In this study we address the problem of the repeatability of autonomic responses in the experimental setting. In healthy volunteers, we measured the heart rate (HR) response to pain anticipation and to pain elicited with galvanic stimulation. After evaluation of pain threshold (T), all subjects underwent the same experimental protocol, whereby a painful stimulus at 1.5T was delivered on the forehead following a warning, while the ECG was continuously recorded. The procedure was repeated three times across a three-week period. The parameters recorded included pain threshold, pain rating, HR response to pain anticipation and HR response to pain. We found a high correlation among the three sessions for all parameters, indicating that, as occurs for pain threshold and pain rating, individual differences in autonomic responses can be reliably reproduced as well, even though significant habituation develops.