Estimation of the accuracy of a surface matching technique for registration of EEG and MRI data

Abstract

Objectives: We developed a method to register EEG and MRI data used for the source reconstruction of electric brain activity. Methods: The method is based on matching of the head surfaces as obtained by 3D scanning after the EEG recording, and by segmentation of MRI data. The registration accuracy was estimated by calculating the residual error of the surface matching and its intra-individual and inter-individual variability. In addition, the test-retest reliability concerning the transformation of electrode positions was studied, to estimate how inaccuracies resulting from the 3D scanning of the head surface translate into registration uncertainty. Results: For 61 measurements, performed on 20 subjects, the average root mean square of the Euclidean distances between the 3D-scanned and the MRI-derived head surfaces amounted to 3.4 mm. An inter-individual standard deviation of 0.24 mm, and an intra-individual standard deviation of 0.003–0.31 mm proved a high inter- and intra-subject stability of the surface matching technique. The variation of transformation results when studying the test-retest reliability amounted to 1.6 mm on average. The maximum error of transformation was smaller than the diameter of the electrodes. Conclusions: The findings suggest that the surface matching technique is a precise method for determination of the transformation of electrode positions and MRI data into a single co-ordinate system and can successfully be used in a routine laboratory setting.