Mild cognitive impairment (MCI) describes an intermediate state between normal aging and dementia (Xue et al., 2018). MCI is a deficit in one or more cognitive domains, with relative preservation of functional independence without dementia (Petersen et al., 1999). The global prevalence of MCI in older adults over 60 years old is 5.0 ~ 36.7% (Dementia Cognitive Disorders Group, 2022). Older adults with MCI are more likely to progress to dementia than their cognitively normal counterparts (Shi-xiang and Yan-jie, 2017). 60.5% of MCI patients will develop dementia within 5 years after the initial diagnosis of MCI (Manly et al., 2008). Compared with irreversible damage of dementia, early intervention in patients with MCI can delay or prevent dementia development and even possibly reverse it to a normal cognitive level (Jin et al., 2017; Pandya et al., 2017). Therefore, MCI may be an important stage for the early detection and intervention of dementia.

There are numerous risk factors for MCI progression, including inheritance, demographic, metabolism, lifestyle, and biological factors. However, the main modifiable risk factor is a leisure activity in lifestyle (Livingston et al., 2020; Chowdhary et al., 2021). Leisure activities are defined as activities in which individuals participate for enjoyment, these activities are independent of work, and mainly include physical, social, and intellectual activities (Verghese et al., 2006; Yang et al., 2015). For retirees, leisure activities are an important part of their lifestyle. Leisure activity engagement may become a crucial part of the adjustment for retirees to maintain cognitive function postretirement (Lee et al., 2018).

Numerous previous studies have found that leisure activity was associated with cognition (Li et al., 2021; Sharifian et al., 2021). Some studies have focused on the association between single leisure activity and cognition and mainly explored the improvement effect of a single activity on cognition. They have found that single activities, such as gardening (Park et al., 2020), square dancing (Zhao et al., 2021), and Tai Chi (Li et al., 2022) can improve cognition in older adults. Additionally, some studies have focused on a certain type of activity, such as physical activity, social activity, and intellectual activity, and mainly analyzed the association between leisure activity and cognition from different dimensions. They have provided evidence that activity, regardless of its type, generally was positively associated with cognition in older adults (Kim et al., 2017; Ingold et al., 2020; Kurita et al., 2020). These above-mentioned studies provide a research basis for exploring the association between leisure activities and cognition and also make important contributions. However, considering that most older adults engage in multiple activities simultaneously (Morrow-Howell et al., 2014), using one activity or a set of certain leisure activity types as a measure may not offer a full picture of an individual's activity engagement patterns. Therefore, it is more realistic to analyze the leisure activity patterns of older adults from the overall activity participation characteristics and then explore the association between different patterns and cognition.

Latent class analysis (LCA) (Qiu, 2008) offers considerable promise in identifying activity patterns. LCA can identify smaller homogeneous subgroups, i.e., classes, based on all types of different leisure activities pursued by older adults. So that we can tailor matching interventions to the unique characteristics and needs of older adults. Fewer studies have investigated activity patterns based on the rationale that older adults can and do participate in multiple activities simultaneously (Amano et al., 2018; Chen et al., 2019; Zhang et al., 2020; Katayama et al., 2021). Moreover, these studies explored the association between each pattern and cognitive levels and did not focus on the MCI stage, a critical status of cognitive decline and cognitive intervention. Furthermore, due to the influence of different cultures, customs, and religious beliefs in different regions, there are certain differences in the types of leisure activities that older adults engage in, resulting in different leisure activity patterns discussed in each study (Amano et al., 2018; Chen et al., 2019; Zhang et al., 2020; Katayama et al., 2021). Therefore, further studies on the association between leisure activity patterns and MCI risk in the Chinese cultural context are warranted. This has important implications for promoting global healthy aging.

By using latent class modeling, the current study aimed to identify leisure activity patterns of older adults in China, explore the influencing factors of different patterns, and further estimate the association between these patterns and the prevalence of MCI. The results of this study would allow health professionals to customize specific interventions for older adults with different characteristics and needs in the future.

A cross-sectional design was used. According to the 10EPV (Events per variable) rule (Peduzzi et al., 1996), each independent variable should have at least ten samples. In our research, there were 30 variables (23 related to items of leisure activities, four related to demographic information, and three related to physical condition). Therefore, the minimum sample size in this study is at least 300. This study was approved by the Ethics Committee of Hebei Medical University (ethical approval number 2021080), and written informed consent was obtained from each subject before the experiment. This study was reported according to the STROBE Statement: Guidelines for Reporting Observational Studies (Cuschieri, 2019).

Rather than focusing on one activity or one type of leisure activity, this study considered various leisure activities as a whole, divided leisure activity patterns according to the overall characteristics, and then explored the association between different leisure activity patterns and cognition. Using 23 indicators of leisure activity, the study used LCA to identify three activity patterns: MDA class, SIA class, and FCA class. Older adults in the FCA class were at high prevalence of MCI.

As expected, the participation in leisure activities of older adults was heterogeneous and assigned to three different latent classes. The largest group was the MDA class (55%). The MDA class showed a high level of participation in almost all areas of leisure activity. This pattern has also been reported in previous studies (Hong et al., 2020; Zhang et al., 2020; Katayama et al., 2021). Of older adults, 21% belonged to the SIA class. The SIA class comprised older adults who had predominantly engaged in intellectual activities. And further results showed that older males were more likely to be classified as the SIA class. This was in line with the findings of previous studies demonstrating that males were more willing to engage in some challenging and competitive intelligence activities (van Uffelen et al., 2017; Hassing, 2020). It may be because males showed significantly higher beta relative power that was responsible for active thinking than females when faced with various challenging tasks (Corsi-Cabrera et al., 1993). A total of 24% of older adults belonged to the FCA class. The FCA class showed a high level of participation in family-focused activities but lower attendance at intellectually demanding activities. This finding was consistent with a study conducted in Singapore (Hong et al., 2020). We further found that older adults with less education were more likely to be classified as FCA class. This may be because the awareness and the sense of the pleasant experience from participation in intellectual activities of older adults with low education levels were not as good as those of older adults with high education levels (Payne et al., 2011).

The results showed that among the three latent classes, older adults in the FCA class had the worst cognitive function and the highest prevalence of MCI. It was difficult to make comparisons between our current study findings and others because of the differences in the measurement of leisure activities, patterns of activity identified, and cognitive assessment tools. However, there was an apparent similarity between our study and others: older adults with high comprehensive participation had better cognitive function (Amano et al., 2018; Chen et al., 2019; Zhang et al., 2020). After controlling for the influence of possible confounding factors, from the analysis of the characteristics of activity participation types, the high prevalence of MCI in the FCA class may be because their participation in intellectual activities was significantly lower than that of the other two classes. The protective effect of intellectual activities on the cognition of the elderly was recognized by many scholars (Yates et al., 2016; Hyun et al., 2020). The cognitive reserve hypothesis posits that leisure activity can protect cognition by increasing cognitive reserve and providing cognitive resilience against brain injury (Stern, 2002). Intellectual activity can protect cognitive function in older adults by increasing the effectiveness and plasticity of neural networks (Stern, 2012; Barker et al., 2021), preventing or slowing the deposition of amyloid-β (Aβ) plaques (Landau et al., 2012). However, as a non-exercise physical activity, there is no consistent conclusion on the relationship between housework and cognition. Yee et al. (2021) have demonstrated that housework was associated with better cognitive function in older adults. Even if the amount of housework was less than recommended by exercise guidelines, the accumulation of these simple activities could increase gray matter volume and improve cognitive function in older adults (Halloway et al., 2019). However, a longitudinal study has shown that housework, such as cooking and gardening, was associated with a decline in spatial ability and memory in older adults (Hassing, 2020). A further study should explore the independent relationship between housework and cognitive function through follow-up to provide evidence in this regard.

Older adults in the FCA class were the key cognitive intervention population for the future. Increasing participation in activities that have greater socialization and cognitive reserve of older adults is the focus of future research, especially increasing participation in activities with different intellectual needs. First, the awareness of activity participation of older adults with low education levels is not enough. In this regard, the community centers should increase publicity efforts in the form of some event pictures, posters, small videos, etc. Second, older adults with low education levels have a low sense of pleasure in participating in intellectually demanding activities. The Match Hypothesis states that the flow experience will be more likely if there is a match between an individual's ability and the capacity requirements for the activity (Payne et al., 2011). As an exceptionally positive state, flow experience can motivate people to perform intellectually demanding activities at a high level and to persevere in these activities over long periods (Csikszentmihalyi et al., 2005). Hence, we can assign a weight to the cognitive component of common activities of the elderly (Karp et al., 2006) and match older adults with different educational levels to participate according to the weight of the cognitive component. For example, gardening, watching TV, participating in outdoor activities, gathering with friends, and participating in some volunteer services are activities with low intellectual needs. Reading, playing chess or cards, traveling, writing, playing mobile games, etc., are activities with high intellectual needs. In the future, activities that match different intellectual needs can be based on the education level and hobbies of older adults. Particularly, older adults with low education level can participate in some activities with lower intellectual needs or replacement activities with similar cognitive components, such as sharing housework tips, growing vegetables, tailoring and repairing, etc. Additionally, previous research has found that older adults with lower education level can obtain stronger cognitive protective effects from activities such as volunteering than those with higher education (Park et al., 2019). The community can provide older adults with low education level the opportunity to participate in volunteer activities and encourage them to actively participate, such as assisting community nucleic acid testing, publicizing garbage sorting, etc., which also has positive significance for their cognition.

This study has some limitations. First, this was a cross-sectional study, limiting the extent of inference about causality. Without longitudinal data, we could not determine whether the presence of MCI predicted participation in a less intellectual activity or whether the lack of intellectual activity led to MCI. Future longitudinal studies are needed to investigate the causal relationship between leisure activity patterns and cognition. Second, the representativeness of the sample needs to be further improved. The nature of the sample could affect the number of latent classes and characteristics. We sampled only older adults in Shijiazhuang City. Moreover, when excluding hearing loss, it was only based on the self-reports of the participants or their family members, which inevitably led to the loss of some samples. Future studies should recruit more broadly representative older adults using objective assessments. Third, leisure activity engagement was self-reported and could be influenced by the recalling bias. Additionally, because of the large number of activities, we did not measure the “total longevity” of leisure activities. We believe that the overall participation of older adults in leisure activities is in a balanced state. However, the cognitive effects of engaging in an activity for 1 year should be different from those lasting 10 years. Finally, this study failed to address other covariates related to biological factors (e.g., cytokines and Aβ). In addition to social and health covariates, future studies should include these factors.

Despite its limitations, this study has important clinical implications. First, this was one of the few studies using LCA to examine leisure activity patterns among older adults. The results showed that a single or a certain type of activity might not fully reflect a person's activity status. As such, we have shown some evidence for looking at patterns of multiple activity domains rather than a single domain. Additionally, the heterogeneity of participation in leisure activities also highlighted the importance of tailoring specific intervention strategies, especially for those at higher risk of cognitive impairment. Understanding the diversity of activity engagement may spur service providers to develop appropriate programs for different characteristics of older adults to maximize the reach of the activity program in the target group, especially for some older adults with low education levels. Community service agencies should strengthen the publicity effect and fully promote the participation of older adults in diverse activities, especially intellectual activities. Finally, it also helps practitioners and policymakers be aware of these heterogeneous behaviors, thereby maximizing the reach and meaning of leisure activities.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

This study was approved by the Ethics Committee of Hebei Medical University (ethical approval number 2021080) and written informed consent was obtained from each subject before the experiment. The patients/participants provided their written informed consent to participate in this study.

YZ and YL conceptualized and designed the study and the investigation. XY, LG, and BC distributed and collected questionnaires together with YZ. YZ drafted the original manuscript. YZ, XX, XM, and XY performed the initial analyses, interpreted the data, and reviewed and revised the manuscript. YL and XX obtained funding for the research and supervised the procedure of the whole investigation to evaluate the project. All authors contributed to the article and approved the submitted version.