Balsis, Carpenter, & Storandt, 2005; Smith-Gamble et al., 2002). Both studies found that individuals with informant-rated personality change at baseline were significantly more likely to be diagnosed with dementia at follow-up occasions. More recently, a study by Yoneda, Rush, Berg, Johansson, and Piccinin (2017) prospectively investigated trajectories of self-reported personality traits using latent growth curve modeling, rather than general personality change using chi square tests. Based on intraindividual trajectories of personality traits in individuals who received a dementia diagnosis after the first measurement occasion, analyses revealed significant increases in neuroticism and no change in extraversion preceding dementia diagnosis. In contrast, personality trait stability over time was found in the individuals who were not diagnosed, suggesting that trajectories of personality are different for individuals who did and did not receive a dementia diagnosis during the observation period. Conceptual replication of this work in additional longitudinal studies of aging would offer additional evidence that personality change is possibly an early indication of dementia, which may be valuable in a clinical setting. Identification of the early signs of dementia can aid in implementing early treatment strategies, planning dementia care services, and facilitating development of screening assessments. Most importantly, although there is not yet a cure for dementia, mounting evidence suggests that progression to dementia may be slowed by adherence to a healthy lifestyle (Di Marco et al., 2014; Middleton & Kristine, 2010; Polidori, Nelles, & Pientka, 2010). Therefore, the sooner that dementia can be identified, the sooner individuals can be educated and, ideally, motivated to commit to a healthier lifestyle. The current work extends Yoneda and colleagues’ research by investigating personality change in two additional affiliates of the Integrative Analysis of Longitudinal Studies on Aging and Dementia (IALSA) network: The Longitudinal Aging Study of Amsterdam (LASA; Huisman et al., 2011) and Einstein Aging Study (EAS; Katz et al., 2012). IALSA was established to enable and encourage replication of research using comparable methods and measures across longitudinal studies of aging and dementia. Further, a coordinated analysis provides the opportunity for immediate conceptual replication and comparison of the results, which protects against Type I errors, allows for evaluation of consistencies and differences between samples, and is important for effective cumulative science (Graham et al., 2017; Hofer & Piccinin, 2009; Hofer & Piccinin, 2010). LASA and EAS were selected based on inclusion of both repeated assessment of personality traits and an indication of definite or probable dementia diagnosis. Our study includes individuals with all types of dementia to increase power within our analyses and because previous research indicates a generalized pattern of personality trait changes in individuals with dementia (Rankin et al., 2003; Torrente et al., 2014). The EAS also provided data for trajectories of personality traits in individuals who were diagnosed with MCI at some point during the study. To our knowledge, no research has examined trajectories of personality traits in individuals with MCI. The rationale for including these analyses is based on robust evidence suggesting that the cognitive impairment observed in MCI represents an early stage of dementia (Petersen et al., 2014). A systematic review including 19 longitudinal studies investigating conversion from MCI to dementia concluded that mild cognitive decline is not a normal part of aging and significantly predicts conversion to dementia (Bruscoli & Lovestone, 2004). Furthermore, Roberts and associates (2014) found that individuals with prevalent or incident MCI (N = 534) have a high risk of progressing to dementia, even when they reverted to normal cognition at some point during the longitudinal study. Therefore, our analyses examining trajectories of personality in individuals with MCI may provide further information regarding personality change in the earliest stages of the disease process, though it is important to note that not all individuals classified with MCI will convert to dementia. The research examining personality change preceding (Yoneda et al., 2017) and following (Dawson et al., 2000; Lykou et al., 2013; Mahoney et al., 2011; Rankin et al., 2003; Torrente et al., 2014) diagnosis of dementia informs our hypotheses: that (a) increases in neuroticism and (b) decreases in extraversion, conscientiousness, agreeableness, and openness will precede diagnosis of dementia. Furthermore, based on retrospective research examining individuals with incident MCI and AD (Lykou et al., 2013), we expect a similar pattern of personality change in individuals who are classified as having MCI. Method Participants and Procedures For both datasets, individuals were not included in the analyses if they were diagnosed with dementia at baseline. Three independent series of latent growth curve models were estimated to examine (a) individuals diagnosed with incident dementia, (b) individuals diagnosed with incident MCI (in EAS), and (c) individuals not diagnosed with dementia (nor MCI, in EAS). For the first set of analyses examining individuals diagnosed with dementia, personality data were excluded after the occasion in which participants were diagnosed based on the following rationales: our primary interest was personality changes preceding diagnosis; self-report following diagnosis has questionable accuracy; personality was generally no longer assessed following diagnosis. For the second set of analyses examining individuals with MCI in the EAS dataset, all available personality data from these individuals were included, resulting in examination of trajectories preceding and following MCI classification. In order to ensure that data were only used to represent one of either MCI or dementia, individuals who were eventually diagnosed with dementia during the study were excluded from the MCI analysis (and vice versa). The third set of analyses examined trajectories of personality traits for individuals who had repeated measurement of personality and did not receive a dementia diagnosis, and, in the case of the EAS, also did not receive diagnosis of MCI. Longitudinal Aging Study Amsterdam participants (LASA; Huisman et al., 2011; Amsterdam, Zwolle and Oss, the Netherlands). The LASA is a population-based cross-sequential longitudinal study consisting of three independent and geographically representative cohorts of older adults, aged 55–85 years. Participants were recruited in 1992–1993 and tested every 3 years following recruitment into the study. Participants were visited in their homes by trained interviewers, who completed a main interview and a medical interview at each occasion. Participants were asked to complete and mail back several questionnaires, such as the Neuroticism scale. Data from individuals with incident probable dementia classification (N = 162) and individuals not diagnosed with dementia during the study (N = 1138) were used in the current analysis. Einstein Aging Study participants (EAS; Katz et al., 2012; New York, United States). The primary objective of the EAS is to study the aging brain and healthy aging, and the risk factors associated with AD. The study began in 1993, and the sample consists of community dwelling adults aged 70 and older who were systematically recruited from Bronx County, New York. After study entry, participants contributed data annually by undergoing physical, cognitive, and psychosocial assessments at a clinical research center (Katz et al., 2012). Due to a limited subsample of individuals classified as having dementia (N = 48), only three waves of personality data were included in the analyses examining the years preceding dementia diagnosis. Additional analyses include five waves of data examining individuals diagnosed with MCI (N = 135) and individuals not diagnosed with either MCI or dementia during the study (N = 602). Measures Each study involved a large battery of measures including assessment of health, functional and mental capacity, well-being, personality, depressive symptoms, and social network. A subsample of these assessments was used in the current analyses. Due to heterogeneity across samples, personality measures and depressive symptoms were not standardized, as standardized scores may be sample-specific. Although the personality and depression scales are not equivalent across samples, unstandardized scores permit investigation of measures of change between different samples (Schumacker & Lomax, 2015). Personality measures Each of the studies assessed personality using different self-report measurements. Although the assessments are not identical, different measurement of the same trait are highly correlated (Luteijn, Starren, & Van Dijk, 2000; McCrae & Costa, 1985), allowing for comparison of the same construct between studies. The LASA used the Neuroticism scale from the Dutch Personality Questionnaire (DPQ; Luteijn et al., 2000), which contains 15 items with three response options: applicable, do not know, and not applicable. Scores on the Neuroticism scale can range from 0 to 30, with higher scores indicating a higher degree of neuroticism. Good internal consistency reliability is reported (α = .85). The EAS assessed the Big Five personality traits using items from the International Personality Item Pool (IPIP; Goldberg, 1999). Each trait scale contains 10 items with five response options ranging from 1 (very inaccurate) to 5 (very accurate). A mean score was computed for each trait scale; thus, scores can range from 1 to 5, with higher scores indicating higher endorsement of that trait. Good internal consistency reliability estimates are reported: neuroticism (α = .76), extraversion (α = .76), agreeableness (α = .72), contentiousness (α = .81), and openness (α = .72). Covariates Diagnosis of dementia and MCI Individuals were included in the analyses examining trajectories of personality traits aligned according to dementia and MCI diagnosis based on their corresponding diagnosis. The LASA did not include a formal dementia diagnosis, but rather a composed variable of probable dementia, determined by taking into account the Mini Mental State Examination (MMSE) score, a telephone administered MMSE, or an Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) (Van den Kommer et al., 2008). Specifically, individuals who scored two or more standard deviations below the mean decline on the MMSE since the last occasion of measurement or at least 28 points on the IQCODE were classified as having probable dementia. Further, continued decline at the next measurement occasion was required. Finally, several factors were considered to obtain information regarding probable diagnosis including reason for participant dropout or cause of death as well as pertinent data from General Practitioners, specialists, and nursing home admittance. The date of probable dementia classification corresponds to the measurement occasion in which participants met the above requirements, only if those participants also demonstrated continued decline at the next wave. Incident MCI diagnosis was not available in LASA. The EAS included a clinical conference in which dementia diagnosis was based on standardized clinical criteria from the DSM-IV (American Psychiatric Association, 1994). Criteria were deliberated according to a comprehensive review of cognitive, functional and neurological assessments. Furthermore, the EAS included criteria to classify amnestic MCI (aMCI) and non-amnestic MCI (naMCI) for participants who were not diagnosed with dementia, but did demonstrate cognitive impairment and had subjective complaints of decline. Specifically, individuals with aMCI demonstrated objective and subjective memory impairment, while individuals with naMCI did not meet memory criteria for aMCI, but scored at least 1.5 SDs below the mean in at least one of the following domains: executive function, language, visuospatial ability, or attention (Katz et al., 2012). The date of dementia or MCI diagnosis was set when the diagnostic team assembled shortly after the in-person examinations. Depressive symptoms Depressive symptoms measured at the first occasion were included as a covariate in this study. Assessment at baseline was used to minimize missingness. Controlling for depression is important because both dementia and neuroticism have been associated with higher levels of depressive symptoms (Nussbaum, 1997; and Kendler, Neale, Kessler, Heath, & Eaves, 1993, respectively). For LASA, depressive symptoms were measured using the self-report Center for Epidemiologic Studies Depressive Scale (CES-D; Radloff, 1977), translated into Dutch. The scale includes 20 items rated on a 4-point scale ranging from 0 (“rarely or none of the time”) to 3 (“most or all of the time”). Participants are asked about the frequency of certain experiences during the past week. Four items are reverse coded prior to summing scores. Total scores range from 0 to 60, with higher scores indicating more depressive symptoms. Good internal consistency reliabilities are reported (α = .87–.90). For EAS, depressive symptoms were measured using the self-report Geriatric Depression Scale (GDS; Yesavage & Sheikh, 1986). The scale includes 15 items with dichotomous response options (yes = 1; no = 0). Participants are asked about the frequency of experiences during the past week. Total scores range from 0 to 15, with higher values indicating more depressive symptoms. Good internal consistency reliabilities are reported (α = .87). Statistical Analysis The nature of a longitudinal research design makes the use of latent growth curve modeling (LGM) desirable. Each personality trait was examined longitudinally with latent growth curve models in Mplus version 7.3 (Muthén and Muthén (1998–2013)) estimating individual trajectories of change in each personality trait with and without covariates. The models were estimated independently in each study. The first set of analyses examined trajectories of personality traits and the preclinical onset of dementia. Individuals were aligned according to the occasion of diagnosis; thus, time was specified as “years-preceding-dementia,” with the occasion in which individuals were diagnosed with dementia specified as time zero (i.e., the intercept), and no measurement of personality after this point. The second set of analyses examined individuals diagnosed with MCI from the EAS. Individuals were aligned chronologically according to the occasion at which they received a classification of MCI (i.e., the intercept, or “0” on each individual’s timeline, is the wave when MCI classification was assigned). As noted, individuals who were diagnosed with incident MCI continued to complete the personality assessments and, for these analyses, the data following diagnosis were included; therefore, the “MCI timeline structured” analyses examined trajectories of personality preceding and after MCI classification. Any individuals who also received a dementia diagnosis during the study were excluded from these analyses, however. The third set of analyses examined individual trajectories of personality for individuals who did not receive a dementia diagnosis during the study, and, for the EAS, were also not classified with MCI. Time was specified as time-in-study and the intercept was specified as the baseline assessment, resulting in the intercept corresponding to a younger age compared with the analyses examining individuals diagnosed with dementia. Although the same number of occasions and set of covariates that were used in the first two series of analyses were used in these models, the analyses aim to provide context rather than a direct comparison due to the necessary difference in the structuring of time and the meaning of the intercept. For EAS, the same number of occasions (five) and set of covariates were used as the analyses examining individuals with MCI in order to maximize the number of waves of data included in these analyses, as structuring according to the same constraints as the analyses examining individuals with dementia would have resulted in three waves of data. For each dataset, sex was included as a dichotomous variable, with male as the reference group (see Table 1 for demographic information). Age and education were measured in years and centered at each sample mean. Centering age and education across samples at a common value may have facilitated interpretability; however, because the mean age is different for each sample, centering at a common value would have resulted in extrapolation. The interaction between age and education was also included to acknowledge expected age-related differences in educational attainment. Across datasets the modal response to depressive symptoms was zero, so depressive symptoms were entered into the models uncentered. Table 1. Demographic Information and Baseline Personality Descriptive Statistics of Participants Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Note. LASA = Longitudinal Aging Study Amsterdam; EAS = Einstein Aging Study; Dem Dx = Individuals diagnosed with dementia; MCI Dx = Individuals diagnosed with MCI; No Dem Dx = Individuals who did not receive a dementia diagnosis during the study; No Dem/MCI Dx = Individuals who did not receive an incident dementia or MCI diagnosis; Time to Dx = time in years to dementia or MCI diagnosis at baseline; Max waves = maximum waves included in analyses; Average waves = average waves of data included in analyses; Depression= depressive symptoms, CES-D (LASA) or GDS (EAS) at Time 1; Wave 1 personality = Means and standard deviations reported from the conditional models; — = not available. View Large Table 1. Demographic Information and Baseline Personality Descriptive Statistics of Participants Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Sample  LASA  LASA  EAS  EAS  EAS  Status  Dem Dx  No Dem Dx  Dem Dx  MCI Dx  No Dem/MCI Dx    N = 162  N = 1138  N = 48  N = 135  N = 602  Variable  M (SD)  M (SD)  M (SD)  M (SD)  M (SD)  Age at first interview  74.36 (6.86)  64.08 (6.76)  80.23 (5.27)  79.29 (5.17)  77.36 (4.80)  Years of education  8.96 (2.37)  9.30 (3.27)  14.15 (3.17)  14.13 (3.52)  14.49 (3.32)  Time to Dx  7.99 (3.30)  —  2.58 (2.54)  —  —  Max waves  4  4  3  5  5  Average waves  2.20 (1.14)  3.80 (0.40)  1.88 (0.89)  2.64 (1.58)  2.42 (1.46)  Depression  8.82 (7.52)  6.60 (6.92)  2.31 (2.39)  2.56 (2.23)  1.78 (1.83)  Personality at Wave 1  DPQ  DPQ  IPIP  IPIP  IPIP   Neuroticism  6.95 (5.87)  5.76 (5.50)  2.15 (0.65)  2.29 (0.68)  2.12 (0.64)   Extraversion  —  —  3.19 (0.62)  3.22 (0.60)  3.37 (0.65)   Conscientiousness  —  —  3.81 (0.68)  3.69 (0.68)  3.83 (0.64)   Agreeableness  —  —  3.94 (0.53)  3.98 (0.53)  4.07 (0.53)   Openness  —  —  3.56 (0.55)  3.60 (0.63)  3.69 (0.64)  Sex  n (%)  n (%)  n (%)  n (%)  n (%)   Female  104 (64.20)  650 (57.78)  24 (50.00)  70 (51.90)  214 (64.45)   Male  58 (35.80)  475 (42.22)  24 (50.00)  65 (48.10)  388 (35.55)  Note. LASA = Longitudinal Aging Study Amsterdam; EAS = Einstein Aging Study; Dem Dx = Individuals diagnosed with dementia; MCI Dx = Individuals diagnosed with MCI; No Dem Dx = Individuals who did not receive a dementia diagnosis during the study; No Dem/MCI Dx = Individuals who did not receive an incident dementia or MCI diagnosis; Time to Dx = time in years to dementia or MCI diagnosis at baseline; Max waves = maximum waves included in analyses; Average waves = average waves of data included in analyses; Depression= depressive symptoms, CES-D (LASA) or GDS (EAS) at Time 1; Wave 1 personality = Means and standard deviations reported from the conditional models; — = not available. View Large For the EAS years preceding dementia analyses that included a maximum of three waves, only linear trajectories were considered. Quadratic trajectories were examined for all other analyses to determine the most appropriate model for trajectory shape. For some personality traits, the quadratic models would not converge; however, in the models that did, the quadratic trajectories revealed higher Bayesian information criteria (BIC), suggesting that the linear models afforded a more parsimonious fit to the data. In addition, the EAS years-preceding-dementia models had some model convergence issues, which may have been due to limited sample size (<100 individuals), or possibly due to the variation in dementia diagnosis between studies. Although standard estimates and errors were reported for the EAS models, a warning that the estimation had reached a saddle point was included in the output. The slope variance was nonsignificant for these models, so analyses were rerun with the slope variance constrained to zero, which resulted in model convergence with no warnings. Therefore, results are reported for the EAS models with the slope variance constrained to zero. Due to larger sample sizes and slope variances, it was not necessary to constrain the slope variances to zero in the analyses examining individuals who were classified with probable dementia in LASA, nor for the analyses examining individuals without a diagnosis in either dataset. An alpha level of .05 was used for all statistical tests. Results Demographic information and baseline personality descriptive statistics are presented in Table 1. Univariate Latent Growth Curve Modeling For each study, univariate growth curve models for each of the personality traits were fitted with and without covariates to examine rates of change in personality traits preceding dementia diagnosis, as well as preceding and following MCI diagnosis. Results are reported for the model with sex, age, education, depressive symptoms at the first occasion, and the interaction between age and education as covariates. All covariates were included in the model as covariates of the intercept (centered at time of dementia or MCI diagnosis) and the slope, allowing both trajectory characteristics to differ across individuals who varied on these characteristics. The following results focus on the main trajectories of individuals diagnosed with dementia and MCI (in the case of EAS), and of individuals who did not receive a diagnosis of dementia (nor MCI, in the case of EAS). Identification and interpretation of significant covariates within each model is included in the supplementary material. Results from the unconditional models are not reported, but this information is available by request. Personality Change in Individuals Diagnosed With Incident Dementia and MCI Neuroticism Both the LASA and EAS datasets included measurements of neuroticism. The conditional models for both studies consistently revealed significant mean linear increases in neuroticism. Figure 1 depicts the results for the average trajectories of neuroticism for each study; that is, for a male participant who entered the study at the mean age with the mean years of education and no depressive symptoms. The intercept (i.e., time = 0) represents the average expected score on the neuroticism scale at time of diagnosis of dementia or of MCI in the case of EAS, and baseline for individuals not diagnosed with MCI or dementia. The slope represents the average change per year between study entry and dementia diagnosis, MCI diagnosis, or time in study, respectively. The average trajectory of neuroticism for individuals with MCI spans the intercept because data were included leading up to and following diagnosis of MCI. The intercept and slope differ substantially between datasets because personality was measured using different scales, and the inter-occasion intervals differ between datasets. Parameter estimates and standard errors for conditional models for trajectories of neuroticism are presented in Table 2. Table 2. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models Personality Trait Neuroticism Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES-D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × Education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Table 2. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models Personality Trait Neuroticism Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Sample  LASA Dem Dx  LASA No Dx  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time metric  Years preceding dementia  Time in study  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  5.214 (1.157)**  3.027 (0.274)**  1.971 (0.151)**  1.850 (0.072)**  1.748 (0.048)**    Female  −0.113 (1.180)  0.325 (0.304)  −0.200 (0.160)  0.148 (0.106)  0.091 (0.052)    Depressive symptoms  0.346 (0.099)**  0.380 (0.027)**  0.140 (0.020)*  0.153 (0.023)**  0.169 (0.014)**    Age  −0.0.15 (0.087)  −0.010 (0.023)  0.005 (0.012)  0.004 (0.009)  −0.001 (0.005)    Education  0.167 (0.176)  −0.083 (0.056)  −0.094 (0.036)*  −0.009 (0.015)  −0.020 (0.008)*    Age × Education  −0.030 (0.025)  0.009 (0.006)  −0.002 (0.004)  −0.001 (0.002)  0.002 (0.001)   Slope  0.258 (0.126)*  0.051 (0.033)  0.075 (0.037)*  0.046 (0.015)*  0.001 (0.010)    Female  −0.116 (0.142)  0.013 (0.035)  −0.092 (0.043)*  0.009 (0.023)  0.002 (0.011)    Depressive symptoms  −0.010 (0.013)  −0.014 (0.003)**  −0.005 (0.008)  −0.008 (0.006)  −0.010 (0.003)*    Age  0.004 (0.010)  0.003 (0.003)  0.000 (0.004)  0.002 (0.002)  0.000 (0.001)    Education  −0.018 (0.021)  0.005 (0.007)  −0.006 (0.012)  −0.001 (0.002)  0.001 (0.002)    Age × Education  −0.001 (0.003)  0.001 (0.001)  −0.005 (0.002)*  −0.001 (0.000)*  0.000 (0.000)  Variance components and fit indices   Intercept  24.824 (7.987)*  15.655 (1.309)**  0.064 (0.034)  0.198 (0.050)**  0.194 (0.021)*   Slope  0.066 (0.123)  0.077 (0.021)**  0.000 (0.000)  0.000 (0.001)  0.001 (0.001)   Cov (IS)  1.062 (0.938)  −0.251 (0.114)*  0.000 (0.000)  0.000 (0.008)  −0.004 (0.004)   Residual  11.517 (2.520)**  8.180 (0.567)**  0.095 (0.031)*  0.128 (0.021)**  0.120 (0.008)**   AIC  2117.025  22314.921  109.685  514.557  1987.202   BIC  2166.426  22395.330  135.882  561.041  2057.606  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES-D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × Education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Figure 1. View largeDownload slide Average trajectories of neuroticism. Note. The left plot shows trajectories of neuroticism based on years preceding dementia diagnosis for individuals with incident dementia from the Longitudinal Aging Study of Amsterdam (LASA) and Einstein Aging Study (EAS) datasets. The left plot also shows the average trajectory of neuroticism based on years preceding and following diagnosis of MCI for individuals from EAS. The right plot shows trajectories of neuroticism for individuals from LASA and EAS who did not receive a dementia diagnosis during the course of the study. Each line represents a male participant who entered the study at the mean age with the mean years of education for each study and no depressive symptoms. Figure 1. View largeDownload slide Average trajectories of neuroticism. Note. The left plot shows trajectories of neuroticism based on years preceding dementia diagnosis for individuals with incident dementia from the Longitudinal Aging Study of Amsterdam (LASA) and Einstein Aging Study (EAS) datasets. The left plot also shows the average trajectory of neuroticism based on years preceding and following diagnosis of MCI for individuals from EAS. The right plot shows trajectories of neuroticism for individuals from LASA and EAS who did not receive a dementia diagnosis during the course of the study. Each line represents a male participant who entered the study at the mean age with the mean years of education for each study and no depressive symptoms. Extraversion, conscientiousness, agreeableness, and openness The EAS protocols included assessment of the four additional personality traits. For conscientiousness, agreeableness, and openness, none of the conditional models revealed significant change, suggesting relative stability in these traits for individuals who were eventually diagnosed with dementia and MCI. The parameter estimates and standard errors for the conditional models of these personality traits are presented in Supplemental Table S1 for both time metrics. Likewise, the extraversion models revealed nonsignificant linear slope means and variances in the years-preceding-dementia models. However, for the MCI-timeline-structured models, both the conditional and unconditional models revealed significant negative linear slope means (p < .01). The results suggest that at the point of MCI classification, a man who entered the study at age 80 with 14 years of education scored an average of 3.3 on the extraversion scale and had been decreasing by .04 per year between study entry and MCI classification. The parameter estimates and standard errors for the conditional extraversion trajectory models are presented in Table 3. Table 3. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models for Personality Trait Extraversion Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES−D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Table 3. Parameter Estimates (and Standard Errors) from Conditional Linear Growth Curve Models for Personality Trait Extraversion Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Sample  EAS Dem Dx  EAS MCI Dx  EAS No Dx  Time matrix  Years preceding dementia  Years surrounding MCI  Time in study  Parameter  ß (SE)  ß (SE)  ß (SE)  Fixed effects   Intercept  3.349 (0.178)**  3.304 (0.095)**  3.579 (0.056)**    Female  0.121 (0.202)  0.140 (0.103)  −0.005 (0.059)    Depressive symptoms  −0.111 (0.036)  −0.066 (0.020)**  −0.104 (0.017)**    Age  −0.017 (0.018)  0.000 (0.010)  −0.002 (0.006)    Education  0.043 (0.029)  −0.006 (0.018)  0.008 (0.008)    Age × Education  −0.004 (0.007)  −0.002 (0.003)  0.001 (0.002)   Slope  −0.069 (0.057)  −0.038 (0.013)*  −0.018 (0.010)    Female  0.086 (0.060)  0.017 (0.017)  0.003 (0.011)    Depressive symptoms  0.003 (0.010)  0.003 (0.004)  0.004 (0.003)    Age  −0.004 (0.006)  0.001 (0.002)  −0.001 (0.001)    Education  0.013 (0.010)  0.000 (0.003)  0.002 (0.002)    Age × Education  0.000 (0.002)  0.000 (0.000)  0.000 (0.000)   Intercept  0.267 (0.071)**  0.238 (0.037)**  0.268 (0.026)**   Slope  0.000 (0.000)  0.001 (0.001)  0.001 (0.001)   Cov (IS)  0.000 (0.000)  0.011 (0.005)*  −0.004 (0.003)   Residual  0.036 (0.011)**  0.088 (0.011)**  0.103 (0.008)**   AIC  109.579  448.387  2014.143   BIC  135.775  494.871  2084.521  Notes. Results are reported with baseline age and education centered at sample mean; AIC = Akaike information criterion; BIC = Bayesian information criterion; Depressive Symptoms = CES−D (LASA) or GDS (EAS) at Time 1; Intercept = estimated trait score at the last available measurement prior to or at time of dementia diagnosis; Slope = estimated rate of change overtime; Age × education = the interaction between age and education. *p < .05. **p ≤ .001. View Large Personality Change in Individuals Not Diagnosed With Dementia Across both datasets, the time-in-study trajectory models for personality traits in individuals who were not diagnosed with dementia revealed limited longitudinal changes. Although slopes in the unconditional models were occasionally significant (two out of six models: neuroticism decreases in LASA and conscientiousness decreases in EAS), once covariates were added to the models, none of the personality trait trajectories identified significant change, suggesting relative stability of all personality traits in individuals who were not diagnosed with dementia (or MCI, in the case of EAS) during the study. The parameter estimates and standard errors for the conditional models are presented in the following tables: neuroticism in Table 2; extraversion in Table 3; and conscientiousness, openness and agreeableness in Supplemental Table S2. Discussion Our previous work (Yoneda et al., 2017) examined trajectories of personality traits using data from Origins of Variance in the Oldest-Old (OCTO-Twin; McClearn et al., 1997). Analyses revealed significant increases in neuroticism and stability in extraversion preceding dementia diagnosis. The primary aim of this study was to replicate these analyses to examine whether consistent changes in personality traits precede dementia diagnosis. Applying a coordinated analysis approach, two additional datasets were selected through the IALSA network based on availability of repeated assessment of personality and classification of dementia. Conceptual replication, utilizing identical statistical models to the extent possible, provides a powerful basis to evaluate replicability, protects against Type I errors, and allows for accelerated accumulation of knowledge (Hofer & Piccinin, 2009). Consistent with our previous work (Yoneda et al., 2017) and supporting our first hypothesis, we identified significant linear increases in neuroticism preceding dementia diagnosis in both EAS and LASA datasets. Furthermore, analyses revealed linear increases in neuroticism in the analyses examining individuals with an incident MCI diagnosis. Finally, consistent with prospective research finding that individuals who were not diagnosed with incident dementia were less likely to show personality change compared with individuals who eventually receive a diagnosis (Balsis et al., 2005; Smith-Gamble et al., 2002), the analyses examining individuals who did not receive a diagnosis during the study revealed stability in neuroticism over time. These findings provide reliable evidence of a consistent pattern of neuroticism increases preceding dementia diagnosis, and, further, suggest that change in neuroticism may occur early in the disease process. Additionally, these results indicate that individuals who remain undiagnosed have markedly different trajectories of neuroticism compared with individuals not diagnosed with incident dementia or MCI. In conducting the current analyses examining trajectories of personality traits in individuals diagnosed with incident dementia, we noted very small slope variance for the EAS years preceding dementia models. Our previous OCTO-Twin models demonstrated similarly low slope variance, and did not converge when all of the covariates were included for the slope; therefore, we had trimmed two of the covariates. However, by constraining the slope variance to zero in the EAS years preceding dementia models, we were able to include all of the covariates on the slope of each personality trait, with no convergence issues. For the purposes of a more meticulous replication (i.e., models in which all covariates of interest were included as predictors of the slope of each personality trait across all three datasets), we re-fitted the OCTO-Twin models to estimate trajectories of neuroticism and extraversion preceding dementia diagnosis with the slope variance constrained to zero. Consistent with the results from the current EAS analyses and our previous OCTO-Twin models, the updated OCTO-Twin models revealed significant increases in neuroticism and stability in extraversion preceding dementia diagnosis, while controlling for all covariates on the slope. Although the results from the updated OCTO-Twin analyses are not presented here, the information is available by request. Assessments of extraversion, conscientiousness, openness, and agreeableness were also available in EAS. Our analyses revealed significant decreases in extraversion only, and solely for individuals with MCI. These results may indicate that individuals with MCI might feel more cognitively challenged in the presence of others, possibly leading to avoidance of social activity. Although consistent with expectations, these results do not explicitly support our second hypothesis that a decrease in extraversion will precede dementia because the finding was seen only in individuals with MCI. The inability to detect significant decreases in extraversion in the individuals with dementia may be accounted for by a variety of reasons, including limited power due to small sample size, or operation of a different process for the individuals with MCI versus those eventually diagnosed with dementia. A further possibility is that individuals with MCI have more insight into the social changes that are occurring. Although self-report does not typically differ from informant-report for individuals with MCI compared with controls (Farias, Mungas, & Jagust, 2005; Ready, Ott, & Grace, 2004), future research including both assessment types may provide unique information about personality changes in individuals with MCI. Finally, our analyses did not support the hypothesis of a decrease in conscientiousness, openness, or agreeableness. The hypotheses were based on retrospective research; thus, several factors regarding methodology could be responsible for the discrepancy between our expectations and findings (see Yoneda et al., 2017 for a more detailed account). Our findings regarding neuroticism increases prior to diagnosis of dementia are consistent with expectations based on neurological research (Mahoney et al., 2011). Although Mahoney and associates (2011) found increases in neuroticism related to neuropathology, their results also indicated that neuropathology was related to decreases in the four personality traits. The discrepancy between our findings and Mahoney and associates’ findings may be due to more substantial personality changes occurring in individuals with FTD, recollection bias, or the timing of personality measurement. Namely, their research assessed changes after diagnosis of FTD while our focus was on change preceding diagnosis; thus, changes in these traits may not yet be measureable, or the brain regions responsible for extraversion, conscientiousness, openness, and agreeableness may not be implicated until the later stages of dementia. Our findings are also consistent with research examining neural correlates of personality in healthy individuals, which finds an association between higher levels of neuroticism and smaller brain volume or reduced cortical thickness (Jackson, Balota, & Head, 2011; Knutson et al., 2001; Wright et al., 2007; Xu & Potenza, 2012). These personality-related neurological findings, in combination with neurological characteristics of dementia such as heightened levels of brain lesions, decreased neuroreactivity, and diminished grey and white matter (Debette & Markus, 2010; Jagust et al., 2008), provide a theoretical context for why increases in neuroticism may be associated with neurodegeneration. However, the comparison is complicated, as our research examines within-person changes and does not include neurological assessments, while the majority of neurology research examines between-person differences. Future research investigating endorsement of neuroticism and intracranial fluid at repeated occasions, which allows evaluation of the amount of brain degeneration since maximum lifetime brain volume, could improve our understanding of personality change, neurology, and dementia. Attrition is a concern in any longitudinal study of aging; however, the majority of missing data across both datasets is due to mortality rather than refusal to participate, which reduces potential selection bias, though could be affected by factors related to mortality. Furthermore, our analyses were limited by the number of individuals who received a dementia or MCI diagnosis. For example, our analyses included individuals with aMCI and naMCI as one group to increase sample size, but trajectories of personality may differ depending on MCI subtype and, importantly, whether individuals eventually convert to dementia. Likewise, although previous research indicates similar personality changes at the domain level after diagnosis of dementia, trajectories of personality in the years leading up to diagnosis may differ depending on the type of dementia. Our analyses could not evaluate whether changes occurred in differing directions for certain personality traits depending on the type of dementia. Future research based on larger samples could extend this conceptual replication by examining trajectories of personality traits according to specific types of dementia or MCI also using LGM. Our syntax is available by request, or we would happily analyze any additional datasets that include assessment of the required constructs. Future research could also apply alternative models to explore a variety of more detailed research questions. Bivariate LGM could examine the association between change in both personality traits and domains of cognitive functioning. Growth mixture modeling (Nylund, Asparouhov, & Muthén, 2007) could examine the possible existence of subpopulations based on change in personality traits. Multistate modeling (Andersen, 1993) could examine change in personality traits corresponding to a series of states such as MCI or dementia diagnosis. Joint growth survival modeling (Tsiatsis & Davidian, 2004) could examine dementia diagnosis or survival as distal outcomes based on change in certain personality traits. A further limitation of this study is the relatively small increases in neuroticism in both datasets. Although these increases are statistically significant, one could question the clinical significance of neuroticism increase, and it may be difficult to detect at the level of the individual patient. However, the magnitude of quantifiable changes is constrained by the properties of the available personality assessments; that is, answers are limited by the nature of the response options. For both datasets, the items used to measure personality traits were relatively limited in quantity; therefore, any measureable change seems potentially meaningful, particularly given the strikingly similar results across datasets. Moreover, significant changes were not detected in the analyses examining individuals who did not receive a diagnosis, which, given the substantially larger samples, would have provided more power to detect significant change. Thus, assessments of personality traits that capture more detail and are more sensitive to variation would enhance this study. For example, The NEO-Personality Inventory Revised (NEO PI-R; Costa & McCrae, 1992) assesses each trait using 40 items answered on a 5-point scale. Future research using the neuroticism scale from the NEO PI-R to assess personality at more frequently administered repeated occasions could improve our understanding of personality change and progression to dementia. Many of this study’s limitations are characteristic of any comparison across independent studies because the study designs are not identical. Although each study included repeated measurement of neuroticism, the measures differed across datasets, and assessments for all five personality traits were available only in EAS. Assessment of depressive symptoms also differed between datasets, with LASA administering the CES-D (Radloff, 1997) and EAS administering the GDS (Yesavage & Sheikh, 1986). In addition, classification of dementia varied across datasets. For EAS, diagnoses were made based on the DSM-IV. Although research comparing different versions of the DSM criteria has found slight differences in rate of diagnosis (Erikinjuntti et al., 1997; Pohjasvaara et al., 1997), there is more similarity in classification using DSM indexes compared with other diagnostic tools (e.g., the Cambridge Mental Disorders in the Elderly), strengthening the replicability of the EAS analyses to those of the OCTO-Twin analyses. The LASA created a “probable dementia” category based on declines on the MMSE or IQCODE, rather than a formal diagnosis, which may have resulted in differential dementia diagnoses. The differences in diagnostic criteria could have affected our results because individuals who may have been classified with dementia in one sample may not have been classified in another sample, and if an individual was not classified with dementia, they were included in the analyses examining individuals who did not receive a dementia diagnosis. Indeed, the studies that utilized similar diagnosis criteria (EAS and OCTO-Twin) also had low slope variance, indicating that trajectories of personality traits may be more similar at the interindividual level for individuals classified according to more formal diagnosis criteria. Finally, the intervals between measurement occasions and available number of occasions also varied between datasets, so quadratic models could not be fitted without additional constraints. Yet, despite the differences in designs, assessments, and diagnostic criteria, considerable similarities were found in the results for both datasets, as well as our previous work using OCTO-Twin data. Heterogeneity in the key features of datasets reinforces significant findings in a coordinated analysis (Hofer & Piccinin, 2009; Lindwall et al., 2012); namely, evidence for generalizability of the findings, based on a consistent pattern of personality change across studies with such different characteristics, is more substantial. The primary strength of this study is the ability to investigate within-person change in personality using data from multiple longitudinal studies of aging. The included studies are heterogeneous in several features including age and cultural background, which offers conceptual rather than identical replication. Despite the differences across datasets, the results revealed a clear pattern of personality trait change, specifically increases in neuroticism, in individuals eventually diagnosed with dementia. The MCI analysis was also consistent with this pattern, and to our knowledge, this was the first study to examine trajectories of personality traits in individuals with incident MCI diagnosis. Overall, this study contributes consistent, replicated findings to the existing literature. A conceptual replication applying a coordinated analysis approach with consistent findings across diverse datasets offers compelling evidence that an increase in neuroticism may indeed be an early indicator of dementia. Supplementary Material Supplementary material are available at The Journals of Gerontology Series B: Psychological and Social Sciences online. Funding This work was supported by the National Institute on Aging (NIA) at the National Institutes of Health (NIH; grant number P01AG043362; 2013–2018) in support of the Integrative Analysis of Longitudinal Studies of Aging (IALSA) research network. Origins of Variance in the Oldest-Old (OCTO-Twin) data collection was funded by the National Institute on Aging at the National Institutes of Health (grant number AG08861); the Swedish Council for Working Life and Social Research; the Adlerbertska Foundation; the Wenner-Gren Foundations; and the Wilhelm and Martina Lundgrens Foundation. The Longitudinal Aging Study Amsterdam (LASA) is largely supported by the Netherlands Ministry of Health, Welfare and Sports, Directorate of Long-Term Care. Einstein Aging Study (EAS) data collection was supported by the National Institutes of Health NIA (National Institute on Aging) Grant AG03949; the Sylvia & Leonard Marx Foundation, and; the Czap Foundation. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official view of the NIH. Conflict of Interest None reported. References Allemand, M., Zimprich, D., & Hertzog, C. ( 2007). Cross-sectional age differences and longitudinal age changes of personality in middle adulthood and old age. Journal of Personality , 75, 323– 358. doi: 10.1111/j.1467-6494.2006.00441.x Google Scholar CrossRef Search ADS   American Psychiatric Association. ( 1987). Diagnostic and statistical manual of mental disorders  ( 3rd ed.). Washington, DC: Author. PubMed PubMed  American Psychiatric Association. ( 1994). Diagnostic and statistical manual of mental disorders  ( 4th ed.). Washington, DC: Author. PubMed PubMed  Andersen, P. K. ( 1993). Statistical models based on counting processes . New York: Springer-Verlag. Google Scholar CrossRef Search ADS   Balsis, S., Carpenter, B. D., & Storandt, M. ( 2005). Personality change precedes clinical diagnosis of dementia of the Alzheimer Type. Journal of Gerontology: Psychological Sciences , 60B, 98– 101. doi: 10.1093/geronb/60.2.P98 Google Scholar CrossRef Search ADS   Berg, A. I., & Johansson, B. ( 2013). Personality change in the oldest-old: Is it a matter of compromised health and functioning? Journal of Personality , 82, 25– 31. doi: 10.1111/jopy.12030 Google Scholar CrossRef Search ADS   Bruscoli, M., & Lovestone, S. ( 2004). Is MCI really just early dementia? A systematic review of conversion studies. International Psychogeriatrics , 16, 129– 140. doi: 10.1017/S1041610204000092 Google Scholar CrossRef Search ADS   Caselli, R., Beach, T., Knopman, D., & Graff-Radford, N. ( 2017). Alzheimer disease: Scientific breakthroughs and translational challenges. Mayo Clinic Proceedings , 92, 978– 994. doi: 10.1016/j.mayocp.2017.02.011 Google Scholar CrossRef Search ADS   Costa, P. T., & McCrae, R. R. ( 1992). Neo Pi-R. The revised NEO personality inventory . Odessa, FL: Psychological Assessment Resources. Dawson, D. V., Welsh-Bohmer, K. A., & Siegler, I. C. ( 2000). Premorbid personality predicts level of rated personality change in patients with Alzheimer disease. Alzheimer Disease and Associated Disorders , 14, 11– 19. doi: 10.1097/ 00002093-200001000-00002 Google Scholar CrossRef Search ADS   Debette, S., & Markus, H. S. ( 2010). The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: Systematic review and meta-analysis. British Medical Journal , 341, 341– 350. http://dx.doi.org/10.1136/bmj.c3666 Google Scholar CrossRef Search ADS   Di Marco, L. Y., Marzo, A., Muñoz-Ruiz, M., Ikram, M. A., Kivipelto, M., Ruefenacht, D.,… Frangi, A. F. ( 2014). Modifiable lifestyle factors in dementia: A systematic review of longitudinal observational cohort studies. Journal of Alzheimer’s Disease: JAD , 42, 119– 135. doi: 10.3233/JAD-132225 Erkinjuntti, T., Ostbye, T., Steenhuis, R., & Hachinski, V. ( 1997). The effect of different diagnostic criteria on the prevalence of dementia. The New England Journal of Medicine , 337, 1667– 1674. doi: 10.1056/NEJM199712043372306 Google Scholar CrossRef Search ADS   Eysenck, H. J., & Eysenck, S. G. G. ( 1964). Manual of the Eysenck personality inventory . London: University of London Press. Farias, S. T., Mungas, D., & Jagust, W. ( 2005). Degree of discrepancy between self and other-reported everyday functioning by cognitive status: Dementia, mild cognitive impairment, and healthy elders. International Journal of Geriatric Psychiatry , 20, 827– 834. doi: 10.1002/gps.1367 Google Scholar CrossRef Search ADS   Floderus-Myrhed, B., Pedersen, N., & Rasmuson, I. ( 1980). Assessment of heritability for personality, based on a short-form of the Eysenck Personality Inventory: A study of 12,898 twin pairs. Behavior Genetics , 10, 153– 162. doi: 10.1007/BF01066265 Google Scholar CrossRef Search ADS   Goldberg, L. R. ( 1999). A broad-bandwidth, public-domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe , Vol. 7 (pp. 7– 28). Tilburg, The Netherlands: Tilburg University Press. Graham, E. K., Rutsohn, J. P., Turiano, N. A., Bendayan, R., Batterham, P. J., Gerstorf, D.,… Mroczek, D. K. ( 2017). Personality predicts mortality risk: An integrative data analysis of 15 international longitudinal studies. Journal of Research in Personality , 70, 174– 186. doi: 10.1016/j.jrp.2017.07.005 Google Scholar CrossRef Search ADS   Hall, C. B., Lipton, R. B., Sliwinski, M., & Stewart, W. F. ( 2000). A change point model for estimating the onset of cognitive decline in preclinical Alzheimer’s disease. Statistics in Medicine , 19, 1555– 1566. doi: 10.1002/(SICI)1097-0258(20000615/30)19:11/12<1555::AID-SIM445>3.0.CO;2–3 Google Scholar CrossRef Search ADS   Hofer, S. M., & Piccinin, A. M. ( 2009). Integrative data analysis through coordination of measurement and analysis protocol across independent longitudinal studies. Psychological Methods , 14, 150– 164. doi: 10.1037/a0015566 Google Scholar CrossRef Search ADS   Hofer, S. M., & Piccinin, A. M. ( 2010). Toward an integrative science of life-span development and aging. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences , 65B, 269– 278. doi: 10.1093/geronb/gbq017 Google Scholar CrossRef Search ADS   Huisman, M., Poppelaars, J., van der Horst, M., Beekman, A. T., Brug, J., van Tilburg, T. G., & Deeg, D. J. ( 2011). Cohort profile: The longitudinal aging study Amsterdam. International Journal of Epidemiology , 40, 868– 876. doi: 10.1093/ije/dyq219 Google Scholar CrossRef Search ADS   Jackson, J., Balota, D. A., & Head, D. ( 2011). Exploring the relationship between personality and regional brain volume in healthy aging. Neurobiology of Aging , 32, 2162– 2171. doi: 10.1016/j.neurobiolaging.2009.12.009 Google Scholar CrossRef Search ADS   Jagust, W. J., Zheng, L., Harvey, D. J., Mack, W. J., Vinters, H. V., Weiner, M. W.,… Chui, H. C. ( 2008). Neuropathological basis of magnetic resonance images in aging and dementia. Annals of Neurology , 63, 72– 80. doi: 10.1002/ana.21296 Google Scholar CrossRef Search ADS   Katz, M. J., Lipton, R. B., Hall, C. B., Zimmerman, M. E., Sanders, A. E., Verghese, J.,… Derby, C. A. ( 2012). Age-specific and sex-specific prevalence and incidence of mild cognitive impairment, dementia, and Alzheimer dementia in blacks and whites: A report from the Einstein Aging Study. Alzheimer Disease and Associated Disorders , 26, 335– 343. doi: 10.1097/WAD.0b013e31823dbcfc Google Scholar CrossRef Search ADS   Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. ( 1993). A longitudinal twin study of personality and major depression in women. Archives of General Psychiatry , 50, 853– 862. doi: 10.1001/archpsyc.1993.01820230023002 Google Scholar CrossRef Search ADS   Knutson, B., Momenan, R., Rawlings, R. R., Fong, G. W., & Hommer, D. ( 2001). Negative association of neuroticism with brain volume ratio in healthy humans. Biological Psychiatry , 50, 685– 690. doi: 10.1016/S0006-3223(01)01220-3 Google Scholar CrossRef Search ADS   van den Kommer, T. N., Comijs, H. C., Dik, M. G., Jonker, C., & Deeg, D. J. ( 2008). Development of classification models for early identification of persons at risk for persistent cognitive decline. Journal of Neurology , 255, 1486– 1494. doi: 10.1007/s00415-008-0942-3 Google Scholar CrossRef Search ADS   Lindwall, M., Cimino, C. R., Gibbons, L. E., Mitchell, M. B., Benitez, A., Brown, C. L.,… Piccinin, A. M. ( 2012). Dynamic associations of change in physical activity and change in cognitive function: Coordinated analyses of four longitudinal studies. Journal of Aging Research , 2012, 493598. doi: 10.1155/2012/493598 Google Scholar CrossRef Search ADS   Luteijn, F., Starren, J., & Van Dijk, H. ( 2000). NPV; Nederlandse Persoonlijkheids Vragenlijst, handleiding (herziene uitgave)  . [DPQ; Dutch Personality Questionnaire, manual (revised edition)]. Lisse: Swets & Zeitlinger. Lykou, E., Rankin, K. P., Chatziantoniou, L., Boulas, C., Papatriantafyllou, O., Tsaousis, I.,… Papatriantafyllou, J. D. ( 2013). Big 5 personality changes in Greek bvFTD, AD, and MCI patients. Alzheimer Disease and Associated Disorders , 27, 258– 264. doi: 10.1097/WAD.0b013e31826e5504 Google Scholar CrossRef Search ADS   Mahoney, C. J., Rohrer, J. D., Omar, R., Rossor, M. N., & Warren, J. D. ( 2011). Neuroanatomical profiles of personality change in frontotemporal lobar degeneration. The British Journal of Psychiatry: The Journal of Mental Science , 198, 365– 372. doi: 10.1192/bjp.bp.110.082677 Google Scholar CrossRef Search ADS   McClearn, G. E., Johansson, B., Berg, S., Pedersen, N. L., Ahern, F., Petrill, S. A., & Plomin, R. ( 1997). Substantial genetic influence on cognitive abilities in twins 80 or more years old. Science (New York, N.Y.) , 276, 1560– 1563. doi: 10.1126/science.276.5318.1560 Google Scholar CrossRef Search ADS   McCrae, R. R., & Costa, P. T. ( 1985). Comparison of EPI and psychoticism scales with measures of the five-factor model of personality. Personality and Individual Differences , 6, 587– 597. doi: 10.1016/0191-8869(85)90008-X Google Scholar CrossRef Search ADS   Middleton, L. E., & Yaffe, K. ( 2010). Targets for the prevention of dementia. Journal of Alzheimer’s Disease: JAD , 20, 915– 924. doi: 10.3233/JAD-2010-091657 Google Scholar CrossRef Search ADS   Mroczek, D. K., & Spiro, A. ( 2003). Modeling intraindividual change in personality traits: Findings from the normative aging study. Journal of Gerontology , 58B, 153– 165. doi: 10.1093/geronb/58.3.P153 Google Scholar CrossRef Search ADS   Muthén, L. K., & Muthén, B. O. ( 1998–2013). Mplus user’s guide  ( 6th ed.). Los Angeles, CA: Muthén & Muthén. Nussbaum, P. D. ( 1997). Late-life depression. In P. D. Nussbaum (ed.), Handbook of Neuropsychology and Aging. Critical Issues in Neuropsychology . Boston, MA: Springer. Google Scholar CrossRef Search ADS   Nylund, K. L., Asparouhov, T., & Muthén, B. O. ( 2007 ). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal , 14, 535– 569. doi: 10.1080/10705510701575396 Google Scholar CrossRef Search ADS   Petersen, R. C., Caracciolo, B., Brayne, C., Gauthier, S., Jelic, V., & Fratiglioni, L. ( 2014). Mild cognitive impairment: A concept in evolution. Journal of Internal Medicine , 275, 214– 228. doi: 10.1111/joim.12190 Google Scholar CrossRef Search ADS   Pohjasvaara, T., Erkinjuntti, T., Vataja, R., & Kaste, M. ( 1997). Dementia three months after stroke. Baseline frequency and effect of different definitions of dementia in the Helsinki Stroke Aging Memory Study (SAM) cohort. Stroke , 28, 785– 792. doi: 10.1161/01.STR.28.4.785 Google Scholar CrossRef Search ADS   Polidori, M. C., Nelles, G., & Pientka, L. ( 2010). Prevention of dementia: Focus on lifestyle. International Journal of Alzheimer’s Disease , 2010, 1– 9. doi: 10.4061/2010/393579 Google Scholar CrossRef Search ADS   Radloff, L. S. 1977. The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement , 1, 385– 401. doi: 10.1177/014662167700100306 Google Scholar CrossRef Search ADS   Rankin, K. P., Kramer, J. H., Mychack, P., & Miller, B. L. ( 2003). Double dissociation of social functioning in frontotemporal dementia. Neurology , 60, 266– 271. doi: 10.1212/ 01.WNL.0000041497.07694.D2 Google Scholar CrossRef Search ADS   Ready, R. E., Ott, B. R., & Grace, J. ( 2004). Patient versus informant perspectives of Quality of Life in Mild Cognitive Impairment and Alzheimer’s disease. International Journal of Geriatric Psychiatry , 19, 256– 265. doi: 10.1002/gps.1075 Google Scholar CrossRef Search ADS   Roberts, R. O., Knopman, D. S., Mielke, M. M., Cha, R. H., Pankratz, V. S., Christianson, T. J. H., … Petersen, R. C. ( 2014). Higher risk of progression to dementia in mild cognitive impairment cases who revert to normal. Neurology , 82, 317– 325. doi: 10.1212/WNL.0000000000000055 Google Scholar CrossRef Search ADS   Schumacker, R. E., & Lomax, R. G. ( 2015). A beginner’s guide to structural equation modeling  ( 4th ed.). New York: Routledge. Small, B. J., Hertzog, C., Hultsch, D. F., & Dixon, R. A. ( 2003). Stability and change in adult personality over 6 years: Findings from the Victoria Longitudinal Study. Journal of Gerontology: Psychological Sciences , 58B, 166– 176. doi: 10.1093/geronb/58.3.P166 Google Scholar CrossRef Search ADS   Smith-Gamble, V., Baiyewu, O., Perkins, A. J., Gureje, O., Hall, K. S., Ogunniyi, A.,… Hendrie, H. C. ( 2002). Informant reports of changes in personality predict dementia in a population-based study of elderly African Americans and Yoruba. The American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry , 10, 724– 732. doi: 10.1097/00019442-200211000-00011 Google Scholar CrossRef Search ADS   Torrente, F., Pose, M., Gleichgerrcht, E., Torralva, T., López, P., Cetkovich-Bakmas, M., & Manes, F. ( 2014). Personality changes in dementia: Are they disease specific and universal? Alzheimer Disease and Associated Disorders , 28, 261– 268. doi: 10.1097/WAD.0000000000000030 Google Scholar CrossRef Search ADS   Tsiatis, A. A., & Davidian, M. ( 2004). Joint modeling of longitudinal and time-to-event data: An overview. Statistica Sinica , 14, 809– 834. Retrieved from http://www.jstor.org/stable/24307417 Wright, C. I., Feczko, E., Dickerson, B., & Williams, D. ( 2007). Neuroanatomical correlates of personality in the elderly. Neuroimage , 35, 263– 272. doi: 10.1016/ j.neuroimage.2006.11.039 Google Scholar CrossRef Search ADS   Xu, J., & Potenza, M. N. ( 2012). White matter integrity and five-factor personality measures in healthy adults. Neuroimage , 59, 800– 807. doi: 10.1016/j.neuroimage.2011.07.040 Google Scholar CrossRef Search ADS   Yesavage, J. A., & Sheikh, J. I. ( 1986). Geriatric depression scale (GDS): Recent evidence and development of a shorter version. Clinical Gerontologist , 5, 165– 173. doi: 10.1300/J018v05n01_09 Google Scholar CrossRef Search ADS   Yoneda, T., Rush, J., Berg, A. I., Johansson, B., & Piccinin, A. M. ( 2017). Trajectories of personality traits preceding dementia diagnosis. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences , 72, 922– 931. doi: 10.1093/geronb/gbw006 © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The Journals of Gerontology Series B: Psychological Sciences and Social SciencesOxford University Press

Published: Mar 28, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off