Geng Integrity Report: Research Progress on Exercise Preventing and Treating Depression via Irisin Regulation

Academic fraud report · Geng Detector

Summary

Verdict: CLEAR — no integrity issues detected. This paper, a review article published as an Epub ahead of print in Chinese General Practice (DOI: 10.12114/j.issn.1007-9572.2025.0454), was assessed across all six standard forensic dimensions. As a narrative review rather than an original experimental study, several fraud-detection dimensions (Western blot splicing, fabricated raw data) are not directly applicable, but full-text and figure scanning still yielded no anomalies. The single Figure 1 is a vector-style mechanistic schematic with no image-reuse, rotation, or PS manipulation痕迹. Table 1 is a qualitative summary of prior literature without falsifiable raw statistics. The publication timeline (received 2026-01-09, revised 2026-03-27, online 2026-06-03) is internally consistent, and the funding project (Hubei Natural Science Foundation Joint Fund, No. 2025AFD575) is temporally plausible. References include up-to-date 2025 sources (e.g., Han CY 2025 in Ann Med; Zhang H 2025 in J Adv Res). Citation-content spot-checks (e.g., refs [29] Korkmaz A 2019, [43] Haghighi AH 2022) matched the original literature. Confidence in the clearance is high, though limited by the inherent scope of review-format papers, which present fewer raw-data checkpoints for forensic analysis.

Verdict

CLEAR — No signs of academic misconduct detected. The article is a review article, which limits the applicability of several experimental-data fraud checks; nevertheless, all six forensic dimensions produced no anomalous findings.

Key findings

Figure integrity (Dims. 1 & 3): The paper contains a single schematic figure (Figure 1: potential mechanisms of exercise-irisin-depression regulation). It is a vector-style mechanistic diagram with no Western blot bands, gels, or microscopy images; no image reuse, rotation, splicing, or PS痕迹 found.
Statistical / data anomalies (Dims. 2 & 4): Table 1 is a qualitative/semi-quantitative summary of prior literature; no Mean±SD, ANOVA, or exact p-values are reported, so no p-hacking, last-digit, or SD consistency tests could be performed. Narrative phrasing ("significantly increased", "no significant change") is appropriate for a review.
Timeline consistency (Dim. 5): Received 2026-01-09, revised 2026-03-27, online first 2026-06-03 — consistent with a standard domestic core-journal workflow ending one day before the report baseline date.
Funding plausibility (Dim. 5): Hubei Natural Science Foundation Joint Fund Project No. 2025AFD575 (approval year 2025) is temporally consistent with a 2026 submission.
Reference recency (Dim. 5): Cites literature as recent as 2025 (e.g., [4] Han CY et al., *Ann Med* 2025; [81] Zhang H et al., *J Adv Res* 2025), consistent with high-quality review practice.
Citation accuracy (Dim. 6): Spot-checks confirmed fidelity: ref [29] Korkmaz A et al. (2019) — "12-week resistance training did not significantly increase plasma irisin, contrasting with aerobic exercise" — matches original findings; ref [43] Haghighi AH et al. (2022) — HIIT-induced 22.22% irisin increase — corresponds to Table 1 entries. No mismatched attributions or fabricated citations detected.

Evidence highlights

DOI: 10.12114/j.issn.1007-9572.2025.0454
Single figure (Figure 1): mechanistic schematic only — no image-manipulation triggers.
Table 1: literature-summary table — no raw statistics to falsify.
Timeline anchors: received 2026-01-09 → revised 2026-03-27 → online 2026-06-03.
Funding project: 2025AFD575 (2025 approval).
Latest references: 2025 publications in *Ann Med* and *J Adv Res*.
Citation spot-check: refs [29] and [43] content consistent with source literature.

Notes

The clearance is partially a function of the article's review format: review articles intrinsically lack raw experimental data, which limits the surface area available for forensic image and statistics checks. This is not a weakness of the review, but a structural caveat on the assessment.
No contact with authors or journal is recommended.
Report is AI-assisted and intended for academic discussion only; final determinations of misconduct should rest with formal institutional investigations.