Cortical evoked potentials (CEP) have been recorded in response to both electrical stimulation (ES) and mechanical stimulation (MS) of the oesophagus. While the optimal parameters for recording reproducible oesophageal CEP to ES have recently been established, they have not yet been determined for MS, and reported CEP to MS show considerable variability. This study aimed to identify the optimal parameters required to record reproducible MS induced CEP. CEP were recorded from the vertex (Cz) in six subjects (one female; age range 23-47 years). MS was performed 5 cm above the lower oesophageal sphincter by rapidly inflating a 2-cm long silicone balloon at a frequency of 0.2 Hz. The rise time to maximum inflation was 165 ms. In order to determine the minimum number of stimuli required to produce optimal signal-to-noise quality, we acquired data in runs of 25, 50, 100 and 300 stimuli and to determine the stimulation intensity that produced the shortest latency and the largest amplitude CEP, we averaged four runs of 50 stimuli at five different intensities ranging from sensory threshold to pain. CEP reproducibility was then assessed in three subjects on three separate occasions using parameters determined from these measurements. We found that optimal signal-to-noise quality was achieved by averaging four runs of 50 stimuli; that P1 latency was shortest and P1-N1 amplitude largest at intensities of 75% and pain threshold and that highly reproducible CEP were obtained in all individuals. We conclude that it is possible to obtain highly reproducible oesophageal CEP to MS which can now be compared to those obtained by ES in order to identify which is most suitable for clinical studies.