Integrity Report: Electromagnetic design and analysis of a bell-shaped superconducting magnet for head fMRI using a novel hybrid optimization method

Academic fraud report · Geng Detector

Summary

This report evaluates the paper by Mei et al. (Supercond. Sci. Technol. 38 (2025) 025009, DOI: 10.1088/1361-6668/ada114) on a bell-shaped superconducting magnet designed for head fMRI using a hybrid optimization method. Overall verdict: questionable but likely honest. One inconsistency was confirmed: Section 4.1 states that Coils 3, 6, and 9 are shield coils carrying reverse current, while Figure 7's caption lists Coils 3, 7, and 9. The actual currents in Table 3 (Coil 3 = -480.29 A, Coil 6 = +480.29 A, Coil 7 = -480.29 A, Coil 9 = -480.29 A) match the Figure 7 caption, indicating that the text misidentified Coil 6 as Coil 7 rather than the reverse. Numerical data, however, are highly self-consistent: 0.5 × 12.57 H × (480.293 A)² = 1,449,806 J ≈ 1.45 MJ, matching the stated stored energy exactly. Floating-point endings, experiment counts, and citation timelines show no anomalies. Pixel-level image analysis was not possible due to lack of high-resolution files. Limitations: image forensics and raw data verification remain pending.

Verdict

🟡 Questionable (likely honest error). One confirmed text-vs-figure contradiction regarding the labeling of shield coils. Numerical data pass internal consistency checks, suggesting the issue is a typographical/copy-editing error rather than fabrication.

Key findings

Text–figure contradiction on shield coils: Section 4.1 names Coils 3, 6, and 9 as reverse-current shield coils; Figure 7 caption names Coils 3, 7, and 9.
Numerical data internally consistent: Stored energy calculation (0.5 × 12.57 H × (480.293 A)² = 1,449,806 J) matches the stated 1.45 MJ exactly.
Table 3 currents favor Figure 7 caption: Coil 3 = -480.29 A, Coil 6 = +480.29 A (positive), Coil 7 = -480.29 A (negative), Coil 9 = -480.29 A; Coils 3, 7, 9 carry the reverse current.
No statistical red flags: Experimental counts in Table 5 (25 trials; successes 23, 17, 11, 8, 5, 2) are arithmetically plausible and trend monotonically with homogeneity tightening.
Timeline and citations clean: Submitted July 2024, accepted December 2024, published January 2025; latest cited works are from 2024 or earlier.
Image forensics not performed: No high-resolution figures were available for splicing or noise-pattern analysis.

Evidence highlights

Direct quote (Section 4.1): "Coils 3, 6, and 9 are shield coils carrying the reverse current."
Direct quote (Figure 7 caption): "Coils 3, 7, and 9 are shield coils carrying the reverse current..."
Table 3 operating currents (Iop): Coil 3 = -480.29 A; Coil 6 = +480.29 A; Coil 7 = -480.29 A; Coil 9 = -480.29 A.
Energy check: E = ½ × 12.57 × (480.293)² = 1,449,806 J ≈ 1.45 MJ (matches paper).
Timeline: Submitted 2024-07; accepted 2024-12; published 2025-01 (Supercond. Sci. Technol. 38 (2025) 025009).
DOI: 10.1088/1361-6668/ada114

Notes

The Coil 6 vs Coil 7 mismatch is best explained by an incomplete global update of the manuscript after the coil numbering scheme was adjusted; the figure caption and underlying data table were updated but the running text was not.
The extraordinary numerical agreement across inductance, current, and stored energy makes ad hoc fabrication of the dataset implausible.
Limitations: original high-resolution figures, FEM raw output files, and experimental raw logs were not available for independent verification. Image-level duplication, stress-field cloud splicing, and circuit-simulation figure integrity therefore remain unverified.
Recommended action: a courteous post (e.g., on PubPeer) flagging the Coil 6/7 typo in Section 4.1. No journal or institutional misconduct referral is warranted on current evidence.