Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Feb 15.
Published in final edited form as: J Affect Disord. 2019 Nov 30;263:310–317. doi: 10.1016/j.jad.2019.11.124

Purpose in Life is a Robust Protective Factor of Reported Cognitive Decline Among Late Middle-Aged Adults: the Emory Healthy Aging Study

Aliza P Wingo 1,2,†,§, Thomas S Wingo 3,†, Wen Fan 3, Sharon Bergquist 4, Alvaro Alonso 5, Michele Marcus 5, Allan I Levey 3,‡, James J Lah 3,‡
PMCID: PMC6989389  NIHMSID: NIHMS1546208  PMID: 31969260

Abstract

Background:

Cognitive abilities tend to decline in advanced age. A novel protective factor of cognitive decline in advanced age is purpose-in-life (PiL), a trait-like tendency to derive life meanings and purpose. However, whether PiL protects against cognitive decline in late-middle-age is unclear. Hence, we examined the association between PiL and perceived cognitive decline, one of the earliest detectable cognitive symptoms before the onset of cognitive impairment. Furthermore, we used a machine learning approach to investigate whether PiL is a robust predictor of cognitive decline when considered with the known protective and risk factors for cognition.

Methods:

PiL was assessed with a 10-item questionnaire and perceived cognitive decline with the Cognitive Function Instrument among 5,441 Emory Healthy Aging Study participants, whose mean age was 63 and 51% were employed. Association between PiL and perceived cognitive decline was examined with linear regression adjusting for relevant confounding factors. Elastic Net was performed to identify the most robust predictors of cognitive decline.

Results:

Greater PiL was associated with less perceived cognitive decline after adjusting for the relevant factors. Furthermore, Elastic Net modeling suggested that PiL is a robust predictor of cognitive decline when considered simultaneously with known protective (education, exercise, enrichment activities) and risk factors for cognition (depression, anxiety, diagnosed medical, mental health problems, smoking, alcohol use, family history of dementia, and others)

Limitation:

This is a cross-sectional study.

Conclusions:

PiL is a robust protective factor of perceived cognitive decline observed as early as middle age. Thus, interventions to enhance PiL merit further investigation.

Cognitive abilities tend to decline in older age, and cognitive decline is a major concern for older adults (Deary et al., 2009; Wilson et al., 2002; Zaninotto et al., 2018). Cognitive decline may lead to overt impairment of cognition and thus a diagnosis of mild cognitive impairment (MCI) or, eventually, dementia (Rajan et al., 2015; Scheltens et al., 2016). Longitudinal studies have shown that subtle decline followed by a slow but more progressive decline of cognitive performance can precede the onset of MCI or dementia by up to 20 years, a period referred to as prodromal phase (Bilgel et al., 2018; Boyle et al., 2017; Li, G. et al., 2017; Rajan et al., 2015).

Early detection of cognitive decline is essential to identify individuals at risk for MCI or dementia so that preventive treatments can be implemented to halt the progression from normal cognition to MCI or dementia (Scheltens et al., 2016). Interestingly, a growing literature suggests that the earliest detectable cognitive symptoms in the prodromal phase is a person‟s perception of his/her decline in cognitive abilities relative to previous level of performance, also known as perceived cognitive decline (Rabin et al., 2017; Scheef et al., 2012; Scheltens et al., 2016). Perceived cognitive decline can capture longitudinal change of cognitive abilities and predict objectively measured cognitive decline in longitudinal studies (Dufouil et al., 2005; Koppara et al., 2015; Rabin et al., 2017). Furthermore, the ability to detect cognitive impairment with neuropsychological tests is best at high levels of impairment, but decreases as people perform closer to the normal ability range, whereas perceived cognitive decline is likely more sensitive (Rabin et al., 2017). Perceived cognitive decline in individuals clinically defined as cognitively normal has been found to be associated with Alzheimer’s disease biomarkers and with higher rate of conversion from normal cognition to MCI or dementia in subsequent years (Jessen et al., 2010; Kaup et al., 2015; Rabin et al., 2017).

Purpose-in-life (PiL) is a recently recognized protective factor against cognitive decline (Boyle et al., 2010a; Boyle et al., 2012; Kim et al., 2019). PiL here refers to a trait-like tendency to derive meaning from life experiences and to possess a sense of direction and purpose in life. Life purpose and meanings have been conceptualized as highly related. For instance, Reker and Wong defined life meaning as “the cognizance of order, coherence, purpose in one’s existence, the pursuit and attainment of worthwhile goals, and an accompanying sense of fulfilment” (Zika and Chamberlain, 1992). Likewise in the more recent psychological literature, the greatest consensus in defining life meaning has centered on two dimensions – coherence (one’s comprehension and sense made of life) and purpose (one’s core aims and aspirations for life) (Martela and Steger, 2016). PiL is potentially modifiable and may be a target for treatment (Breitbart et al., 2012; Kissane et al., 2019; Park et al., 2019). Higher PiL has been associated with a slower decline in cognitive performance over time in advanced age in two longitudinal cohorts, the Rush Memory and Aging Project (MAP) and the Health and Retirement Study (Boyle et al., 2012; Kim et al., 2019). Additionally, higher baseline PiL was associated with lower incidence of MCI and of Alzheimer’s dementia in the Rush MAP cohort (Boyle et al., 2010b). These findings point to an intriguing association between higher PiL and cognitive stability in advanced age. Participants in the prior studies tended to be older (mid-70s) and retired (Boyle et al., 2010a; Boyle et al., 2012; Kim et al., 2019). Hence, the relationship between PiL and cognitive decline is unclear among employed, younger adults in their 50s and 60s.

This present study examines whether PiL is associated with perceived cognitive decline, which may be among the earliest manifestations of cognitive symptoms before the expression of MCI or dementia, in primarily mid-life adults. Furthermore, we used a machine learning approach to investigate whether PiL is a robust predictor of perceived cognitive decline when considered with the known protective factors (i.e. education, exercise, enrichment activities) and risk factors (depression, anxiety, smoking, alcohol use, diagnosed medical and mental health problems, and family history of dementia) for cognition simultaneously (Becker et al., 2018; Bellou et al., 2016; Bhattarai et al., 2019; Milgram et al., 2006; Scheltens et al., 2016). We hypothesized that greater PiL is a robust predictor of less perceived cognitive decline among late-middle-aged individuals.

Methods

The Emory Healthy Aging Study (EHAS)

Participants in this study were from the EHAS, a longitudinal, on-line study launched in November of 2014 that aims to better understand factors that contribute to healthy aging and identify markers that can predict common age-related diseases such as Alzheimer’s disease, cardiovascular disease, diabetes mellitus, and others. To date, the study has enrolled 20,523 consented community-based registrants from the greater Atlanta area and follows these individuals over time. EHAS participants complete surveys and health history questionnaires that include their family’s medical history, age, sex, race, education, employment status, marital status, income, and many other factors as described below. All study procedures were approved by the Emory Institutional Review Board.

Psychological factors

Purpose-in-life (PiL) refers to a trait-like tendency to derive meaning from life experiences and to possess a sense of direction and purpose in life. It was assessed with a modified 10-item measure derived from Ryff’s and Keyes’s scale of psychological well-being, which has been evaluated psychometrically (Abbott et al., 2010; Abbott et al., 2006; Ryff and Keyes, 1995). For this 10-item PiL scale, the Cronbach’s coefficient alpha was 0.73 indicating a moderate level of internal consistency (Barnes et al., 2007). Some of these items are i) I feel good when I think of what I’ve done in the past and what I hope to do in the future; ii) I live life one day at a time and do not really think about the future; iii) I have a sense of direction and purpose in life. Participants rated their level of agreement with these items using a 5-point scale, 1=strongly disagree, 2=disagree, 3=neutral, 4=agree, and 5=strongly agree. Ratings of items that were negatively worded were reverse-coded so that higher scores on all individual items indicate greater PiL. We took the average of the ratings for these 10 items to represent each participant’s PiL, with higher scores indicating greater levels of meaning, purpose, and direction in life.

Current depressive symptoms were assessed with the eight-item Patient Health Questionnaire depression scale (PHQ-8) (Kroenke et al., 2009). The PHQ-8 is a valid measure of current depression in the general population (Kroenke et al., 2009). Response to each item ranges from 0 to 3, with higher score indicating more severity, and the total score on the PHQ-8 ranges from 0 – 24. A PHQ-8 score of 10 or more indicates clinically significant current depression (Kroenke et al., 2009). We excluded participants who did not respond to more than one question from the PHQ-8. In addition, to take into account potential missing response in one question on the PHQ-8, we took the average of the responses to these eight items to represent each participant’s current depressive symptoms. Hence, our participants’ PHQ-8 scores range from 0 to 3, with a score of 1.25 representing clinically significant current depression.

Current anxiety symptoms were evaluated with the GAD-7 scale (Spitzer et al., 2006), a well-validated tool for assessing symptoms of generalized anxiety disorder. The GAD-7 has been validated both in the general population (Lowe et al., 2008) and primary care setting (Spitzer et al., 2006), showing excellent internal consistency (Cronbach α = 0.92) and very good test-retest reliability (intraclass correlation = 0.83). Scores from the GAD-7 range from 0 to 21, with higher scores reflecting more anxiety symptoms. Participants who did not respond to more than one item on the GAD-7 were excluded from our analysis. To take into consideration potential non-response to one item on the GAD-7, we took the average of the responses for the seven items to represent current anxiety score for each participant; hence, our participants’ GAD-7 scores range from 0 to 3, with a score cutoff of 0.7, 1.4, and 2.1 representing mild, moderate, and severe anxiety, respectively (Spitzer et al., 2006).

Behavioral health

The following self-reported behavioral health variables were used: ever smoked cigarettes, current alcohol use, current strenuous exercise, current non-strenuous exercise, current walking, and enrichment activities. Smoking was treated as a binary variable of yes or no in our analyses. Alcohol use in the last 12 months was assessed with respect to frequency and amount. Frequency of moderate to strenuous exercise (aka exercise frequency here), frequency of exercise through hobbies such as slow dancing, bowling, or golfing (aka exercise through hobby here), and frequency of walking were assessed. Enrichment activities such as reading newspapers, magazines, books, writing letters or e-mails, and playing board games were assessed and treated as a semi-quantitative variable with higher value indicating more enrichment activities that a participant performed.

Medical health

The following self-reported medical health conditions were collected and used in our analyses: high blood pressure, stroke or transient ischemic attack, heart attack, coronary heart disease, congestive heart failure, atrial fibrillation, high cholesterol, pulmonary embolism, deep vein thrombosis, blood clots, other circulatory or vascular system problem, migraine, mild cognitive impairment or memory loss, Alzheimer’s disease, other dementia, concussion, traumatic brain injury, spinal cord injury, seizures, Parkinson’s disease, amyotrophic lateral sclerosis, multiple sclerosis, other nervous system problems, tuberculosis, hepatitis B or C, HIV or AIDS, kidney disease, chronic lung disease, sleep apnea, thyroid problems, diabetes, organ transplant, pulmonary fibrosis, and other disease. For our analyses, the sum of all positive responses was used to indicate the burden of medical problems for each participant.

Mental health

Participants were asked if a doctor has ever diagnosed them with any of the following conditions: Post-traumatic stress disorder, bipolar disorder, depression, schizophrenia, other mental health disorder. For analysis, the sum of all positive responses was used to indicate the burden of mental health problems for each participant.

Family history

Maternal and paternal history of dementia was collected and used in our analyses. Family history of dementia for each parent was assessed in a systematic fashion. Participants were asked if their biological father was ever diagnosed with MCI or AD, memory loss, confusion or other dementia. If a participant answered “yes” to any of these questions, paternal history of dementia was coded as yes; otherwise, it was coded as no. Likewise, participants were asked if their biological mother was ever diagnosed with MCI or AD, memory loss, confusion or other dementia. If a participant answered “yes” to any of these questions, then maternal history of dementia was coded as yes; otherwise, it was coded as no.

Perceived cognitive decline

We used the Cognitive Function Instrument (CFI) to assess perceived memory decline and cognitive decline that interferes with daily functioning. The CFI has been shown to be a sensitive and reliable instrument in tracking early decline in cognitive function in older adults who do not have cognitive impairment at baseline in longitudinal studies (Amariglio et al., 2015; Li, C. et al., 2017). Specifically, perceived memory decline was assessed using six questions from the CFI. Some of these items are: Please think about your current experiences compared to one year ago: 1) Do you feel that your memory has declined substantially? 2) Do others tell you that you tend to repeat questions over and over? 3) Have you been misplacing things more often? Response to each question was based on a 3-point scale, with 0=no, 1=maybe, 2=yes. For each participant, the average score from these six questions was used to represent his/her perceived memory decline, with higher scores reflecting more perceived memory decline.

Cognitive decline that interferes with daily activities was assessed with eight questions from the CFI. Some of them are: Please think about your current experiences compared to one year ago: 1) Do you have more trouble driving (drive more slowly, tend to get lost, have accidents)? 2) Do you have more difficulty managing money? 3) Are you less involved in social activities? Response to each question was based on a 3-point scale, with 0=no, 1=maybe, 2=yes. The average score from these eight questions was used to represent each participant’s perceived cognitive decline, with higher scores reflecting worse cognitive decline that interfere with daily functioning.

Statistical analysis

Linear regression was used to assess for association between PiL and perceived memory decline, adjusting for age, sex, education, employment status, current depressive symptoms, current anxiety symptoms, the number of diagnosed medical problems, and the number of diagnosed mental health problems. A second linear regression model was performed to assess for association between PiL and cognitive decline that interferes with daily functioning, adjusting for the same covariates listed above.

Next, Elastic Net was used to select independent predictors and estimate their effects for perceived cognitive decline. Elastic Net is a machine learning algorithm for regularization of regression models combining Ridge and Lasso, which balances the model selection so that it is not overly complex (i.e. overfitted) or too simple (i.e. underfit) (Friedman et al., 2010). In other words, Elastic Net fits a model with all the provided variables (sometimes called features) but constrains the coefficients by shrinking the uninformative ones to zero. Data from the EHAS was divided into a training dataset and a test dataset using random sampling with a ratio of 2:1. Then we performed cross validation of the training dataset to obtain different models with different coefficient α values, ranging from 0 to1 (interval = 0.1). For each α value, we applied the corresponding model on our test dataset to predict the outcome. Then we calculated mean square error (MSE) or mis-classification rate for each model. We declared the best model to be the one with the lowest MSE. In the best model, the coefficient estimates for particular variables indicate that these variables are robust predictors for the outcome. Variables without a coefficient estimate are not considered robust predictors of the outcome. Our primary analyses were to determine which variables best predicted i) perceived memory decline, and ii) cognitive decline interfering with daily activities, respectively. The predictors we included in the Elastic Net models were PiL, age, sex, education, employment status, marital status, income, current depressive symptoms, current anxiety symptoms, walk frequency, exercise frequency, frequency of exercising via hobbies, enrichment activities, smoking status, average sleep duration a night, frequency of alcohol use, typical amount of alcohol use, the number of diagnosed medical problems, the number of diagnosed mental health problems, the number of medications currently taking, maternal history of dementia, and paternal history of dementia. Due to different rates of missingness, sample sizes varied when including different variables in the model. We standardized the scales of all the predictors to provide standardized coefficients from the Elastic Net models. Of note, all the standardized coefficients for the robust predictors from Elastic Net models may be biased upward due to the nature of regularization (Friedman et al., 2010). All Elastic Net analyses were performed with the R ‘glmnet’ package version 3.5.1 (Friedman et al., 2010).

Results

Characteristics of the Emory Healthy Aging Study (EHAS) participants

A total of 5,441 EHAS participants had data on PiL, perceived memory decline, and cognitive decline interfering with daily functioning available for analysis. Their mean age was 63, 71% were female, 84% were Caucasian, 70% were married, mean education was 17 years, 51% were employed, and incomes ranged from $40,000 or less to $150,000 or more per year (Table 1 and Supplementary Table 1). With regard to age range, 4% of the participants were younger than 40, 46% were between 40 and 65, and 49% were older than 65.

Table1:

Characteristics of the Emory Healthy Aging Study participants (N=5441)

Variable Mean (SD), Median, Range
Age 63.5 (11.3), 65, [20 – 96]
Education level (years) 16.9 (2.8), 16, [8 – 22]
Female sex (N, %) 3837 (70.6%)
Ethnicity (N, %)
 White 1883 (83.7%)
 African Americans 206 (9.1%)
 Others 161 (7.2%)
Employment (N, % employed) 2747 (50.5%)
Marital status (N, %)
 Married 3779 (69.6%)
 Separated 1167 (21.5%)
 Never married 484 (8.9%)
Income (N, %)
 $39,999 or less 473 (9.1%)
 $40,000 to $59,999 632 (12.2%)
 $60,000 to $99,999 1360 (26.3%)
 $100,000 to $149,999 1264 (24.4%)
 $150,000 or more 1453 (28.0%)
Exercise frequency (N, %)
 None 2008 (36.9%)
 1–2 days/week 1374 (25.2%)
 3–4 days/week 1464 (26.9%)
 5 or more days/week 590 (10.9%)
Exercise through hobby (N, %)
 None 3023 (56.9%)
 1–2 days/week 1580 (29.7%)
 3–4 days/week 493 (9.3%)
 5 or more days/week 219 (4.1%)
Walk frequency (N, %)
 Rarely/never 503 (9.3%)
 1–4 times a month 1010 (18.7%)
 2 to 6 times a week 2749 (50.9%)
 7 or more time a week 1136 (21.0%)
Sleep duration (N, %)
 6 or fewer hours a night 1614 (29.7%)
 7 to 8 hours a night 3410 (62.7%)
 9 or more hous a night 413 (7.6%)
Enrichment activity/ 2.8 (0.7), 2.8, [0 – 4]
Purpose in life 4.0 (0.6), 4.1, [1.2 – 5.0]
Depressive symptoms 0.3 (0.4), 0.2, [0 – 3]
Anxiety symptoms 0.3 (0.4), 0.1, [0 – 3]
Ever smoke (yes, N, %) 1814 (33.4%)
Alcohol use frequency (N, %)
 No drinks in past year 654 (12.5%)
 1 to 11 times in past year 883 (17.1%)
 1 to 3 times a month 844 (16.3%)
 1 to 2 times a week 914 (17.7%)
 3 to 4 times a week 681 (13.2%)
 5 to 7 times a week 1208 (23.4%)
Number of diagnosed medical problems 2.4 (2.1), 2, [0 – 34]
Number of diagnosed mental health problems 0.4 (0.7), 0, [0 – 5]
Number of medications taken for medical or mental health problems 1.4 (1.3), 1, [0 – 21]
Memory compared to 10 years ago (N, %)
 Same, better 2496 (48.2%)
 Worse or much worse 2685 (51.8%)
Perceived memory decline 0.4 (0.4), 0.3, [0 – 2]
Cognitive decline affecting functioning 0.1 (0.3), 0, [0 – 2]
Open in a new tab

Among these participants, 25% reported exercising once to twice a week, 27% reported exercising 3 to 4 days a week, and 11% more than five days a week. Participants reported engaging in a mean of 3 enrichment activities a week. With regard to tobacco and alcohol use, 33% of the participants reported having smoked cigarettes in the past, and 37% reported drinking alcohol three or more times a week (Table 1 and Supplementary Table 1). Participants reported a mean of 2.4 diagnosed medical problems, a mean of 0.4 diagnosed mental health problem, and taking a mean of 1.4 medications for these problems (Table 1). Most participants did not have clinically significant current depressive or anxiety symptoms based on the score ranges for these measures (Table 1).

PiL among the EHAS participants ranged from 1.2 to 5, with a mean of 4, and median of 4.1, which is slightly higher than the mean PiL score of the Rush Memory and Aging cohort (mean PiL of 3.7; mean age at enrollment of 80) (Boyle et al., 2010a; Boyle et al., 2012) and consistent with prior studies showing that PiL declines slightly with advancing age.

Among these participants, 52% reported that their memory was worse compared to ten years before (Table 1). However, their median score for perceived memory decline was 0.3, reflecting minimal perceived memory decline (Table 1). Likewise, their median score for cognitive decline that interferes with functioning was 0 indicating that at least 50% of the participants did not have cognitive decline that affected their daily functioning (Table 1). Indeed, 62% of the participants reported no cognitive decline that interfered with daily functioning (i.e. having a score of 0), consistent with a younger age range and still-employed status in these participants.

To investigate whether PiL is a robust predictor of perceived cognitive decline as early as middle age, we performed Elastic Net on a subset of participants aged 40 to 65, whose characteristics are presented in Supplementary Table 2. In this subset of middle-aged participants, mean age was 57, median age 58, and age ranged between 40 and 65.

Pairwise correlation between purpose in life and other sociodemographic characteristics

From pairwise correlations we found that higher PiL was associated with younger age, more years of education, being married, being employed, higher income, more enrichment activities, more frequent walk, more frequent exercise, longer sleep duration a night, less depressive symptoms, less anxiety symptoms, never smoked, and drinking alcohol 3 to 4 times a week (Table 2 and Supplementary Table 1). Likewise, higher PiL was associated with having fewer diagnosed medical problems, fewer diagnosed mental health problems, less perceived memory decline, less cognitive decline that interferes with daily functioning, and less memory decline compared to 10 years ago (Table 2 and Supplementary Table 1). Notably, PiL was not associated with sex or race (Table 2 and Supplementary Table 1).

Table 2:

Pairwise correlations between purpose-in-life and other characteristics among EHAS participants

Characteristics p-value* Direction of association
Age 0.002 Higher PiL ~ lower age
Sex 0.133 N/A
Race 0.632 N/A
Education (years) < 2.2e–16 Higher PiL ~ more years of education
Marital status < 2.2e–16 Higher PiL ~ being married
Employment < 2.2e–16 Higher PiL ~ being employed
Income < 2.2e–16 Higher PiL ~ higher income level
Enrichment activities 3.2e–10 Higher PiL ~ more enrichment activities
Walk frequency < 2.2e–16 Higher PiL ~ higher walk frequency
Exercise frequency < 2.2e–16 Higher PiL ~ higher exercise frequency
Exercise through hobby 6.7e–09 Higher PiL ~ more frequent exercise
Sleep duration (hours) 2.5e–07 Higher PiL ~ longer sleep duration
Depressive symptoms < 2.2e–16 Higher PiL ~ lower depressive symptoms
Anxiety symptoms < 2.2e–16 Higher PiL ~ fewer anxiety symptoms
Ever smoke 1.1e–10 Higher PiL ~ never smoked
Alcohol use 1.5e–06 Higher PiL ~ drink alcohol a few times a week
Number of diagnosed medical problems < 2.2e–16 Higher PiL ~ fewer diagnosed medical problems
Number of diagnosed mental health problems < 2.2e–16 Higher PiL ~ fewer diagnosed mental health problems
Number of medications currently taking 7.9e–06 Higher PiL ~ fewer medications
Perceived memory decline < 2.2e–16 Higher PiL ~ less perceived memory decline
Cognitive decline interfering with functioning < 2.2e–16 Higher PiL ~ less cognitive decline that interferes with functioning
Memory worse compared to 10 years ago < 2.2e–16 Higher PiL ~ less perceived memory decline compared to 10 years ago
Open in a new tab
*

For continuous variables, Spearman correlation test was used. For categorical variables, ANOVA was used. See Supplementary Table for correlation coefficients.

Among participants aged 40 to 65, the pairwise associations between PiL and education, marital status, employment, income, enrichment activities, exercise, current depressive and anxiety symptoms, number of diagnosed medical and mental health problems, number of medications currently taking, perceived memory decline, perceived cognitive decline interfering with daily functioning are similar to those seen in all the participants (Supplementary Table 3).

Association between purpose in life and perceived cognitive decline

We found that higher PiL was associated with less perceived memory decline after adjusting for age, sex, education, current depressive symptoms, current anxiety symptoms, number of diagnosed medical problems, number of diagnosed mental health problems, and employment status (β = −0.03; p=0.018, N=4536). Likewise, higher PiL was associated with less cognitive decline that interferes with daily functioning after adjusting for the same above-mentioned covariates (β = −0.05; p = 8.2E-14, N=4534). While employment status was not predictive of perceived memory decline (β = −0.02; p=0.15), it predicted cognitive decline that interfere with daily functioning (β = −0.03; p = 2.3E-06). These findings suggest that higher PiL is associated with less perceived memory decline regardless of the employment status. However, for more severe cognitive decline such as cognitive decline that interferes with functioning, being employed was associated with less cognitive decline.

Identifying robust predictors of perceived cognitive decline using Elastic Net

Since PiL was associated with the known protective factors of cognition, such as education, exercise, enrichment activities, and inversely associated with known risk factors of cognition including current depressive and anxiety symptoms, diagnosed medical and mental health problems, and alcohol and smoking habit, we sought to address whether PiL is a proxy for these protective factors or is an independent robust protective factor for cognitive stability. Thus, we used Elastic Net, a machine learning approach, to select the most robust risk and protective factors for the outcomes of i) perceived memory decline and ii) cognitive decline that interferes with functioning. We tested the following variables as potential predictors: PiL, age, sex, education, employment status, marital status, income, current depressive symptoms, current anxiety symptoms, walk frequency, exercise frequency, frequency of exercising via hobbies, enrichment activities, smoking status, average sleep duration a night, frequency of alcohol use, typical amount of alcohol use, the number of diagnosed medical problems, the number of diagnosed mental health problems, the number of medications taking currently, maternal history of dementia, and paternal history of dementia.

Elastic Net modeling suggests that PiL is a robust predictor of perceived memory decline and of cognitive decline that interferes with daily functioning (Tables 3 and 4). Particularly, we found the following important predictors for perceived memory decline in the order of higher to lower magnitude of effect - current depressive symptoms, current anxiety symptoms, number of enrichment activities, PiL, number of diagnosed medical problems, number of medications taking currently, number of diagnosed mental health problems, exercise frequency, and employment status (Table 3). For cognitive decline that interferes with daily functioning, we found the following robust predictors in the order of higher to lower effect size - current depressive symptoms, PiL, number of diagnosed medical problems, employment status, number of enrichment activities, current anxiety symptoms, number of diagnosed mental health problems, frequency of walking, and frequency of exercise (Table 4).

Table 3:

Robust predictors of perceived memory decline from Elastic Net (N=2737, α=0.3). Variables with corresponding coefficients are considered important predictors of perceived memory decline.

Variable Standardized coefficient
Purpose-in-life −0.014
Age -
Sex -
Education (years) -
Employment status −0.003
Current depressive symptoms 0.102
Current anxiety symptoms 0.035
Frequency of walking -
Frequency of exercise −0.004
Number of enrichment activities −0.017
Smoking status -
Sleep duration a night -
Marital status -
Income -
Frequency of exercising through hobbies -
Frequency of alcohol use -
Typical alcohol drink count per day -
Number of diagnosed medical problems 0.013
Number of medications currently taking 0.013
Number of diagnosed mental health problems 0.004
Maternal history of dementia -
Paternal history of dementia -
Open in a new tab

“-“ denotes coefficient approaching 0 or non-significant predictor

Table 4:

Robust predictors of perceived cognitive decline that interferes with daily functioning from Elastic Net (N = 2736, α = 1.0). Variables with corresponding coefficients are considered important predictors of cognitive decline that interferes with daily functioning.

Variable Standardized coefficient
Purpose-in-life −0.019
Age -
Sex -
Education (years) -
Employment status −0.009
Current depressive symptoms 0.110
Current anxiety symptoms 0.004
Frequency of walking −0.001
Frequency of exercise 0.0009
Number of enrichment activities −0.006
Smoking status -
Sleep duration a night -
Marital status -
Income -
Frequency of exercising through hobbies -
Frequency of alcohol use -
Typical alcohol drink count per day -
Number of diagnosed medical problems 0.013
Number of medications currently taking -
Number of diagnosed mental health problems 0.003
Maternal history of dementia -
Paternal history of dementia -
Open in a new tab

“-“ denotes coefficient approaching 0 or non-significant predictor

Likewise, in the subset of participants aged 40 to 65, PiL is a robust predictor of perceived memory decline (Supplementary Table 4) and perceived cognitive decline interfering with daily activities (Supplementary Table 5). Interestingly, for perceived memory decline in middle age, all the predictors we provided were considered important based on Elastic Net, and PiL was ranked fourth in its effect size among the 22 predictors (Supplementary Table 4). The magnitude of effect of PiL on perceived memory decline was less than those of current depressive symptoms, marital status, and current anxiety symptoms based on the standardized coefficients from Elastic Net (Supplementary Table 4). For perceived cognitive decline that interferes with daily functioning in middle age, PiL was ranked third among the 12 robust predictors (Supplementary Table 5). The salient predictors of perceived cognitive decline that interferes with functioning in the order of higher to lower magnitude of effect were current depressive symptoms, employment status, PiL, current anxiety symptoms, number of enrichment activities, number of diagnosed mental health problems, number of diagnosed medical problems, frequency of exercise, income, frequency of walking, typical alcohol drink counts per day, and education (Supplementary Table 5).

Discussion

PiL has been suggested to be a novel protective factor for cognitive decline in advanced age in two longitudinal community-based cohorts of older (baseline mean age of 79 and 73, respectively), retired individuals (Boyle et al., 2012; Kim et al., 2019). Here, we investigated whether PiL is also a mitigating factor for cognitive decline among community-based participants with a younger age range (mean age of 63) and 50.5% were employed. The age range of fifties and sixties tends to coincide with the prodromal phase before the onset of MCI or dementia, when some subtle cognitive changes may emerge. During this time period, mild cognitive difficulties may be difficult to distinguish from normal ability using neuropsychological testing due to issues related to test sensitivity and the adequacy of normative data (Rabin et al., 2017). Hence, we evaluated perceived cognitive decline because it has been suggested to be one of the earliest detectable cognitive symptoms during the prodromal phase. We found that higher PiL was indeed associated with less perceived memory decline and less reported cognitive decline that interferes with daily functioning after adjusting for potential confounding factors.

Next, we used a machine learning approach, Elastic Net, to examine whether PiL is simply a proxy for the known protective factors of cognition (i.e. education, exercise, enrichment activities) or an independent protective factor of cognitive decline. We found that higher PiL was an independent, robust predictor of less perceived memory decline and less reported cognitive decline that interferes with daily functioning when considered alongside the known protective factors (education, exercise, enrichment activities) and known risk factors for cognition (depression, anxiety, diagnosed medical problems, mental health problems, smoking, alcohol use, and family history of dementia). These findings hold true in the subset of middle-age participants aged 40 to 65. Together, our findings suggest that PiL is an important protective factor for cognitive decline among employed as well as retired persons as early as middle age.

Beyond being a robust protective factor of cognitive resilience against the effects of aging, PiL has been conceptualized as an important building block for psychological resilience and psychological well-being (Chen et al., 2019; Rutten et al., 2013; Ryff and Keyes, 1995). PiL is not simply the absence of depression; it has been shown to decrease the risk of depression (Wood and Joseph, 2010), post-traumatic stress disorder (Shrira et al., 2015), suicide ideation (Harlow et al., 1986; Kachadourian et al., 2019), alcohol and substance use (Martin et al., 2011; Roos et al., 2015), as well as risk of stroke (Kim et al., 2013b; Yu et al., 2015) and cardiovascular disease (Boehm et al., 2016; Kim et al., 2013a). PiL is a multidimensional construct influenced by both genetic and environmental factors (Keyes et al., 2010). Additionally, PiL is potentially modifiable and may be a target for treatment (Breitbart et al., 2012; Kissane et al., 2019; Park et al., 2019). For instance, a meta-analysis of randomized controlled trials evaluating effects of psychosocial interventions on meaning/purpose in adults with cancer found significant improvements in meaning/purpose with small to medium effect sizes (Park et al., 2019). Our study highlights the potential benefit of enhancing one’s tendency for deriving purpose and meaning in life as a means of protecting against cognitive decline.

Our findings also shed light on the other important predictors of perceived memory decline and cognitive decline that interferes with daily functioning. For perceived memory decline, the important predictors in the order of higher to lower magnitude of effect include current depressive symptoms, current anxiety symptoms, number of enrichment activities, PiL, number of diagnosed medical problems, number of medications taking currently, number of diagnosed mental health problems, frequency of exercise, and employment status. The predominant predictors for perceived cognitive decline that interferes with daily activities in the order of higher to lower effect size include current depressive symptoms, PiL, number of diagnosed medical problems, employment status, number of enrichment activities, current anxiety symptoms, number of diagnosed mental health problems, frequency of walking, and frequency of exercise.

Our findings are consistent with observations that higher PiL was associated with better performance in episodic memory, executive function, and overall cognition in a cross-sectional study of 3489 middle-aged adults (mean age of 56) recruited by the Midlife in the US study (Lewis et al., 2017). While the Midlife in the US study shows an association between higher PiL and better objectively tested cognitive performance in mid-life, our findings suggest that higher PiL is an important predictor of less perceived memory decline as early as in middle age. Taken together, both studies suggest that PiL may be a robust protective factor for cognition as early as late middle age.

Interpretation of our findings should take into consideration the study’s limitations. First, perceived memory decline was not based on cognitive testing but on self-perception. Therefore, perceived memory or cognitive decline can be influenced by mood state such as current depressive or anxiety symptoms. This concern is tempered by our adjustment for current depressive and anxiety symptoms in our regression models and Elastic Net models. Second, this is a cross-sectional study and thus we cannot establish a causal relationship between higher PiL and less perceived memory decline. There is a possibility that decline in cognition could reduce one’s sense of purpose-in-life. Third, the generalizability of the study may be limited given that the majority of the participants were women, Caucasian, and highly educated.

Our study has some notable strengths. For instance, to our knowledge, this is the first study to examine the relationship between PiL and perceived cognitive decline, which constitutes one of the earliest detectable cognitive symptoms in the prodromal phase before the onset of MCI or dementia. In addition, we leveraged a machine learning approach to identify robust predictors of perceived cognitive decline. Furthermore, we performed the analyses in a subset of participants aged 40 to 65, of whom 73% were employed to provide insights into the important effects of PiL on perceived cognitive decline in an earlier life window, i.e. middle age, and among employed persons.

In conclusion, our findings together with the extant literature suggest that PiL is an important and novel protective factor against cognitive decline among middle-aged and advanced aged individuals regardless of their employment status. Future studies are needed to investigate the underlying mechanisms for this protective effect.

Supplementary Material

1

HIGHLIGHTS.

  • Greater purpose in life (PiL) associated with lower self-reported cognitive decline

  • Protective effects of PiL on cognition were observed as early as middle age

  • Machine learning algorithm revealed PiL is a robust protective factor of cognition

Acknowledgements

We express our gratitude to the Emory Healthy Aging Study (EHAS) participants for their time and participation. We gratefully acknowledge the effort and support of the EHAS staff and Investigators.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interest: None

REFERENCES

  1. Abbott RA, Ploubidis GB, Huppert FA, Kuh D, Croudace TJ, 2010. An Evaluation of the Precision of Measurement of Ryff’s Psychological Well-Being Scales in a Population Sample. Social indicators research 97(3), 357–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Abbott RA, Ploubidis GB, Huppert FA, Kuh D, Wadsworth ME, Croudace TJ, 2006. Psychometric evaluation and predictive validity of Ryff’s psychological well-being items in a UK birth cohort sample of women. Health and quality of life outcomes 4, 76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Amariglio RE, Donohue MC, Marshall GA, Rentz DM, Salmon DP, Ferris SH, Karantzoulis S, Aisen PS, Sperling RA, 2015. Tracking early decline in cognitive function in older individuals at risk for Alzheimer disease dementia: the Alzheimer’s Disease Cooperative Study Cognitive Function Instrument. JAMA neurology 72(4), 446–454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Barnes LL, Wilson RS, Bienias JL, de Leon CF, Kim HJ, Buchman AS, Bennett DA, 2007. Correlates of life space in a volunteer cohort of older adults. Experimental aging research 33(1), 77–93. [DOI] [PubMed] [Google Scholar]
  5. Becker E, Orellana Rios CL, Lahmann C, Rucker G, Bauer J, Boeker M, 2018. Anxiety as a risk factor of Alzheimer’s disease and vascular dementia. The British journal of psychiatry : the journal of mental science 213(5), 654–660. [DOI] [PubMed] [Google Scholar]
  6. Bellou V, Belbasis L, Tzoulaki I, Middleton LT, Ioannidis JP, Evangelou E, 2016. Systematic evaluation of the associations between environmental risk factors and dementia: An umbrella review of systematic reviews and meta-analyses. Alzheimer’s & dementia : the journal of the Alzheimer’s Association. [DOI] [PubMed] [Google Scholar]
  7. Bhattarai JJ, Oehlert ME, Multon KD, Sumerall SW, 2019. Dementia and Cognitive Impairment Among U.S. Veterans With a History of MDD or PTSD: A Retrospective Cohort Study Based on Sex and Race. Journal of aging and health 31(8), 1398–1422. [DOI] [PubMed] [Google Scholar]
  8. Bilgel M, An Y, Helphrey J, Elkins W, Gomez G, Wong DF, Davatzikos C, Ferrucci L, Resnick SM, 2018. Effects of amyloid pathology and neurodegeneration on cognitive change in cognitively normal adults. Brain : a journal of neurology 141(8), 2475–2485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Boehm JK, Chen Y, Williams DR, Ryff CD, Kubzansky LD, 2016. Subjective well-being and cardiometabolic health: An 8–11year study of midlife adults. J Psychosom Res 85, 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Boyle PA, Buchman AS, Barnes LL, Bennett DA, 2010a. Effect of a purpose in life on risk of incident Alzheimer disease and mild cognitive impairment in community-dwelling older persons. Arch Gen Psychiatry 67(3), 304–310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Boyle PA, Buchman AS, Bennett DA, 2010b. Purpose in life is associated with a reduced risk of incident disability among community-dwelling older persons. Am J Geriatr Psychiatry 18(12), 1093–1102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Boyle PA, Buchman AS, Wilson RS, Yu L, Schneider JA, Bennett DA, 2012. Effect of purpose in life on the relation between Alzheimer disease pathologic changes on cognitive function in advanced age. Arch Gen Psychiatry 69(5), 499–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Boyle PA, Yang J, Yu L, Leurgans SE, Capuano AW, Schneider JA, Wilson RS, Bennett DA, 2017. Varied effects of age-related neuropathologies on the trajectory of late life cognitive decline. Brain. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Breitbart W, Poppito S, Rosenfeld B, Vickers AJ, Li Y, Abbey J, Olden M, Pessin H, Lichtenthal W, Sjoberg D, Cassileth BR, 2012. Pilot randomized controlled trial of individual meaning-centered psychotherapy for patients with advanced cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 30(12), 1304–1309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chen Y, Kim ES, Koh HK, Frazier AL, VanderWeele TJ, 2019. Sense of Mission and Subsequent Health and Well-Being Among Young Adults: An Outcome-Wide Analysis. American journal of epidemiology 188(4), 664–673. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Deary IJ, Corley J, Gow AJ, Harris SE, Houlihan LM, Marioni RE, Penke L, Rafnsson SB, Starr JM, 2009. Age-associated cognitive decline. British medical bulletin 92, 135–152. [DOI] [PubMed] [Google Scholar]
  17. Dufouil C, Fuhrer R, Alperovitch A, 2005. Subjective cognitive complaints and cognitive decline: consequence or predictor? The epidemiology of vascular aging study. Journal of the American Geriatrics Society 53(4), 616–621. [DOI] [PubMed] [Google Scholar]
  18. Friedman J, Hastie T, Tibshirani R, 2010. Regularization Paths for Generalized Linear Models via Coordinate Descent. Journal of statistical software 33(1), 1–22. [PMC free article] [PubMed] [Google Scholar]
  19. Harlow LL, Newcomb MD, Bentler PM, 1986. Depression, self-derogation, substance use, and suicide ideation: lack of purpose in life as a mediational factor. J Clin Psychol 42(1), 5–21. [DOI] [PubMed] [Google Scholar]
  20. Jessen F, Wiese B, Bachmann C, Eifflaender-Gorfer S, Haller F, Kolsch H, Luck T, Mosch E, van den Bussche H, Wagner M, Wollny A, Zimmermann T, Pentzek M, Riedel-Heller SG, Romberg HP, Weyerer S, Kaduszkiewicz H, Maier W, Bickel H, 2010. Prediction of dementia by subjective memory impairment: effects of severity and temporal association with cognitive impairment. Archives of general psychiatry 67(4), 414–422. [DOI] [PubMed] [Google Scholar]
  21. Kachadourian LK, Tsai J, Harpaz-Rotem I, Southwick SM, Pietrzak RH, 2019. Protective correlates of suicidality among veterans with histories of posttraumatic stress disorder and major depressive disorder: Results from the National Health and Resilience in Veterans Study. Journal of affective disorders 246, 731–737. [DOI] [PubMed] [Google Scholar]
  22. Kaup AR, Nettiksimmons J, LeBlanc ES, Yaffe K, 2015. Memory complaints and risk of cognitive impairment after nearly 2 decades among older women. Neurology 85(21), 1852–1858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Keyes CL, Myers JM, Kendler KS, 2010. The structure of the genetic and environmental influences on mental well-being. Am J Public Health 100(12), 2379–2384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kim ES, Sun JK, Park N, Kubzansky LD, Peterson C, 2013a. Purpose in life and reduced risk of myocardial infarction among older U.S. adults with coronary heart disease: a two-year follow-up. J Behav Med 36(2), 124–133. [DOI] [PubMed] [Google Scholar]
  25. Kim ES, Sun JK, Park N, Peterson C, 2013b. Purpose in life and reduced incidence of stroke in older adults: ‘The Health and Retirement Study’. J Psychosom Res 74(5), 427–432. [DOI] [PubMed] [Google Scholar]
  26. Kim G, Shin SH, Scicolone MA, Parmelee P, 2019. Purpose in Life Protects Against Cognitive Decline Among Older Adults. The American journal of geriatric psychiatry : official journal of the American Association for Geriatric Psychiatry. [DOI] [PubMed] [Google Scholar]
  27. Kissane DW, Lethborg C, Brooker J, Hempton C, Burney S, Michael N, Staples M, Osicka T, Sulistio M, Shapiro J, Hiscock H, 2019. Meaning and Purpose (MaP) therapy II: Feasibility and acceptability from a pilot study in advanced cancer. Palliative & supportive care 17(1), 21–28. [DOI] [PubMed] [Google Scholar]
  28. Koppara A, Wagner M, Lange C, Ernst A, Wiese B, Konig HH, Brettschneider C, Riedel-Heller S, Luppa M, Weyerer S, Werle J, Bickel H, Mosch E, Pentzek M, Fuchs A, Wolfsgruber S, Beauducel A, Scherer M, Maier W, Jessen F, 2015. Cognitive performance before and after the onset of subjective cognitive decline in old age. Alzheimer’s & dementia (Amsterdam, Netherlands) 1(2), 194–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kroenke K, Strine TW, Spitzer RL, Williams JB, Berry JT, Mokdad AH, 2009. The PHQ-8 as a measure of current depression in the general population. Journal of affective disorders 114(1–3), 163–173. [DOI] [PubMed] [Google Scholar]
  30. Lewis NA, Turiano NA, Payne BR, Hill PL, 2017. Purpose in life and cognitive functioning in adulthood. Neuropsychology, development, and cognition. Section B, Aging, neuropsychology and cognition 24(6), 662–671. [DOI] [PubMed] [Google Scholar]
  31. Li C, Neugroschl J, Luo X, Zhu C, Aisen P, Ferris S, Sano M, 2017. The Utility of the Cognitive Function Instrument (CFI) to Detect Cognitive Decline in Non-Demented Older Adults. Journal of Alzheimer’s disease : JAD 60(2), 427–437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Li G, Larson EB, Shofer JB, Crane PK, Gibbons LE, McCormick W, Bowen JD, Thompson ML, 2017. Cognitive Trajectory Changes Over 20 Years Before Dementia Diagnosis: A Large Cohort Study. Journal of the American Geriatrics Society 65(12), 2627–2633. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lowe B, Decker O, Muller S, Brahler E, Schellberg D, Herzog W, Herzberg PY, 2008. Validation and standardization of the Generalized Anxiety Disorder Screener (GAD-7) in the general population. Medical care 46(3), 266–274. [DOI] [PubMed] [Google Scholar]
  34. Martela F, Steger MF, 2016. The three meanings of meaning in life: Distinguishing coherence, purpose, and significance. The Journal of Positive Psychology 11(5), 531–545. [Google Scholar]
  35. Martin RA, MacKinnon S, Johnson J, Rohsenow DJ, 2011. Purpose in life predicts treatment outcome among adult cocaine abusers in treatment. J Subst Abuse Treat 40(2), 183–188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Milgram NW, Siwak-Tapp CT, Araujo J, Head E, 2006. Neuroprotective effects of cognitive enrichment. Ageing research reviews 5(3), 354–369. [DOI] [PubMed] [Google Scholar]
  37. Park CL, Pustejovsky JE, Trevino K, Sherman AC, Esposito C, Berendsen M, Salsman JM, 2019. Effects of psychosocial interventions on meaning and purpose in adults with cancer: A systematic review and meta-analysis. Cancer 125(14), 2383–2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rabin LA, Smart CM, Amariglio RE, 2017. Subjective Cognitive Decline in Preclinical Alzheimer’s Disease. Annual review of clinical psychology 13, 369–396. [DOI] [PubMed] [Google Scholar]
  39. Rajan KB, Wilson RS, Weuve J, Barnes LL, Evans DA, 2015. Cognitive impairment 18 years before clinical diagnosis of Alzheimer disease dementia. Neurology 85(10), 898–904. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Roos CR, Kirouac M, Pearson MR, Fink BC, Witkiewitz K, 2015. Examining temptation to drink from an existential perspective: Associations among temptation, purpose in life, and drinking outcomes. Psychology of addictive behaviors : journal of the Society of Psychologists in Addictive Behaviors 29(3), 716–724. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Rutten BPF, Hammels C, Geschwind N, Menne-Lothmann C, Pishva E, Schruers K, van den Hove D, Kenis G, van Os J, Wichers M, 2013. Resilience in mental health: linking psychological and neurobiological perspectives. Acta Psychiatrica Scandinavica 128(1), 3–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Ryff CD, Keyes CL, 1995. The structure of psychological well-being revisited. J Pers Soc Psychol 69(4), 719–727. [DOI] [PubMed] [Google Scholar]
  43. Scheef L, Spottke A, Daerr M, Joe A, Striepens N, Kolsch H, Popp J, Daamen M, Gorris D, Heneka MT, Boecker H, Biersack HJ, Maier W, Schild HH, Wagner M, Jessen F, 2012. Glucose metabolism, gray matter structure, and memory decline in subjective memory impairment. Neurology 79(13), 1332–1339. [DOI] [PubMed] [Google Scholar]
  44. Scheltens P, Blennow K, Breteler MM, de Strooper B, Frisoni GB, Salloway S, Van der Flier WM, 2016. Alzheimer’s disease. Lancet (London, England) 388(10043), 505–517. [DOI] [PubMed] [Google Scholar]
  45. Shrira A, Shmotkin D, Palgi Y, Soffer Y, Hamama Raz Y, Tal-Katz P, Ben-Ezra M, Benight CC, 2015. How Do Meaning in Life and Positive Affect Relate to Adaptation to Stress? The Case of Firefighters Following the Mount Carmel Forest Fire. The Israel journal of psychiatry and related sciences 52(3), 68–70. [PubMed] [Google Scholar]
  46. Spitzer RL, Kroenke K, Williams JBW, Lowe B, 2006. A Brief Measure for Assessing Generalized Anxiety Disorder: The GAD-7. Arch Intern Med 166(10), 1092–1097. [DOI] [PubMed] [Google Scholar]
  47. Wilson RS, Beckett LA, Barnes LL, Schneider JA, Bach J, Evans DA, Bennett DA, 2002. Individual differences in rates of change in cognitive abilities of older persons. Psychology and aging 17(2), 179–193. [PubMed] [Google Scholar]
  48. Wood AM, Joseph S, 2010. The absence of positive psychological (eudemonic) well-being as a risk factor for depression: a ten year cohort study. J Affect Disord 122(3), 213–217. [DOI] [PubMed] [Google Scholar]
  49. Yu L, Boyle PA, Wilson RS, Levine SR, Schneider JA, Bennett DA, 2015. Purpose in life and cerebral infarcts in community-dwelling older people. Stroke; a journal of cerebral circulation 46(4), 1071–1076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Zaninotto P, Batty GD, Allerhand M, Deary IJ, 2018. Cognitive function trajectories and their determinants in older people: 8 years of follow-up in the English Longitudinal Study of Ageing. Journal of epidemiology and community health 72(8), 685–694. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zika S, Chamberlain K, 1992. On the relation between meaning in life and psychological well-being. British journal of psychology (London, England : 1953) 83 (Pt 1), 133–145. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1546208-supplement-1.docx (24.3KB, docx)

RESOURCES