Standard subtraction with calculated coefficients increases EFC pair agreement in the tested MR-ART tables.

In 436 MRI scans from 148 participants in OpenNeuro ds004173 version 1.0.2, standard subtraction with coefficients estimated without expert grades retained an increase in entropy focus criterion (EFC) pair agreement when both test participants were excluded from coefficient fitting. Mean EFC agreement was 0.830346 with calculated coefficients and 0.717721 with raw values, a paired increase of 0.112625 (pointwise 95% participant-bootstrap interval 0.016559 to 0.201491). The comparison used grade-3 versus grade-2 scans within each of two instructed-motion conditions and gave the conditions equal weight. The comparison with full subtraction was 0.002055 (interval -0.011660 to 0.017838), which does not establish equivalence or a benefit over full subtraction. Calculated coefficients gave lower mean agreement than raw values for coefficient of joint variation and total signal-to-noise ratio. This same-dataset result is limited to one scanner, one protocol, imperfect single standard scans, and eight missing motion scans; it does not establish clinical utility, repeatability, causality, or performance at other sites.

magnetic resonance imaging · image quality · entropy focus criterion · standard subtraction · pair agreement · bootstrap · MR-ART