Quantifying MR head motion in the Rhineland Study - A robust method for population cohorts

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Population Health

Background:

• Head motion during Magnetic Resonance (MR) acquisition degrades image quality and can bias neuromorphometric analysis.
• Quantifying head motion has neuroscientific and clinical applications, including controlling for motion in statistical analyses and as a variable in neurological studies.
• The accuracy of markerless optical head tracking for motion quantification is underexplored, with no prior quantitative analysis in large, general population cohorts.

Purpose of the Study:

• To develop and validate a robust registration method for markerless optical head tracking using depth camera data.
• To assess the method's accuracy in estimating small head movements in compliant participants.
• To establish an analysis pipeline for computing head motion scores for downstream analyses and apply it to a large population cohort.

Main Methods:

• A novel registration method was developed to align depth camera data for sensitive head motion estimation.
• The method was validated against functional MRI (fMRI) motion traces, independently acquired breathing signals, and image-based quality metrics from structural T1-weighted MRI.
• An analysis pipeline was created to compute motion scores, which was applied to data from the Rhineland Study, a large population cohort.

Main Results:

• The developed method outperformed the vendor-supplied algorithm in all three validation experiments.
• Head motion significantly increased over the duration of the scan session in the Rhineland Study cohort.
• Age and body mass index (BMI) were replicated as motion correlates, with weak but significant interactions observed with within-session motion increase, age, BMI, and sex.

Conclusions:

• Markerless optical head tracking provides a sensitive and accurate method for quantifying head motion during MR acquisition.
• The developed pipeline enables the integration of motion data into large-scale population studies, improving the reliability of neuroimaging analyses.
• Camera-based motion scores show high correlation with fMRI motion estimates, suggesting their utility as a surrogate measure for controlling motion in statistical analyses.