Optimization of the Navigated TMS Mapping Algorithm for Accurate Estimation of Cortical Muscle Representation Characteristics

• Published in: Brain sciences Vol. 9; no. 4; p. 88
• Main Authors: Sinitsyn, Dmitry O, Chernyavskiy, Andrey Yu, Poydasheva, Alexandra G, Bakulin, Ilya S, Suponeva, Natalia A, Piradov, Michael A
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 19.04.2019; MDPI
• Subjects: Accuracy; Algorithms; bootstrapping; Brain research; cortical muscle representation; Data analysis; Experiments; Magnetic fields; Magnetic resonance imaging; Mapping; Methods; Morphology; Motor evoked potentials; Muscles; navigated transcranial magnetic stimulation; Sampling; Software; TMS motor mapping; Transcranial magnetic stimulation; variability; Finland; TMS motor mapping; bootstrapping; accuracy; cortical muscle representation; navigated transcranial magnetic stimulation; variability
• ISSN: 2076-3425; 2076-3425
• DOI: 10.3390/brainsci9040088

Navigated transcranial magnetic stimulation (nTMS) mapping of cortical muscle representations allows noninvasive assessment of the state of a healthy or diseased motor system, and monitoring changes over time. These applications are hampered by the heterogeneity of existing mapping algorithms and the lack of detailed information about their accuracy. We aimed to find an optimal motor evoked potential (MEP) sampling scheme in the grid-based mapping algorithm in terms of the accuracy of muscle representation parameters. The abductor pollicis brevis (APB) muscles of eight healthy subjects were mapped three times on consecutive days using a seven-by-seven grid with ten stimuli per cell. The effect of the MEP variability on the parameter accuracy was assessed using bootstrapping. The accuracy of representation parameters increased with the number of stimuli without saturation up to at least ten stimuli per cell. The detailed sampling showed that the between-session representation area changes in the absence of interventions were significantly larger than the within-session fluctuations and thus could not be explained solely by the trial-to-trial variability of MEPs. The results demonstrate that the number of stimuli has no universally optimal value and must be chosen by balancing the accuracy requirements with the mapping time constraints in a given problem.