Does exercise slow Parkinson's disease?

Does exercise slow the disease down, or does it only mask symptoms?

Clinical trials do not show that exercise stops brain cells from dying. Although physical activity improves movement, no human trial proves that it slows the underlying disease or preserves dying dopamine neurons. A 2023 systematic review of 49 trials involving 2,104 participants examined pooled data from six trials with 469 participants tested in the "off" state, an assessment conducted after withholding medication overnight to reduce drug masking. Exercise produced a modest score improvement, but the evidence was rated as low certainty; better scores could easily reflect temporary physical conditioning rather than salvaged neurons.

Biological markers and brain scans remain inconclusive. The 2023 review pooled four trials measuring blood levels of brain-derived neurotrophic factor, a protein that supports neuron survival, but found an increase supported only by very low certainty evidence. A 2025 randomised trial by Lorek and colleagues assigned 30 participants to 12 weeks of aerobic interval training or usual care, using fluorodopa positron emission tomography, an imaging scan that visualises how the brain converts dopamine precursors. While the 15 exercising participants improved their movement, the trial was far too small to establish whether training prevented dopamine terminals from degenerating.

This piece is about one half of the question. For what exercise has been shown to do, and how strong that evidence is, see What exercise has been shown to do, and what it hasn't.

Few trials test whether benefits survive once training stops. In the 2023 review, only nine of 49 trials tracked participants after exercise ended, across breaks of 2.5 to 96 weeks, and conflicting study designs prevented pooling the results. The SPARX phase 2 trial gave 128 participants six months of treadmill exercise and met a non-futility threshold, a statistical target designed to justify running a larger study rather than to prove clinical efficacy. A phase 3 trial evaluating 370 participants over 12 months, logged under the trial registration NCT04284436, has not yet published its results.

Can brain scans show if exercise protects dopamine cells?

Brain scans do not show that exercise saves dopamine cells. Dopamine transporter imaging, such as DaT-SPECT scans that measure proteins recycling dopamine on nerve endings, tracks the rate at which Parkinson's damages those cells over time. While exercise helps people move more easily day to day, scans tracking these transporter proteins show that physical activity does not slow or stop the loss of the neurons themselves.

The clearest tracking data in people found no cellular preservation. An observational study tracking 173 people with early Parkinson's over two years found that dopamine transporter binding fell at the same rate no matter how much participants exercised. Two small randomised trials, one with 25 scanned participants and another with 30, did show positive scan changes after training programmes, but they measured temporary dopamine release and dopamine production rather than whether dying nerve endings were kept alive.

Better movement does not mean the underlying disease has slowed. A 2023 systematic review combining 49 exercise trials with 2,104 participants found that exercise improved motor scores when participants were off medication, but found no robust evidence of slowed disease progression. As researchers noted in a trial registration for an aerobic exercise study, nobody has yet proved with biological markers that exercise protects brain cells in living human patients.

Does harder exercise produce a different result?

Harder exercise can keep motor symptoms stable for longer, but no trial shows that it slows the disease itself. Head-to-head trials measure how a person moves during a clinic assessment, not whether dopamine-producing brain cells are surviving. Current research shows that high-intensity workouts stabilise short-term movement scores better than gentler workouts, but neither has been proved to alter the underlying biological course.

The strongest trial to date was designed to rule out dead ends, not to prove efficacy. In the SPARX Phase 2 randomised trial of 128 untreated participants, walking on a treadmill at 80% to 85% of maximum heart rate (43 participants) limited worsening on a clinician motor rating scale to 0.3 points over six months, compared to 3.2 points of worsening with usual care (40 participants). That high-intensity workout cleared the benchmark in a futility trial, a design built to decide whether a larger trial is worth running rather than to prove a treatment works. Moderate exercise at 60% to 65% of maximum heart rate (45 participants) failed that test. Smaller trials show less clear differences: a 2024 randomised trial of 29 participants by Kathia and colleagues compared ten weeks of high-intensity interval cycling to moderate cycling and found no statistically significant difference between the groups on motor scores (p = 0.51).

Whether higher intensity delays long-term progression remains an open question. A trial registration, recorded under identifier NCT04284436, describes an ongoing Phase 3 study tracking 370 unmedicated participants over 12 months. Until trials of that duration report their results, the evidence confirms only that vigorous workouts can steady day-to-day motor scores over several months, leaving true disease slowing unproven.

What happens when people in trials stop exercising?

Some physical gains linger for weeks after training stops, but others vanish almost immediately. A washout period is a planned pause where exercise stops so researchers can see whether improvements survive. In a 20-week double-blind randomised trial of 40 people, those who completed 10 weeks of sensory-focused exercise preserved their gains in walking speed, stride length, and overall disease severity after a 10-week break. Participants assigned to boxing lost their walking speed and stride gains as soon as training stopped, although both groups retained improvements in quality of life.

Certain motor skills can persist for up to two months, though the evidence remains limited to small cohorts. In a trial registered under NCT01636297, an 8-week high-intensity cycling programme produced improvements in reaction time and movement speed that lasted through 4 and 8 weeks of rest. A single-blind trial evaluating attentional focus during walking similarly found that motor scores on the Unified Parkinson’s Disease Rating Scale held at an 8-week follow-up. These lasting gains suggest exercise may alter underlying neurological processing rather than just masking symptoms, but very few trials track participants after the sweat dries.

Most exercise trials never include a washout period at all, leaving durability untested. Most studies measure participants only at their final exercise session, confirming that immediate physical gains occur without checking whether they endure. To detect true durability, a washout must run beyond 8 weeks, yet nobody has established how long a pause is needed to clear temporary symptomatic effects in people whose baseline progression differs. By relying on subjective clinical rating scales rather than continuous movement sensors, small trials risk missing the precise moment when benefits fade away.

Why is it so difficult for trials to prove disease slowing?

Standard clinical tests measure physical fitness rather than surviving brain cells. In a systematic review of 49 randomised controlled trials with 2,104 participants, exercise improved motor scores, but the certainty of this evidence was low. The gains in movement speed came from stronger muscles and better coordination, not proven changes in disease biology. Testing participants in the "off" state, after withholding Parkinson's medication overnight, removes short-term drug action but leaves muscle power and cardiovascular stamina intact.

Stopping exercise to test for lasting change causes participants to drop out. Researchers sometimes use a wash-out period, a trial phase where an intervention stops completely to see whether benefits persist after the treatment ends. In a systematic review evaluating trial designs for disease modification, 20 out of 44 Parkinson's trials used a wash-out phase. Halting an effective physical routine causes symptoms to deteriorate; in one evaluated trial, only 38 of 54 participants managed to complete the wash-out period.

Direct biological markers inside the brain are rarely tracked. The review of 49 trials found no evidence that exercise altered brain volume, inflammation, or markers of oxidative stress. Studies measuring dopamine receptor binding or brain-derived neurotrophic factor were too small and varied to pool into a reliable meta-analysis. Without clear biological indicators that change alongside physical gains, trials cannot separate a conditioned body from a protected brain. The trial registration NCT06941012 has not yet reported.

What this does not show

The evidence does not prove that exercise protects dopamine neurons or alters underlying brain pathology. In a systematic review of 49 trials involving 2,104 participants by Li and colleagues, exercise failed to reduce biological markers of inflammation or oxidative stress, and failed to halt brain tissue loss. Furthermore, the Phase 2 SPARX trial by Schenkman and colleagues was a futility trial, built to decide whether a larger trial was worth running rather than to prove that vigorous exercise works. That trial tested 128 participants over six months: high-intensity treadmill training produced an exploratory signal in 43 participants, but moderate-intensity exercise failed to beat usual care.

The definitive trial has not yet reported its findings. Confirmation rests with SPARX3, which is an ongoing clinical trial registration (NCT04284436) of 370 participants comparing treadmill intensities over 12 months. Everyday mobility claims also remain unproven across common routines. In a meta-analysis of 54 studies covering 2,828 participants, neither balance and gait training nor dance increased walking speed at any volume tested. No exercise type in that analysis achieved a minimal clinically important difference, which is the threshold where a change becomes noticeable and meaningful to a person in daily life.

Most exercise trials in Parkinson's disease leave their outcomes unpublished. A systematic audit by Silva and colleagues evaluated 236 registered clinical trials of exercise for Parkinson's disease and found that only 70 (29.7%) ever reported results, with only 61 (25.8%) appearing in peer-reviewed journals. The published literature is weakened by open-label designs that measure surrogate endpoints instead of actual disease progression. Adjacent therapies have faced formal retractions: a meta-analysis on integrated Chinese and Western medicine by Wang and colleagues was retracted in 2023, and an unblinded drug trial by Huang and colleagues was retracted in 2024.

This evidence outlines the boundary between temporary physical conditioning and unproven disease slowing, but it cannot determine which training routines match your specific mobility and health. Discuss these trial limits with your neurologist or physiotherapist to design an activity plan that targets your daily symptoms safely.

Sources

• Does Exercise Attenuate Disease Progression in People With Parkinson's Disease? A Systematic Review With Meta-Analyses [10.1177/15459683231172752] (tier 2, Synthesis) — 10.1177/15459683231172752
• A review of clinical trial designs used to detect a disease-modifying effect of drug therapy in Alzheimer's disease… [10.1186/s12883-016-0606-3] (tier 2, Synthesis) — 10.1186/s12883-016-0606-3
• Optimal dose and type of exercise improve walking velocity in adults with Parkinson's disease: a systematic review… [10.1038/s41598-025-85456-7] (tier 2, Synthesis) — 10.1038/s41598-025-85456-7
• Effect of power training on physical functional performance of patients with Parkinson's disease: A systematic… [10.1371/journal.pone.0314058] (tier 2, Synthesis) — 10.1371/journal.pone.0314058
• Reporting and methodological quality of clinical trials on exercise therapy for Parkinson's disease [10.1016/j.parkreldis.2019.11.011] (tier 2, Synthesis) — 10.1016/j.parkreldis.2019.11.011
• The role of virtual reality on outcomes in rehabilitation of Parkinson's disease: meta-analysis and systematic… [10.1007/s10072-019-04144-3] (tier 2, Synthesis) — 10.1007/s10072-019-04144-3
• The Effect of the Integrated Chinese and Western Medicine for the Treatment of Parkinson's Disease: A Meta-Analysis [10.1155/2022/4134931] (RETRACTED, tier 2, Synthesis) — 10.1155/2022/4134931
• Exercise-Induced Neuroplasticity in Parkinson's Disease: A Metasynthesis of the Literature [10.1155/2020/8961493] (tier 2, Synthesis) — 10.1155/2020/8961493
• 12-Week Aerobic Interval Training Boosts Neuroplasticity and Motor Function in Parkinson's Disease: Insights From… [10.1177/15459683251360729] (tier 3, Randomised trial) — 10.1177/15459683251360729
• Boxing vs Sensory Exercise for Parkinson's Disease: A Double-Blinded Randomized Controlled Trial [10.1177/15459683211023197] (tier 3, Randomised trial) — 10.1177/15459683211023197
• Effect of High-Intensity Treadmill Exercise on Motor Symptoms in Patients With De Novo Parkinson Disease: A Phase 2… [10.1001/jamaneurol.2017.3517] (tier 3, Randomised trial) — 10.1001/jamaneurol.2017.3517
• Exercise response in Parkinson's disease: insights from a cross-sectional comparison with sedentary controls and a… [10.1136/bmjopen-2017-017194] (tier 3, Randomised trial) — 10.1136/bmjopen-2017-017194
• Exercise increases caudate dopamine release and ventral striatal activation in Parkinson's disease [10.1002/mds.27865] (tier 3, Randomised trial) — 10.1002/mds.27865
• High-intensity interval versus moderate-intensity continuous cycling training in Parkinson's disease: a randomized trial [10.1152/japplphysiol.00219.2024] (tier 3, Randomised trial) — 10.1152/japplphysiol.00219.2024
• Can Dual Task Walking Improve in Parkinson's Disease After External Focus of Attention Exercise? A Single Blind… [10.1177/1545968317746782] (tier 3, Randomised trial) — 10.1177/1545968317746782
• Efficacy of pramipexole combined with levodopa for Parkinson's disease treatment and their effects on QOL and serum… [10.1177/0300060520922449] (RETRACTED, tier 3, Randomised trial) — 10.1177/0300060520922449
• Association of Physical Activity and APOE Genotype With Longitudinal Cognitive Change in Early Parkinson Disease [10.1212/wnl.0000000000011852] (tier 4, Other trial) — 10.1212/wnl.0000000000011852
• Evaluating dopamine transporter imaging as an enrichment biomarker in a phase 2 Parkinson's disease trial [10.1186/s12883-021-02470-8] (tier 4, Other trial) — 10.1186/s12883-021-02470-8
• Simplified quantification of [18F]FE-PE2I PET in Parkinson's disease: Discriminative power, test-retest reliability… [10.1177/0271678x20958755] (tier 4, Other trial) — 10.1177/0271678x20958755
• High intensity aerobic exercise improves information processing and motor performance in individuals with… [10.1007/s00221-020-06009-0] (tier 4, Other trial) — 10.1007/s00221-020-06009-0
• RETRACTED: A systematic genomic screen of 755 early-onset Parkinson’s patients from the 100,000 Genomes Project… [10.21203/rs.3.rs-9290362/v1] (tier 8, Preprint (not peer reviewed)) — 10.21203/rs.3.rs-9290362/v1
• Study in Parkinson Disease of Exercise [NCT04284436] (tier 9, Trial registration (no results yet)) — nct04284436
• Partnered Dance Aerobic Exercise as a Neuroprotective, Motor and Cognitive Intervention in Parkinson's Disease [NCT04122690] (tier 9, Trial registration (no results yet)) — nct04122690
• Moderate Versus High Volume Light-Moderate Intensity Exercise for People With Moderate Parkinson's Disease [NCT06088355] (tier 9, Trial registration (no results yet)) — nct06088355
• Validation of α-synuclein Modifications in Parkinson's dIsoRder Evolution [NCT06941012] (tier 9, Trial registration (no results yet)) — nct06941012