Longitudinal changes in motor-evoked potential variability in Parkinson’s disease

Abstract

Introduction: Motor-evoked potential (MEP) amplitude is intrinsically variable across repeated transcranial magnetic stimulation (TMS) trials. This variability is often treated as measurement noise, but it may also reflect fluctuations in corticospinal excitability, synaptic responsiveness, and motor system stability. In Parkinson’s disease (PD), where cortical excitability and inhibitory mechanisms evolve with disease progression, changes in trial-to-trial MEP variability could provide information not captured by mean amplitude alone. In a previous study that followed a cohort of patients with PD over 4.5 years, clinical and neurophysiological changes were identified, but response variability was not specifically examined. Therefore, this secondary analysis investigated whether intraindividual MEP variability changed over time and whether these changes were associated with motor deterioration.

Objective: To determine whether intraindividual MEP amplitude variability changes longitudinally in Parkinson’s disease and whether these changes are associated with motor deterioration, particularly across the pre- and post-pandemic assessment periods.

Methodology: This secondary analysis included 22 patients with Parkinson’s disease followed across four assessment periods: T1 (September 2018–June 2019), T2 (June 2019–March 2020), T4 (March 2021–March 2022), and T5 (March 2022–March 2023). MEPs were recorded bilaterally from the first dorsal interosseous muscle using single-pulse TMS. For each hemisphere and period, intraindividual MEP variability was calculated from 10 responses elicited at 130% of the resting motor threshold using the coefficient of variation. Longitudinal changes were assessed with linear mixed-effects models including period, hemisphere, and their interaction as fixed effects and participant as a random intercept. Associations with MDS-UPDRS Part III scores were examined using mixed models and Spearman correlations of changes between T2 and T4. Sensitivity analyses included robust variability estimates and leave-one-out testing.

Results: The linear mixed-effects model showed no significant effect of assessment period, (F_3,116.26 = 1.57, p = 0.199), hemisphere (more vs. less affected, F_1,109.01 = 0.006, p = 0.937), or period-by-hemisphere interaction, (F_3,108.93 = 0.14, p = 0.935). Although the MEP coefficient of variation increased by approximately 16% from T2 to T4 in both hemispheres, this change was not statistically significant. Concurrent MEP variability was also not associated with MDS-UPDRS Part III scores. However, from T2 to T4, the increase in variability in the less affected hemisphere was strongly correlated with motor deterioration (ρ = 0.715, p<0.001, n = 18); this association remained stable in the leave-one-out analysis (ρ = 0.662–0.775).

Conclusions: MEP amplitude variability did not show significant overall longitudinal changes or consistent hemispheric differences. However, the increase in variability in the less affected hemisphere from T2 to T4 was strongly associated with motor worsening, suggesting that changes in trial-to-trial corticospinal variability may capture clinically relevant progression not reflected by mean MEP amplitude alone. Given the small sample and exploratory design, these findings require confirmation in larger prospective cohorts using more MEP trials, standardized medication states, and direct statistical comparison of the associations observed in the more and less affected hemispheres.