Clinical NeuroscienceFeasibility of automated analysis and inter-examiner variability of cortical silent period induced by transcranial magnetic stimulation

- Transcranial magnetic stimulation induced silent period (cSP) is commonly analysed manually.
- A new automated routine for cSP analysis was compared with manually analysed cSPs.
- Data from a large-population of healthy subjects were reanalysed by two of the authors.
- Inter-examiner effect in cSP analysis was found quite large.
- Automated routine was found valid and at least as reliable as the manual analysis.

Cortical silent period (cSP) is a short interruption in electromyography (EMG) during active muscle contraction induced with transcranial magnetic stimulation (TMS). The cSP is a measure of cortical inhibition and is believed to represent inhibitory interneuron effects on excited motor cortical areas. Several pathological conditions and pharmacological manipulations induce changes to cSP duration indicating alterations in intracortical inhibition. At present, it is common to manually analyse the cSP duration from measured EMG. However, to avoid inter-examiner effects on cSP interpretation and detection, as well as to allow for quick measurement online, automated routine would be preferable. In this study, we evaluate the feasibility of a straight-forward cSP detection routine based on analysing the rectified first derivative of the EMG signal following TMS. Previously measured cSPs of 54 healthy subjects were reanalysed manually by two of the authors and using the automated routine. Furthermore, we recruited one subject for whom the cSPs were induced with several stimulation intensities, and those cSPs were analysed manually by two of the authors as well as using the automated routine. We found that cSPs were detected correctly by the automated cSP detection routine, and agreement with manually analysed subject-specific mean cSPs was excellent (ICC = 0.992, p < 0.001). The inter-examiner variability was similar to the variability between manual and automated analysis. Hence, we believe the introduced cSP detection routine would be feasible for online cSP detection, in such a way that is presently used to detect the motor evoked potentials.