Depression can cause a variety of problems such as low quality of life and disability in the elderly. Studies on the dose-response relationship between physical activity and depression symptoms are still insufficient at present.

This study is designed to analyze the relationships of different-intensity physical activity durations and total physical activity volume with depressive symptoms.

The data were derived from the baseline survey of the Community-based Cohort Study on Nervous System Diseases conducted during 2018-2019. Information on demographic characteristics and physical activity was investigated by questionnaire. The Geriatric Depression Scale-30 (GDS-30) was used to screen whether the participant had depressive symptoms. A total of 5 558 participants aged 55 years and above from four provinces of China with complete information on demographic characteristics, physical activity, and GDS-30 were included in the final analysis. Light physical activity (LPA) duration, moderate-to-vigorous physical activity (MVPA) duration, and total physical activity volume were calculated. Multiple logistic regression models were used to analyze the relationships of different physical activity durations and total physical activity volume with depressive symptoms, and restricted cubic spline models for 5 quantile knots (P₅, P₅₀, P₇₅, P₉₀, and P₉₅) were employed to evaluate the dose-effect relationships, using zero as reference.

The prevalence of depression symptoms was 10.2% (567/5 558). The median LPA duration, MVPA duration, and total physical activity volume were 8.2 h·week^-1, 7.0 h·week^-1, and 31.5 MET·h·week^-1 in the participants with depressive symptoms, which were lower than those in the participants without (P < 0.05). Compared with the residents without LPA, LPA >10.5 h·week^-1 was associated with a 47% lower risk of having depressive symptoms (OR=0.53, 95% CI: 0.33-0.85) after selected confounder adjustment. The risk of having depressive symptoms was gradually decreased with increased time spent in LPA (P_trend < 0.05, non-linear P>0.05). The odds ratio of having depressive symptoms was 2.35 (95% CI: 1.50-3.68) times higher in the people with >24.5 h·week^-1 of MVPA than those without MVPA. With the increase of MVPA duration and total physical activity volume, the risk of having depressive symptoms first decreased and then increased (P_trend < 0.05, non-linear P < 0.05).

There is a near linear relationship between the duration of LPA and the risk of having depressive symptoms within a certain range. There is a near U-shaped dose-effect relationship of MVPA duration and total physical activity volume with the risk of having depressive symptoms.