This study employed a questionnaire survey method to collect data from 860 medical students enrolled at a university in Liaoning Province, China. The primary objective was to investigate the current status of physical exercise and suicidal ideation among medical students. To achieve this, the following scales were utilized: Symptom Checklist 90 (SCL-90), Physical Activity Registration Scale-3 (PARS-3), Positive and Negative Suicide Ideation (PANSI) scale, and the Emotion Regulation Questionnaire-Short (ERQ-S). The collected data were subsequently subjected to analysis to explore the relationships between emotional regulation and suicidal ideation within the medical student population. This analysis aimed to shed light on the potential influence of emotional regulation on suicidal ideation, serving as a theoretical basis for the development of prevention and intervention strategies tailored to reduce suicidal ideation among medical students, thereby enhancing their mental well-being.

All statistical analyses were performed using SPSS 21.0 (IBM Corp., Armonk, NY, USA). Continuous variables were tested for normality via Shapiro–Wilk test. Normally distributed data are expressed as mean ± standard deviation (SD); non-normally distributed data are reported as median (interquartile range, IQR). Statistical differences were analyzed using independent samples t-tests for gender comparisons. One-way ANOVA followed by Scheffe’s post hoc tests for comparisons across exercise intensity groups, contingent on verifying homogeneity of variance. Pearson correlation analysis and linear regression were employed to examine relationships between physical exercise and various factors, including mental health, emotional regulation, and suicidal ideation among medical students. A two-way ANOVA was conducted for each SCL-90 dimension to examine main effects of gender and exercise intensity, and their interaction. Tukey HSD post-hoc tests followed significant main effects. The significance level of p < 0.05 was considered statistically significant.

The questionnaire was rigorously designed through a comprehensive review of domestic and international literature, ensuring alignment with research objectives and adherence to principles of clarity, logical sequencing, and unambiguous wording. Prior to formal deployment, a pre-investigation involving 50 medical students was conducted to assess feasibility, identify ambiguities, and optimize question structure; this pre-test phase also informed sample size estimation for the main study based on effect size variability observed in pilot data. During the data collation stage, we meticulously reviewed the collected questionnaires. 8 questionnaires exhibiting inconsistencies in responses, a repeated selection of the same option throughout the survey, or patterns of systematically selected options were excluded from the analysis. The questionnaire of SCL-90, PARS-3, PANSI and ERQ-S demonstrated a high level of reliability, with a Cronbach’s α coefficient of 0.910, 0.890, 0.895 and 0.880, respectively, underscoring its consistency and dependability. Harman’s one-way test was applied to assess common method bias, revealing that four factors with eigenvalues exceeding 1 were identified. The cumulative variance explained by the first factor was 30.94%, which fell below the critical threshold of 40%, indicating no significant issue with common method bias.

All research procedures adhered to relevant ethical guidelines and regulations. This study received approval from the medical ethics review board of China Medical University [No.2023–107, Dated 06Mar2023]. Written informed consent was obtained from all participating students, ensuring their voluntary participation and protection of their rights. Informed consent was acquired from all study subjects before their involvement in the research.

A total of 860 questionnaires were distributed using a convenient sampling method, with 852 valid questionnaires successfully recovered, resulting in an effective response rate of 99.07%. The sample consisted of 389 males and 463 females. In terms of academic year, the distribution included 195 students from Grade I, 172 from Grade II, 188 from Grade III, 148 from Grade IV, and 149 from Grade V. Additionally, there were 356 students hailing from urban areas and 496 from rural areas.

PARS-3 assessment revealed a mean total score of 28.67 ± 5.54 among medical students. Significant gender differences were observed, with male students demonstrating higher physical activity levels (31.24 ± 7.51) than female students (26.12 ± 4.62), p < 0.001, Cohen’s d = 0.87 [95% CI: 0.68–1.06]. This large effect size (d > 0.8) indicates clinically meaningful gender-based activity disparities. Exercise intensity distribution analysis classified participants into three tiers: low-intensity (n = 442, 51.9%), moderate-intensity (n = 221, 25.9%), and high-intensity (n = 189, 22.2%) exercise groups.

Gender-based SCL-90 comparisons showed males scored significantly lower than females in obsessive-compulsive symptoms (1.54 ± 0.42 vs. 1.62 ± 0.37; p < 0.001), interpersonal sensitivity (1.49 ± 0.26 vs. 1.58 ± 0.41; p < 0.001), depression (1.63 ± 0.22 vs. 1.73 ± 0.34; p < 0.001), and anxiety (1.52 ± 0.17 vs. 1.68 ± 0.13; p < 0.001). Non-significant differences emerged in somatization, hostility, terror, paranoia, and psychosis (Table 1).

When assessing the impact of physical exercise on mental health, it is observed that medical students engaged in high-and medium-intensity physical exercise had lower scores for depression and psychosis in SCL-90 compared to those engaged in low-intensity exercise (p < 0.05). Additionally, students with high-intensity exercise had lower anxiety scores in the SCL-90 than those with medium-intensity exercise (p < 0.05). No significant differences were found in the scores for other factors (p > 0.05, Table 2). Further analysis revealed a negative correlation between the level of physical activity and the scores for depression (Standard regression coefficient = −0.236, t = 3.146, p < 0.001) and psychosis in the SCL-90 (Standard regression coefficient = −0.207, t = 3.125, p < 0.001).

Applying established diagnostic thresholds (NSI ≥ 1.63 and PSI ≤ 3.33) (Lei et al., 2021), 69 students (8.10%) were classified as having active suicidal ideation. No statistically significant differences were observed in suicide ideation scores or total scores concerning gender, grade, or place of origin. After controlling demographic variables such as gender, grade, and birthplace, Table 3 showed a significant positive correlation between physical exercise and emotional regulation (r = 0.124, p < 0.01), as well as a significant negative correlation between physical exercise and suicidal ideation (r = −0.117, p < 0.01). Additionally, a significant negative correlation existed between emotional regulation and suicidal ideation (r = −0.236, p < 0.01). Two-way ANOVA revealed significant main effects of gender on SCL-90 factors and no significant interactions emerged (p > 0.05), indicating exercise benefits are gender-invariant (Supplementary Table 1).

Physical exercise had a significant negative predictive effect on suicidal ideation (β = −0.126, p < 0.001). Emotional regulation also exhibited a positive relationship with physical exercise (β = 0.114, p < 0.001) (Figure 1a).

In the regression analysis with suicidal ideation as the predictor variable and physical exercise as the outcome variable, physical exercise negatively predicted suicidal ideation (β = −0.052, p < 0.05). Importantly, emotional regulation was introduced as a mediator, and it significantly negatively predicted suicidal ideation (β = −0.381, p < 0.001). While the predictive impact of physical exercise on suicidal ideation decreased in the presence of emotional regulation, it remained statistically significant (Figure 1b). This indicates that emotional regulation partially mediated the relationship between physical exercise and suicidal ideation, with the mediating effect accounting for 38.1% of the total effect.

Exercise intensity demonstrates a dose–response effect on mental health: high/medium-intensity exercisers showed clinically meaningful reductions in depression and psychosis versus low-intensity counterparts, with anxiety relief scaling to intensity levels. These findings suggest that physical exercise plays a beneficial role in promoting the mental health of medical students. Engaging in regular physical activity can enhance self-confidence, improve interpersonal skills, combat loneliness, and bolster students’ perseverance while alleviating academic pressures and preventing negative emotions. Physical exercise serves as a valuable tool for relaxation and mental well-being.

Physical exercise reduces suicidal ideation both directly and indirectly: while it alleviates established risk factors like depression and sleep disorders (Lee et al., 2013; Sewall et al., 2020). On a biological level, physical exercise can influence neurotransmitter levels, such as dopamine and serotonin, along with endorphin release, thereby enhancing overall brain function. This, in turn, enables individuals to better regulate negative emotions and foster positive ones, ultimately reducing the risk of suicide (Grasdalsmoen et al., 2020). Additionally, physical exercise’s connection to sleep quality plays a crucial role in its impact on mental health (Vancampfort et al., 2018). Moreover, outdoor exercise within a natural environment, with increased exposure to sunlight, has been associated with improved mental well-being (Corazon et al., 2019).

Our findings confirm that increased physical exercise significantly predicts reduced suicidal ideation among medical students, with intensity-dependent effects demonstrating dose–response characteristics. The establishment of consistent exercise habits emerges as a critical protective factor for mental health, corroborating prior evidence that physical activity reduces suicidality across educational stages (Southerland et al., 2016). Crucially, we observed that: (1) stronger emotional regulation capacity inversely correlates with suicidal ideation, aligning with evidence that maladaptive regulation strategies elevate risk (Law et al., 2015) and (2) emotional regulation partially mediates exercise’s protective effect, explaining its potency as an indirect intervention pathway. This mediation mechanism integrates physiological and psychological benefits into a unified protective framework, substantiating emotion regulation as a primary conduit through which exercise mitigates suicide risk.

Despite our significant findings, this study has limitations. It was conducted at a single medical school in Liaoning Province, potentially limiting the generalizability of our conclusions. While our models controlled for gender, future studies with larger gender-stratified samples could further explore sex-specific patterns in exercise-mediated emotional regulation pathways, particularly given documented disparities in mental health vulnerability and activity patterns. Additionally, due to time and sample size constraints, potential confounding variables, such as peer relationships and the psychological impact of evolving epidemic situations, were not assessed. Future research should expand its scope to encompass a broader range of variables and collect more extensive data to confirm and build upon our results (Jang and Park, 2019).