Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Medicine Science

Potential Fabrication In Research Images Threatens a Key Theory of Alzheimer's Disease (science.org) 99

Posted by EditorDavid from the mind-games dept.
A 37-year-old junior professor in Tennessee "identified apparently altered or duplicated images in dozens of journal articles," reports Science magazine.

But that was just the beginning for Matthew Schrag, whose sleuthing then "drew him into a different episode of possible misconduct, leading to findings that threaten one of the most cited Alzheimer's studies of this century and numerous related experiments." The first author of that influential study, published in Nature in 2006, was an ascending neuroscientist: Sylvain Lesné of the University of Minnesota, Twin Cities. His work underpins a key element of the dominant yet controversial amyloid hypothesis of Alzheimer's, which holds that A clumps, known as plaques, in brain tissue are a primary cause of the devastating illness, which afflicts tens of millions globally. In what looked like a smoking gun for the theory and a lead to possible therapies, Lesné and his colleagues discovered an A subtype and seemed to prove it caused dementia in rats.

If Schrag's doubts are correct, Lesné's findings were an elaborate mirage....

A 6-month investigation by Science provided strong support for Schrag's suspicions and raised questions about Lesné's research. A leading independent image analyst and several top Alzheimer's researchers — including George Perry of the University of Texas, San Antonio, and John Forsayeth of the University of California, San Francisco (UCSF) — reviewed most of Schrag's findings at Science's request. They concurred with his overall conclusions, which cast doubt on hundreds of images, including more than 70 in Lesné's papers. Some look like "shockingly blatant" examples of image tampering, says Donna Wilcock, an Alzheimer's expert at the University of Kentucky. The authors "appeared to have composed figures by piecing together parts of photos from different experiments," says Elisabeth Bik, a molecular biologist and well-known forensic image consultant. "The obtained experimental results might not have been the desired results, and that data might have been changed to ... better fit a hypothesis...."

Schrag's work, done independently of Vanderbilt and its medical center, implies millions of federal dollars may have been misspent on the research — and much more on related efforts. Some Alzheimer's experts now suspect Lesné's studies have misdirected Alzheimer's research for 16 years. "The immediate, obvious damage is wasted NIH funding and wasted thinking in the field because people are using these results as a starting point for their own experiments," says Stanford University neuroscientist Thomas Südhof, a Nobel laureate and expert on Alzheimer's and related conditions.

Lesné did not respond to requests for comment....

Some Alzheimer's experts see a failure of skepticism, including by journals that published the work.
Schrag has warned America's National Institutes of Health that the suspect work "not only represents a substantial investment in [NIH] research support, but has been cited ... thousands of times and thus has the potential to mislead an entire field of research."

And Harvard neurologic disease professor Dennis Selkoe told Science "There are certainly at least 12 or 15 images where I would agree that there is no other explanation" than manipulation. Selkoe's bigger worry, he says, is that the Lesné episode might further undercut public trust in science during a time of increasing skepticism and attacks. But scientists must show they can find and correct rare cases of apparent misconduct, he says. "We need to declare these examples and warn the world."
Thanks to Slashdot reader Crypto Fireside for sharing the story!
This discussion has been archived. No new comments can be posted.

Potential Fabrication In Research Images Threatens a Key Theory of Alzheimer's Disease More

Potential Fabrication In Research Images Threatens a Key Theory of Alzheimer's Disease

Comments Filter:

  • Perverse incentives (Score:5, Insightful)

    by Mononymous ( 6156676 ) on Saturday July 23, 2022 @10:50PM (#62728486)

    It's all about impactful published results, how much your work is cited, not how correct it is.
    Fortunately, most people seem to resist the temptation.

    • Re: (Score:3)

      by Mal-2 ( 675116 )

      Except you'll get caught, unless you're exceptionally lucky and get it more or less right in spite of cutting corners. The more you get cited, the sooner a discrepancy may be noticed. That's what I don't get. If someone is just publishing crap to meet a quota, they're not going to particularly want others to cite it...

      • Re:Perverse incentives (Score:4, Interesting)

        by quintessencesluglord ( 652360 ) on Saturday July 23, 2022 @11:33PM (#62728548)

        How long did the fat/cholesterol heart disease link go on for? And how much damage followed in its wake?

        One of my instructors use to work in a lab but couldn't produced the desired results, so was let go and ended up in teaching (great teacher by the way).

        Per him, fudging results is the norm instead of an outlier.

        • Re: (Score:3)

          by ufgrat ( 6245202 )

          There's still a link, there's just some question as to what causes cholesterol issues. What's not a question is that if your arteries harden due to cholesterol / plaque buildup, they'll have a harder time functioning, and if you get blocked arteries in your heart from the same type of build up, you will have heart problems.

          But you're right, many of our current issues with food was the direct result of the anti-fat campaigns that started in the late 70's, and resulted in most of the flavor in our foods bein

        • Re: (Score:3)

          by Bongo ( 13261 )

          Yes, and for me, hearing about this years ago is what made me realise, of shit you can't just blindly "trust the science". There's a lot of subtle incentives going on, and all scientists are human beings, and we all suffer biases whether we realise it or not. Most people are not whistleblowers -- one psychologist estimates only 1 in a 1000 has a spine. We all work in systems and if the system is heading in a certain direction, who are you to stop it? Where's the math person who pointed out that sub-prime wa

        • One of the most interesting papers I've read on the subject - and a deep look into how bad results can fill up the literature even with good intentions - was Is there a publication bias in behavioral intranasal oxytocin research on humans? Opening the file drawer of one lab [core.ac.uk]. After half a decade of doing research and getting published, they realized that if they added all of their research up, the positive results that got published and the negative results that they couldn't get any journals interested in,

      • Re:Perverse incentives (Score:5, Insightful)

        by ufgrat ( 6245202 ) on Sunday July 24, 2022 @01:35AM (#62728658)

        It's apparently taken 16 years for anyone to notice this particular fudging of results. And let's not forget the MMR vaccine causes autism-- oh wait, no it doesn't, that was made up too.

        Ego and vanity are a limitless resource for some people.

      • Not necessarily. Scientists are no different than other people. They have biases just like all people. If a particular study of science has become too homogeneous (especially on a contentious topic), other scientists might not be too critical in reviewing studies that match the popular opinion (consensus). There's a real big danger in letting branches of science become bubbles of thought where anyone that raises questions to the consensus is silenced and kicked off the island. True science is about alw

      • Re: (Score:2)

        by ceoyoyo ( 59147 )

        The quota includes citations.

        People who do this kind of stuff don't really seem to think about the long term consequences if they're wrong. They're sure they're right, but that experiment just turned out a little bit the wrong way, but they can totally demonstrate the point properly by just swapping this little bit of image with this other one....

        It doesn't even have to be on purpose. There's a lot of unconscious bias too. There's a well known effect in particle physics research where anomalies are underrep

    • Those that resist don't get funding (Score:4, Insightful)

      by FeelGood314 ( 2516288 ) on Saturday July 23, 2022 @11:22PM (#62728532)
      My girlfriend does blood flow analysis of brains and the she can't replicate anything that anyone has done in her very narrow area of expertise. She's gotten so good (and fast) at it that many other researchers ask her to do parts of their studies. It might be a narrow, biased sample size but it's also 0 and 20 for replication. Unfortunately she hasn't discovered anything that will get funding. She had something that looked promising in sex difference in strokes but discovered that most of the difference was a sex difference in response to the anesthetic. Had she ignored that mistake she would have been published.

      • It is a major failure of science funding that people like your girlfriend find it very hard to get funding. A flashy but wrong paper in Nature, Cell or the Lancet would do far more for her career than making 100 papers right. Career wise she would have been much better off publishing the wrong paper. I get that science funding aims to give money to people who actually do work in order to avoid waste, but in service of avoiding waste, it's produced a lot. There are better alternatives to the current system.

      • I don't get why the hell that result couldn't be published -- it's a potential confounder in dozens or hundreds of other studies, and could have saved billions of dollars in research grants chasing down that gremlin over and over.

        • It happens because research doesn't just affect what we believe, it also affects what we do, and what we do makes money (or doesn't.) And if your study doesn't help anyone make money, then no one is going to be disposed towards funding more of your studies, because the people with the money have only one goal... more

        • Journals donâ(TM)t like to print articles that find no results. There, in a nutshell, is everything wrong with modern science. Itâ(TM)s like the dangerous groupthink that cleared Challenger to fly: concerns are squashed down and only good news flows upwards. So in science publishing, discoveries are published and get funding; replication studies that fail donâ(TM)t get published.

          The amount of unreproducible science out there is staggering- and the implications are frightening. This is but th

          • Re: (Score:2)

            by ceoyoyo ( 59147 )

            Journals donâ(TM)t like to print articles that find no results.

            This is true, and also the way it should be. It's quite rare for someone to actually power an experiment to show meaningful negative results, and even rarer for someone to actually do the stats necessary. Journals quite rightly reject "we did an experiment and our p-value was not less than 0.05 therefore there is no effect." This is wrong, despite how often you hear people say it.

            There is a publication bias effect when your results contradi

    • Every time this comes up, I plug Elizabeth Bik. She has an extremely developed ability to spot fabrications in images.

      There's small community of people like Elizabeth Bik who spot these fake results. Funding them is more important than funding actual research, because, as this example shows, fraudulent research slows down progress for everyone. We would be farther along if the well-meaning scientists weren't going down the wrong paths.

    • > It's all about impactful published results, how much your work is cited, not how correct it is.

      Yes - and that's true whether the scientist likes it or not.
      You can be 100% accurate in your research ad how you present it. If you're very careful and you don't make any mistakes, you'll get correct results - such as results showing that eating tons of junk food tends to make you fat. And you'll not get published. Nobody will read your paper.

      The standard in science is that there is only a 5% chance of the r

      • Re:That's why most published research is wrong (Score:4, Insightful)

        by phantomfive ( 622387 ) on Sunday July 24, 2022 @12:10AM (#62728588) Journal

        The trick (one trick) is to be shocking in your paper title, surprising in your abstract, and boring in the actual paper.

        • Re: (Score:2)

          by colfer ( 619105 )

          Compare and contrast the Ted Talk formula. The google results on that topic are both comedic parodies and sincere advice, with much overlap. So I'll skip the boring description!

      • Re: (Score:2)

        by Ed_1024 ( 744566 )

        So true. XKCD has a classic strip on that as well: https://xkcd.com/882/ [xkcd.com]

        I always thought that 1/20 was a rather low bar, considering the ~1/3,500,000 for a discovery.

        • > I always thought that 1/20 was a rather low bar

          I know what you mean. It's also measuring the probability of the wrong thing. It's backwards.

          It measures the probability of getting the result you got.
          But you already KNOW that you got that result! The probability you'll get that result is actually 100% - you already got it. What you want to compute is the odds of the null hypothesis, given the result you got. P does it backwards - the odds of getting that result, given the null hypothesis. They are ver

    • Re: (Score:2)

      by Bongo ( 13261 )

      Fortunately, most people seem to resist the temptation.

      “It is simply no longer possible to believe much of the clinical research that is published, or to rely on the judgment of trusted physicians or authoritative medical guidelines. I take no pleasure in this conclusion, which I reached slowly and reluctantly over my two decades as an editor of The New England Journal of Medicine.”

      -- former Editor-in-Chief at the New England Journal of Medicine

      • Re: (Score:3)

        by colfer ( 619105 )

        All about the benjamins, and throw in the archaic structure of university medical research labs.

        The Anil Potti scandal at Duke was eye-opening. Briefly, personalized medicine for cancer diagnosis was a Theranos-level hot topic, and a mid-level researcher faked his data. A third-year med student who knew stats raised a red flag and was pressured to keep quiet by numerous high-ranking people at Duke, including the late-career lead researcher, who knew nothing about stats but held god-like status in the depart

  • Make it a misdemeanor or massively fineable to deliberately and knowingly fabricate data (the only reason I don't say make it a felony is because it could be weaponized to suppress research). That's a type of fraud that affects many people. Many grad students have spent years on some research project only to find out one of the key papers they were basing their work on was not reproducible despite being published in a reputable journal. A big problem is that journals accept papers based on trust. There is a

    • Think long and hard about who judges criminality, and thus, who would be defining "correct" or "incorrect" science. Welcome to unintended consequences, otherwise known as "the road to hell is paved with good intentions."

      I think we have to hope that being publicly discredited as a fraud at worst and sloppy at best would be enough to discourage most bad actors.

    • Re:Make it a crime (Score:5, Informative)

      by pz ( 113803 ) on Sunday July 24, 2022 @01:44AM (#62728668) Journal

      It already is a crime to commit scientific fraud when funded by governmental institutions like the NIH and the NSF. There is an entire department called the Office for Scientific Integrity devoted entirely to this exact issue.

      The level of trust inherent in scientific publishing has been going down, slowly. I was dinged for potential plagarism in one submission recently because I had repeated one of the phrases we commonly use to describe our stimuli: I was accused of plagarizing myself from an earlier paper. Unfortunately, the flunky assigned to screening my submission lacked the ability to see the issue, understanding only that the automated tool picked up a violation and that was verboten, so we had to rewrite that phrase in order to proceed.

      Bringing it back on-topic, there has been a lot of evidence that amyloid plaques are not causal for AD that has been growing over the years, including, importantly, a string of failures of treatments that attack the plaques to produce therapeutic benefit, even in mouse models. While it isn't my field, I do follow AD research and the current shock is over the scientific misconduct, not the implications. The beta amyloid theory was already on its last legs.

      Now, the plaques are still problematic in that they do appear to be strongly associated with AD. But a lot of understanding has been developing on their origins and, importantly, their links to a protein called tau and the tangles it forms in AD.

      • Re: (Score:2)

        by MrKaos ( 858439 )

        Now, the plaques are still problematic in that they do appear to be strongly associated with AD. But a lot of understanding has been developing on their origins and, importantly, their links to a protein called tau and the tangles it forms in AD.

        I've always wondered if they are a consequence of cognitive processes that occur either in trauma or personality disorders.

  • If they find out Lesne faked data and did all this damage, his sentence should be to live and work with people suffering from dementia for the rest of his life. No holidays, no time off. Seven days a week,16 hours a day changing diapers, reading to them, cleaning them, helping them dress and eat, cleaning, cooking, doing laundry, and anything else overworked, underpaid nursing staff need done until eventually they take him away in a box.

  • Good explanation for results (Score:3)

    by Todd Knarr ( 15451 ) on Saturday July 23, 2022 @11:32PM (#62728546) Homepage

    This would neatly explain why every attempt to target amyloid plaques has failed to produce results in the clinic.

    • No, it doesn't. This deals with a specific type of amyloid called AB*56. The amyloid hypothesis itself is independent of this. Well, that model could be flawed too (though I suspect it isn't) .. but if it is, it would be for reasons unrelated to the alleged data fabricated by the study in question.

      • So why has no treatment produced any results, not even the ones that have been verified to reduce amyloid plaque levels? If X correlates to Y but successfully targeting X doesn't make any difference in Y, to me that sounds like X is a result or side-effect of Y rather than a cause and you need to stop messing around with X and go looking for another cause.

        https://www.science.org/content/blog-post/athira-and-alzheimer-s

  • Because the number of papers published in journals is the number one measure of scientific progress for most academic institutions. Most researchers are too busy moving out of the next paper instead of actually progressing science

  • To ensure consistency in our deliberations, we must steadfastly remain skeptical of the claims of the skeptic who found reasons to be skeptical of the claims of the original researcher.

    In other words, it's skeptics all the way down!

  • Selkoe's bigger worry, he says, is that the Lesné episode might further undercut public trust in science during a time of increasing skepticism and attacks.

    Gee, ya think?

    People who are skeptical aren't necessarily "anti-science"; they are just realistic that scientists are human like everyone else.

  • Old vs new science (Score:4)

    by Reiyuki ( 5800436 ) on Sunday July 24, 2022 @07:53AM (#62729040)
    Old science was about exploring the nature of the reality we live in.

    New science is about twisting reality to fit the worldviews we already hold.

    • Trying to cram new data into existing theory is not a recent phenomenon. You can ask how the "authorities" bent over backwards to rationalize Galileo's observations of an imperfect cosmos in Sidereus Nuncius. Or the centuries of attempts to rationalize observed retrograde motion within the geocentric model (orbits upon orbits upon orbits).

      We always try to understand new, sometimes conflicting data into existing models and theories (e.g., ad hoc modifications to existing theory). When enough of those conflic

      • Remember that during Copernicus/Galileo's time, the Ptolemaic model was more popular, more complete, and more accurate.

        You can build a very accurate, predictive model with completely inverted assumptions

        • That's my point, actually. The old model did work well, so when new data came around that poked holes in that model, the reflex is to fit the new data into the old model, not alter the model foundations. Or question the validity of the new data (some suggested, for example, that the sunspots were actually imperfections in the telescope lenses).

          • Gotcha. Ya, today the silliness seems to lie with virtual particles, that exist more as mathematical bridges than concrete entities (especially when 3-body problems enter the picture)

    • Re: (Score:1)

      by Tablizer ( 95088 )

      No, old science was full of riff-raff also. There's a break even point between spending our research efforts policing accuracy and in spending it on new research. Too much spent on the first means we don't get as much fresh research, and too much on the second means we get too many bad results.

      • I think the core is when a science disconnects from practice/reality.

        Discovery of the relationship between dielectricity and magnetism gave us modern Electrical Engineering, while discovering and mapping numerous quarks/gluons/leptons and various other subatomic has resulted in minimal practical benefit, if any.

  • Intentionally misleading the scientific community on health research for deadly diseases that affect millions of people is a form of mass murder. In the case of Alzheimer's, a day's delay in finding cures could result in the deaths of over 300 people. In addition to the deaths this guy may have caused, he also caused funding (which likely runs into many billions of dollars counting corporate research that may have been mislead), to be directed down a false path. That causes deaths also both in stealing fund

  • So many of these fraud cases appear to be related to 2D image manipulation (e.g., western blots). It seems like studies that produce these images need to supply verifiably raw images as supporting info (granted back in 2006 supporting info was less commonly used than it is now). How to ensure raw images/data are truly unadultered is difficult. I'm sure there are people here that know more about this than me, but I could imagine encrypted date/time stamps being injected into images upon capture, that can be

    • The way to do this is to embed a private key into the camera itself and have it sign the file. The resulting file has the image data, the signature, and a certificate with the public key that is signed by the manufacturer. You then need to keep the original image around in any files that get passed around. If the manipulations are straightforward transforms then your idea of keeping track of them and storing them as a small set of instructions in the file should work. The other route is to store the full or

  • Confirmation? (Score:3)

    by groobly ( 6155920 ) on Sunday July 24, 2022 @11:03AM (#62729376)

    Did anyone try to duplicate the results? Apparently not, because you don't get grants to duplicate work that's already been done. Therein lies the problem, folks.

    • Re: (Score:2)

      by caseih ( 160668 )

      Yes the article says there were attempts to replicate the AB*56 research but only a few researchers felt they had found any evidence to confirm that AB*56 existed at all, and none of them could isolate it like the two original authors claimed.

  • If Lesne intentionally doctored images, I hope they jail his ass. My father has Alzheimers, and it's a horrible disease. This sets back research by taking other researchers down a dead end.

  • And, in spite of there being no evidence of a benefit, the FDA approved Biogen's Alzheimer's drug. Sounds to me like some people got paid off.
  • By the way, I can tell you for sure that you should pay attention to the site https://westcoastsupply.cc/products/pre-rolls/ [westcoastsupply.cc] on which I have been buying weed for a very long time and at a very pleasant price, I myself did not expect that this could be and with high quality everything. I wish you good luck!

Slashdot Top Deals

news: gotcha

Close