Methodology on Co-Registration of MRI and Optoelectronic Motion Capture Marker Sets: In-Vivo Wrist Case Study

Abstract

S kin-surface mounted markers provide incomplete spatial information of the underlying-bone. A new methodology is developed combining optoelectronic motion capture MOCAP and imaging modalities to co-register the positions of underlying-bone and external markers. Skin surface-mounted markers, utilized in MR imaging, were coated with reflective material to collect spatial data in passive infra-red optoelectronic MOCAP system. Two-link jig mechanisms were designed to mount-on marker sets; these were rotated in increments through 180° of angular rotation at pre-determined angles. The rotations were recorded within the MOCAP system and 3T MRI scanner under a 3D STIR short tau-inversion recovery sequence. A 3D in-silico model was built for the co-registration of marker centroids' on a 1 to 1 scale. Differences were calculated from the co-registered data obtained from these two systems using the same set of markers. Root mean square error RMSE and angular rotation was less than 1.5 mm in translation and 1° respectively in-vitro. Concordance Correlation Coefficients CCC was calculated as 0.9788 to 1 . Mean-Difference plots showed good agreement. Next, adduction/abduction movements of the natural wrist joint were investigated in six healthy subjects. MOCAP data was collected for three sets of motions, and MRI scans were repeated twice to derive within-subject repeatability data. Within-subject, the maximum RMSE for wrist angular rotations was 1.28° and 1.30° respectively in vivo. Pearson correlation coefficient was calculated for adduction and abduction as 0.70 and 0.71 respectively. Paired Student-t test identified systematic differences. The used methodology established the way to analyze the relationship between the bone and external markers

Keywords

• joint angle
• optoelectronic motion capture systems
• cross sectional imaging modalities
• image registration
• wrist kinematics.