Real‐time motion detection of functional MRI data

The objective of this work was to implement a motion-detection algorithm on a commercial real-time functional magnetic resonance imaging (fMRI) processing package for neurosurgical planning applications. A real-time motion detection module was implemented on a commercial real-time processing package. Simulated functional data sets with introduced translational, in-plane rotational, and through-plane rotational motion were created. The coefficient of variation (COV) of the center of intensity was used as a motion quantification metric. Coefficients of variation were calculated before and after image registration to determine the effectiveness of the motion correction; the limits of correctability were also determined. The motion-detection module allowed for real-time quantification of the motion in an fMRI experiment. Along with knowledge of the limits of correctability, this enables determination of whether an experiment needs to be reacquired while the patient is in the scanner. This study establishes the feasibility of using real-time motion detection for presurgical planning fMRI and establishes the limits of correctable motion.