Automated TMS hotspot-hunting using a closed loop threshold-based algorithm

Abstract

BackgroundAlthough neuronavigation is increasingly used for optimizing coil positioning, the inter-session reliability of hotspot location remains unsatisfactory, probably due to the variability of motor evoked potentials (MEPs) and residual investigator bias. PurposeTo increase the reliability and accuracy of hotspot location we introduce a novel automated hotspot-hunting procedure (AHH). MethodsAHH is based on resting motor thresholds (RMTs) instead of MEP amplitudes. By combining robotic coil positioning with a closed loop target search algorithm AHH runs independently from the investigator. AHH first identifies all targets with an RMT below a defined intensity of stimulator output (MEP-positive) and then locates the motor hotspot of a target muscle by measuring RMTs at all identified MEP-positive targets. Results were compared to robotic MEP amplitude TMS mapping (MAM) using a 7×7 predefined target grid and suprathreshold intensities and manual hotspot search (MHS). Sequence of stimulation was randomized from pulse to pulse in AHH and MAM. Each procedure was tested in 8 subjects. ResultsInter-session CoG shift was significantly reduced with AHH (1.4mm (SEM: 0.4)) as compared to MAM (7.0mm (SEM: 1.8)) (p=0.018) and MHS (9.6mm (SEM: 2.2)) (p=0.007). No statistical difference was observed between MAM and MHS. RMTs were reliable between sessions. ConclusionOur method represents the first fully automated, i.e. investigator-independent, TMS hotspot-hunting procedure. Measuring RMTs instead of MEP amplitudes leads to significantly increased accuracy and reliability of CoG locations. Moreover, by assessing thresholds AHH is the first procedure to fulfill the original hotspot definition.