Geng integrity review: 'Intermittent Exercise Activates NRG1-PI3K/Akt Pathway to Inhibit Cardiomyocyte Apoptosis in MI Rats' (doi:10.19582/j.cnki.11-3785/g8.2016.06.011)

Academic fraud report · Geng Detector

Summary

Overall verdict: highly suspicious (high-but-not-conclusive concern). The review identifies unusually small standard deviations in two tables of an in vivo rat intermittent-exercise / myocardial-infarction study (Tan et al., Journal of Beijing Sport University, 2016). In Table 2, the exercise (ME) and inhibitor (MEA) groups report LVSP SDs of ±1.94 and ±1.55 mmHg, versus ±6.36 (sham) and ±5.45 (MI), which is biologically implausible for n=12 live-animal haemodynamic measurements. In Table 1, Caspase 3 activity means swing across an order of magnitude (≈800 to ≈2941) while SDs stay tightly clustered (≈23–62), inconsistent with the heterogeneity expected from tissue homogenate assays. Western blot image authenticity could not be assessed from the available text alone. Methodological details (animal ventilator, qPCR cycler, AG1478 inhibitor) are temporally plausible for a 2016 manuscript, so no obvious low-level fabrication was detected in the Methods. Confidence is moderate-to-high for the numerical concerns; image-based verification remains a key unresolved step.

Verdict

🟠 Highly suspicious. Two quantitative anomalies (implausibly tight LVSP SDs in Table 2 and suspiciously stable SDs across a wide mean range for Caspase 3 activity in Table 1) are consistent with fabricated or simulated data, but the available source PDF did not contain extractable image data to confirm duplication/manipulation. Formal determination requires institutional investigation and access to raw records.

Key findings

Table 2 (cardiac haemodynamics): ME group LVSP = 125.66 ± 1.94 mmHg; MEA group LVSP = 111.77 ± 1.55 mmHg. Sham SD = ±6.36; MI SD = ±5.45. SDs < 2 mmHg across n = 12 live, exercised, cannulated rats are not biologically plausible.
Table 1 (Caspase 3 activity): means range from ≈800 to 2941.03 while SDs remain tightly clustered (S ±23.21, MI ±62.60, ME ±39.56, MEA ±57.23), giving implausibly low coefficients of variation for tissue-homogenate enzyme assays.
Western blot figures (Figs. 1–3) reporting NRG1, ErbB2/4, PI3K/Akt and apoptosis-related protein expression could not be pixel-checked from the provided text, leaving image authenticity unverified.
Methods (ALC V8 ventilator, Bio-Rad CFX96, AG1478) are temporally consistent with 2015–2016 practice, ruling out only the most naïve forms of fabrication.

Evidence highlights

doi: 10.19582/j.cnki.11-3785/g8.2016.06.011
Table 2 LVSP: ME 125.66 ± 1.94; MEA 111.77 ± 1.55; S ±6.36; MI ±5.45; n = 12 per group (as stated in the paper).
Table 1 Caspase 3: MI mean 2941.03, SD ±62.60 (CV ≈ 2.1%); S mean ≈ 800, SD ±23.21 (CV ≈ 2.9%).
Intermittent treadmill training duration: 8 weeks.

Notes

Limitations: review was performed from extracted text/table values only; raw haemodynamic traces (e.g., PowerLab exports), uncropped Western blots, and original lab notebooks were not available.
Recommended next steps: request raw traces and full-resolution original blots from corresponding author Tian Zhenjun; raise a PubPeer comment focused on Tables 1–2 SDs; report to the editorial office of 北京体育大学学报 for third-party statistical review; cross-check prior theses from the same laboratory for systematic irregularities.
This report is AI-assisted and intended for academic discussion only; ultimate findings of misconduct must come from an official institutional investigation.