Abstract
Background
Episodic memory naturally deteriorates with age, and its deficits are widely recognized as the most significant feature and the most sensitive indicator of cognitive decline. It has been suggested that adopting a healthy lifestyle can play a protective role in preserving episodic memory. This study aimed to systematically examine the relationship between lifestyle factors (social activities, leisure activities, physical activities, internet use, smoking, alcohol drinking, and sleep quality) and episodic memory in middle-aged and older adults.
Methods
The current study included 10,392 participants from the Chinese Health and Retirement Longitudinal Survey. A linear mixed model was used to explore the associations between lifestyle factors and episodic memory performance and the age- and sex-specific effects of the association.
Results
Low-frequency alcohol drinking, higher engagement in social, leisure, and physical activities, increased internet use, and improved sleep quality were associated with better episodic memory performance in middle-aged and older adults. Stratified analyses demonstrated that internet use significantly correlated with episodic memory performance in middle-aged adults but not in older adults. On the other hand, sleep quality showed a significant association with episodic memory performance in women but not in men.
Conclusions
This study highlights the association between various lifestyle factors and episodic memory performance, with variations observed based on age and sex. Adopting healthy lifestyle factors can have positive effects on episodic memory in middle-aged adults, emphasizing the importance of adhering to healthy lifestyles from middle age onwards to counteract episodic memory decline.
Keywords: Lifestyle factors, Episodic memory, Linear mixed model, Middle-aged adults, Older adults
Introduction
Older adults commonly experience declines in various cognitive abilities, and deficits in episodic memory are recognized as a prominent and sensitive indicator of cognitive decline [1, 2]. Episodic memory, involving the conscious recollection of personal experiences, holds significance in distinguishing between typical age-related cognitive decline and cognitive decline associated with pathological conditions [3]. Substantial declines in episodic memory are often regarded as early signs of neurodegenerative disorders [4].
Among the numerous factors influencing episodic memory decline, lifestyle factors have gained attention due to their modifiable nature [5]. Engaging in a healthy lifestyle that includes leisure, social, and physical activities may directly protect against episodic memory impairment by enhancing the utilization of brain networks and serving as indirect compensation [6]. For example, maintaining an intellectually stimulating lifestyle has been associated with better episodic memory and reduced risk of Alzheimer’s disease in older adults [7]. Higher-quality social activities have been found to predict better episodic memory in older adults [8]. While studies have examined the association between specific lifestyle factors and episodic memory, comprehensive investigations into the impact of multiple lifestyle factors on episodic memory are limited. Additionally, some studies had small and non-representative samples, which restricted the generalizability of their findings.
Middle age is a critical period in which individuals often experience subtle or significant changes in health, including cardiovascular and hormonal alterations, as well as shifts that can influence the adoption of a healthy lifestyle [9]. Age-related differences in episodic memory can already be detected in middle-aged adults [10]. Engaging in intellectual, physical, and social activities during midlife has been shown to predict cognitive performance in later life, and this prediction remains independent of factors such as years of education, occupation, and late-life lifestyles [11]. While previous studies have primarily focused on the impact of lifestyle factors on episodic memory in older adults, the evidence regarding middle-aged adults is relatively limited.
Both cross-sectional and longitudinal analyses indicated that males exhibit lower episodic memory performance and experience steeper age-related decline compared to females in middle-aged and older adults [12]. Moreover, lifestyle factors have shown differential effects on episodic memory in men and women. For instance, moderate alcohol drinking was associated with a reduced rate of decline in episodic memory among older women compared to older men [13]. In a longitudinal study, higher dietary diversity at baseline was predictive of better episodic memory in older women four years later, while higher dietary quality was associated with better episodic memory in older men during follow-up [14]. These findings highlight the potential necessity and significance of further investigating sex differences when examining the association between lifestyle factors and episodic memory.
Although a large number of studies have previously investigated the association between lifestyle factors and episodic memory, many of them focused on specific types of lifestyle factors with relatively small sample sizes. In this study, we aimed to systematically examine the relationship between multiple lifestyle factors and episodic memory in middle-aged and older adults using a nationally representative follow-up sample of over 10,000 participants. We hypothesized that unhealthy lifestyle factors such as drinking and smoking would have negative impacts on episodic memory, whereas healthy lifestyle factors such as engagement in social activities, leisure activities, and physical exercise, maintaining good sleep quality, and internet use would have positive impacts. Moreover, we further investigated age and sex differences in these associations. We hypothesized that the associations between lifestyle factors and episodic memory would differ by sex (male and female) and age (middle-aged adults and older adults) groups.
Methods
Participants
Participants were from the Chinese Health and Retirement Longitudinal Survey (CHARLS), a prospective cohort study of Chinese residents aged 45 years and older. The CHARLS national baseline survey (wave 1) was conducted in 2011. Follow-up surveys were conducted every two to three years. There have been five waves of data collection to date: wave 2 in 2013, wave 3 in 2015, wave 4 in 2018, and wave 5 in 2020. Ethics approval was obtained from the institutional review board of Peking University, and written informed consent was obtained from all participants. Further details on the study design, sampling, and measures can be found in previous publications [15].
Out of the initial baseline sample of 17,298 participants aged 45 and older, 623 were excluded due to memory-related diseases and stroke, resulting in a final sample size of 16,675 participants at baseline. Follow-up rates were as follows: 2,341 participants lost to follow-up at wave 2, 1,495 at wave 3, 1,465 at wave 4, and 982 at wave 5. The final sample for this study included 10,392 subjects. Detailed demographic information for included and lost participants is listed in Table 1.
Table 1.
Characteristics of demographic variables of the included participants and lost to follow up participants at baseline
| Variables | Participants of all ages | Middle-aged adults | Older adults | p | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Included (n = 10,392) |
Lost to follow-up (n = 6,283) |
p | Included (n = 5,368) |
Lost to follow-up (n = 2,752) |
p | Included (n = 5,024) |
Lost to follow-up (n = 3,531) |
p | |||
| Age | 57.97 ± 8.62 | 61.14 ± 11.25 | < 0.001 | 51.37 ± 4.18 | 50.91 ± 3.81 | < 0.001 | 65.03 ± 6.22 | 69.12 ± 8.28 | < 0.001 | < 0.001 | |
| Female | 5568 (53.61%) | 3019 (48.48%) | < 0.001 | 2920 (54.45%) | 1353 (49.18%) | < 0.001 | 2648 (52.71%) | 1666 (47.25%) | < 0.001 | 0.076 | |
| Education | < 0.001 | < 0.001 | < 0.001 | < 0.001 | |||||||
| Illiterate | 2850 (27.45%) | 1682 (26.85%) | 1040 (19.39%) | 401 (14.56%) | 1810 (36.08%) | 1281 (36.40%) | |||||
| Sishu/home school and below | 4188 (39.67%) | 2325 (37.12%) | 1901 (35.44%) | 875 (31.77%) | 2287 (45.59%) | 1450 (41.20%) | |||||
| Elementary school | 2198 (21.17%) | 1260 (20.12%) | 1573 (29.33%) | 802 (29.12%) | 625 (12.46%) | 458 (13.02%) | |||||
| High school and above | 1145 (13.92%) | 997 (15.92%) | 850 (15.85%) | 667 (24.22%) | 295 (5.88%) | 330 (9.38%) | |||||
| Married | 9359 (90.06%) | 5181 (82.46) | < 0.001 | 5113 (95.25%) | 2563 (93.13%) | < 0.001 | 4246 (84.51%) | 2618 (74.14) | < 0.001 | < 0.001 | |
| Rural Hukou | 8671 (83.46%) | 4281 (68.22%) | < 0.001 | 4544 (84.68%) | 1877 (68.28%) | < 0.001 | 4127 (82.16%) | 2404 (68.18%) | < 0.001 | 0.001 | |
| Hypertension | 2248 (21.83%) | 1638 (26.54%) | < 0.001 | 901 (16.95%) | 486 (18.04%) | 0.221 | 1347 (27.04%) | 1152 (33.13%) | < 0.001 | < 0.001 | |
| Dyslipidemia | 874 (8.61%) | 551 (9.07%) | 0.316 | 222 (8.38%) | 407 (7.76%) | 0.335 | 467 (9.52%) | 329 (9.60%) | 0.899 | 0.001 | |
| Diabetes | 501 (4.88%) | 391 (6.37%) | < 0.001 | 220 (4.15%) | 116 (4.32%) | 0.772 | 281 (5.66%) | 275 (7.76%) | < 0.001 | < 0.001 | |
| IADL | 0.61 ± 1.79 | 1.15 ± 2.83 | < 0.001 | 0.40 ± 1.45 | 0.47 ± 1.67 | < 0.001 | 0.84 ± 2.07 | 1.67 ± 3.38 | < 0.001 | < 0.001 | |
| Depression | 17.07 ± 7.47 | 15.99 ± 8.51 | < 0.001 | 16.40 ± 7.46 | 14.73 ± 8.39 | 0.070 | 17.78 ± 7.42 | 16.97 ± 8.49 | < 0.001 | < 0.001 | |
| Employment status | < 0.001 | < 0.001 | < 0.001 | < 0.001 | |||||||
| Retired | 890 (8.61%) | 1050 (17.03%) | 213 (3.99%) | 221 (8.22%) | 677 (13.54%) | 829 (23.84%) | |||||
| Employed | 7588 (73.78%) | 3261 (52.89%) | 4467 (83.65%) | 2000 (74.38%) | 3121 (62.41%) | 1261(36.27%) | |||||
| Unemployed | 1863 (18.02%) | 1855 (30.08%) | 660 (12.36%) | 468 (17.40%) | 1203 (24.06%) | 1387 (38.89%) | |||||
| Episodic memory | 4.07 ± 1.66 | 4.10 ± 1.75 | 0.254 | 4.34 ± 1.65 | 4.56 ± 1.70 | < 0.001 | 3.78 ± 1.62 | 3.74 ± 1.60 | 0.377 | < 0.001 | |
| Smoking | < 0.001 | 0.132 | < 0.001 | ||||||||
| Never smoking | 6374 (63.25%) | 3532 (60.21%) | 3381 (65.55%) | 1592 (63.25%) | 2993 (60.83%) | 1940 (57.93%) | |||||
| Ever smoking | 747 (7.41%) | 586 (9.99%) | 294 (5.70%) | 158 (6.28%) | 453 (9.21%) | 428 (12.78%) | |||||
| Current smoking | 2957 (29.34%) | 1784 (30.41%) | 1483 (28.75%) | 767 (30.47) | 1474 (29.96) | 981 (29.29%) | |||||
| Alcohol drinking | 0.875 | 0.005 | 0.215 | < 0.001 | |||||||
| Not drinking | 6895 (66.61%) | 4100 (66.25%) | 3472 (64.96%) | 1659 (61.24%) | 3423 (68.36%) | 2441 (70.14%) | |||||
| Low frequency drinking | 826 (7.98%) | 494 (7.98%) | 468 (8.76%) | 262 (9.67%) | 358 (7.15%) | 232 (6.67%) | |||||
| High frequency drinking | 2631 (25.42%) | 1595 (25.77%) | 1405 (26.29%) | 788 (29.09%) | 1226 (24.49%) | 807 (23.19%) | |||||
| Social activities | 0.45 ± 0.63 | 0.42 ± 0.62 | 0.003 | 0.47 ± 0.65 | 0.44 ± 0.64 | 0.042 | 0.42 ± 0.62 | 0.40 ± 0.62 | 0.118 | 0.005 | |
| Leisure activities | 0.33 ± 0.78 | 0.36 ± 0.83 | 0.050 | 0.35 ± 0.76 | 0.37 ± 0.79 | 0.195 | 0.32 ± 0.80 | 0.35 ± 0.86 | 0.091 | 0.125 | |
| Physical activities | 0.12 ± 0.56 | 0.23 ± 0.77 | < 0.001 | 0.11 ± 0.53 | 0.24 ± 0.77 | < 0.001 | 0.13 ± 0.59 | 0.23 ± 0.77 | < 0.001 | < 0.001 | |
| Internet use | 0.05 ± 0.36 | 0.11 ± 0.54 | < 0.001 | 0.07 ± 0.42 | 0.19 ± 0.72 | < 0.001 | 0.03 ± 0.26 | 0.04 ± 0.35 | 0.010 | < 0.001 | |
| Sleep quality | 1.96 ± 1.19 | 1.94 ± 1.19 | 0.241 | 2.02 ± 1.17 | 2.00 ± 1.17 | 0.656 | 1.90 ± 1.20 | 1.89 ± 1.20 | 0.596 | < 0.001 | |
Notes: IADL, instrumental activities of daily living. The p in the last column represents the differences between included middle-aged adults and older adults.
Measures
Episodic memory
Participants were asked to immediately recall as many Chinese nouns as possible from a list of ten words (immediate word recall). Researchers recorded the total number of words recalled by each participant in immediate memory [16]. It is important to note that the immediate recall of the word list was assessed only once in wave 1, wave 2, and wave 3. However, in wave 4 and wave 5, three immediate recall measures were conducted, with the first immediate recall test being the same as in the previous waves. To ensure consistency from waves 1 to wave 5 in the data analysis, only the first immediate recall of wave 4 and wave 5 is included in the current study.
Lifestyle factors
In CHARLS, respondents were asked to indicate activities they had participated in over the last month. Activities with a participation rate of at least 2% were included [17]. Social activities included interacting with friends and providing free help to family, friends, or neighbors. Leisure activities included playing mahjong, chess, cards, or joining a club of associations. Physical activities included participating in sports, social or other types of clubs. Internet use included the specific use status of the network. For each activity, participants were asked to answer whether they participated and how often on a four-point scale. The higher the score is, the more often the participant participates.
For sleep quality, participants were asked to report the frequency of sleep disturbances in the last week. The question is scored on a four-point scale. The higher the score is, the better the quality of sleep [18].
Two questions (“Have you ever smoked?” and “Do you still have the habit or have you totally quit?”) were used to measure smoking. According to the answers, participants were classified as never-smokers, ever-smokers, or current-smokers. For alcohol drinking, participants were asked “Have you ever drunk alcohol in the past? How often?”, According to the answers, participants were classified as never-drinkers, low-frequency alcohol drinkers (i.e., less than once a month), or high-frequency alcohol drinkers (i.e., more than once a month).
Confounding variables
Demographic status and health status associated with episodic memory were used as confounding variables [19]. Demographic status includes age, sex, marital status (married and nonmarried), education (illiterate, sishu/home school and below, elementary school, high school and above), and hukou status (rural and nonrural). Health status was composed of chronic diseases (including hypertension, hyperlipidemia, and diabetes), and instrumental activities of daily living (IADL). The IADL includes 5 items required for independent living (e.g., shopping, cooking) and is measured on a 4-point scale: no difficulty, difficult but still able to do, difficult and need help, and unable to do. In addition, depression and employment status were also included as confounding variables. The depression was evaluated through the Center for Epidemiologic Studies Depression scale (CES-D). The higher the score, the more depressed the participant is. The employment status was classified as employed, unemployed and retired.
Statistical analyses
Linear mixed model analysis
Participants were divided into middle-aged adults (45–64 years) and older adults (65 years and older) according to their age at the last follow-up. We used t tests and chi-square tests to examine the baseline (2011/12 wave) differences between the middle-aged and older participants. Linear mixed model (LMM), a type of regression analysis that processes longitudinal data using all available data during follow-up, addresses differences in the length of follow-up and accounts for the fact that repeated measurements of the same individual are correlated [20]. The LMM was applied to longitudinally explore the association between lifestyle factors and episodic memory in middle-aged and older adults at five separate time points over a 9-year period. Specifically, follow-up time was set as repeated measures, episodic memory as dependent variables, smoking and drinking as categorical independent variables, and other lifestyles factors as continuous variables. Each participant was set as a random effect, and all lifestyle factors were set as fixed effects. Only the main effects of each lifestyle factor were considered. Due to complexity of interpreting numerous factors, the interaction between lifestyle factors is not included in the model. The Restricted Maximum Likelihood (REML) method, handling missing data under the assumption that the data are missing at random, was used for parameter estimation. Cases with missing values were not excluded to ensure the likelihood function was computed using all available data.
Associations between lifestyle factors and episodic memory were analyzed in four different models that controlled for different potential confounders. In Model 1, age, follow-up time, and sex were adjusted. Model 2 was additionally adjusted for education, marital status, and hukou status on the basis of Model 1. Model 3 was additionally adjusted for hypertension, dyslipidemia, diabetes, and IADL on the basis of Model 2. Model 4 was additionally adjusted for depression and employment status on the basis of Model 3. The model fit was evaluated using the Akaike information criterion (AIC) and Bayesian information criterion (BIC), and the lower the AIC and BIC values were, the better the fit. The statistical analyses were performed using SPSS statistical software package version 22.0, and all p values were two-sided.
Exploratory analysis
Given the inclusion of both harmful and beneficial lifestyle factors, further exploratory analysis was conducted using data from wave 5 (the latest wave of data) to examine the effects of different types of lifestyle factors on episodic memory. Specifically, lifestyle factors were first classified into three categories. Smoking and high-frequency alcohol drinking were classified as harmful factors; social activities, leisure activities, physical activities, and internet use were classified as beneficial factors; ever smoking, low-frequency alcohol drinking, and sleep quality were classified as neutral factors due to their controversial effects on cognitive functions in previous studies [21–23]. Participants were then divided into three categories based on their lifestyle type. Those who had at least one harmful lifestyle factor without any beneficial lifestyle factors are classified as harmful lifestyle group (HLG), those who had at least one beneficial lifestyle factor without any harmful lifestyle factors are classified as beneficial lifestyle group (BLG), and those who had at least one harmful lifestyle factor and one beneficial lifestyle factor are classified as mixed lifestyle group (MLG). One-way analysis of variance (ANOVA) was conducted to examine whether there were significant differences on episodic memory among these three groups. Post hoc independent t-tests were employed to investigate the intergroup differences.
Results
Sample characteristics
At baseline, compared with middle-aged adults, older adults were significantly older, more depressed, less independent in daily life, had poor episodic memory, engaged in more physical activities, engaged in fewer social activities, used the internet less, and had worse sleep quality (ps < 0.01). Older adults had significantly higher rates of hypertension, dyslipidemia, diabetes, retirement, and not drinking (ps < 0.01), while rates of high school education and above, married, rural Hukou, and never smoking were significantly lower (ps < 0.01). There were no significant group differences in gender and leisure activities (ps ≥ 0.05) (Table 1).
We also compared differences in demographic variables between the included participants and participants lost to follow-up. For all participants, compared to those who were lost to follow-up, the included participants were significantly younger, more depressed, more independent in daily life, engaged in more social activities, participated in fewer physical activities, and less internet use (ps < 0.01). Among the included participants, the proportions of female, married, rural Hukou, and never smokers were significantly higher (ps < 0.001), while the proportion of high school education or above, hypertension, diabetes, and retirees were significantly lower (ps < 0.001). There were no significant group differences in dyslipidemia, alcohol drinking, episodic memory, engagement in leisure activities, and sleep quality (ps ≥ 0.05). For middle-aged participants, compared with those who were lost to follow-up, those included were significantly older, more independent in daily life, engaged in more social activities and fewer physical activities, used less internet use, and had lower episodic memory scores (ps < 0.05). Among the included middle-aged participants, the proportion of female, married, rural Hukou, and non-drinkers were significantly higher (ps < 0.01), while the proportion of high school education or above, and retirees was significantly lower (ps < 0.001); there were no significant group differences in the hypertension, dyslipidemia, diabetes, depression, smoking, leisure activities, and sleep quality (ps ≥ 0.05). For older participants, the included participants were significantly younger, more depressed, more independent in daily life, engaged in fewer physical activities, and less internet use compared to those lost to follow-up (ps < 0.05). Among the included participants, the proportion of female, married, rural Hukou, and non-smokers was significantly higher (ps < 0.001), while the proportion of high school education or above, hypertension, diabetes, and retirees were significantly lower (ps < 0.001). There were no significant group differences in dyslipidemia, alcohol drinking, episodic memory, social activities, leisure activities, and sleep quality (ps ≥ 0.05) (Table 1).
Association between lifestyle factors and episodic memory
LMM analyses showed that AIC and BIC decreased gradually as the number of covariates in the four models increased, indicating that Model 4 was the best-fitting model. For the overall participants, low-frequency alcohol drinkers have better episodic memory performance than never-drinkers (B = 0.101, 95% CI: [0.030, 0.173], p = .005). The frequency of participation in social activities (B = 0.098, 95% CI: [0.069, 0.127], p < .001), leisure activities (B = 0.095, 95% CI: [0.069, 0.121], p < .001), physical activities (B = 0.114, 95% CI: [0.084, 0.143], p < .001), internet use (B = 0.171, 95% CI: [0.130, 0.212], p < .001), and sleep quality (B = 0.024, 95% CI: [0.006, 0.042], p = .010) are positively associated with episodic memory performance (Table 2). Smoking and high-frequency drinking are not significantly associated with episodic memory performance (Table 2).
Table 2.
Association between lifestyle factors and episodic memory in middle-aged and older adults (n = 10,392)
| Variables | Model 1a | Model 2b | Model 3c | Model 4d | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| B | 95% CI | p | B | 95% CI | p | B | 95% CI | p | B | 95% CI | p | |
| Intercept | 1.064 | 0.739 | 0.730 | 0.715 | ||||||||
| Smoking | ||||||||||||
| Never smoking | Reference | Reference | Reference | Reference | ||||||||
| Ever smoking | 0.044 | -0.081, 0.090 | 0.919 | 0.017 | -0.064, 0.099 | 0.676 | 0.043 | -0.040, 0.126 | 0.313 | 0.051 | -0.032, 0.134 | 0.232 |
| Current smoking | -0.117 | -0.192, -0.426 | 0.002 | -0.062 | -0.132, 0.007 | 0.080 | -0.046 | -0.116, 0.025 | 0.206 | -0.034 | -0.104, 0.036 | 0.343 |
| Alcohol drinking | ||||||||||||
| Never drinking | Reference | Reference | Reference | Reference | ||||||||
| Low frequency drinking | 0.138 | 0.068, 0.209 | < 0.001 | 0.104 | 0.034, 0.173 | 0.003 | 0.097 | 0.026, 0.168 | 0.008 | 0.101 | 0.030, 0.173 | 0.005 |
| High frequency drinking | 0.010 | -0.048, 0.070 | 0.744 | -0.003 | -0.058, 0.052 | 0.907 | -0.020 | -0.076, 0.036 | 0.474 | -0.020 | -0.076, 0.036 | 0.480 |
| Social activities | 0.113 | 0.085, 0.141 | < 0.001 | 0.111 | 0.083, 0.138 | < 0.001 | 0.102 | 0.073, 0.131 | < 0.001 | 0.098 | 0.069, 0.127 | < 0.001 |
| Leisure activities | 0.157 | 0.131, 0.182 | < 0.001 | 0.112 | 0.087, 0.136 | < 0.001 | 0.105 | 0.079, 0.131 | < 0.001 | 0.095 | 0.069, 0.121 | < 0.001 |
| Physical activities | 0.172 | 0.143, 0.201 | < 0.001 | 0.123 | 0.095, 0.151 | < 0.001 | 0.125 | 0.096, 0.155 | < 0.001 | 0.114 | 0.084, 0.143 | < 0.001 |
| Internet use | 0.281 | 0.242, 0.319 | < 0.001 | 0.169 | 0.131, 0.207 | < 0.001 | 0.176 | 0.135, 0.217 | < 0.001 | 0.171 | 0.130, 0.212 | < 0.001 |
| Sleep quality | 0.791 | 0.063, 0.095 | < 0.001 | 0.070 | 0.055, 0.086 | < 0.001 | 0.059 | 0.043, 0.076 | < 0.001 | 0.024 | 0.006, 0.042 | 0.010 |
Notes: B represents an unstandardized coefficient.a AIC = 119,033.598, BIC = 119,054.291; b AIC = 115,523.492, BIC = 115,544.159;
c AIC = 105,809.399, BIC = 105,829.894; d AIC = 105,604.962, BIC = 105,621.454.
Age- and sex-specific association between lifestyle factors and episodic memory
We further performed LMM analyses by sex in middle-aged and older adults to explore age- and sex-specific associations between lifestyle factors and episodic memory, and the results showed that Model 4 had the lowest AIC and BIC for all conditions. Therefore, we report the results of Model 4 here. Smoking and drinking were not significantly associated with episodic memory performance in middle-aged women, middle-aged men, older women, and older men (ps > 0.05). The frequency of participation in social activities, leisure activities, and physical activities were significantly associated with episodic memory performance in all four subgroups (ps < 0.05) (Table 3). Frequency of internet use and sleep quality also presented different association patterns with episodic memory performance in the age- and sex-stratified analyses. The association between the frequency of internet use and episodic memory performance was significant in both middle-aged women and men (ps < 0.001), but the association is not significant in both older women and men (ps > 0.05). Sleep quality is significantly associated with episodic memory performance in both middle-aged women and older women (ps < 0.05); however, the association is not significant in middle-aged men and older men (ps > 0.05) (Table 3).
Table 3.
Age- and sex-stratified analyses of the relationship between lifestyle factors and episodic memory
| Variables | Middle-aged women (n = 2,920) a |
Middle-aged men (n = 2,443) b |
Older women (n = 2,648) c |
Older men (n = 2,376) d |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| B | 95% CI | p | B | 95% CI | p | B | 95% CI | p | B | 95% CI | p | |
| Intercept | 0.778 | 0.710 | 0.673 | 0.653 | ||||||||
| Smoking | ||||||||||||
| Never smoking | Reference | Reference | Reference | Reference | ||||||||
| Ever smoking | 0.185 | -0.162, 0.291 | 0.307 | -0.037 | -0.171, 0.098 | 0.201 | 0.031 | -0.213, 0.284 | 0.805 | 0.091 | -0.036, 0.218 | 0.158 |
| Current smoking | 0.064 | -0.170, 0.540 | 0.577 | -0.074 | -0.198, 0.040 | 0.592 | 0.114 | -0.074, 0.302 | 0.233 | -0.071 | -0.187, 0.045 | 0.228 |
| Alcohol drinking | ||||||||||||
| Never drinking | Reference | Reference | Reference | Reference | ||||||||
| Low frequency drinking | 0.084 | -0.048, 0.216 | 0.212 | 0.102 | -0.034, 0.239 | 0.461 | 0.122 | -0.040, 0.284 | 0.630 | 0.124 | -0.017, 0.266 | 0.085 |
| High frequency drinking | -0.028 | -0.161, 0.106 | 0.687 | 0.036 | -0.060, 0.133 | 0.140 | -0.034 | -0.175, 0.106 | 0.141 | -0.069 | -0.166, 0.027 | 0.159 |
| Social activities | 0.085 | 0.036, 0.134 | 0.001 | 0.080 | 0.016, 0.144 | 0.015 | 0.127 | 0.072, 0.182 | < 0.001 | 0.132 | 0.066, 0.197 | < 0.001 |
| Leisure activities | 0.088 | 0.039, 0.136 | < 0.001 | 0.065 | 0.012, 0.117 | 0.015 | 0.123 | 0.067, 0.179 | < 0.001 | 0.107 | 0.058, 0.156 | < 0.001 |
| Physical activities | 0.087 | 0.038, 0.135 | < 0.001 | 0.123 | 0.049, 0.197 | 0.001 | 0.128 | 0.069, 0.187 | < 0.001 | 0.078 | 0.013, 0.143 | 0.019 |
| Internet use | 0.168 | 0.107, 0.229 | < 0.001 | 0.163 | 0.094, 0.231 | < 0.001 | 0.091 | -0.057, 0.239 | 0.229 | 0.107 | -0.021, 0.234 | 0.101 |
| Sleep quality | 0.039 | 0.007, 0.070 | 0.017 | -0.021 | -0.063, 0.022 | 0.336 | 0.037 | 0.002, 0.071 | 0.038 | 0.010 | -0.031, 0.052 | 0.619 |
Notes: Models were controlled for age, gender, follow time, education, marriage, hukou, hypertension, dyslipidemia, diabetes, IADL, depression and employment status. B represents an unstandardized coefficient. aAIC = 35,623.579, BIC = 35,637.876; bAIC = 20,797.267, BIC = 21,810.535; cAIC = 28,793.391, BIC = 28,807.260; dAIC = 20,384.291, BIC = 20,397.511.
Influence of different types of lifestyle factors on episodic memory
In the exploratory analysis, HLG included 4,303 participants, BLG included 2,549 participants, and MLG included 1,895 participants. One-way ANOVA suggested significant differences on episodic memory among HLG, BLG and MLG (F(2, 8744) = 101.22, p < .001). Independent t-tests showed that episodic memory was significantly lower in HLG than in BLG (t = -11.43, p < .001) and MLG (t = -11.80, p < .001). There was no significant group difference between BLG and MLG (t = -1.34, p = .257).
Discussion
This study investigated the association between multiple lifestyle factors and episodic memory in a nationwide population-based survey of middle-aged and older adults. Results found that low-frequency alcohol drinking, higher levels of social activities, leisure activities, physical activities, internet use, and better sleep quality were associated with better episodic memory. Age- and sex-stratified analyses showed that the beneficial effects of lifestyle factors on episodic memory varied by age and sex.
The current study found that higher leisure and social activities improved episodic memory in middle-aged and older adults, consistent with previous findings [24, 25]. According to the Scaffolding Theory of Aging and Cognition (STAC), leisure and social activities, as forms of neural enrichment, can directly enhance or maintain brain structure and function by fostering efficient connectivity, boosting cortical thickness, synaptic density, and other markers of brain health [26, 27]. These mechanisms may underlie episodic memory enhancement due to increased participation in leisure and social activities. Increased participation in physical activities was associated with better episodic memory, potentially due to its positive impact on the cardiovascular system. Previous research has shown that physical activity can enhance the production of neurotrophic and vascular growth factors, which contributes to the maintenance of cardiovascular fitness [28]. Moreover, physical activity induces adaptive changes in brain structure and function, supporting cognitive function in older adults [29]. Internet use, particularly social media, serves as cognitive stimulation and promotes cognitive health through increased engagement and social interaction in middle-aged and older adults [30]. Access to information and services online reduces social isolation and enhances cognitive function, reducing the risk of cognitive impairment in this population [30, 31].
The current study revealed a positive correlation between sleep quality and episodic memory, consistent with previous research. A cross-sectional study supported this by demonstrating a positive link between sleep quality and cognitive function [32]. Poor sleep quality can hinder sleep-dependent memory consolidation and result in impaired memory performance [33, 34]. This relationship may be attributed to the impact of sleep quality on brain aging, as inadequate sleep can contribute to brain atrophy and exacerbate cognitive decline [34, 35]. Insufficient sleep disrupts circadian rhythms regulated by gene expression in various regions of the brain such as the frontal, thalamic, hypothalamic, and locus coeruleus, consequently affecting hippocampal function [36, 37], which in turn leads to cognitive decline.
We also found that individuals who engaged in low-frequency alcohol drinking exhibited better performance in episodic memory compared to non-drinkers in the overall participants. A recent study also reported a lower rate of cognitive impairment among low-frequency drinkers [38]. Low-frequency alcohol drinking may help mitigate the risk of stroke and congestive heart failure while promoting brain health [39]. However, it should be noted that the association became insignificant in all four subgroups in the age- and sex-stratified analyses. Therefore, the association between low-frequency alcohol drinking and episodic memory should be interpreted with caution.
There was an age-specific effect of internet use. While internet use exhibited a significant association with episodic memory in middle-aged women and men, this relationship was not significant among older individuals. Age differences exist regarding the use of internet. Middle-aged individuals generally demonstrate greater proficiency in internet use compared to older adults, leading to more effective and engaging online experiences that can positively impact cognitive function such as episodic memory [40]. Conversely, older adults may encounter greater challenges in navigating the internet and exhibit lower rates of internet adoption compared to other age groups [41], potentially undermining the cognitive benefits of internet use. There was an sex-specific effect on the sleep quality. While sleep quality shows a significant association with episodic memory in middle-aged women and older women, the association was not significant in middle-aged and older men. Recent research on sleep deprivation supports these findings, as women exhibited impaired working memory performance after one night of sleep loss, while men were unaffected [42]. These results suggest that women’s cognitive function may be more sensitive to sleep quality. This perspective offers insight into the sex-specific association between sleep quality and cognitive function in women. However, given the limited number of studies exploring sex-specific differences in sleep quality and cognitive function in this population, further research is warranted in the future.
While the current study included both beneficial and harmful lifestyle factors, it revealed extensive positive effects of lifestyle factors on episodic memory. Surprisingly, no significant negative impacts of harmful lifestyle factors, such as current smoking and high-frequency alcohol drinking, were observed in the overall participants or in specific age- and sex-specific groups. However, the exploratory analysis using the wave 5 data revealed negative impacts of harmful lifestyle factors on episodic memory. Specifically, episodic memory was significantly lower in HLG than in BLG and MLG, while no significant difference was observed between the BLG and MLG, suggesting that participation in beneficial lifestyle factors may counteract harmful factors in damaging episodic memory.
Several limitations of the present study should be mentioned. Firstly, previous research has indicated that dietary diversity plays a significant role in lifestyle and may influence cognitive function in middle-aged and older adults [43]. Unfortunately, the database utilized in this study did not include dietary diversity as a factor, preventing us from examining its association with episodic memory. Secondly, the classification of CHARLS data on smoking and alcohol drinking is very broad. Therefore, in this study, smoking and alcohol drinking could only be analyzed as categorical variables. Using a more detailed classification, further investigation of the effects of smoking and drinking on episodic memory is warranted for future studies.
Conclusion
In this extensive population-based study, we observed that low-frequency alcohol drinking, increased engagement in social, leisure, and physical activities, higher levels of internet use, and improved sleep quality were linked to enhanced episodic memory in middle-aged and older adults. These findings underscore the importance of adopting a healthy lifestyle starting from middle age to preserve cognitive performance and prevent cognitive decline. Further interventional studies are needed to establish the causal relationship between lifestyle factors and cognitive function in this population.
Acknowledgements
The authors would like to thank the Centre for Healthy Aging and Development at Peking University for making the CHARLS data available to the public.
Author contributions
P. Wang wrote the paper. C. Zhou performed data collection and data analysis. H.-J. Li conceived the idea and wrote the manuscript.
Funding
This work was supported by the National Natural Science Foundation of China (31871143), the China Postdoctoral Science Foundation (2023M740297), and Postdoctoral Fellowship Program of CPSF (GZB20230074). The funders played no role in the design, conduct, or reporting of this study.
Data availability
Data used in the current study was publicly available and could be obtained at http://charls.pku.edu.cn/.
Declarations
Ethics approval and consent to participant
Ethics approval was obtained from the institutional review board of Peking University, and written informed consent was obtained from all participants. All methods were carried out in accordance with guidelines and regulations of the Declaration of Helsinki.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Hartshorne JK, Germine LT. When does cognitive functioning peak? The asynchronous rise and fall of different cognitive abilities across the life span. Psychol Sci. 2015;26(4):433–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Moscovitch M, Winocur G. The neuropsychology of memory and aging. In: Fregus IM, Craik TA, Salthouse, editors. The handbook of aging and cognition. Hillsdale, NJ, US: Lawrence Erlbaum Associates, Inc; 1992. pp. 315–72. [Google Scholar]
- 3.Bäckman L, Small BJ, Fratiglioni L. Stability of the preclinical episodic memory deficit in Alzheimer’s disease. Brain. 2001;124:96–102. [DOI] [PubMed] [Google Scholar]
- 4.Boraxbekk CJ, Lundquist A, Nordin A, Nyberg L, Nilsson LG, Adolfsson R. Free recall episodic memory performance predicts dementia ten years prior to clinical diagnosis: findings from the Betula Longitudinal Study. Dement Geriatr Cogn Dis Extra. 2015;5(2):191–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Song S, Stern Y, Gu Y. Modifiable lifestyle factors and cognitive reserve: a systematic review of current evidence. Ageing Res Rev. 2022;74:101551. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Arenaza-Urquijo EM, Wirth M, Chetelat G. Cognitive reserve and lifestyle: moving towards preclinical Alzheimer’s disease. Front Aging Neurosci. 2015;7. [DOI] [PMC free article] [PubMed]
- 7.Hertzog C, Kramer AF, Wilson RS, Lindenberger U. Enrichment effects on adult cognitive development: can the functional capacity of older adults be preserved and enhanced? Psychol Sci Public Interest. 2008;9(1):1–65. [DOI] [PubMed] [Google Scholar]
- 8.Luo M, Edelsbrunner PA, Siebert JS, Martin M, Aschwanden D. Longitudinal within-person associations between quality of social relations, structure of social relations, and cognitive functioning in older age. J Gerontol Ser B-Psychol Sci Soc Sci. 2021;76(10):1960–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bookheimer SY, Salat DH, Terpstra M, Ances BM, Barch DM, Buckner RL, et al. The lifespan human Connectome Project in aging: an overview. NeuroImage. 2019;185:335–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Van Etten E, Sumida C, Graves L, Holden H, Lopez F, Mustafa A, et al. Age-related differences in memory for who, when, and where are detectable in middle-aged adults. Aging Neuropsychol Cogn. 2022;29(5):733–45. [DOI] [PubMed] [Google Scholar]
- 11.Chan D, Shafto M, Kievit R, Matthews F, Spink M, Valenzuela M, et al. Lifestyle activities in mid-life contribute to cognitive reserve in late-life, independent of education, occupation, and late-life activities. Neurobiol Aging. 2018;70:180–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lundervold AJ, Wollschlager D, Wehling E. Age and sex related changes in episodic memory function in middle aged and older adults. Scand J Psychol. 2014;55(3):225–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wardzala C, Murchison C, Loftis JM, Schenning KJ, Mattek N, Woltjer R, et al. Sex differences in the association of alcohol with cognitive decline and brain pathology in a cohort of octogenarians. Psychopharmacology. 2018;235(3):761–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Milte CM, Ball K, Crawford D, McNaughton SA. Diet quality and cognitive function in mid-aged and older men and women. BMC Geriatr. 2019;19(1):361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhao Y, Hu Y, Smith JP, Strauss J, Yang G. Cohort Profile: the China Health and Retirement Longitudinal Study (CHARLS). Int J Epidemiol. 2014;43(1):61–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lei X, Smith JP, Sun X, Zhao Y. Gender differences in cognition in China and reasons for change over time: evidence from CHARLS. J Econ Ageing. 2014;4:46–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kesavayuth D, Liang Y, Zikos V. An active lifestyle and cognitive function: evidence from China. J Econ Ageing. 2018;12:183–91. [Google Scholar]
- 18.Beard JR, Si Y, Liu Z, Chenoweth L, Hanewald K. Intrinsic capacity: validation of a new WHO concept for healthy aging in a longitudinal Chinese study. J Gerontol Ser A-Biol Sci Med Sci. 2022;77(1):94–100. [DOI] [PubMed] [Google Scholar]
- 19.Li W, Sun N, Kondracki A, Sun W. Sex, sleep duration, and the association of cognition: findings from the China Health and Retirement Longitudinal Study. Int J Environ Res Public Health. 2021;18:19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Proust-Lima C, Dartigues J-F, Jacqmin-Gadda H. Misuse of the linear mixed model when evaluating risk factors of cognitive decline. Am J Epidemiol. 2011;174(9):1077–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ashare RL, Falcone M, Lerman C. Cognitive function during nicotine withdrawal: implications for nicotine dependence treatment. Neuropharmacology. 2014;76(0):581–91. Pt B(0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zhang R, Shen L, Miles T, Shen Y, Cordero J, Qi Y, et al. Association of low to moderate alcohol drinking with cognitive functions from middle to older age among US adults. JAMA Netw Open. 2020;3(6):e207922. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kong J, Zhou L, Li X, Ren Q. Sleep disorders affect cognitive function in adults: an overview of systematic reviews and meta-analyses. Sleep Biol Rhythms. 2023;21(2):133–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mao C, Li Z-H, Lv Y-B, Gao X, Kraus VB, Zhou J-H, et al. Specific leisure activities and cognitive functions among the oldest-old: the Chinese longitudinal healthy longevity survey. J Gerontol Ser A-Biol Sci Med Sci. 2020;75(4):739–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Marioni RE, Proust-Lima C, Amieva H, Brayne C, Matthews FE, Dartigues J-F, et al. Cognitive lifestyle jointly predicts longitudinal cognitive decline and mortality risk. Eur J Epidemiol. 2014;29(3):211–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Park DC, Reuter-Lorenz P. The adaptive brain: aging and neurocognitive scaffolding. Annu Rev Psychol. 2009;60:173–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Reuter-Lorenz PA, Park DC. How does it STAC up? Revisiting the scaffolding theory of aging and cognition. Neuropsychol Rev. 2014;24(3):355–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hoetting K, Roeder B. Beneficial effects of physical exercise on neuroplasticity and cognition. Neurosci Biobehav Rev. 2013;37(9):2243–57. [DOI] [PubMed] [Google Scholar]
- 29.Lu G, Zhixiong M, University BS. Effects of physical activity on cognitive function of the elderly. Sport Sci Res. 2014.
- 30.Han M, Tan XY, Lee R, Lee JK, Mahendran R. Impact of social media on health-related outcomes among older adults in Singapore: qualitative study. JMIR Aging. 2021;4(1):e23826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhang K, Kim K, Silverstein NM, Song Q, Burr JA. Social media communication and loneliness among older adults: the mediating roles of social support and social contact. Gerontologist. 2021;61(6):888–96. [DOI] [PubMed] [Google Scholar]
- 32.Ord AS, Slogar S-M, Sautter SW. Lifestyle factors, cognitive functioning, and functional capacity in older adults. Int J Aging Hum Dev. 2022;94(4):387–414. [DOI] [PubMed] [Google Scholar]
- 33.Pace-Schott EF, Spencer RMC. Sleep-dependent memory consolidation in healthy aging and mild cognitive impairment. Curr Top Behav Neurosci. 2015;25:307–30. [DOI] [PubMed] [Google Scholar]
- 34.Yaffe K, Falvey CM, Hoang T. Connections between sleep and cognition in older adults. Lancet Neurol. 2014;13(10):1017–28. [DOI] [PubMed] [Google Scholar]
- 35.Sexton CE, Storsve AB, Walhovd KB, Johansen-Berg H, Fjell AM. Poor sleep quality is associated with increased cortical atrophy in community-dwelling adults. Neurology. 2014;83(11):967–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kimiwada T, Sakurai M, Ohashi H, Aoki S, Tominaga T, Wada K. Clock genes regulate neurogenic transcription factors, including NeuroD1, and the neuronal differentiation of adult neural stem/progenitor cells. Neurochem Int. 2009;54(5–6):277–85. [DOI] [PubMed] [Google Scholar]
- 37.Stranahan AM. Chronobiological approaches to Alzheimer’s disease. Curr Alzheimer Res. 2012;9(1):93–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ran LS, Liu WH, Fang YY, Xu SB, Li J, Luo X, et al. Alcohol, coffee and tea intake and the risk of cognitive deficits: a dose-response meta-analysis. Epidemiol Psychiatr Sci. 2021;30:e13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.O’Keefe JH, Bhatti SK, Bajwa A, DiNicolantonio JJ, Lavie CJ. Alcohol and cardiovascular health: the dose makes the poison…or the remedy. Mayo Clin Proc. 2014;89(3):382–93. [DOI] [PubMed]
- 40.Wang Y, Chen H. Effects of mobile internet use on the health of middle-aged and older adults: evidences from China health and retirement longitudinal study. BMC Public Health. 2024;24(1):1490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Neves BB, Fonseca JRS, Amaro F, Pasqualotti A. Social capital and internet use in an age-comparative perspective with a focus on later life. PLoS ONE. 2018;13(2). [DOI] [PMC free article] [PubMed]
- 42.Rångtell FH, Karamchedu S, Andersson P, Liethof L, Olaya Búcaro M, Lampola L, et al. A single night of sleep loss impairs objective but not subjective working memory performance in a sex-dependent manner. J Sleep Res. 2019;28(1):e12651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Yin Z, Fei Z, Qiu C, Brasher MS, Kraus VB, Zhao W, et al. Dietary diversity and cognitive function among elderly people: a population-based study. J Nutr Health Aging. 2017;21(10):1089–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data used in the current study was publicly available and could be obtained at http://charls.pku.edu.cn/.
