Cycling After Balance Practice Boosts Motor Learning in Parkinson's Patients

Motor learning is fundamental to rehabilitation in Parkinson's disease (PD), yet both the encoding and consolidation of motor memories are impaired in this population. Cortico-striatal circuit dysfunction reduces synaptic plasticity—including long-term potentiation in the primary motor cortex—making it harder for people with PD (pwPD) to acquire and retain new motor skills. Pharmacological approaches have had limited success, and the dopamine overdose hypothesis even suggests medications may worsen procedural learning in some contexts. Non-pharmacological strategies are therefore urgently needed.

This pilot randomized controlled trial investigated whether combining motor skill practice with cardiovascular exercise (CVE) over multiple sessions could enhance motor learning in pwPD. Twenty-four participants with mild-to-moderate PD (Hoehn & Yahr 2.1–2.3) were randomized to a CVE group (EXE, n=12) or a seated rest control group (REST, n=12). Both groups practiced a novel dynamic balance task on a balance board across six weekly sessions. Immediately after each practice block, the EXE group performed 20 minutes of moderate-intensity stationary cycling (targeting ~60–70% of peak heart rate), while the REST group sat quietly. A cardiopulmonary exercise test established baseline VO2peak, and motor performance was quantified as time-in-balance (TIB) across practice trials.

Using a linear mixed model accounting for individual variability and disease severity (MDS-UPDRS III), the researchers found that both groups improved significantly across sessions. However, the EXE group demonstrated significantly greater performance gains in sessions 4 and 5 (p<0.05), translating to 48.9% total improvement versus 28.9% in REST. These findings were robust to sensitivity analyses excluding an outlier participant and to models incorporating age as a covariate. Contrary to the secondary hypothesis, between-session offline consolidation did not differ significantly between groups; instead, a non-significant trend toward greater within-session online learning drove the group difference. Exploratory analyses revealed a positive association between baseline VO2peak and performance gains, suggesting that cardiorespiratory fitness moderates the learning benefit of CVE.

The mechanistic interpretation centers on CVE-induced upregulation of catecholamines (including dopamine) and neurotrophic factors such as BDNF, along with enhanced motor cortex excitability—all of which may lower the threshold for synaptic plasticity during skill encoding. The delayed emergence of group differences (sessions 4–5 rather than early sessions) is consistent with a cumulative neurobiological priming effect, where repeated CVE gradually creates a more favorable neuroplastic environment. The finding that online rather than offline learning was enhanced is noteworthy and may reflect compensatory engagement of cortical circuits when striatal pathways are compromised.

Several limitations temper these conclusions. The sample was small (n=24), limiting statistical power and generalizability. All participants were relatively high-functioning (mild-to-moderate PD, mean MoCA ~26). The study did not include neuroimaging or biomarker measures to confirm the proposed neurobiological mechanisms. Blinding of participants was not possible given the nature of the intervention, introducing potential expectancy effects. Future trials with larger samples, active control conditions, biomarker outcomes, and longer follow-up periods are needed to confirm and extend these findings.