IN MOST cases of peer review, scientists review a paper for free while the publisher makes a huge profit margin. Two of the world’s biggest academic publishers are Springer Nature and Elsevier: their recent profit margins range from 28 to 38 per cent. The fees charged by these publishers are generally paid for by research grants; in Britain, that is mostly public money. The status quo of science is worth a lot of money — for some. 

However, not everything that is peer-reviewed is valid. A recent book published this month, Charles Piller’s Doctored, covers the story of fraud in research on Alzheimer’s disease — all of which had been peer-reviewed.

One of the leading theories in the late 20th century about Alzheimer’s was that the accumulation of amyloid proteins was to blame. Amyloids, named after their resemblance to starch molecules, clump together into tangles of gunk in the brain.

It was a compelling hypothesis: this gunk was impairing the brain’s functioning. Therefore, it made sense to try to target the amyloid proteins to stop them forming the tangles. Unfortunately, many experimental drugs with this goal failed to halt the progression of Alzheimer’s. Could the amyloid hypothesis be wrong?

Then in 2006, a French neuroscientist called Sylvain Lesne published a paper in Nature. Lesne and his colleagues injected one particular amyloid protein into rats and showed that the rats immediately developed memory problems. It seemed like a conclusive result: an amyloid protein was able to cause Alzheimer-like symptoms. The amyloid hypothesis received a major boost and the study was cited thousands of times.

Only in the 2020s did the story unravel. A community of “image sleuths” who had become experts in finding manipulated images looked at Lesne’s paper. They found that key images in it had been faked: somebody had photoshopped the desired results rather than reporting them honestly. They found similar problems in Lesne’s other papers. The other co-authors on Lesne’s paper eventually withdrew their support. Now that Piller’s book on the scandal is out, Lesne has resigned from his professorship at the University of Minnesota.

Lesne was not the only Alzheimer’s researcher implicated. Last year, Piller published an article about Eliezer Masliah, a neuroscientist and director of the National Institute for Aging, with funding for Alzheimer’s research under his remit. Similar to Lesne, his published papers turned out to contain numerous cases of image fraud. One neuroscientist who reviewed the dossier of problematic images was stunned. They told Piller their reaction was: “I was falling from a chair, basically.”

The scandal has revealed other cases. Astonishingly, one of the neuroscientists who had first raised concerns about Lesne discovered that his own academic mentor had also faked images on papers they had published together.

“I’m nauseated talking about it,” he told Piller.

The proliferation of doctored images may well have progressed Alzheimer’s research along a bogus path — although even some supporters of the amyloid hypothesis have told Piller they never believed Lesne’s 2006 paper anyway. But either way, the sorry debacle highlights that the peer review process cannot reliably catch fraud. What is the solution to such subtle subterfuge?

It’s tempting, but dangerous, to see the solution to the problem as greater punishments for fraudsters. In his own words, Piller has said what’s needed is for the institutions of science to “get serious, fast, about policing fraud and fakery.” It’s not clear exactly what he means, but the choice of word is telling — and concerning.

While it seems reasonable that a scientist with an egregious record of fraud should be fired after considering due process, that can happen already. Tougher “policing” would mean giving institutions more authority over scientists and their work, leading to an even more top-heavy structure. The Alzheimer’s examples are particularly clear here: the successful frauds were professors, mentors and directors of institutes. 

Though it is tempting to appeal to a policeman to step in and discipline the frauds, it is precisely the top-down exercise of power that supports dishonesty. Science at its best is a network of people who work together by challenging each other. 

The fraud cases Piller covers highlight that science as currently practised in universities can be gamed. So it makes sense to look again at the game. 

A commodity fetishism around “high-impact” papers directly helps frauds: knowledge becomes a relationship of papers and journals, not something nurtured, sustained and developed through social relations.

Richard Feynman once said that “the first principle is that you must not fool yourself — and you are the easiest person to fool.” It is for just this reason that science has always been a social endeavour. The hardest person to fool is a colleague who cares: someone who is not you, but who has the time, inclination and enthusiasm to engage with your work.

When leading scientists are railroaded into becoming administrators at the top of a pyramid of labour they have no time to engage with, frauds can prosper. 

One of the most important testing grounds for scientific truth is the research group. If this group becomes a relationship of a boss and employees, the inequalities and concealments natural to the antagonism of subordination grow. The pursuit of profit — here, increased grant income or professional success through achieving the “correct” result — becomes the focus.

It’s naive to only blame bad incentives for the existence of frauds. Often frauds have deep-seated psychological reasons for their actions. But minimising the disproportionate rewards of “honest” success — flashy publications, gigantic grants, powerful positions — would do a great deal to minimise those for fraud too. If science were more equal, being a fraud would be far less attractive.

Anecdotal reports suggest that across science, journal editors are struggling to find peer reviewers. It’s a worrying trend, exacerbated by burnout and a feeling that it’s not worth it. If scientists don’t have time to hold each other to account, the scientific community is harmed.

It’s notable that the image sleuths who caught the Alzheimer’s frauds were also giving their labour for free, but nobody else was profiting beyond the value of a better understanding. They worked round the clock for something bigger: an idea.  

At a time when Donald Trump’s policies are attacking federal scientists in the US, many feel uneasy talking about a case where science got it wrong. But science is robust where it tends towards relational democracy.

Calling for more policing in response to fraud risks playing into the hands of authoritarians. Scientists need to be freely able to think outside mainstream accepted notions, to test each other’s theories, and to be happy to be found wrong. More liberty will make it easier for scientists to discover the truth  — not turning science into a police state.