"A good integrity day"
September 27, 2024 5:17 PM Subscribe
Neuroscientist Dr. Eliezer Masliah was prolific, publishing around 800 papers on Alzheimer's and Parkinson's. His work good enough for long enough that in 2016, he was appointed director of the Division of Neuroscience at the National Institute on Aging.
Yesterday, Science published a report casting substantial credible doubt on 132 of those papers, and on Masliah specifically, as the sole common author among the papers cited.
More specifically, the papers reportedly reuse imagery, sometimes from a control, sometimes from unrelated research papers, often with a few locations copy pasted or slightly different croppings. Software tool Imagetwin was used to find these duplicate images.
Medical chemist and blogger Derek Lowe re-iterates in plain language what the report means for drug trials:
(Post title quotes Elisabeth Bik)
Yesterday, Science published a report casting substantial credible doubt on 132 of those papers, and on Masliah specifically, as the sole common author among the papers cited.
More specifically, the papers reportedly reuse imagery, sometimes from a control, sometimes from unrelated research papers, often with a few locations copy pasted or slightly different croppings. Software tool Imagetwin was used to find these duplicate images.
Medical chemist and blogger Derek Lowe re-iterates in plain language what the report means for drug trials:
As the article details, this all has some direct drug discovery implications, particularly for an antibody called prasinezumab which targets alpha-synuclein. All four of the fundamental papers about prasinezumab (as cited on the web site of its developer, Prothena) are full of manipulated images, unfortunatelyThe NIH statement on the matter is brief but indicates that "Dr. Masliah is not serving in the capacity of director of the NIA DN."
(Post title quotes Elisabeth Bik)
We had all been wondering how a drug that didn't work ever got that far in the approval process. I would never have wanted to believe that this could be the reason.
posted by hydropsyche at 5:42 PM on September 27 [18 favorites]
posted by hydropsyche at 5:42 PM on September 27 [18 favorites]
This is also what we call "science-by-Photoshop".
posted by They sucked his brains out! at 5:43 PM on September 27 [4 favorites]
posted by They sucked his brains out! at 5:43 PM on September 27 [4 favorites]
Gandy was disturbed, for example, that Masliah and colleagues seem to have used the same image of a mitochondrion, a cellular energy-producing structure, in two articles on different topics published 2 years apart in different journals. “The bus driver could see that they are identical,” Gandy says.
While some of this inevitably falls through the cracks here and there, what I don't get is how this gets past review stage at the scale described here, as reviewers are usually colleagues — people in the same field — people who are authoring papers in the same journals, reading the same papers (and especially from high impact authors, whose work you would see more frequently and prominently).
I could see how you could get away with it once or twice, but I wonder if this will put some kind of spotlight on the peer review process — maybe some of it, unfortunately, could well do with the same kinds of automation used for undergraduate students, looking here for evidence of falsification (e.g., duplication) instead of plagiarism.
posted by They sucked his brains out! at 5:52 PM on September 27 [1 favorite]
While some of this inevitably falls through the cracks here and there, what I don't get is how this gets past review stage at the scale described here, as reviewers are usually colleagues — people in the same field — people who are authoring papers in the same journals, reading the same papers (and especially from high impact authors, whose work you would see more frequently and prominently).
I could see how you could get away with it once or twice, but I wonder if this will put some kind of spotlight on the peer review process — maybe some of it, unfortunately, could well do with the same kinds of automation used for undergraduate students, looking here for evidence of falsification (e.g., duplication) instead of plagiarism.
posted by They sucked his brains out! at 5:52 PM on September 27 [1 favorite]
This wasn't some two-bit junior prof in a publish or perish mode in a 'developing' country,
That's a really weird way to frame this. Developing countries are expected to be less ethical than developed ones? What in the history of developed countries makes you reach this conclusion?
posted by signal at 6:30 PM on September 27 [13 favorites]
That's a really weird way to frame this. Developing countries are expected to be less ethical than developed ones? What in the history of developed countries makes you reach this conclusion?
posted by signal at 6:30 PM on September 27 [13 favorites]
read the link i posted - which is a good example of the pressure that scientists are under in those places and the results that occur - and note that I put the word 'developing' in quotes for a reason....(as well as 'elite').
posted by lalochezia at 6:33 PM on September 27 [15 favorites]
posted by lalochezia at 6:33 PM on September 27 [15 favorites]
I don't see how your prejudices about "those places", whatever that refers to, is a positive or useful way to address the clear ethical shortcomings of "developed places".
posted by signal at 6:38 PM on September 27 [4 favorites]
posted by signal at 6:38 PM on September 27 [4 favorites]
Previously, on influential neuroscientists falsifying Alzheimer's research [feat. Sylvan Lesné]
posted by Iris Gambol at 6:42 PM on September 27 [3 favorites]
posted by Iris Gambol at 6:42 PM on September 27 [3 favorites]
Rich people steal more than poor people, so I don't suspect poor people of theft the same way I do rich people. But if a poor person does steal, I have a little more sympathy for those struggling to keep food and shelter than I do for those thieves who have plenty already.
This is why this jackass seems especially vile to me.
posted by SaltySalticid at 7:20 PM on September 27 [10 favorites]
This is why this jackass seems especially vile to me.
posted by SaltySalticid at 7:20 PM on September 27 [10 favorites]
With current brain scans & imaging we know more today, but know less as one cannot dissect a living brain.
That's quitter talk. You can do it! I believe in you!
posted by Kyol at 7:35 PM on September 27 [9 favorites]
That's quitter talk. You can do it! I believe in you!
posted by Kyol at 7:35 PM on September 27 [9 favorites]
In related news, Francesca Gino's claims of defamation against DataColada and Harvard Business School have been dismissed, after her own research was found to contain massive data fraud.
posted by The Pluto Gangsta at 8:15 PM on September 27 [13 favorites]
posted by The Pluto Gangsta at 8:15 PM on September 27 [13 favorites]
Every scientific paper needs to be reproduced twice by unrelated institutions.
People whose work isn't reproduced, or who don't reproduce others work, should be treated as failing basic standards.
This means papers need reproduction instructions. And unreproduced papers should have a name like "preprint" or "scientific draft" or "guess".
Yes this would result in 2x fewer papers "fully published" (assuming reproduction is half as hard as initial publication). And scientists who only puvlish drafts and don't confirm others research should be treated as feeeloaders.
This will at least require networks of fraudsters, which.
posted by NotAYakk at 8:23 PM on September 27 [2 favorites]
People whose work isn't reproduced, or who don't reproduce others work, should be treated as failing basic standards.
This means papers need reproduction instructions. And unreproduced papers should have a name like "preprint" or "scientific draft" or "guess".
Yes this would result in 2x fewer papers "fully published" (assuming reproduction is half as hard as initial publication). And scientists who only puvlish drafts and don't confirm others research should be treated as feeeloaders.
This will at least require networks of fraudsters, which.
posted by NotAYakk at 8:23 PM on September 27 [2 favorites]
Obviously this person needs to be burned at the stake, but this reflects badly on everyone. If a guy can just make shit up for a couple of decades and not get caught then your system is broken. Some people need to engage in some serious soul-searching while the rest of us are toasting marshmallows over his smoldering corpse.
posted by It's Never Lurgi at 8:41 PM on September 27 [7 favorites]
posted by It's Never Lurgi at 8:41 PM on September 27 [7 favorites]
I think progress on Parkinson's has been glacial because the attitude until recently was 'yes, we’ll take anything seriously and fund it . . . as long as it’s not chemicals! It can never, ever be chemicals!'
And I have a strong suspicion this guy played right into that, and it kept the money flowing and him in his position.
posted by jamjam at 9:19 PM on September 27 [3 favorites]
And I have a strong suspicion this guy played right into that, and it kept the money flowing and him in his position.
posted by jamjam at 9:19 PM on September 27 [3 favorites]
Whew, I had to spend a few frantic minutes checking current state-of-the-art and nobody I know was caught up in this.
...which does nobody anywhere any good (aside from me and my ability to sleep tonight), I realize, and I therefore offer purely as a statement of how seriously fucked this whole thing is and can we please start brutally punishing scientific fraud?
[A very good friend of mine forcibly shut down a drug investigation study because it wasn't producing good results. The PI wanted to find an excuse to keep going, and was extremely grumpy (borderline litigious) when my friend suggested they were going to pull rank (vastly more respected) and would shut down the study regardless of what the PI wanted. So ... maybe make sure there's always someone on the team who doesn't give a shit about what the PI wants or doesn't want? Also, wtf is it with neurodegenerative ailments and drug fakery? Just the combo of desperate and hard-to-replicate, or do the fakers actually seek it out as being safe territory?]
posted by aramaic at 10:50 PM on September 27 [6 favorites]
...which does nobody anywhere any good (aside from me and my ability to sleep tonight), I realize, and I therefore offer purely as a statement of how seriously fucked this whole thing is and can we please start brutally punishing scientific fraud?
[A very good friend of mine forcibly shut down a drug investigation study because it wasn't producing good results. The PI wanted to find an excuse to keep going, and was extremely grumpy (borderline litigious) when my friend suggested they were going to pull rank (vastly more respected) and would shut down the study regardless of what the PI wanted. So ... maybe make sure there's always someone on the team who doesn't give a shit about what the PI wants or doesn't want? Also, wtf is it with neurodegenerative ailments and drug fakery? Just the combo of desperate and hard-to-replicate, or do the fakers actually seek it out as being safe territory?]
posted by aramaic at 10:50 PM on September 27 [6 favorites]
The Pluto Gangster . . .A fly in the ointment on the Francesca Gino lawsuit. While it's clear that Gino was/is a Black Hat, her lawyer claims that she was treated more harshly than Black Hatted members of the IVP = Ivy League Patriarchy. Andrew T. Miltenberg, an attorney for Gino, wrote in a statement that the decision “clearly demonstrates Harvard treated Professor Gino differently from other misconduct investigations and their own stated policies.” and the Federal Judge agrees there is a case to answer. Data Colada get out of jail free, though; so that's a win.
posted by BobTheScientist at 11:42 PM on September 27 [4 favorites]
posted by BobTheScientist at 11:42 PM on September 27 [4 favorites]
Mu Yang has been posting her dossier findings on PubPeer. This example seems especially egregious.
posted by lock robster at 12:44 AM on September 28 [1 favorite]
posted by lock robster at 12:44 AM on September 28 [1 favorite]
Rich people steal more than poor people,
I didn't want to cllog the thread with the obvious. $$$ was the reason for his years doing this.
posted by thomcatspike at 12:51 AM on September 28 [2 favorites]
I didn't want to cllog the thread with the obvious. $$$ was the reason for his years doing this.
posted by thomcatspike at 12:51 AM on September 28 [2 favorites]
That's a really weird way to frame this. Developing countries are expected to be less ethical than developed ones? What in the history of developed countries makes you reach this conclusion?
Non-western countries have had very serious and long-standing problems with massive levels of academic fraud at almost every level and anybody even tangentially affiliated with in the english speaking academic world has had to deal with it. It's more ubiquitous than the Nigerian prince banking transfer request email scams.
What the current iteration of fraud within Western academic institutions is showing is that the model of unpaid peer review and lackadaisical but immensely profitable scientific publishing business is both unsecure and inaccurate as well as unsustainable.
Then there is the academic funding model which has a trickle up superstar funding formula that incentivises spectacle (often fraud) and mostly doesn't fund the grunt work of review and replication.
What's encouraging is that lots of big name frauds are finally getting caught! So bad actors can be removed and the systematic failures that they exploited which are getting uncovered can potentially be patched if the will to do it can be mustered.
posted by srboisvert at 3:38 AM on September 28 [10 favorites]
Non-western countries have had very serious and long-standing problems with massive levels of academic fraud at almost every level and anybody even tangentially affiliated with in the english speaking academic world has had to deal with it. It's more ubiquitous than the Nigerian prince banking transfer request email scams.
What the current iteration of fraud within Western academic institutions is showing is that the model of unpaid peer review and lackadaisical but immensely profitable scientific publishing business is both unsecure and inaccurate as well as unsustainable.
Then there is the academic funding model which has a trickle up superstar funding formula that incentivises spectacle (often fraud) and mostly doesn't fund the grunt work of review and replication.
What's encouraging is that lots of big name frauds are finally getting caught! So bad actors can be removed and the systematic failures that they exploited which are getting uncovered can potentially be patched if the will to do it can be mustered.
posted by srboisvert at 3:38 AM on September 28 [10 favorites]
While some of this inevitably falls through the cracks here and there, what I don't get is how this gets past review stage at the scale described here, as reviewers are usually colleagues — people in the same field — people who are authoring papers in the same journals, reading the same papers (and especially from high impact authors, whose work you would see more frequently and prominently).
I could see how you could get away with it once or twice, but I wonder if this will put some kind of spotlight on the peer review process — maybe some of it, unfortunately, could well do with the same kinds of automation used for undergraduate students, looking here for evidence of falsification (e.g., duplication) instead of plagiarism.
Peer review is designed to look for badly done science and poorly explained papers and to help the authors improve them until they're read for publication. It is not designed to look for outright fakery. If the journal has a series of questions for a reviewer to answer about a paper, it's "Do the results support the conclusions?" not "Is this paper fake?"
posted by hydropsyche at 3:50 AM on September 28 [15 favorites]
I could see how you could get away with it once or twice, but I wonder if this will put some kind of spotlight on the peer review process — maybe some of it, unfortunately, could well do with the same kinds of automation used for undergraduate students, looking here for evidence of falsification (e.g., duplication) instead of plagiarism.
Peer review is designed to look for badly done science and poorly explained papers and to help the authors improve them until they're read for publication. It is not designed to look for outright fakery. If the journal has a series of questions for a reviewer to answer about a paper, it's "Do the results support the conclusions?" not "Is this paper fake?"
posted by hydropsyche at 3:50 AM on September 28 [15 favorites]
And this is one of the reasons why peer review - which is built on the good faith assumption of "looking for error" rather than "looking for malfeasance" - is broken as currently configured.
This is one place where automated/machine learning analysis of papers datasets and images will be useful, to flag some of these obvious fraud cases earlier to be analyzed by a real person later. But the work involved is huge......and..... who pays for all of this?
posted by lalochezia at 5:01 AM on September 28 [1 favorite]
This is one place where automated/machine learning analysis of papers datasets and images will be useful, to flag some of these obvious fraud cases earlier to be analyzed by a real person later. But the work involved is huge......and..... who pays for all of this?
posted by lalochezia at 5:01 AM on September 28 [1 favorite]
Mod note: One removed. Please remember the Content Policy and avoid wishing harm on others.
posted by Brandon Blatcher (staff) at 6:12 AM on September 28 [3 favorites]
posted by Brandon Blatcher (staff) at 6:12 AM on September 28 [3 favorites]
The good news is that thanks to generative fill in Photoshop, future fraud will be nigh undetectable!
Which is just f’n great, isn’t it
posted by caution live frogs at 6:33 AM on September 28 [1 favorite]
Which is just f’n great, isn’t it
posted by caution live frogs at 6:33 AM on September 28 [1 favorite]
If the journal has a series of questions for a reviewer to answer about a paper, it's "Do the results support the conclusions?" not "Is this paper fake?"
I just refereed a paper earlier this week and (in addition to the longform report) one of the questions I had to answer was "Is this paper sound and not misleading?" with possible answers of Yes, Maybe, and No. (This was separate from the question about whether the paper was well organized and clear.)
But then, this was for an academic-society-run journal in a mostly non-monetizable field.
posted by heatherlogan at 6:56 AM on September 28 [3 favorites]
I just refereed a paper earlier this week and (in addition to the longform report) one of the questions I had to answer was "Is this paper sound and not misleading?" with possible answers of Yes, Maybe, and No. (This was separate from the question about whether the paper was well organized and clear.)
But then, this was for an academic-society-run journal in a mostly non-monetizable field.
posted by heatherlogan at 6:56 AM on September 28 [3 favorites]
We basically expect scientists to have an infinite supply of good faith, public spirit, capacity for boring grunt work, honesty in the face of negative consequence. Within a system that rewards absolutely none of the above. Instead, an honestly reported negative result from a well-designed and well executed study can tank your career.
So, not making excuses for this guy, but in addition to coming up with better ways to check each others' work, we might also think about misaligned incentives. Because the kind of person that will abuse a flawed system for personal gain will probably not just stop existing.
posted by kleinsteradikaleminderheit at 7:29 AM on September 28 [15 favorites]
So, not making excuses for this guy, but in addition to coming up with better ways to check each others' work, we might also think about misaligned incentives. Because the kind of person that will abuse a flawed system for personal gain will probably not just stop existing.
posted by kleinsteradikaleminderheit at 7:29 AM on September 28 [15 favorites]
« Older When Did SFF Get Too Big? | Sea robins can taste with their feet Newer »
posted by thomcatspike at 5:38 PM on September 27