It was a rare moment in cannabis research: On September 8, 2020, the journal Neuropsychopharmacology retracted a paper it had published six years earlier, titled “Cannabinoid type 1 receptor availability in the amygdala mediates threat processing in trauma survivors.” Among the authors of the retracted paper is Alexander Neumeister, a disgraced ex-NYU professor who studied the endocannabinoid system as part of the neurology of PTSD.
Many cannabis studies have been controversial. Some are clearly wrong. But exceedingly few reports involving cannabis or cannabinoids end up retracted – permanently marked as papers not meeting the standards of science that we expect.
What sort of malpractice merits a retraction? Guidelines are set out by the Committee on Publication Ethics (COPE). According to the COPE recommendations, journal editors should consider a retraction for reasons, including:
- Honest error that invalidates a result (e.g. a coding error produced totally wrong statistics)
- Fabrication or falsification of data (e.g. photoshopping images)
- Plagiarism (including excessive self-plagiarism)
- Undisclosed conflicts of interest
- Copyright infringement
- Manipulated peer review, etc.
Oftentimes, particularly in the case of honest error or authorship disputes, an article will be corrected rather than retracted. And if the journal is aware of an issue, but an investigation into the author(s) is ongoing, editors may use an expression of concern to warn readers about an unresolved concern.
Theft & Fraud
Alexander Neumeister, a NYU trauma researcher, was caught embezzling money, initiating a string of investigations into his misdeeds. In 2016, eight medical trials run by Neumeister were shut down and he was fired from his position at NYU. According to an article in The New York Times, “In several instances … Dr. Neumeister had falsified documents by signing a fellow investigator’s name on reports.”
Though his sentencing was not about cannabis per se, the privilege of being a white professional was evident at the conclusion of his trial. Medscape reporters described the bizarre situation whereby Neumeister would avoid facing the music, so to speak, by actually playing music: “In June 2018, [Neumeister] pleaded guilty to the theft of $87,000, after which a judge sentenced Neumeister — a classically trained pianist — to play piano for ‘an hour at least twice weekly for the next three years at group facilities in Bridgeport, New Haven, Hartford, and Waterbury.’”
Even during the 2018 trial, the consequences of his misconduct hadn’t yet leached into his scientific publications. It took another two years before the US Office of Research Integrity (ORI), a division of the department of Health and Human Services, addressed this issue. On Jan. 7, 2020, the ORI disclosed “that Respondent [Alexander Neumeister] engaged in research misconduct by intentionally, knowingly, and/or recklessly falsifying and/or fabricating data in the clinical records of research supported by six (6) NIMH grants, resulting in the inclusion of falsified and/or fabricated research methods and results in four (4) published papers… . “1
Among the papers red-flagged by the ORI was the 2014 Neuropsychopharmacology article on how threat processing in trauma survivors was mediated by CB1 cannabinoid receptors in the amygdala.
Neumeister, according to the ORI, “misrepresented the characteristics of the subjects entered in the research record by:
- combining data from multiple subjects to represent single subjects to justify financial payments
- changing and/or instructing his staff to change, omit, or ignore clinical and psychiatric assessment data contained in the electronic and/or written research records…”
Although Neumeister always collaborated with other researchers, only he is assigned blame for the misconduct. The punishment? The ORI concluded: Neumeister must “exclude himself voluntarily for a period of two (2) years … from any contracting or subcontracting with any agency of the United States Government.”
The meager consequence of the ORI investigation parallels his peculiar community-service sentencing.
Mums the Word
Neumeister settled the ORI’s allegations without admitting guilt, but his corrupt behavior left little room for misinterpretation. Yet it wasn’t until eight months after the ORI’s damning findings were released that Springer Nature, the publisher of Neuropsychopharmacology, retracted the paper. What took so long?
“The editors have retracted this article. An investigation conducted by the US Office of Research Integrity (ORI) concluded this article includes ‘falsified and/or fabricated research methods and results.’
The authors A Neumeister, H Huang, M Zheng, R E Carson, M N Potenza, R H Pietrzak, and D Piomelli have chosen not to comment on this retraction. The authors S Corsi-Travali, S-F Lin, and S Henry have not responded to correspondence about this retraction.”
As Retraction Watch wryly noted, “The notice itself sounds a lot like a child who says ‘I’m invisible because my eyes are closed.’”
The radio silence from Neumeister’s coauthors raises questions about scientific integrity. It is all the more concerning when considering that one of authors who refused to comment is Daniele Piomelli, editor-in-chief of Cannabis and Cannabinoid Research. Published by Mary Ann Liebert, CCR is perhaps the preeminent scientific journal specializing on the endocannabinoid system, as well as the plant that led to its discovery. Numerous high-quality articles have appeared in CCR, which has been endorsed by well-regarded organizations such as the International Cannabinoid Research Society and the Society of Cannabis Clinicians.
To be clear: Project CBD is not suggesting that Piomelli or Neumeister’s other coauthors engaged in unethical behavior that caused the retraction of the CB1 paper in Neuropsychopharmacology. But would Piomelli also remain silent if misconduct arose in the journal he edits? Since its launch in late 2015, Cannabis and Cannabinoid Research has not issued a single retraction. Neither has CCR marked any articles with an expression of concern.
Is this because cannabis researchers are exceptionally well-behaved and mistake-free? Actually, a dearth of retractions suggests a lack of enforcement. By not investigating misconduct, a science journal won’t have to acknowledge that poor scholarship was published in its pages. But this avoidance behavior engenders the problem, allowing it to fester as bad actors slip under the radar and continue to commit fraud.
A Badge of Honor
Retractions are not necessarily a sign that the institution of science has gone wrong. They are, in fact, an integral aspect of self-correction, to which the scientific method aspires. It is not enough to perform “good” experiments to identify what is true. Equally important is weeding through studies to make sure that poor research does not pollute the literature, identifying what we now know is false.
In fact, papers can be retracted merely for being wrong. Consider a 2009 study purporting to show that curcumin and resveratrol bind to the CB1 receptor and inhibit its activation. Four months after its publication, the authors at the University of Arkansas, Little Rock retracted the article, stating:
“Subsequent studies in our laboratory and data obtained from three additional independent labs… have failed to replicate these initial findings… As such, we are retracting our article from publication in the Journal of Pharmacology and Experimental Therapeutics (JPET)…
The authors would like to extend their sincere apology to JPET and the scientific community as a whole. It is our hope that the swift correction of our initial report by presentation of findings conducted by four independent laboratories will help to minimize any future ramifications resulting from this very unfortunate situation.”
The Little Rock researchers also propose why this could have occurred, pointing to possible contamination with a commonly-used CB1 inhibitor. A notice like this does more than enough to correct the record. It should be a badge of honor, not of shame.
Even after a retraction, publicity is essential to ensure the work is no longer cited. Not all citation managers alert the users when they attempt to reference a retracted article. (Zotero is a notable exception, linking to a database compiled by Retraction Watch.) Quietly brushing misconduct under the rug makes it even harder to correct the record.
Researchers at the University College of London recently cited Neumeister’s fraudulent CB1 study in a systematic review on “The Effectiveness of Cannabinoids in the Treatment of Posttraumatic Stress Disorder (PTSD).” And two of the 30 citations to Neumeister’s falsified work were garnered between the ORI report and the publication of the retraction notice. By dithering for eight months before retracting the Neuropsychopharmacology article, Springer Nature allowed falsehoods to propagate further into the scientific literature.
Salami Slicing & Other Scams
A handful of cannabis-related retractions can be found by searching the Retraction Watch database. When unethical behaviors are caught, such as the falsification or fabrication of data, authors will often respond that (1) a student no longer with the university performed the experiments in question, and (2) they no longer have access to the original data. These defenses have the same hollow ring as repeating “I don’t recall” in a court room to evade responsibility.
Some researchers engage in a practice euphemistically known as “salami slicing.” It means cutting up a data set into many pieces and publishing each result separately, as if they were all from separate experiments. The evaluation of academics is largely based on metrics like the impact factor (which roughly measures how many citations one’s papers accrue), as well as the sheer number of papers they’ve published. Salami slicing is a way of gaming the system by analyzing one’s data to find the minimal publishable units from one set of experiments. The diced-up papers may even cite each other, further bolstering the impact factor metric.
This practice appears to be the impetus for two retractions from Harvard researchers in 2011. The first paper described the role of the endocannabinoid system in bone marrow stem cells, followed a few months later by another paper on the influence of just the CB2 cannabinoid receptor. Both papers were retracted within the year, with the second withdrawal notice citing duplication “of data, text, and images that are nonessential to the paper.”
But retractions are not always the result of misconduct on the authors’ part. Consider, for example, what happened with a 2017 study, “Confirmed marijuana use and lymphocyte count in black people living with HIV,” published in Elsevier’s Drug and Alcohol Dependence, and subsequently retracted.
The authors of this paper had the misfortune of choosing to utilize a certain assessment tool, called the Morisky Medication Adherence Scale (MMAS), which was copyrighted by the UCLA professor Donald Morisky in 2006. Morisky has become infamous for harassing researchers.
A typical case might go like this. The MMAS is employed by scientists who want to assess if cannabis users are more or less likely to take their conventional anti-retroviral drugs to treat HIV. So they email Morisky for permission to use his patented 8-question scale, but after weeks of silence they decide to go ahead without permission – it’s just a research tool, after all.
But only days after the final publication of their article, an email arrives from Morisky’s lawyer giving the authors three options: retract the article, face a lawsuit, or pay exorbitant fees – up to tens of thousands of dollars, for use of the MMAS.
The authors retracted the 2017 article, with a notice that read, in part: “It is important to note that the retraction of this article is not the result of any research misconduct on the part of the authors or that of the team. The retraction relates to the published version of the article that contains the MMAS-8 scale. The journal will publish a revised version that does not contain the tool or any references to it.”
Some researchers have even reported that Morisky and his lawyer upped their fee after agreeing to a payment. This predatory practice leans heavily on the stigma of retractions to coerce a settlement. And now Morisky has been sued by his lawyer, and they’re in a legal wrangle over control of the MMAS-8 scale.
Adrian Devitt-Lee, Project CBD’s chief science writer, is pursuing a PhD in math at the University College of London.
© Copyright, Project CBD. May not be reprinted without permission.
Of the four Neumeister papers red-flagged by Office of Research Integrity, two were published by the Journal of the American Medical Association, JAMA, as listed below:
- Association of in vivo k-opioid receptor availability and the transdiagnostic dimensional expression of trauma-related psychopathology. JAMA Psychiatry 2014 Nov;71(11):1262-70. Erratum in: JAMA Psychiatry 2014 Nov 7;71(11):1301.
- Association of posttraumatic stress disorder with reduced in vivo norepinephrine transporter availability in the locus coeruleus. JAMA Psychiatry 2013 Nov;70(11):1199-205.
- Linking plasma cortisol levels to phenotypic heterogeneity of posttraumatic stress symptomatology. Psychoneuroendocrinology 2014 Jan;39:88-93.
- Cannabinoid type 1 receptor availability in the amygdala mediates threat processing in trauma survivors. Neuropsychopharmacology 39(11):2519-28, 2014 Oct.
JAMA was quick to respond to the ORI, retracting both papers the following month. In their retraction notice, the ORI findings against Neumeister were quoted directly. The third paper, “Linking plasma cortisol levels to phenotypic heterogeneity of posttraumatic stress symptomatology,” was published by Elsevier, a company renowned for its non-informative retraction notices. After procrastinating for six months until July 4, 2020, Elsevier issued a two-sentence retraction notice citing the ORI without describing any reasons for the retraction.
A flawed study by Harvard scientists claims that marijuana smokers get into more car accidents even when they’re not high. Stoned driving should be considered when assessing the public health impact of cannabis, but not disproportionate to the established harm.
Bias is baked into scientific research about cannabis. How can you tell what is worth reading, and what isn’t?
The research is in. Scientists have discovered a cure for cannabis addiction — and it turns out to be cannabis! Adrian Devitt-Lee explores how a new JAMA study illustrates a larger problem in the scientific community: Hypothesizing After the Results are Known.