Skip to main content
BMC Sports Science, Medicine and Rehabilitation logoLink to BMC Sports Science, Medicine and Rehabilitation
. 2026 May 13;18:322. doi: 10.1186/s13102-026-01728-3

The effects of exercise on generalized anxiety symptoms in adults: a systematic review and meta-analysis

Haiman Su 1, Lili Li 1, Jun Wu 1, Manh Cuong Nguyen 2, Hui Ding 3, Chen Wei 4, Zhenyu Zhang 5,
PMCID: PMC13360853  PMID: 42129849

Abstract

Objective

This study aimed to systematically evaluate the interventional effect of exercise on symptoms of generalized anxiety.

Methods

This systematic review and meta analysis was prospectively registered on PROSPERO (Registration Number: No. CRD420251146169). A comprehensive literature search was conducted across seven electronic databases (Embase, Web of Science, PubMed, The Cochrane Library, Wanfang, VIP, and the China National Knowledge Infrastructure) from their respective inceptions to June 2025. This search specifically targeted randomized controlled trials (RCTs) examining the effects of exercise interventions on generalized anxiety symptoms. The inclusion criteria were defined as follows: adult participants (aged 18 years or older) presenting with generalized anxiety symptoms, without any major comorbidities. The PEDro scale was utilized to assess the methodological quality of the included trials. All statistical analyses, including the primary meta analysis, subgroup evaluations, sensitivity analyses, and publication bias assessments, were performed using Stata 17.0 software. The standardized mean difference (SMD) with a 95% confidence interval (CI) was utilized as the primary effect measure. Finally, the GRADE framework was employed to evaluate the overall certainty of the evidence.

Results

A total of 10 studies (involving 2408 participants) were included. The meta analysis revealed that exercise interventions provided suggestive evidence for a potential reduction in generalized anxiety symptoms (SMD=-0.502,95%CI[-0.830,-0.174], P = 0.003). Preliminary subgroup analyses observed a potential positive trend in middle aged and older adults (SMD=-1.063,95%CI[-1.831,-0.295], P = 0.007) compared with younger adults (SMD=-0.388,95%CI[-0.698,-0.057], P = 0.022). Mind body exercise (SMD=-0.680,95%CI[-0.914,-0.445], P < 0.001) and aerobic exercise (SMD=-0.543,95%CI[-1.064,-0.022], P = 0.041) demonstrated potentially beneficial trends, whereas current evidence remains insufficient to determine the efficacy of resistance exercise (SMD=-0.302,95%CI[-0.964,0.361], P = 0.372). Regarding exercise parameters, potential positive trends were observed for frequencies of ≥ 4 sessions per week (SMD=-1.063,95%CI[-1.831,-0.295], P = 0.007), single session durations of 21 to 40 min (SMD=-1.079, 95%CI[-1.575,-0.582], P < 0.001), and interventions lasting ≥ 12 weeks (SMD=-0.868,95%CI[-1.255,-0.481], P < 0.001) or 8 weeks (SMD=-0.780,95%CI[-1.315,-0.246], P = 0.004). Evidence was insufficient to support 6 week programs (SMD=-0.061,95%CI[-0.517,0.395], P = 0.794) or ≤ 20 min sessions (SMD=-0.022, 95%CI[-0.618,0.574], P = 0.943). What was subjectively classified as light intensity exercise demonstrated a potential positive trend (SMD=-0.915,95%CI[-1.402,-0.427], P < 0.001). While leave one out sensitivity analyses confirmed a consistent beneficial direction (SMD ranging from − 0.43 to -0.61, all P < 0.05), the overall interpretability is severely limited by exceptionally high heterogeneity (I²=90.37%).

Discussion

The included studies demonstrated acceptable methodological quality (PEDro scores 6 to 8); however, limitations in blinding and subjective intensity classifications introduce substantial bias. In conclusion, while current findings provide suggestive evidence that exercise may assist in ameliorating generalized anxiety symptoms, the exceptionally high heterogeneity dictates that these results must be interpreted with low confidence rather than as definitive efficacy. Furthermore, while analyses across the literature indicate that specific exercise parameters (such as higher frequencies, longer intervention durations, and mind body modalities) exhibit potential positive trends, the specific strata demonstrating the largest effects are often derived from severely underpowered subgroups. Consequently, these observations must be viewed strictly as hypothesis generating rather than optimal clinical prescriptions, highlighting the critical need for robust future trials.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13102-026-01728-3.

Keywords: Generalized Anxiety, Exercise, Systematic Review, Meta-analysis

Introduction

Generalized Anxiety Disorder (GAD) is a common mental disorder in the anxiety spectrum. Analysis of World Health Organization survey data indicates that GAD has a lifetime prevalence of 3.7% and a 12-month prevalence of 1.8%, with the highest prevalence in high-income countries (5.0%) and the lowest in low-income countries (1.6%) [1]. Its prevalence increases with economic development, making it a relatively common subtype among anxiety disorders [2]. Its primary characteristic is persistent and excessive worry, often lasting for months or longer. Patients often exhibit a variety of somatic symptoms, such as fatigue, muscle tension, and restlessness, alongside psychological symptoms including difficulty concentrating and irritability [3, 4]. Patients with GAD present with persistent abnormalities in blood parameters and may experience lifelong disease burden. Its chronic nature can cause long-term damage to physical health [5] and lead to cardiovascular issues and immune dysregulation [6, 7]. These symptoms often result in significant functional impairment in daily life, thereby severely affecting the quality of life [2, 4]. Epidemiological studies have shown that GAD is prevalent in both community and clinical populations, with a higher incidence in women than in men [8]. Over half of individuals with GAD meet the diagnostic criteria for major depressive disorder in their lifetime (WHO, 2022), significantly increasing their risk of comorbid depression [9]. By amplifying the perception of somatic symptoms and reducing treatment adherence, GAD exacerbates patients’ vulnerability to comorbidities, consequently increasing the healthcare and socioeconomic burden [10]. In summary, GAD not only affects the psychological state but also has a profound effect on overall health and social functioning, making the improvement of generalized anxiety a pressing issue that demands immediate attention.

Exercise is widely recognized as a cost-effective and efficacious alternative non-pharmacological therapy for mental health management. Current evidence suggests that exercise has a positive effect on reducing symptoms of various mental disorders [11].Compared with pharmacotherapy, exercise has fewer side effects, avoiding the risks of medication-induced cognitive dullness and dependency and being more amenable to community-wide implementation [12]. Research indicates that regular exercise can improve patient treatment adherence and alleviate core symptoms of anxiety, depression, and GAD by modulating neuroplasticity, particularly by elevating levels of neurotrophic factors such as BDNF. This mechanism underscores the potential advantages of exercise interventions in comorbidity management [13, 14]. Exercise not only effectively reduces anxiety symptoms but also possesses transdiagnostic applicability [9]. This intervention modality demonstrates unique advantages in long-term management, particularly for patients who do not respond adequately to pharmacotherapy, making physical activity a viable strategy [11]. Therefore, as an adjunctive approach, exercise intervention holds promise for addressing the limitations of existing GAD treatments and providing patients with a more comprehensive management strategy .

By reviewing previous studies [1521], we have found that exercise intervention holds potential in alleviating anxiety. However, key challenges remain, including significant population heterogeneity, insufficient focus on GAD as a specific disorder, and contradictory or insufficient evidence regarding the optimization of exercise prescription parameters (type and dosage). To address these gaps, my research plan includes the following specific measures: restricting the study population to adults with GAD symptoms to enhance population homogeneity; conducting quantitative analysis on precise exercise prescriptions for GAD subgroups to establish evidence-based guidance on “dose-response” relationships. This study aims to evaluate the overall effect of exercise intervention on GAD symptoms and further analyze the impact of factors such as age, severity of the condition, as well as exercise type, frequency, session time, duration, and intensity on treatment efficacy. The goal is to develop a more precise exercise regimen and provide personalized clinical guidance for healthcare professionals. It is worth noting that many randomized controlled trials (RCTs) on ‘anxiety’ include mixed diagnostic groups. To ensure sample homogeneity and the specificity of our findings for GAD, we applied strict inclusion criteria focusing solely on GAD populations.

Method

Study design

This study was conducted and reported in accordance with the PRISMA guidelines for meta-analyses [22] and has been registered in the International Prospective Register of Systematic Reviews (PROSPERO), registration number (No. CRD420251146169 ).

Literature search strategy

The literature search was performed independently by two researchers across the following databases: Embase, Web of Science, PubMed, The Cochrane Library, Wanfang Database, VIP, and China National Knowledge Infrastructure (CNKI), with the aim of identifying randomized controlled trials (RCTs) investigating the effects of exercise on generalized anxiety. The search period spanned from the inception of each database to June 2025. This was supplemented by a manual search of the reference lists of included studies. To ensure comprehensiveness, the following search strategies were employed:

English Search Strategy: (Exercise OR physical exercise OR physical activity OR aerobic exercise OR resistance exercise OR multicomponent exercise OR Mind-body Exercise OR Sports) AND (Generalized Anxiety Disorder OR GAD OR Overanxious Disorder OR Generalized Anxieties) AND (randomized controlled trial OR Randomized OR Controlled OR Trial).

Chinese Search Strategy: (运动 OR 体育锻炼 OR 有氧运动 OR 体力活动 OR 抗阻运动 OR 身心运动 OR 锻炼 OR 体育) AND (广泛性焦虑) AND (随机对照实验 OR 随机 OR 对照 OR 实验).

These search strategies were carefully designed to capture relevant literature while maintaining methodological rigor. To optimize the balance between search sensitivity and specificity, and to rigorously preserve diagnostic homogeneity, primary search terms were purposefully restricted to GAD-specific keywords. Expanding the search to encompass broader terminologies (e.g., ‘anxiety disorders’) would have significantly reduced search specificity by introducing a massive volume of literature on non-target anxiety subtypes or subclinical stress. However, to ensure no relevant data were overlooked, trials investigating broader anxiety disorders were carefully evaluated during the full-text screening phase; studies providing independent, extractable data for a formally diagnosed GAD subgroup were deemed eligible for inclusion.

Inclusion and exclusion criteria

Inclusion criteria

(1) Participants: Individuals identified with generalized anxiety. Specifically, this encompasses patients with a formal clinical psychiatric diagnosis of GAD based on standard diagnostic criteria, as well as individuals identified as having significant generalized anxiety symptoms through validated screening tools [23]. (2) Intervention: Any form of physical exercise, including aerobic exercise, resistance training, mind-body exercises, and other structured physical activity programs. The intervention must have a duration of at least 6 weeks, with clearly defined exercise frequency, session time, and/or intensity.(3) Control Group: The control group must be a no-treatment (wait-list) control, placebo control, or usual care group that does not include an exercise component.(4) Outcomes: At least one of the following outcomes must be reported: changes in anxiety symptoms measured by validated scales, such as the Generalized Anxiety Disorder 7-item Scale (GAD-7), Hamilton Anxiety Rating Scale (HAM-A), Penn State Worry Questionnaire (PSWQ), State-Trait Anxiety Inventory (STAI), or the 36-Item Short Form Health Survey (SF-36).(5) Study Design: RCT.

Exclusion criteria

(1) Interventions not involving physical exercise or a structured physical activity program, such as pharmacological treatments or psychotherapy alone.(2) Studies where the experimental group combined exercise with other intervention modalities (e.g., exercise plus cognitive therapy).(3) Studies without a control group.(4) Duplicate publications, or studies with incomplete data that could not be obtained or verified.(5) Literature that did not report outcomes specific to generalized anxiety symptoms or adverse events related to the exercise intervention.(6) Non-peer-reviewed articles, such as commentaries, editorials or conference abstracts.

Literature screening, data extraction, and quality assessment

Literature screening and data extraction

After retrieving the relevant literature, the records were imported into EndNote to remove duplicates. Subsequently, two researchers independently performed literature screening and data extraction in a blinded manner. The extracted information included the first author’s name, publication year, country, baseline characteristics of the study population (age, sex, and disease duration), intervention details, and outcome measures.

Quality assessment

The methodological quality of the included studies was assessed using the PEDro scale [24]. This scale comprises 10 items: “random allocation,” “concealed allocation,” “baseline comparability,” “blinding of subjects,” “blinding of therapists,” “blinding of assessors,” “adequate follow-up (> 85%),” “intention-to-treat analysis,” “between-group statistical comparisons,” and “point measures and variability data.” Each satisfied criterion scored 1 point; otherwise, 0 points. The total score is 10. A score < 4 indicates low quality, 4–5 fair quality, 6–8 good quality, and 9–10 excellent quality. Only studies with fair quality or higher were included in this review. Simultaneously, the GRADEpro system was used to evaluate the quality of evidence for each outcome measure, categorizing the evidence quality for each outcome into four levels: high, moderate, low, and very low. Quality ratings were performed independently by two researchers. Any discrepancies were resolved through discussion with a third researcher until consensus was reached [25, 26].

Data analysis

Stata 17.0 software was used to perform the meta-analysis, subgroup analysis, sensitivity analysis and assessment of publication bias. Heterogeneity was tested for all included outcome measures based on sample size, mean values, and standard deviations of improvement scores pre- and postintervention. As all the included outcome measures were continuous variables, the mean difference (MD) was used for outcomes measured with the same method and unit, whereas the standardized mean difference (SMD) was applied for outcomes measured with different methods or units. Heterogeneity was assessed using the P-value and I² statistics. A P-value < 0.05 and I² > 50% indicated significant heterogeneity among studies, warranting the use of a random-effects model. Otherwise, a fixed-effects model was used. The results of the meta-analysis are presented with 95% confidence intervals (95% CI).If a single study includes multiple independent intervention groups (e.g., the aerobic exercise group and resistance exercise group in the study by Matthew et al.), each intervention group will be treated as an independent study entry for analysis to ensure consistency in the unit of analysis and accuracy of the analysis.

The classification of exercise intensity (e.g., ‘light’, ‘moderate’) was based on the descriptive terminology used in the original studies (e.g., ‘walking’ was typically categorized as light-intensity, ‘jogging’ as moderate-intensity), as objective measures (e.g., %HRmax, METs) were rarely reported.

Results

Literature search results

A total of 3,443 potentially relevant records were identified through database searches: PubMed (n = 50), The Cochrane Library (n = 2,273), Embase (n = 301), Web of Science (n = 652), China National Knowledge Infrastructure (CNKI, n = 15), Wanfang Database (n = 75), and VIP Database (n = 77). The records were imported into EndNote X9 for management. After removing duplicates, 3,096 unique records remained. Following a preliminary screening of titles and abstracts, 111 studies were deemed eligible for a full-text review. After a detailed assessment of the full texts, 10 studies met the inclusion criteria and were included in the final analysis. The literature screening process is shown in Fig. 1.

Fig. 1.

Fig. 1

Flowchart of literature screening

Characteristics of included studies

The 10 included studies (comprising 14 study entries) involved 2,408 participants, with 1,285 assigned to the intervention groups and 1,123 to the control groups. The study population primarily consisted of individuals with generalized anxiety across different age groups (including young adults and middle-aged and older adults). The included studies were conducted in China, the United States, Ireland, and Uganda. In all included studies, the intervention group received an exercise program (specifically aerobic, mind-body, and resistance exercises), while the control group received usual care. The duration of the exercise interventions across the studies ranged from 6 weeks to 1 year, the intervention frequency varied from 1 to 14 sessions/week, and the single session time lasted from 16 to 120 min. The basic characteristics of the included studies are shown in Table 1.

Table 1.

Characteristics of included literatures

Study Country n(T/C) Gender (M/F)/n Age(T/C)/years Intervention type Intervention protocol Assessment
T C
Yang 2023 [27] China

T485,

C287

377/108

145/

142

18–22/18–22 Aerobic 2 sessions/w, 25 min training/session, 8 w
Zhang2016 [28] China

T30,

C32

5/25 5/27

47.9 ± 12.4/

48.03 ± 8.8

Mind-body 2 sessions/w, 25 min training/session, 8 w
Kathleen 2025 [29] Uganda

T421,

C413

232/189 159/254 19/19 Aerobic 2 sessions/w, 16 min effective training/session, 6 w
Matthew2012① [30] America

T10,

C10

0/10 0/10 18–37/18–37 Resistance 2 sessions/w, 16 min training/session, 6 w
Matthew2012② [30] America

T10,

C10

0/10 0/10 18–37/18–37 Aerobic 2 sessions/w, 16 min effective training/session, 6 w
Gordon 2021 [31] Ireland

T12,

C15

4/8 6/9

25.4 ± 4.9/

25.4 ± 4.9

Resistance 2 sessions/w, 16 min training/session, 6 w
Brett 2020 [32] Ireland

T14,

C14

5/9 5/9 26 ± 2.6 Resistance 14 sessions/w, 30 min/session, 1 year
Matthew2011① [33] America

T10,

C10

0/10 0/10 18–37/18–37 Resistance 3 sessions/w: 1 in-class (120 min) + 2 after-class (30 min/session), 16 w
Matthew2011② [33] America

T10,

C10

0/10 0/10 18–37/18–37 Aerobic 3 sessions/w: 1 in-class (120 min) + 2 after-class (30 min/session), 16 w
Matthew2016① [34] America

T10,

C10

0/10 0/10 18–37/18–37 Resistance 2 sessions/w, 25 min training/session, 8 w
Matthew2016② [34] America

T10,

C10

0/10 0/10 18–37/18–37 Aerobic 2 sessions/w, 25 min training/session, 8 w
Liu2020 [35] China

T40,

C40

26/14 25/15 60–74/60–74 Aerobic 2 sessions/w, 16 min effective training/session, 6 w
Huang 2023① [36] China

T103,

C131

0/103 0/131 18–20/18–20 Mind-body 2 sessions/w, 16 min training/session, 6 w
Huang 2023② [36] China

T120,

C131

0/120 0/131 18–20/18–20 Aerobic 2 sessions/w, 16 min effective training/session, 6 w

M/F: male/female;①GAD-7 ②HAMA ③PSWQ ④STAI-Y2 ⑤STAI-TRAIT⑥SF-36

Quality assessment of included studies

All 10 studies (comprising 14 study entries) were RCTs. The PEDro scale scores ranged from 6 to 8, indicating overall good quality of the included studies. All included studies implemented random allocation and met the criteria for “random allocation,” “baseline similarity,” “between-group statistical comparison,” and “point estimates and variability.” 6 studies satisfied the criterion for “intention-to-treat analysis,” and four studies met the requirement for “assessor blinding.” The PEDro scores ranged from 6 to 8, with an average score of 6.7. No low-quality studies were identified; consequently, the overall quality of the included literature was considered good (Table 2).

Table 2.

Evaluation of the quality of the included literatures

Study 1 2 3 4 5 6 7 8 9 10 11 TS
Yang 2023 [27] 6
Zhang 2016 [28] 6
Kathleen 2025 [29] 6
Matthew 2012 [30] 8
Gordon 2021 [31] 7
Brett 2020 [32] 7
Matthew 2011 [33] 8
Matthew 2016 [34] 6
Liu 2020 [35] 6
Huang 2023 [36] 6

1 eligibility criteria, 2 allocation of randomization, 3 concealed allocation, 4 similarity baseline, 5 subject blinding, 6 therapist blinding, 7 assessor blinding, 8 more than 85% retention, 9 intention-to-treat analysis, 10 between-group comparisons, 11 point and variability measures, TS, total score.

Meta-analysis

The results of the meta analysis are shown in Fig. 2. The pooled effect size was SMD=-0.502 ,[95%CI-0.830,-0.174], P = 0.003), providing suggestive evidence for a potential reduction in generalized anxiety symptoms following exercise intervention, with a moderate effect size. However, the I² statistic was 90.37% (I²>75%), suggesting exceptionally high heterogeneity among the 14 studies. Consequently, despite the statistical significance, this overall pooled effect must be interpreted with low confidence rather than as definitive efficacy.

Fig. 2.

Fig. 2

Forest plot

Subgroup analysis

To explore potential sources of heterogeneity, this study conducted exploratory subgroup analyses based on age, severity of illness, exercise type, intensity, frequency, session time, duration, literature quality, and assessment scales, as shown in Table 3.

Table 3.

Subgroup analysis

Subgroup No of studies Hedges’s g 95%CI P-Value I2
Over all 14 -0.502 [-0.830, -0.174] 0.003 90.37
Age
    Middle-aged and older adults 2 -1.063 [-1.831, -0.295] 0.007 79.02
    Young adults 12 -0.388 [-0.698, -0.057 0.022 90.11
Illness Severity
    Moderate 9 -0.443 [-0.986, 0.100] 0.110 87.91
    Mild 5 -0.460 [-0.718, -0.203] 0.000 69.57
Type
    Resistance 5 -0.302 [-0.964, 0.361] 0.372 69.30
    Aerobic 7 -0.543 [-1.064, -0.022] 0.041 95.38
    Mind-body 2 -0.680 [-0.914, -0.445] 0.000 0.00
Intensity
    Moderate 11 -0.361 [-0.744, -0.023] 0.065 90.54
    Light 3 -0.915 [-1.402, -0.427] 0.000 75.97
Frequency
    1–3 12 -0.401 [-0.745, -0.057] 0.022 90.11
    ≥ 4 2 -1.063 [-1.831, -0.295] 0.007 79.02
Session time
    21–40 3 -1.079 [-1.575, -0.582] 0.000 42.14
    ≤ 20 6 -0.022 [-0.618, -0.574] 0.943 64.84
    ≥ 41 5 -0.623 [-0.997, -0.249] 0.001 93.30
Duration
    6 7 -0.061 [-0.517, 0.395] 0.794 68.10
    8 2 -0.780 [-1.315, -0.246] 0.004 0.00
    ≥ 12 5 -0.868 [-1.255, -0.481] 0.000 89.42
Literature quality
    ≤ 6 8 -0.402 [-0.968,0.164] 0.164 97.00
    > 6 6 -0.606 [-0.939,-0.272] 0.000 0.00
Assessment scale
    Disorder-specific scale 8 -0.700 [-1.029,-0.371] 0.000 90.37
    Trait anxiety scale 4 -0.660 [-1.060,-0.260] 0.001 0.00
    Quality of life scale 2 0.949 [0.320,1.579] 0.003 0.00

The results indicated that exercise was effective for all age groups. In an exploratory analysis, its effect appeared greater in middle-aged and older adults than in young adults. It must be noted that the middle-aged and older subgroup included only two studies with a small sample size and limited statistical power; therefore, these results are strictly exploratory, should not be viewed as definitive exercise prescriptions, and require further validation from more studies. Exercise showed a significant effect for individuals with mild severity (P < 0.001), but the effect was not significant for those with moderate severity (P = 0.110).

Regarding exercise type, preliminary findings suggested thatmind-body exercise demonstrated a potentially large effect (SMD=-0.680, P < 0.001), followed by aerobic exercise (SMD=-0.543, P = 0.041). However, the mind-body subgroup consisted of only two studies, which severely limits statistical power and inherently increases the risk of type I error. Consequently, these specific subgroup results must be interpreted strictly as hypothesis generating rather than robust clinical evidence. Furthermore, the effect of resistance exercise did not reach statistical significance (P = 0.372). Notably, the resistance exercise subgroup included only five studies, several of which had very small sample sizes (for instance, 10 participants per group) and relatively short intervention durations (such as 6 weeks). Therefore, rather than definitively concluding that resistance exercise is ineffective, we consistently state that current evidence is strictly insufficient to determine the true effectiveness of resistance training.

In an exploratory subgroup analysis of intensity, light exercise intensity showed a larger effect (SMD=-0.915) compared to moderate intensity. Since most of the included studies did not report objective intensity indicators (e.g., %HRmax, METs, RPE), the classification of ‘light’ and ‘moderate’ intensity in this study was primarily based on the descriptive definitions from the original studies. This limitation reduces the objectivity and consistency of exercise intensity classification. Consequently, these findings may be influenced by classification bias, and strong conclusions regarding optimal exercise intensity should be avoided.

An exploratory analysis of frequency indicated that high-frequency intervention (≥ 4 times/week) might have a more pronounced effect; however, this subgroup contained only two studies, increasing the risk of false-positive findings and preventing it from being recommended as a definitive exercise prescription. A session time of 21–40 min was identified as the optimal duration range, with an effect size significantly higher than other groups (P < 0.001). Interventions lasting ≤ 20 min were essentially ineffective. Interventions lasting ≥ 12 weeks showed a significant effect (SMD=-0.868, P < 0.001), followed by 8-week interventions (SMD=-0.780, P = 0.004). The effect of the 6-week intervention group did not reach statistical significance (SMD=-0.061, P = 0.794), suggesting that evidence for the benefits of short-term exercise intervention is currently insufficient.

Studies were divided into low risk of bias (PEDro score > 6) and high risk of bias (PEDro score ≤ 6) groups. The results showed that the low risk of bias group (score > 6) had a highly statistically significant pooled effect size (SMD=-0.606, P < 0.001) with no heterogeneity within the group (I²=0.00%). The high risk of bias group (score ≤ 6) had a smaller and non-significant pooled effect size (SMD=-0.402, P = 0.164) along with extremely high heterogeneity within the group (I²=97.00%). When the analysis was limited to the more methodologically rigorous low risk of bias studies, heterogeneity was significantly reduced, almost disappearing, and a more precise, statistically powerful effect estimate was obtained. This further demonstrates that the positive intervention effect of exercise on generalized anxiety is stable and consistent in high-quality studies, indicating that the overall conclusion is not driven by low-quality research.

Subgroup analysis was performed based on the conceptual foundation of the assessment scales, with the following results: Disorder-specific scale group (GAD-7, HAMA, PSWQ): (SMD=-0.700, P < 0.001), significantly effective. Trait anxiety scale group (STAI-Trait): (SMD=-0.660, P = 0.001), also significantly effective with no heterogeneity within the group (I²=0.00%). Quality of life scale group (SF-36): (SMD = 0.949, P = 0.003), significantly effective but in the opposite direction.

Sensitivity analysis

To investigate whether the heterogeneity between studies was driven by individual studies, a leave one out sensitivity analysis was performed by successively excluding each study to examine its impact on the overall effect size and heterogeneity (Table 4). The pooled effect size for all studies was SMD=-0.50,95%CI[-0.83,-0.17], P < 0.05, with I²=90.37%. The results after excluding individual studies showed a range of SMD from − 0.43to-0.61, with 95% CIs ranging from-0.74 to -0.11 and I² values between 86.28% and 91.70%. All P values remained statistically significant (P < 0.05). This indicates that no single outlier study was identified as the primary source of the high heterogeneity (as I² consistently remained ≥ 86.28%). Therefore, this persistently high, unexplained variance confirms that the substantial heterogeneity is deeply rooted in broader clinical and methodological diversity across the included trials (such as mixed diagnostic populations and subjective intensity classifications) rather than individual statistical anomalies. Consequently, while the beneficial direction remained stable across permutations, the precise magnitude of the overall pooled effect is highly susceptible to specific trial characteristics and must be interpreted with low confidence.

Table 4.

Exclusion studies involved the exclusion of the sensitivity effects of individual studies

Exclusion study I2 Hedges P
Yang 2023 [27] 87.39% -0.53[-0.89, -0.18] < 0.05
Zhang2016 [28] 91.66% -0.48[-0.84, -0.12] < 0.05
Kathleen 2025 [29] 84.67% -0.43[-0.76, -0.11] < 0.05
Matthew2012① [30] 91.67% -0.50[-0.88, -0.15] < 0.05
Matthew2012② [30] 91.70% -0.50[-0.85, -0.14] < 0.05
Gordon 2021 [31] 91.67% -0.48[-0.83, -0.13] < 0.05
Brett 2020 [32] 91.59% -0.48[-0.83, -0.12] < 0.05
Matthew2011① [33] 91.69% -0.50[-0.85, -0.15] < 0.05
Matthew2011② [33] 91.70% -0.50[-0.85, -0.14] < 0.05
Matthew2016① [34] 86.28% -0.61[-0.89, -0.32] < 0.05
Matthew2016② [34] 87.99% -0.59[-0.89, -0.29] < 0.05
Liu2020 [35] 88.49% -0.43[-0.74, -0.12] < 0.05
Huang 2023① [36] 90.80% -0.48[-0.84, -0.12] < 0.05
Huang 2023② [36] 90.60% -0.51[-0.87, -0.15] < 0.05

Publication bias

Assessment of publication bias indicated no evidence of significant bias (t = 1.45, p = 0.172 > 0.05), Fig. 3.

Fig. 3.

Fig. 3

Funnel Plot

Quality of evidence assessment

The overall quality of the evidence was rated as moderate (Table 5). Serious limitations were identified: most studies did not provide detailed reports on blinding procedures or adequately describe the allocation concealment. These methodological shortcomings may introduce potential measurement bias in the outcome assessment. Substantial heterogeneity existed in the characteristics of the included populations, such as differences between clinically diagnosed generalized anxiety and subclinical symptoms, and age ranges spanning from young adults to middle-aged and older adults. The intervention protocols were not standardized, with significant variations in exercise type (mind-body/aerobic/resistance), intensity, and frequency, which reduced the generalizability of the results.

Table 5.

GRADE quality of evidence evaluation

Outcome Number of study Quality of evidence evaluation Quality rating
Risk of bias Indirectness Inconsistency Publication bias Quality rating
GAD 14 Serious Not Serious Not Serious Not Serious Not Serious Moderate

Adverse events

None of the included studies provided information regarding adverse events or their safety monitoring protocols. Consequently, we are unable to determine whether adverse events simply did not occur, or if they were undocumented due to a lack of systematic safety reporting. Furthermore, because the original trials completely lacked explicit methodological descriptions regarding safety outcomes, it remains entirely unclear whether any potential adverse events were actively monitored by researchers or merely passively reported by participants.

Discussion

The findings of this study provide suggestive evidence that exercise interventions may assist in ameliorating symptoms of generalized anxiety, which aligns with previous literature [15]. In particular, preliminary subgroup analyses observed a potential positive trend for mind body exercise and aerobic exercise [21]. Inconsistencies also exist within the broader literature. Studies [16, 18] found that aerobic exercise did not show a significant effect on anxiety disorders, which may be attributed to differences in the design of control conditions. When compared to control groups with scheduled exercise time, aerobic exercise did not yield significant differences [18]. Study [16] noted that “the reported effect sizes were influenced by the type of control condition, with trials using waitlist/placebo controls reporting large effects, whereas other trials showed no effects.” Overall, these primary findings offer suggestive empirical support for exercise as an adjunctive strategy in managing generalized anxiety, highlighting the critical need to carefully consider specific exercise modalities and trial designs in future clinical applications. Mechanistically, research indicates that exercise may assist in reducing anxiety by modulating the dopamine system and improving cerebral blood flow regulation [37]. Such biological responses may vary by life stage and exercise intensity. For instance, while high intensity exercise confers long term benefits for adolescents [38], our exploratory analysis observed a potential positive trend for mild intensity exercise in adults. However, considering the subjective nature of intensity definitions, these observations must be interpreted cautiously as hypothesis generating rather than definitive foundations for clinical protocols.

We found that exercise type and parameters may be critical factors influencing intervention outcomes. Exploratory subgroup analysis revealed that both mind-body and aerobic exercises produced significant effects. However, regarding resistance exercise, we strictly caution that our findings must not be interpreted as evidence of no effect; rather, current evidence is simply insufficient to determine its effectiveness. While dedicated meta analyses on resistance training have shown it can significantly improve anxiety symptoms irrespective of training parameters [18], the lack of statistical significance in our specific analysis is heavily confounded by limited statistical power. The inconsistency in our findings likely stems from differences in study populations, intervention durations, and the severely limited number of included studies in this specific subgroup. Specifically, our analysis was based on only five small scale studies, with some individual trials containing as few as 10 participants. Therefore, rather than definitively concluding that resistance exercise is ineffective, we emphasize that further verification through well powered, large scale RCTs is strictly needed. The aerobic exercise subgroup included seven studies with participants ranging from 18 to 74 years of age, showing substantial variations in intervention protocols. These factors contributed to the considerable heterogeneity within this subgroup. Despite this heterogeneity, the pooled effect size remained statistically significant. Research indicates that aerobic exercise can increase the expression of brain-derived neurotrophic factor (BDNF) and promote hippocampal neurogenesis, synaptic remodeling, and neural plasticity, thereby optimizing emotional regulation and improving cognitive flexibility. These mechanisms address core GAD symptoms, such as excessive worry [39]. Additionally, by inhibiting sympathetic nervous system hyperactivity, reducing systemic inflammation, and regulating hypothalamic-pituitary-adrenal (HPA) axis function, exercise reduces anxiety-related physiological arousal [12, 40], In our exploratory analysis, mind-body exercises demonstrated a large effect size. Theoretically, these modalities uniquely combine the neurobiological effects of exercise with cognitive and behavioral components. By enhancing prefrontal cortex (PFC) inhibitory control over the amygdala and modifying excessive responses to threat stimuli, they directly address GAD’s core symptom of uncontrollable worry [39, 41].Mind-body exercises, which emphasize conscious regulation and cognitive restructuring, effectively activate brain regions associated with higher-order executive functions (such as the PFC) [12]. However, it is crucial to emphasize that this mind-body subgroup contained only two studies. Consequently, this limits statistical power, and this finding must be interpreted as strictly exploratory rather than a definitive superiority over aerobic exercise.

Regarding exercise intensity, preliminary subgroup analyses observed a potential positive trend for interventions categorized as light intensity rather than moderate intensity exercises. Mechanistically, this observation aligns with the underlying pathophysiology of the disorder. Patients with GAD typically exhibit sympathetic nervous system hyperactivity and parasympathetic inhibition. Preliminary evidence suggests that light intensity exercise may assist in activating the parasympathetic nervous system, potentially counteracting sympathetic excitation [6] while reducing HPA axis related cortisol levels and alleviating stress responses [42]. However, despite this theoretical rationale, study [17] indicated that high intensity exercise might yield more pronounced effects. The discrepancy between these findings may stem from differences in population characteristics (general clinical anxiety in [17] vs. GAD in our study) or, fundamentally, the subjective definitions of intensity. As noted previously, our classification of intensity lacked objective physiological indicators (e.g., %HRmax, METs) and relied entirely on descriptive labels from the original authors. This inevitably introduces severe classification bias, precluding any definitive comparisons between intensity levels. Furthermore, while these exploratory analyses indicated potentially positive effects in the middle aged and older population, this age subgroup also comprised only two studies. Therefore, we explicitly refrain from designating any exercise intensity as an “optimal” prescription. Any interpretations regarding specific exercise intensities and age specific sensitivities must be highly conservative, strictly viewed as hypothesis generating, and require robust validation in future trials using standardized intensity monitoring.

For exercise frequency, high frequency exercise (≥ 4 sessions/week) demonstrated greater effects than low frequency exercise (1 to3 sessions/week) in our subgroup analysis. This observation suggests that high frequency exercise may be particularly suitable for middle aged and older adults, as age related decline in neural plasticity might require more frequent stimulation to maintain its effects [43]. However, younger adults may face challenges in adhering to high frequency protocols due to academic and work pressures. Their higher baseline neural plasticity may allow low frequency programs (1 to 3 sessions/week) to effectively alleviate anxiety while being more sustainable and carrying lower neural stress risks [44].Regarding intervention duration, programs ≥ 12 weeks showed significant effects, followed by 8 week programs, whereas 6 week programs did not reach statistical significance. This stability may derive from two factors: the time required for internalization and automation of behavioral habits through repeated reinforcement which may establish stable coping strategies that reduce individual response variability [45]; and the remodeling of HPA axis function and restoration of autonomic nervous system (ANS) balance (sympathetic/parasympathetic) which appears to be a gradual process where long term intervention (≥ 12 weeks) is crucial for reestablishing physiological homeostasis [6, 46]. While short term interventions (e.g., 6 weeks) may provide immediate relief, their effects appear to be less stable, with greater residual variability in physiological indicators.We also identified illness severity as a potential moderator of the intervention outcomes. Exercise interventions demonstrated significant efficacy in patients with mild anxiety, whereas those with moderate anxiety showed non significant improvements. Functional magnetic resonance imaging (fMRI) studies have revealed that generalized anxiety disorder is associated with hyperactivity and dysregulation in the right prefrontal cortex (PFC) [47]. In mild cases, PFC dysregulation may represent a “window of neuroplasticity” where appropriate exercise could promote neural circuit reorganization and activate the BDNF/TrkB pathway, thereby restoring normal PFC function [48, 49] and ultimately leading to clinical improvement. Additionally, patients with mild anxiety may exhibit higher sensitivity to exercise induced changes in endogenous neurotransmitter levels. These findings underscore the potential role of exercise in ameliorating mild anxiety. In contrast, patients with moderate anxiety often present with neuroendocrine disturbances, such as HPA axis dysfunction, and may demonstrate neuropathological changes, including amygdala hyperactivity and PFC hypofunction. These alterations might elevate the threshold for neuroplasticity, making conventional exercise interventions insufficient to activate the necessary molecular pathways, such as BDNF expression and synaptic plasticity, thereby compromising neural remodeling [50]. Consequently, exercise alone may be inadequate to reverse pathological states in moderate cases, requiring a combination of pharmacological or psychological therapies for comprehensive treatment efficacy [51].

Furthermore, situating our findings within the broader mental health literature highlights the potential transdiagnostic effects of exercise [52]. Beyond generalized anxiety, physical activity has demonstrated efficacy in ameliorating symptoms across a spectrum of psychiatric conditions, including major depressive disorder and panic attacks [53]. This transdiagnostic utility is likely mediated by shared neurobiological pathways, such as the generalized reduction of systemic neuroinflammation and the universal enhancement of neuroplasticity across distinct brain regions [54]. Therefore, conceptualizing exercise not merely as a disorder specific intervention, but rather as a broad spectrum transdiagnostic mental health strategy, may optimize its integration into routine psychiatric care and provide more holistic benefits for patients presenting with complex comorbidities [55].

Limitations

Several limitations of this meta analysis should be acknowledged. First, regarding the literature search and population, our primary strategy purposefully restricted keywords to GAD specific terminology to maintain strict diagnostic homogeneity and high search specificity. While we extracted independent GAD subgroup data from broader trials when encountered during full text screening, we did not systematically search broader anxiety terminologies to actively identify hidden subgroups. This approach introduces a potential selection bias, as it likely excluded relevant, broader clinical trials that may have contained GAD specific subgroups. More importantly, the included studies encompassed a mixed population; some studies utilized formal psychiatric diagnostic interviews, while others relied solely on validated screening scales to identify participants. This mixing of clinically diagnosed patients with individuals experiencing subclinical anxiety symptoms introduces substantial clinical heterogeneity. This remains a major limitation that significantly affects the clinical applicability of our findings. Due to the limited total number of included studies, we were unable to conduct a statistically powered stratified analysis to differentiate the intervention effects between these two distinct populations, necessitating a more cautious interpretation of the results. Additionally, the inclusion criteria required participants to have no major comorbidities to accurately estimate the independent effect of exercise. While methodologically sound, this limits the generalizability of the findings to more common clinical populations with comorbidities.

Second, from a statistical perspective, a major limitation is the exceptionally high heterogeneity observed in the overall analysis (I² = 90.37%), which substantially weakens the interpretability of the pooled effect. While we conducted sensitivity and subgroup analyses to explore this variance, they did not fully explain the sources of heterogeneity. The limited number of included studies precluded the use of meta-regression to further investigate continuous covariates without risking severely underpowered results. This substantial, unexplained heterogeneity inherently limits the strength of causal inference. Consequently, the overall pooled effect provides only suggestive evidence rather than definitive efficacy, and these findings must be interpreted with low confidence. Furthermore, given the small number of included studies, the assessment of publication bias (e.g., funnel plot inspection) may also be underpowered, which is a standard caveat in such meta-analyses.

Third, regarding subgroup analyses and exercise parameters, several subgroups specifically those analyzingmind-body exercise, high frequency interventions, and middle aged/older adults contained a very small number of studies (often N = 2). This limits statistical power and creates a high risk of type I error (false positives). Furthermore, the classification of exercise intensity (e.g., light vs. moderate) relied primarily on descriptive terminology from the original studies rather than objective physiological indicators (such as %HRmax, METs, or RPE), which inevitably introduces significant classification bias. Therefore, any interpretations regarding optimal intensity must be highly conservative. Consequently, these specific subgroup findings must be interpreted strictly as hypothesis-generating only, rather than robust evidence for definitive clinical exercise prescriptions, and require validation in future large-scale studies.

Finally, at the trial design level, blinding participants and therapists is inherently challenging in exercise interventions, which may lead to performance and detection biases, potentially inflating the effect size. Moreover, most studies did not adequately describe allocation concealment. Lastly, this study did not quantitatively analyze the specific interaction between exercise and anti-anxiety medications (e.g., SSRIs). Given that exercise and medications may produce synergistic effects through shared neurobiological mechanisms, future high quality RCTs should focus on exploring these interactions, standardizing the reporting of objective exercise intensity, and obtaining individual participant data to provide more precise clinical guidance.

Conclusion

In conclusion, while the current meta analysis provides suggestive evidence that exercise interventions may assist in ameliorating symptoms of generalized anxiety, the exceptionally high unexplained heterogeneity and mixed diagnostic populations dictate that these overall findings must be interpreted with low confidence rather than as definitive efficacy. Our comprehensive subgroup analyses revealed that the intervention effects appear to be modulated by various exercise parameters and patient characteristics. Specifically, potential positive trends were observed for mind body exercises, interventions subjectively classified as light intensity, programs with higher training frequencies, moderate single session times, and prolonged overall durations. Additionally, middle aged and older adults, as well as patients with mild anxiety, appeared to show greater responsiveness to the interventions.However, we strongly caution against synthesizing these variables into an “optimal” exercise protocol. While the overall analysis encompasses a robust literature base, the specific parameter strata demonstrating the most pronounced effects are often derived from subgroups with severely limited sample sizes and rely heavily. This inherently introduces substantial classification bias and a high risk of type I error.Consequently, these specific regimen parameters and age related observations must be viewed strictly as hypothesis generating only, rather than reliable foundations for clinical prescriptions. Robust, large scale, and highly homogeneous randomized controlled trials are urgently needed before definitive, personalized exercise guidelines for generalized anxiety can be established.

Supplementary Information

Supplementary Material 1. (27.5KB, docx)

Acknowledgements

We gratefully acknowledge the support of the studies included in the analysis.

Registration and protocol

This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for meta-analyses and has been prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) under registration number No. CRD420251146169.

Abbreviations

GAD-7

Generalized Anxiety Disorder 7-item scale

HAMA

Hamilton Anxiety Rating Scale

PSWQ

Penn State Worry Questionnaire

STAI-Y2

State-Trait Anxiety Inventory Form Y-2

STAI-Trait

State-Trait Anxiety Inventory (Trait subscale)

SF-36

36-Item Short Form Health Survey

SMD

Standardized Mean Difference

CI

Confidence Interval

RCT

Randomized Controlled Trial

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

Authors’ contributions

H. S. (Haiman Su) conceived and designed the study, performed the statistical analysis using Stata/SE 17.0, interpreted the results, and drafted the initial manuscript; L. L. (Lili Li) and J. W. (Jun Wu) contributed to the development of the research protocol, led the process of literature screening and data extraction, and performed the methodological quality assessment of included studies using the PEDro scale; M. N. (Manh Cuong Nguyen), H. D. (Hui Ding), and C. W. (Chen Wei) coordinated the data collection process, reviewed the quality of included studies, and provided support for statistical analysis and interpretation; Z. Z. (Zhenyu Zhang) provided overall supervision of the project, reviewed and revised the final version of the manuscript, and ensured the scientific integrity of the study. All authors read and approved the final version of the manuscript and agreed to be accountable for all aspects of the work.

Funding

The authors declare that this study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

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.Ruscio AM, Hallion LS, Lim CC, W, et al. Cross-sectional Comparison of the Epidemiology of DSM-5 Generalized Anxiety Disorder Across the Globe [J]. JAMA Psychiatry. 2017;74(5):465–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Pollack MH. Refractory generalized anxiety disorder [J]. J Clin Psychiatry. 2009;70(Suppl 2):32–8. [DOI] [PubMed] [Google Scholar]
  • 3.Gorman JM. Generalized anxiety disorder [J]. Clin Cornerstone. 2001;3(3):37–46. [DOI] [PubMed] [Google Scholar]
  • 4.Demartini J, Patel G, Fancher TL. Generalized Anxiety Disorder [J]. Ann Intern Med. 2019;170(7):Itc49–64. [DOI] [PubMed] [Google Scholar]
  • 5.Jamali Z, Hosseiniara R, Khalili P, et al. The association between major depressive disorder and generalized anxiety disorder with hematological indices in the Rafsanjan youth cohort study [J]. Sci Rep. 2025;15(1):17396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Verdonk C, Teed AR, White EJ, et al. Heartbeat-evoked neural response abnormalities in generalized anxiety disorder during peripheral adrenergic stimulation [J]. Neuropsychopharmacology. 2024;49(8):1246–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Ma Z, Zhao M, Zhao H, et al. Causal role of immune cells in generalized anxiety disorder: Mendelian randomization study [J]. Front Immunol. 2023;14:1338083. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Locke AB, Kirst N. Diagnosis and management of generalized anxiety disorder and panic disorder in adults [J]. Am Fam Physician. 2015;91(9):617–24. [PubMed] [Google Scholar]
  • 9.Solmi M, Basadonne I. Exercise as a transdiagnostic intervention for improving mental health: An umbrella review [J]. J Psychiatr Res. 2025;184:91–101. [DOI] [PubMed] [Google Scholar]
  • 10.Skou ST, Nyberg M, Dideriksen M, et al. Exercise therapy and self-management support for individuals with multimorbidity: a randomized and controlled trial. Nat Med. 2025;31(9):3176–82. [DOI] [PMC free article] [PubMed]
  • 11.Ashdown-Franks G, Firth J, Carney R, et al. Exercise as Medicine for Mental and Substance Use Disorders: A Meta-review of the Benefits for Neuropsychiatric and Cognitive Outcomes [J]. Sports Med. 2020;50(1):151–70. [DOI] [PubMed] [Google Scholar]
  • 12.Byrne GJ. Interventions for generalized anxiety disorder [J]. Curr Opin Psychiatry. 2023;36(2):134–9. [DOI] [PubMed] [Google Scholar]
  • 13.Saramago P, Gega L. Digital Interventions for Generalized Anxiety Disorder (GAD): Systematic Review and Network Meta-Analysis [J]. Front Psychiatry. 2021;12:726222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Stein MB. Neurobiology of generalized anxiety disorder [J]. J Clin Psychiatry. 2009;70(Suppl 2):15–9. [DOI] [PubMed] [Google Scholar]
  • 15.Stubbs B, Vancampfort D. An examination of the anxiolytic effects of exercise for people with anxiety and stress-related disorders: A meta-analysis [J]. Psychiatry Res. 2017;249:102–8. [DOI] [PubMed] [Google Scholar]
  • 16.Bartley CA, Hay M, Bloch MH. Meta-analysis: aerobic exercise for the treatment of anxiety disorders [J]. Prog Neuropsychopharmacol Biol Psychiatry. 2013;45:34–9. [DOI] [PubMed] [Google Scholar]
  • 17.Aylett E, Small N. Exercise in the treatment of clinical anxiety in general practice - a systematic review and meta-analysis [J]. BMC Health Serv Res. 2018;18(1):559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Gordon BR, Mcdowell CP, Lyons M, et al. The Effects of Resistance Exercise Training on Anxiety: A Meta-Analysis and Meta-Regression Analysis of Randomized Controlled Trials [J]. Sports Med. 2017;47(12):2521–32. [DOI] [PubMed] [Google Scholar]
  • 19.Ensari I, Greenlee TA, Motl RW, et al. Meta-analysis of acute exercise effects on state anxiety: an update of randomized controlled trials over the past 25 years [J]. Depress Anxiety. 2015;32(8):624–34. [DOI] [PubMed] [Google Scholar]
  • 20.Tian J, Kong Q, Yu H, et al. Comparative Efficacy of Acute Exercise Intervention on Anxiety in Adolescents and Young Adults: A Network Meta-Analysis [J]. Scand J Med Sci Sports. 2024;34(10):e14735. [DOI] [PubMed] [Google Scholar]
  • 21.Gonçalves-Bradley DC Millerkj, Areerob P, et al. Comparative effectiveness of three exercise types to treat clinical depression in older adults: A systematic review and network meta-analysis of randomised controlled trials [J]. Ageing Res Rev. 2020;58:100999. [DOI] [PubMed] [Google Scholar]
  • 22.Page Mj, Bossuyt P Moherd. PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews [J]. BMJ. 2021;372:n160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Spitzer Rl, Kroenke K, Williams JB, et al. A brief measure for assessing generalized anxiety disorder: the GAD-7 [J]. Arch Intern Med. 2006;166(10):1092–7. [DOI] [PubMed] [Google Scholar]
  • 24.Ludyga S, Gerber M, Pühse U, et al. Systematic review and meta-analysis investigating moderators of long-term effects of exercise on cognition in healthy individuals [J]. Nat Hum Behav. 2020;4(6):603–12. [DOI] [PubMed] [Google Scholar]
  • 25.Huang Q, Ren XY, Zhang R, et al. Application of GRADE in Chinese clinical practice guidelines/expert consensus [J]. Chin J Evid Based Med. 2021;21(12):1457–62. [Google Scholar]
  • 26.Zeng XT, Leng WD, Li S, et al. How to Understand and Use GRADE system Correctly? A Briefly Outline [J]. Chin J Evid Based Med. 2011;11(9):985–90. [Google Scholar]
  • 27.Yang SQ, Wang WF, Fu JH, et al. Effect of aerobics course on improving college students’ emotional state [J]. China J Health Psychol. 2023;31(5):771–5. [Google Scholar]
  • 28.Zhang J, Zhang WW, Shen H. The Effect of Baduanjin Practice on the Generalized Anxiety Disorder [J]. Chin J Sports Med. 2016;35(3):231–3. [Google Scholar]
  • 29.Latimer K, Larok R, Nyeko JP, et al. A sport-for-protection program reduces anxiety and depression in youth affected by displacement: A randomized controlled trial of the Game Connect program in Uganda [J]. J Affect Disord. 2025;376:84–91. [DOI] [PubMed] [Google Scholar]
  • 30.Herring MP, Jacob ML, Suveg C, et al. Feasibility of exercise training for the short-term treatment of generalized anxiety disorder: a randomized controlled trial [J]. Psychother Psychosom. 2012;81(1):21–8. [DOI] [PubMed] [Google Scholar]
  • 31.Gordon BR, Mcdowell CP, Lyons M, et al. Resistance exercise training among young adults with analogue generalized anxiety disorder [J]. J Affect Disord. 2021;281:153–9. [DOI] [PubMed] [Google Scholar]
  • 32.O’sullivan D, Gordon BR, Lyons M, et al. Effects of resistance exercise training on depressive symptoms among young adults: A randomized controlled trial [J]. Psychiatry Res. 2023;326:115322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Herring MP, Jacob ML, Suveg C, et al. Effects of short-term exercise training on signs and symptoms of generalized anxiety disorder [J]. Ment Health Phys Act. 2011;4(2):71–7. [Google Scholar]
  • 34.Herring MP, Johnson KE, O’connor PJ. Exercise training and health-related quality of life in generalized anxiety disorder [J]. Psychol Sport Exerc. 2016;27:138–41. [Google Scholar]
  • 35.Liu QL, Huang HL. Therapeutic effect of fitness exercise on elderly patients with mild to moderate generalized anxiety disorder [J]. Chin J Gerontol. 2020;40(17):3704–6. [Google Scholar]
  • 36.Huang X, Sun T. The effect of yoga classes on the mental health of female college students [J]. J Sports Med Phys Fit. 2023;63(11):1244–50. [DOI] [PubMed] [Google Scholar]
  • 37.Juarez EJ, Samanez-Larkin GR, Exercise. Dopamine, and Cognition in Older Age [J]. Trends Cogn Sci. 2019;23(12):986–8. [DOI] [PubMed] [Google Scholar]
  • 38.Miller K, Morley C, Fraser BJ, et al. Types of leisure-time physical activity participation in childhood and adolescence, and physical activity behaviours and health outcomes in adulthood: a systematic review [J]. BMC Public Health. 2024;24(1):1789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Smith PJ, Merwin RM. The Role of Exercise in Management of Mental Health Disorders: An Integrative Review [J]. Annu Rev Med. 2021;72:45–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Degeorge KC Groverm, Streeter GS. Generalized Anxiety Disorder and Panic Disorder in Adults [J]. Am Fam Physician. 2022;106(2):157–64. [PubMed] [Google Scholar]
  • 41.Yan H, Han Y, Shan X, et al. Common and exclusive spontaneous neural activity patterns underlying pure generalized anxiety disorder and comorbid generalized anxiety disorder and depression [J]. J Affect Disord. 2023;331:82–91. [DOI] [PubMed] [Google Scholar]
  • 42.Mehta UV, Wilt GE, Roscoe C, et al. The Association between Multiple Environmental Exposures and Symptoms of Generalized Anxiety Disorder in a Prospective, US-Based Cohort Study [J]. Environ Health Perspect. 2025;133(6):67008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Mcdowell CP, Wilson KE, Monroe DC, et al. Physical activity partially mediates associations between Big personality traits and incident generalized anxiety disorder: Findings from the irish longitudinal study on ageing [J]. J Affect Disord. 2020;277:46–52. [DOI] [PubMed] [Google Scholar]
  • 44.Ndabi JS, Simwanza A, Chukwuorji JC, et al. Physical activity and anxiety during the COVID-19 pandemic in Tanzania: insights for public health policy in low-income contexts [J]. Front Public Health. 2024;12:1483153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Jacobson NC. Digital phenotyping of generalized anxiety disorder: using artificial intelligence to accurately predict symptom severity using wearable sensors in daily life [J]. Transl Psychiatry. 2022;12(1):336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Chen Y, Cui Q, Fan YS, et al. Progressive brain structural alterations assessed via causal analysis in patients with generalized anxiety disorder [J]. Neuropsychopharmacology. 2020;45(10):1689–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Luber BM Daviss. Using neuroimaging to individualize TMS treatment for depression: Toward a new paradigm for imaging-guided intervention [J]. NeuroImage. 2017;148:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Andrzejewski ME, Mckee BL, Baldwin AE, et al. The clinical relevance of neuroplasticity in corticostriatal networks during operant learning [J]. Neurosci Biobehav Rev. 2013;37(9 Pt A):2071–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Bouayad-Gervais K, Minogianis EA, Lévesque D, et al. The self-administration of rapidly delivered cocaine promotes increased motivation to take the drug: contributions of prior levels of operant responding and cocaine intake [J]. Psychopharmacology. 2014;231(21):4241–52. [DOI] [PubMed] [Google Scholar]
  • 50.Costa AS Haegera, Romanzetti S, et al. Effect of a multicomponent exercise intervention on brain metabolism: A randomized controlled trial on Alzheimer’s pathology (Dementia-MOVE) [J]. Alzheimers Dement (N Y), 2020;6(1):e12032. [DOI] [PMC free article] [PubMed]
  • 51.Lavie CJ. Milani RV. Generalized anxiety disorder [J]. N Engl J Med. 2004;351(21):2239. [DOI] [PubMed] [Google Scholar]
  • 52.Singh B, Olds T, Curtis R, et al. Effectiveness of physical activity interventions for improving depression, anxiety and distress: an overview of systematic reviews. Br J Sports Med. 2023;57(18):1203–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Ashdown-Franks G, Firth J, Carney R, et al. Exercise as Medicine for Mental and Substance Use Disorders: A Meta-review of the Benefits for Neuropsychiatric and Cognitive Outcomes. Sports Med. 2020;50(1):151–70. [DOI] [PubMed] [Google Scholar]
  • 54.Fernández-Rodríguez R, Álvarez-Bueno C, Martínez-Ortega IA, et al. Immediate effect of high-intensity exercise on brain-derived neurotrophic factor in healthy young adults: A systematic review and meta-analysis. J Sport Health Sci. 2022;11(3):367–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Stubbs B, Vancampfort D, Hallgren M, et al. EPA guidance on physical activity as a treatment for severe mental illness: a meta-review of the evidence and Position Statement from the European Psychiatric Association (EPA), supported by the International Organization of Physical Therapists in Mental Health (IOPTMH). Eur Psychiatry. 2018;54:124–44. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1. (27.5KB, docx)

Data Availability Statement

The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.


Articles from BMC Sports Science, Medicine and Rehabilitation are provided here courtesy of BMC

RESOURCES