Dementia-MOVE
Longitudinal effects of physical activity on sleep in association with hippocampal-amygdala subfield structure and functional connectivity in Alzheimer's disease
Clinical Question
Does a 6-month multimodal exercise intervention preserve hippocampal-amygdala subfield structure and connectivity, and how do these relate to sleep parameters in patients on the Alzheimer's disease continuum over 18 months?
Bottom Line
A 6-month multimodal exercise program transiently preserved hippocampal-amygdalar subfield volumes and improved hippocampus-amygdala functional connectivity during the active intervention phase, but disease progression resumed after intervention ended, supporting the need for sustained continuous physical activity for lasting benefits in AD.
Major Points
- A 6-month multimodal exercise intervention transiently preserved hippocampal-amygdalar subfield volumes
- The intervention improved hippocampus-amygdala functional connectivity during the active phase
- Multimodal sleep assessment (PSQI, diaries, wearables) related to MTL structure and connectivity
- Physical activity and sleep are linked to MTL integrity and cognition in AD
- Long-term benefits require sustained intervention; disease progression resumes after intervention ends
Design
Study Type: Randomized controlled trial (substudy analyzing structural, functional connectivity, and brain-behavior outcomes)
Randomization: 1
Allocation: Randomization performed with WINPEPI software, stratified by age and fitness level (VO2max or 6MWT distance); not stratified by clinical staging. In case of odd numbers, stratification favored the intervention group.
Follow-up Duration: 18 months (T1 baseline, T2 3 months, T3 6 months immediately after end of 6-month intervention, T4 18 months / 12 months after end of intervention)
Centers: 1
Countries: Germany
Sample Size: 46
Analyzed: 46
Analysis: Longitudinal effects evaluated using linear mixed-effects models; structural analysis included all participants; functional analysis included 23 intervention and 19 control participants at baseline (16 controls at T3) after motion-based QC exclusion. Sensitivity analyses excluded 2 participants without CSF biomarkers and 2 with obstructive sleep apnea.
Power Calculation: Sample size based on prior intervention studies and power analyses for the primary trial outcome (metabolic MRI), estimating 22 participants per group. No separate a priori power calculation for present structural, functional connectivity, and brain-behavior analyses.
Registration: ClinicalTrials.gov NCT03939286
Inclusion Criteria
- Age 50-80 years
- Diagnosis of mild cognitive impairment (MCI) or mild dementia
- Fulfilling both clinical and biomarker criteria of AD alongside the AD continuum according to AT(N) NIA-AA criteria
- Pathological amyloid accumulation (A+) based on CSF Aβ1-42 or Aβ42/40 ratio, or in cases where CSF unavailable, diagnosis based on neuroimaging, neuropsychological findings and expert clinical consensus per national dementia guidelines
Arms
| Field | Multimodal exercise intervention + psychoeducation | Control |
|---|---|---|
| N | 25 | 21 |
| Intervention | 6-month supervised multimodal exercise sessions combining aerobic endurance, resistance, and coordination training; weekly 60-min group sessions delivered by trained exercise therapists following a standardized protocol; participants also instructed to perform moderate- to high-intensity home-based training ≥30 min/week plus stretching/toning ≥15 min twice/week. Also received monthly structured psychoeducational sessions. | Monthly structured psychoeducational sessions (~1 h) covering diagnostic procedures, therapeutic options, cognitive training, lifestyle factors, depression, sleep, communication, and social-medical aspects of dementia care, plus a separate nutrition session. Designed as active control per FINGER-informed approach. |
| Duration | 6 months | 6 months |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| Substudy outcomes (not designated as primary in the source): hippocampal and amygdala subfield volumes and hippocampus-amygdala functional connectivity analyzed longitudinally. The parent Dementia-MOVE trial's primary outcome was metabolic MRI; no separate a priori power calculation was performed for the present structural, connectivity, or brain-behavior analyses. | Primary | Psychoeducation control group progressed with hippocampal atrophy over follow-up | Intervention group showed increased hippocampus-amygdala connectivity and attenuated subfield atrophy after 6-month intervention; both groups progressed in cognitive decline and hippocampal atrophy by 1 year later | ||
| Associations between hippocampal/amygdalar volumes and connectivity with sleep parameters (PSQI, diary, wearables) | Secondary | Hippocampal and amygdalar volumes and connectivity were associated with sleep parameters | |||
| Associations between hippocampal/amygdalar volumes and connectivity with activity parameters | Secondary | Volumes and connectivity were associated with activity parameters | |||
| Associations between hippocampal/amygdalar volumes and connectivity with cognitive parameters | Secondary | Volumes and connectivity were associated with cognitive parameters | |||
Subgroup Analysis
Sensitivity analyses performed excluding 2 participants without CSF biomarker confirmation and separately excluding 2 participants from the intervention group with diagnosed obstructive sleep apnea.
Criticisms
- Small sample size (n=46) with no separate a priori power calculation for structural, functional connectivity, or brain-behavior analyses
- Single-country (Germany), single-center study limiting generalizability
- No rs-fMRI data at T4 due to shortened MRI protocol, limiting long-term functional connectivity assessment
- Wearable (FitBit) data collected only for ~2 weeks mid-intervention
- Randomization not stratified by clinical staging
- Benefits appeared transient, with disease progression resuming after intervention ended
- Findings link sleep parameters to MTL measures but do not demonstrate that exercise causally improved sleep
Based on: Dementia-MOVE (Alzheimer's & Dementia, 2026)
Authors: Dadsena R, Costa AS, David S, ..., Haeger A
Citation: Alzheimers Dement. 2026 Aug 25;22(8):e71651. doi: 10.1002/alz.71651
Content summarized and formatted by NeuroTrials.ai.