EEG–MRI Co-registration and Sensor Labeling Using a 3D Laser Scanner

Laurent Koessler (1, 2) , Thierry Cecchin (2) , Olivier Caspary (2) , Adnane Benhadid (3) , Hervé Vespignani (1, 2) , Louis Maillard (1, 2)

1 Service de neurologie [CHRU Nancy]
2 CRAN - Centre de Recherche en Automatique de Nancy
3 IADI - Imagerie Adaptative Diagnostique et Interventionnelle

This paper deals with the co-registration of an MRI scan with EEG sensors. We set out to evaluate the effectiveness of a 3D handheld laser scanner, a device that is not widely used for co-registration, applying a semi-automatic procedure that also labels EEG sensors. The scanner acquired the sensors' positions and the face shape, and the scalp mesh was obtained from the MRI scan. A pre-alignment step, using the position of three fiducial landmarks, provided an initial value for co-registration, and the sensors were automatically labeled. Co-registration was then performed using an iterative closest point algorithm applied to the face shape. The procedure was conducted on five subjects with two scans of EEG sensors and one MRI scan each. The mean time for the digitization of the 64 sensors and three landmarks was 53 s. The average scanning time for the face shape was 2 min 6 s for an average number of 5,263 points. The mean residual error of the sensors co-registration was 2.11 mm. These results suggest that the laser scanner associated with an efficient co-registration and sensor labeling algorithm is sufficiently accurate, fast and user-friendly for longitudinal and retrospective brain sources imaging studies.

Mots clés: Multimodal co-registration; Laser scanner; Iterative closest point algorithm; Surface fitting; Brain sources imaging

Domaines: Ingénierie biomédicale

Laurent Koessler, Thierry Cecchin, Olivier Caspary, Adnane Benhadid, Hervé Vespignani, et al.. EEG–MRI Co-registration and Sensor Labeling Using a 3D Laser Scanner. Annals of Biomedical Engineering, 2011, 39 (3), pp.983-995. ⟨10.1007/s10439-010-0230-0⟩. ⟨hal-00545773⟩