Accusations of doctored images and manipulated Alzheimer's disease research may tarnish the University of Minnesota, but a bigger question looms amid the race for a cure.

What of the landmark U Alzheimer's discoveries remains valid?

Researchers questioning whether images in U studies were doctored said they could undermine a key discovery in 2006: a protein, called abeta star 56, that independently caused memory loss in rats and looked like the long-awaited smoking gun behind Alzheimer's. The leader of the U research, Dr. Karen Ashe, countered that a colleague, Sylvain Lesné, was wrong to alter images, but she defended the discovery.

"While the editing of select images should not have occurred, the adjustments are non-material, inconsequential and have no bearing on the research findings," she said.

Investigations by the U and National Institutes of Health — which funded much of the research — will assess wrongdoing by Lesné or other authors, while scientific journals determine whether the studies with suspect images require corrections or retractions.

Behind the controversy is a vexing neurological disease that afflicts 6 million Americans and is expected to grow with an aging population. The condition inhibits thinking cells, neurons, from performing cognitive or memory functions, or from conveying signals that tell muscles and organs what to do.

While multiple papers are in question, the 2006 study in the journal Nature is gaining the most attention because it discovered abeta star 56, or Aβ*56. Some researchers were dismissive because of struggles to replicate the findings, but there is little question of the study's impact. The paper has been cited thousands of times by scientists who have used it as a foundation for follow-up Alzheimer's research.

"We wouldn't be where we actually are today in terms of understanding," without this study and related research, said Maria Carrillo, chief science officer for the Alzheimer's Association. However, as the organization readied its convention in San Diego next week, she said the scheduled presentations were proof that research has moved beyond this discovery.

Rooting out academic improprieties remains important, though, she said. "We are self-policing. If we can't count on that, then everything unravels."

The U paper built on the theory that the disease was linked to amyloids, proteins that can abnormally build up as plaques and perhaps obstruct neurons. The researchers targeted soluble forms, rather than hardened plaques, that could accumulate for years in advance of the dementia symptoms that come with age.

U researchers found a correlation between Aβ*56 and cognitive problems in middle-aged mice genetically bred to produce amyloid plaques. They then purified the protein and injected it in young rats, which consequently showed memory problems based on their inability to navigate a water maze.

At the time, the discovery that "Aβ*56 impairs memory independently of plaques or neuronal loss" was hailed by Nature as a "star suspect" in the search for Alzheimer's treatments. Today, the paper is tagged with a warning to treat its results with caution until the review of disputed images is complete.

Much scrutiny is of Western blots, which use electrical charges to separate proteins and a chemical process to create visual representations of them. The size and thickness of the chemical bands produced on film correspond to the amount of protein and, by extension, whether it is implicated in a disease.

A blow-up of one blot in the Nature paper showed bands proving the presence of Aβ*56 in rats exhibiting memory loss. However, Dr. Matthew Schrag, an Alzheimer's researcher in Tennessee, found linear discolorations around the bands suggesting they may have been cut and pasted. Some bands also appeared duplicated. Another blot showed clusters of identical peripheral dots around the bands that suggested photo editing.

Schrag, conducting the review outside his work at Vanderbilt University, published his concerns on the PubPeer academic website and contributed to a Science magazine investigation of Lesné in July . Expert reviewers corroborated the concerns.

"This is a very sad example of human fragility and malfeasance," said Dr. Dennis Selkoe, a neuroscientist at Harvard Medical School. The proponent of the amyloid link to Alzheimer's agreed that some U images appeared manipulated.

The Science article suggested that Lesné had manipulated images before he joined Ashe's team as a research assistant in 2002 and was promoted as a U assistant professor with his own lab in 2009. A supervisor of Lesné's doctoral education at the University of Caen Normandy in France told the magazine he withdrew a paper before publication because he questioned images Lesné produced.

Lesné did not reply to requests to comment for this story.

The U has confronted this problem before, ordering a recall in 2008 of a landmark paper about adult stem cells after finding it contained manipulated images.

Schrag said he found no studies with manipulated images in which Ashe was an author without Lesné, but that the concerns extend beyond their 16-year-old paper. He found signs of manipulated images in a 2013 study in the journal Brain in which the U researchers affirmed their findings in human tissue about Aβ*56 as a precursor for Alzheimer's. Images issued this year as a correction look so different that Schrag wonders if they came from the same experiment.

Blots mean little to the untrained eye, but they are the essence of research, said Elisabeth Bik, a San Diego microbiologist turned forensic image consultant. She agreed some images in Lesné's papers appear manipulated.

"A science paper is not like a children's book, where the images are just there to enlighten the whole story," she said. "It's different. The images, in my opinion, are the data."

The now-disputed Nature paper influenced years of research. Federal funding increased for Alzheimer's in general, but particularly for studies targeting amyloids.

The research was necessary because amyloids are part of the Alzheimer's puzzle, but the heightened focus slowed studies of other key pieces, said Dr. Ronald Petersen, director of the Mayo Clinic Alzheimer's Disease Research Center. Immune reactions and cardiovascular disease also influence Alzheimer's along with tau, a protein that can build up abnormally inside neurons.

The disparity shows in drug development. Aduhelm received federal clearance last year as an Alzheimer's treatment that breaks down amyloid plaques, though some doctors dispute whether it also slows cognitive decline. Three monoclonal antibody infusions are midway through clinical trials; all target amyloids.

Trials of other compounds targeting amyloids have failed. Ashe said it is unfair to tag the disputed U papers for such failures because they involved classes of amyloid protein that were different and easier to replicate than Aβ*56.

Ashe said she expressed doubts that drugs targeting those proteins would work and that she isn't wedded to amyloids as the primary cause of Alzheimer's. Her research has explored tau and other potential causes.

Carrillo of the Alzheimer's Association said limited funding years ago forced conservative judgments to support research in areas such as amyloids where there was early evidence. Increases have promoted bolder exploration, and she expects Alzheimer's treatments targeting tau and inflammation to emerge right behind the current wave targeting amyloids.

Aβ*56 has been "fairly irrelevant" to current drug studies, she said, so the idea that the U controversy could undercut ongoing discoveries is "overdone and overstated."