The recording of somatosensory evoked potentials (SEPs) is a non-invasive routine clinical testing procedure in neurology. For trigeminal nerve stimulation, however, SEPs have not received a widespread clinical attention. A variety of protocols and procedures have been used to record trigeminal SEPs (TSEPs). Differences encountered include the stimulation mode, site and frequency, the recording electrode position and data acquisition parameters. This has resulted in a diversity of recorded TSEP signals, making comparisons almost impossible. The general picture shows a number of short latency waves (within 3 ms) of peripheral origin, followed by at least two longer latency waves (12-15 ms and 19-22 ms). Furthermore, potential waves with a very long latency (> 100 ms) follow when the response is produced by painful stimulation. The origin of the long and very long latency waves is still a matter of debate. In order to allow reliable data interpretation and comparisons between the outcome of different studies, a standardized protocol should be applied for TSEP recordings. By providing an overview, this paper aims to mark a step forward in the harmonization of TSEP protocols with respect to the neural processes of interest. Further studies should also encounter the potential application of other neuroimaging techniques, such as functional magnetic resonance imaging or positron emission tomography, preferably in combination with TSEP recordings.