College students undergo a critical developmental transition from adolescence to adulthood. Although these emerging adults strive to achieve high levels of intellect, ambition, and self-esteem (Pedrelli et al., 2015), they still have to face typical challenges for this developmental stage, such as finding an accommodation, engaging in relationships, coping with competitive and academic difficulties, financial stress, and struggling to make important decisions (Parker et al., 2004). Consequently, a considerable number of college students develop stress-related mental disorders (e.g., depression and anxiety), which limit their daily activities and academic performance (Fam, 2018; Sobocki et al., 2006). Specifically, stress, anxiety, and depression are typically elicited in individuals (including college students) when faced with an unpredictable or challenging situation or event (Ketata et al., 2021). Additionally, individuals have reported feeling emotionally worse during the COVID-19 pandemic as compared to how they felt pre-pandemic (Chi et al., 2020; Chi et al., 2022; Chi, Liang, et al., 2021; Hossain et al., 2020). For instance, during the COVID-19 pandemic the prevalence of stress, anxiety, and depression in the global general population reached 29.6%, 31.9%, and 33.7%, respectively (Salari et al., 2020). Furthermore, college students in emerging adulthood are more likely to present these emotional states because they have difficulty adjusting during this developmental transition period (Arnett & Tanner, 2006); this may be exacerbated during the COVID-19 pandemic (Wang et al., 2020). A recent meta-analysis investigating the psychological effects of COVID-19 suggests a relatively high prevalence of stress (23%), anxiety (29%), depression (37%) among college students (Wang et al., 2021). Thus, searching for protective factors in this particular cohort of emerging adults is urgently needed, which allows for a timely implementation of interventions to effectively alleviate negative emotions as the current strategies and vaccines cannot fully protect against the globally continuous COVID-19 pandemic.

One of the most frequently studied protective factors against stress, anxiety, and depression is physical activity (PA) (Anderson & Shivakumar, 2013; Chi, Liang, et al., 2021; Moljord et al., 2014). PA is defined as a type of bodily movement that consumes energy from skeletal muscles and (US Department of Health & Human Services, 1996) has beneficial psychological effects (Miles, 2007). Specifically, the positive links of PA with stress, anxiety, and depression have been well documented across different age groups and individuals with or without chronic illnesses (Dinas et al., 2011; Kandola et al., 2019; Rethorst et al., 2009; ter Riet et al., 2012). Such psychological benefits from regular engagement in PA can be attributed to changes occurring at different levels of analysis: (i) molecular level; (ii) brain structure and function; and, (iii) psycho-social factors (Stillman et al., 2020). From a molecular perspective, brain-derived neurotrophic factor (BDNF) which is the most abundant neurotrophic factor in the brain, has been shown to be associated with PA level (Huang et al., 2014). There is evidence to suggest that decreased BDNF levels, especially in the hippocampus, are correlated with stress-induced depression and anxiety (Altar, 1999; Duman & Monteggia, 2006). In terms of brain structure and function, the exercise-induced changes in the hypothalamic-pituitary-adrenal (HPA) axis are observed to play a critical role in the manifestations of depression and anxiety symptoms (Anderson & Shivakumar, 2013). With respect to psychological factors, researchers have recently paid great attention to resilience and have proposed that PA reduces negative emotion through improved resilience (Moljord et al., 2014).

While there are many different definitions for psychological resilience (Sisto et al., 2019), the fundamental concept is the ability to maintain or regain mental health despite adverse conditions (Afek et al., 2021). Resilience especially comes into play in dealing with and tolerating challenging or troublesome situations, such as stress, anxiety, and depression. Empirical studies have documented that resilience is negatively correlated with the aforementioned emotional states, and it can maintain or even improve the mental health of college students (Hartley, 2011). Therefore, exploring how to build resilience may help individuals suffering from negative emotional states, such as depression and anxiety. In addition, cardiorespiratory fitness, whether achieved through regular PA and/or structured physical exercise program, can promote resilience (Li et al., 2020). In this context, the following pathways might mediate the positive effect of PA and/or physical fitness on resilience-related parameters: (i) stress-buffering effects on negative emotions, (ii) elevating neurophysiological adaptations to external stressors, (iii) enhancing an anti-inflammatory state, and (iv) increasing neuroplasticity and growth factor expression (Silverman & Deuster, 2014). Specifically, researchers have hypothesized that PA, especially physical exercise as structured, planned, and repetitive form of PA, serves as an sporadic stressor on the physical body that can maximize adaptation to other types of stressors (e.g., psychological stressors) (Budde et al., 2016; Caspersen et al., 1985). In terms of the biological profile, physiological toughening refers to an initially elevated catecholamine response, followed by a quick recovery and decreased HPA axis responses (similar to the mechanism of PA in improving depression and anxiety). Such physiological changes were observed to associate with enhanced performance during challenging/stressful situations, increased tolerance to stressors (e.g., improved resilience), and increased emotional stability (e.g., reduced anxiety and depression). Collectively, PA may help individuals neurophysiological adapt to external stressors and ultimately reduce stress-induced anxiety and depression (Sothmann et al., 1996). Taken together, relatively high levels of regular PA are positively associated with mental health among adolescents (Chi, Liang, et al., 2021). Therefore, resilience, among other factors, may be presumably an important mediator between PA and negative emotional states especially in individuals being at risk of developing mental disorders (Hegberg & Tone, 2015).

To better understand the association between PA and negative emotional states, researchers incorporated “affective responses to exercise” into current theoretical models of exercise behavior (Ekkekakis et al., 2005b, 2008; Stych & Parfitt, 2011). Exercise tolerance is defined as an ability to continuously perform exercise plans despite exercise intensity beyond the physiological threshold even when the activity is experienced as unpleasant/uncomfortable (Ekkekakis et al., 2005a). According to a previous study, exercise tolerance is positively associated with vigorous leisure-time PA (Teixeira et al., 2021). Furthermore, there is some overlap in the definition of exercise tolerance and psychological resilience. Both of them involve persistence and adaptation in difficult situations, but exercise tolerance is related to high exercise intensity and psychological resilience is related to real dilemmas. Despite the lack of research on exercise tolerance and resilience, a few studies have revealed a potential association between these two factors. As Ekkekakis et al. (2007) found in a response to a circuit training, the exercise-induced hypoalgesia (resulting in tolerance to high exercise intensity) was achieved through changes in pain tolerance levels, which seems to be comparable to the mechanism of PA to improve resilience.

The current evidence provides some indications that PA, exercise tolerance, resilience and negative emotional states are related in a way; that increased exercise can lead to higher levels of exercise tolerance and resilience thereby reducing negative emotional states. Therefore, the present study aimed to investigate the serial multiple mediating effects of exercise tolerance and resilience between PA level and negative emotional states in a cohort of Chinese college students. In this study we hypothesized that PA improves individuals' ability to persist and adapt themselves in uncomfortable activity situations, which can be transferred to troublesome situations in daily life (i.e., psychological resilience). In other words, PA can improve psychological resilience, with exercise tolerance as a mediating factor. Further, we want to investigate whether the effects of PA level on depression and anxiety were mediated by the psychological factors operationalized via exercise tolerance and resilience. Specifically, we hypothesized that: (1) PA level would be negatively correlated with stress, anxiety, and depressive symptoms, and (2) exercise tolerance and psychological resilience mediate the association between PA level and negative emotional states.

This study aimed to gain a better understanding regarding the possible associations between PA level and emotional states, and the potential role of exercise tolerance and resilience as mediators of such a relationship. Firstly, our findings reveal a significant and negative association between PA level and emotional states. This finding is consistent with previous studies and supported the positive effect of a relatively high PA level on emotional states (Anderson & Shivakumar, 2013; Dinas et al., 2011; Kruk et al., 2019; Teixeira et al., 2013). More importantly, the promising results of the present study helped to explain how PA results in reductions of negative emotional states (stress, anxiety, and depression), mainly due to the serial mediation effects of exercise tolerance and resilience. These results are discussed below.

The psychological mechanisms of PA-induced effects on negative emotions have been examined in recent decades. Of note, results of the present study showed that associations of resilience with PA and negative emotional states (stress, anxiety, and depression). Specifically, resilience was positively associated with PA level and negatively with stress-related negative emotions, although the correlation coefficient for the former is relatively low. Rutter (1985) suggested that psychological resilience protects individuals against challenging situations primarily through four potential pathways: (a) decreasing risk impact; (b) minimizing negative chain reactions; (c) building up psychological abilities like stronger self-esteem and self-efficacy; and (d) creating opportunities for adaptive responses. There are some studies available that explored the association between resilience and these two variables (PA level and negative emotional states) and have obtained similar results to the present study (Moljord et al., 2014; Stych & Parfitt, 2011). Simultaneously, some studies addressed that exercise activates physical responses in order to enhance resilience, which could indirectly reduce the negative impact of stress on negative emotions through its mediating effect (Anyan & Hjemdal, 2016; Ma et al., 2019). In other words, resilience represents a dynamic process. While being physically active, the individual exerts a top-down regulation, so as to achieve the internal balance (internal steady state), which, in turn, attenuates the negative emotions caused by the physiological reaction originating by being physical active (Belcher et al., 2021). Therefore, resilience is likely to play an important role in exercise to relieve negative emotions.

The current study also found that the exercise tolerance related to resilience was another significant mediator of the association between PA level and emotional states, extending previous findings about this relationship. Resilience did not independently mediate the association between PA level and emotional states, which indicated that exercise tolerance may play a more important role in the link between regular PA and emotional states. Specifically, college students with a higher PA level also show a higher level of exercise tolerance, which, in turn, led to less stress, anxiety, and depression. However, previous studies on exercise tolerance have mainly focused on patient populations in order to explore ways to improve exercise tolerance. For example, Burtscher et al. (2010) found that the beneficial effects of interval hypoxia training on exercise tolerance seem to be greater in patients with coronary artery disease or chronic obstructive pulmonary disease when compared to healthy participants, as higher exercise tolerance is associated with lower mortality. Similarly, another study provided evidence that spinal anesthesia enhances exercise tolerance in patients with chronic obstructive pulmonary disease (Gagnon et al., 2012). Although there is no study focusing on the relationship between exercise tolerance and emotional states, and it has been suggested that regular exercise can improve exercise tolerance (Ferguson et al., 2007). In this context, this is a new idea that increasing PA levels, which increases exercise tolerance and ultimately reduces stress, anxiety and depression.

Lastly, our finding supports our hypothesis that PA exerts an influence on emotional states among college students is mediated by exercise tolerance and resilience. In particular, the path was PA → exercise tolerance → resilience → emotional states. We hold the opinion that usually when performing some high intensity exercise, when the force is exhausted, individuals need to keep persevering, which is like facing a setback, and in such a process hones their will, thereby developing their resilience. Such resilient students tend to perceive the challenging and unpleasant conditions as non-permanent, and thus strengthen defenses against negative emotions about the pandemic, and experience positive emotions. From a practical perspective, university administrators should create a physically active environment by increasing the accessibility of sports facilities and incorporating exercise sessions (recommended by ACSM) into a curriculum, which could help students get into the habit of exercise for improving mental health. The results of this study provide a new perspective for the prevention of mental disorders within a campus environment.

Given the increasingly serious situation of depression, anxiety and stress among college students, the relationship between such negative emotional states and PA is receiving increased attention from scholars and practitioners. To better understand the relationship between PA and negative emotional states, a number of psychological factors need to be taken into account such as exercise tolerance and resilience. The present study suggests that PA can predict negative emotional states indirectly through exercise tolerance and resilience. The results of this study encourage college students to get into the habit of regular PA which provides a new perspective for the prevention of mental disorders within a campus environment.