The correct definition (taken from http://www.medilexicon.com) of the pre-auriciular point is “a point of the posterior root of the zygomatic arch lying immediately in front of the upper end of the tragus”. The zygomatic arch or cheek bone is the skull bone in front of the ear as depicted in this figure
and the tragus is a small backward-pointing eminence situated in front of the ear canal.
The approximate position of the pre-auriciular point is indicated in the figure above by the green point. It can be palpated (i.e. felt with the finger tip) the best if the subject moves his jaw by opening and closing his or her mouth.
The problem with using the LPA and RPA according to this definition is that it can be really difficult to localize them precisely on the anatomical MRI. Mislocalization of these points can lead to severe misregistration between the MRI and MEG/EEG, and consequently affect the accuracy of source localization.
Because of the difficulty of localizing the LPA and RPA points in both EEG/MEG and in the anatomical MRI, various labs have adopted slightly different conventions for the fiducial points for the ears. Some labs use the junction between the tragus and the helix, marked with the red dot in the figure of the ear above, which can be located more precisely both anatomically and on the MRI slices. Other possibilities that are less frequently uses are the tip of the tragus, the ear lobe or the mastoid.
At the Donders Centre in Nijmegen we use ear molds with a hole in them (see below) to attach the markers: In the MEG scanner we insert a small tube into the hole (the tube is also used for auditory stimulation) and attach the MEG localizer coil to the tube. In the MRI scanner we use the same ear molds, but rather insert a custom-made marker with a small drop of vitamine E into the hole. The position thereby obtained with the MEG localizer coils is as precisely as possible reproduced in the MRI, given the movement that is allowed by the ear molds. We have various sizes of ear molds, both at the MEG and MRI scanner, and subjects should use the same size in both scanners.
When using different fiducial locations in the MEG and the MRI, the difference in the coordinates (which can be 5-20 mm) has to be taken into account in the coregistration procedure, otherwise the source localization would suffer from the systematic coregistration error. Luckily, most research labs acquire the EEG/MEG and the MRI using the same fiducial locations.
The consequence of different fiducial locations in different labs is that the terms “LPA” and “RPA”, although used in software such as FieldTrip and other EEG/MEG tools, do not always refer to the same anatomical landmarks. In your analysis you have to take care that the positions consistently refer to the same landmarks, whether they are in front of the ear, on the tragus or in line with the ear canal. So whenever the software uses LPA and RPA, you have to be aware of your lab convention.
Note that for the nasion, where at the Donders Centre we also place one of the MEG localizer coils, we do not use a MRI marker. The nasion is easy to identify in the anatomical MRI images.