BACKGROUND: Electroencephalography (EEG) can capture the cortical response evoked by transcranial magnetic stimulation (TMS). The TMS pulse provokes a large artefact, which obscures the cortical response in the first milliseconds after TMS. Removing this artefact remains a challenge.

METHODS: We delivered monophasic and biphasic TMS to a melon as head phantom and to four healthy participants and recorded the pulse artefact at 5 kHz with a TMS-compatible EEG system. Pulse delivery was either synchronized or non-synchronized to the clock of the EEG recording system. The effects of synchronization were tested at 10 and 20 kHz using the head phantom. We also tested the effect of a soft sheet placed between the stimulation coil and recording electrodes in both human and melon.

RESULTS & CONCLUSION: Synchronizing TMS and data acquisition markedly reduced trial-to-trial variability of the pulse artefact in recordings from the phantom or from the scalp. Reduced trial-to-trial variability was also observed at high sampling frequencies. The use of a soft sheet reduced the variability in recordings on the head phantom, but not in human participants. Effective reduction of the trial-to-trial variability renders it possible to create an artefact template for off-line filtering. Template-based subtraction of the artefact from the EEG signals is a prerequisite to effectively recover the immediate physiological response in the stimulated cortex and inter-connected areas.