Reproducibility of Single-Pulse, Paired-Pulse, and Intermittent Theta-Burst TMS Measures in Healthy Aging, Type-2 Diabetes, and Alzheimer’s Disease

Background: Transcranial magnetic stimulation (TMS) can be used to assess neurophysiology and the mechanisms of cortical brain plasticity in humans in vivo. As the use of these measures in specific populations (e.g., Alzheimer's disease; AD) increases, it is critical to understand their reproducibility (i.e., test-retest reliability) in the populations of interest. Objective: Reproducibility of TMS measures was evaluated in older adults, including healthy, AD, and Type-2 diabetes mellitus (T2DM) groups. Methods: Participants received two identical neurophysiological assessments within a year including motor thresholds, baseline motor evoked potentials (MEPs), short- and long-interval intracortical inhibition (SICI, LICI) and intracortical facilitation (ICF), and MEP changes following intermittent theta-burst stimulation (iTBS). Cronbach's α coefficients were calculated to assess reproducibility. Multiple linear regression analyses were used to investigate factors related to intraindividual variability. Results: Reproducibility was highest for motor thresholds, followed by baseline MEPs, SICI and LICI, and was lowest for ICF and iTBS aftereffects. The AD group tended to show higher reproducibility than T2DM or controls. Intraindividual variability of baseline MEPs was related to age and variability of RMT, while the intraindividual variability in post-iTBS measures was related to baseline MEP variability, intervisit duration, and Brain-derived neurotrophic factor (BDNF) polymorphism. Conclusion: Increased reproducibility in AD may reflect pathophysiological declines in the efficacy of neuroplastic mechanisms. Reproducibility of iTBS aftereffects can be improved by keeping baseline MEPs consistent, controlling for BDNF genotype, and waiting at least a week between visits. Significance: These findings provide the first direct assessment of reproducibility of TMS measures in older clinical populations. Reproducibility coefficients may be used to adjust effect- and sample size calculations for future studies.