Optimal exercise modality and dose for cognitive function in individuals with Alzheimer’s disease and mild cognitive impairment : a systematic review and metaanalysis

다국어 초록 (Multilingual Abstract)

Alzheimer’s disease (AD) and mild cognitive impairment (MCI) are the most common cognitive disorders in older adults. Both conditions are associated with increased disability risk, impaired quality of life, and higher mortality. Physical exercise is a safe, low-cost, and widely accessible non-pharmacological intervention with growing evidence supporting its beneficial effects on cognitive function.
To investigate the dose–response relationships between various exercise modalities and their effectiveness in improving cognitive function in individuals with AD and MCI.
PubMed, Web of Science, Embase, the Cochrane Library, Medline, and Scopus were systematically searched up to October 2025 without restrictions on publication year. Randomized controlled trials (RCTs) were included. Data were analyzed using R software, with the MBNMAdose package to compute standardized mean differences (SMD) and 95% credible intervals (95% CrI).
A total of 45 RCTs involving 2,736 participants were included. A significant nonlinear dose–response relationship between exercise and global cognitive function was observed in individuals with AD and MCI. In the overall population, the optimal effective dose of exercise ranged from 520 to 1200 MET-min/week, with the peak cognitive benefit occurring at approximately 1200 MET-min/week. Multicomponent exercise (ME) demonstrated the greatest improvement in cognitive function among all exercise modalities and showed a clear advantage at higher dose levels (around 1200 MET-min/week). Subgroup analyses indicated that the optimal dose for cognitive improvement was approximately 1200 MET-min/week in individuals with AD, whereas a lower optimal dose of approximately 930 MET-min/week was observed in those with MCI.
Meta-regression analysis revealed a significant interaction between intervention duration and disease stage, suggesting that individuals with MCI are more responsive to longer intervention periods.
Physical exercise exhibits a clear dose–response relationship with cognitive improvement in individuals with AD and MCI. ME is recommended as the primary training modality (e.g., 4 timesper week, 60 minutes per session). For individuals with MCI, extending the intervention duration beyond this framework may further maximize cognitive benefits. Further high-quality randomized controlled trials are warranted to confirm the safety and long-term effectiveness of this dose range.

국문 초록 (Abstract)

알츠하이머병과 경도인지장애는 노인 인구에서 가장 흔한 인지기능 저하질환으로 기능 장애 위험 증가, 삶의 질 저하 및 사망률 증가와 밀접한 관련이있다. 신체활동은 비약물적 중재로 비교적 안전하고 저비용으로 인지기능의 저하를억제할 수 있는 것으로 보고된다.
본 연구의 목적은 알츠하이머병 및 경도인지장애 환자 각각의 다양한 운동유형과 운동량에 따른 인지기능 개선의 용량–반응 관계를 규명하는 데 있다.
문헌 검색은 PubMed, Web of Science, Embase, Cochrane Library, Medline, Scopus 데이터베이스를 대상으로 2025 년 10 월까지 발표 연도에 제한 없이 수행하였다. 사전에 정의된 포함 기준에 따라 무작위 대조군 연구를 선정하였으며, R소프트웨어의 MBNMAdose 패키지를 이용하여 표준화 평균차와 95% 신뢰구간을산출하였다.
총 45 편의 RCT 가 포함되었으며, 전체 대상자는 2,736 명이었다. 알츠하이머병및 경도인지장애 환자에서 신체활동과 전반적 인지기능 간에는 유의한 비선형용량–반응 관계가 관찰되었다. 전체 대상자 분석에서 최적의 유효 운동량은 주당520–1200 MET-min 으로 나타났으며 약 1200 MET-min/week 에서 인지기능 개선효과가 최고점을 보였다. 운동 유형별 분석 결과, 복합운동은 모든 운동 유형 중
가장 큰 인지기능 개선 효과를 나타냈으며 특히 약 1200 MET-min/week 의 고용량구간에서 인지기능 개선 효과가 더 크게 나타났다. 하위 분석에서는 알츠하이머병환자에서 약 1200 MET-min/week 가 최적 운동량으로 나타났고 경도인지장애환자에서는 약 930 MET-min/week 의 더 낮은 최적 운동량이 확인되었다. 메타회귀분석 결과 중재 기간과 질환 단계 간의 유의한 상호작용이 관찰되었으며 이는 경도인지장애 환자에서 중재 기간이 길수록 인지기능 개선 효과가 더 크게 나타남을 보여준다.
결론적으로, 신체활동은 알츠하이머병 및 경도인지장애 환자의 인지기능개선에 있어 명확한 용량–반응 관계를 보이는 것으로 나타났다. 복합운동은 주요권장 운동 형태로 제시되며(예: 주 4 회, 회당 60 분), 특히 경도인지장애 환자에서는 해당 운동 처방을 장기간 유지할 경우 인지기능 개선 효과를 더욱 극대화할 수 있는 것으로 나타났다. 향후 연구에서는 본 연구에서 제시한 운동량 범위의 안전성과 장기적 효과를 추가적으로 검증할 필요가 있을 것으로 사료된다.

목차 (Table of Contents)

• CHAPTER Ⅰ 1
• Introduction 1
• 1. Significance 1
• 2. Purpose 4
• 3. Hypotheses 4
• CHAPTER Ⅱ 5
• Literature Review 5
• 1. Alzheimer's Disease and Mild Cognitive Impairment 5
• 1.1 Definition and Epidemiology 5
• 1.2 Diagnostic Criteria 8
• 1.3 Pathophysiology 9
• 2. Exercise on Cognitive Function 10
• 3. Biological Mechanisms Linking Exercise and Cognitive Function in Cognitive Decline 11
• 4. Dose-response Relationship in Exercise Intervention 13
• 5. Need for a Bayesian Model Based Dose-response Network Meta-analysis 14
• CHAPTER Ⅲ 16
• Methods 16
• 1. Study Design 16
• 2. Data Sources and Search Strategy 16
• 3. Eligibility Criteria (PICO) 16
• 3.1 Population 17
• 3.2 Intervention 17
• 3.3 Control Group (Comparison) 18
• 3.4 Outcomes 18
• 3.5 Study Design 18
• 4. Study Selection and Data Extraction 19
• 5. Risk of Bias 21
• 6. Outcome Measurement 21
• 7. Primary Bayesian Dose-response Network Meta-analysis Model 22
• 8. Ranking of Exercise-Dose Interventions 23
• 9. Network Consistency, Fit and Heterogeneity Checks 24
• 10. Subgroup Analysis 25
• 11. Meta-regression 25
• 12. Sensitivity Analysis 26
• 13. Statistical Analysis 26
• CHAPTER Ⅳ 28
• Results 28
• 1. Study Characteristics 28
• 2. Risk of Bias 28
• 3. Dose-response Network Meta-analysis 31
• 3.1 Connectivity 31
• 3.2 Consistency 32
• 3.3 Transitivity 33
• 3.4 Selection of Models 36
• 3.4.1 Model Candidates and Rationale for Comparison 36
• 3.4.2 Linear and Non-linear Functions and Models Fit Comparison 36
• 3.4.3 Models fit comparison 38
• 3.5 Dose-response Relationships 41
• 3.6 Ranking of Effectiveness of Total and Different Exercise 45
• 4. Subgroup Analysis 47
• 5. Meta-regression (Covariates) 50
• 6. Exercise Prescription Recommendation 51
• 7. Sensitivity Analysis 53
• CHAPTER Ⅴ 54
• Discussion 54
• 1. Interpretation and Implications of the Main Findings 54
• 2. Strengths 58
• 3. Limitations 59
• 4. Future Research Directions 59
• CHAPTER Ⅵ 61
• Conclusion 61
• References 62
• Appendix 75