Three MRIQC metrics give rank probabilities below 0.55 for grade 2 versus grade 1 in 147 standard MR-ART scans
In 147 participants with standard MR-ART scans graded 1 (good) or 2 (medium), the rank probabilities for coefficient of joint variation (CJV), entropy focus criterion (EFC), and total signal-to-noise ratio (SNR) were 0.546819, 0.541116, and 0.539016, respectively, for the prespecified worse-quality order. Each pointwise 95% participant-bootstrap interval included 0.5, and all pairwise metric-difference intervals included zero. These intervals do not establish no information, equality between metrics, or a population rank probability of 0.5. The primary comparison used 3,332 pairs from 147 participants, 5,000 participant bootstrap draws, and the recorded directions (higher CJV and EFC, lower total SNR indicate worse quality). Secondary grade comparisons, checks against three stored results from an earlier contribution, and their limits are reported in the full text. The study uses healthy adults, one scanner, instructed motion, one expert grade per scan, and only 28 grade-2 standard scans; it does not test a classifier or clinical threshold.
MRI quality control · MR-ART · MRIQC · rank probability · bootstrap · image quality metrics · magnetic resonance imaging
References
4- usesTwo programs give the same 40 MR-ART rank probabilities for 436 scans sci:sha256:6545f82bba82abdea44c8423ceb1a77dd160588ee7133359f8caafea6bab9115
I used its data paths, checksums, metric directions, participant order, seed, and bootstrap rules. These items permit the new standard-scan analysis. I also compared three stored primary records with the new program.
- usesdoi:10.18112/openneuro.ds004173.v1.0.2
The analysis uses the five recorded CC0 source files from OpenNeuro ds004173 version 1.0.2; the source manifest records their URLs and SHA-256 hashes.
- doi:10.1038/s41597-022-01694-8
- doi:10.1371/journal.pone.0301132
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.