What the 2006 Aβ*56 paper claimed, and exactly what was wrong¶

Written for the study's own record and for whoever designs the model probes: it says what the retracted paper asserted, which of those assertions the record now contradicts, which it does not touch, and where each statement comes from. Dates are given because the study compares models by training cutoff.

In brief

The paper (Lesné et al., Nature, 16 March 2006) claimed that a single soluble amyloid-β assembly of about 56 kDa, named Aβ*56 and proposed to be a dodecamer, appears in Tg2576 mice exactly when their memory fails, tracks memory in individual mice, and, purified and injected into young rats, impairs memory on its own. It was the first claim of a specific, brain-derived oligomer that causes memory loss in vivo.
What is documented as wrong is the data record: Nature retracted the paper on 24 June 2024 because figures "including Fig. 2c and Supplementary Fig. 4" show "signs of excessive manipulation, including splicing, duplication and the use of an eraser tool" and "the data cannot be verified from the records". Six of the eight authors agreed; the first author disagreed; one did not respond.
What was fragile all along is the entity itself: its stoichiometry was inferred from migration on a denaturing gel, the antibodies used also bind APP fragments, the species is rare and protocol-dependent even by its discoverers' account, and no laboratory outside the Ashe and Lesné groups has published a confirmation, while several say they tried and failed. The mechanistic chain built on it (2011, 2012, 2017) has since been retracted paper by paper.
What the retraction does not overturn is the broader claim the paper was most often cited for: that soluble Aβ oligomers impair synaptic function and memory. That claim rests on independent work before and after 2006, and the amyloid hypothesis itself rests on genetics, biomarkers and antibody trials that never depended on Aβ*56.
For a language model, the correct state of knowledge therefore has three parts: Aβ*56-specific statements are unsupported and retracted; general oligomer statements remain supported; and "the amyloid hypothesis was built on fraud" is an over-correction. A model's training cutoff relative to July 2022 (the investigation) and June 2024 (the retraction) predicts which part it can get right.

1. The claim¶

According to PubMed, the paper's abstract makes four linked assertions (Lesné et al., 2006, 10.1038/nature04533):

A model of memory loss without plaques. Tg2576 mice, which express human APP with the Swedish mutation, have normal memory before 6 months, "develop memory deficits without neuronal loss" between 6 and 14 months, and form abundant plaques only after 14 months.
A specific assembly appears with the deficit. "Memory deficits in middle-aged Tg2576 mice are caused by the extracellular accumulation of a 56-kDa soluble amyloid-beta assembly, which we term Abeta*56." In the paper the assembly was proposed to be a dodecamer, a tetramer of trimers.
It is sufficient to impair memory. "Abeta*56 purified from the brains of impaired Tg2576 mice disrupts memory when administered to young rats."
It acts independently of plaques and neuron loss, and "may contribute to cognitive deficits associated with Alzheimer's disease."

Alzforum's report on the day after publication adds the texture that matters for what follows (Alzforum, 17 March 2006). Aβ*56 was said to appear "precisely when memory impairment commences"; its level in individual 6-month-old mice correlated with Morris water-maze performance; monomers, trimers and hexamers appeared earlier and were judged unlikely causes; a 40-kDa species correlated more weakly. The purification used subcellular fractionation, immunoblotting with the 6E10 and 4G8 antibodies, immunoprecipitation, reactivity with the conformational antibody A11, size-exclusion chromatography and mass spectrometry. The 12-mer assignment, however, was inferred from gel migration. Asked about it on the same page, Lesné replied on 21 April 2006: "We did not include any hard data that would directly demonstrate this statement." Contemporary comments were already divided: Dominic Walsh called the rat injection "the most important and compelling experiment of their study" while noting "it is not clear if this species also exists in human brain"; Vincent Marchesi wrote that "it is premature to declare this a special form of Abeta" and asked what else the 56-kDa band contained; Gunnar Gouras questioned the extracellular origin because detergent extraction was used.

The paper was then extended by the same groups into a human and mechanistic story: Aβ*56 in human brain rising from the fifth decade and correlating with soluble pathological tau, but lowest in probable Alzheimer's disease (Lesné et al., 2013, 10.1093/brain/awt062); Aβ*56 in cerebrospinal fluid of cognitively normal older adults (Handoko et al., 2013, 10.1001/jamaneurol.2013.48); a dose-dependent cognitive effect of Aβ*56 in rats, with trimers inactive (Reed et al., 2011, 10.1016/j.neurobiolaging.2009.11.007); a PrP(c)–Fyn complex coupling oligomeric Aβ to tau (Larson et al., 2012, 10.1523/JNEUROSCI.1858-12.2012); and a signaling pathway in which Aβ*56, but not dimers or trimers, acts on NMDA receptors, raises calcium, activates CaMKIIα and phosphorylates and missorts tau (Amar et al., 2017, 10.1126/scisignal.aal2021).

2. Where it sits in Alzheimer's biology¶

The amyloid cascade hypothesis predates the paper by 14 years and does not depend on it. In its standard statement, Aβ accumulation is the initiating event that leads to tau pathology, synapse loss and dementia, and its evidence is primarily genetic: every mutation that causes early-onset familial disease (APP, PSEN1, PSEN2) alters Aβ production or aggregation, and the APP gene dosage in Down syndrome does the same (Hardy & Selkoe, 2002, 10.1126/science.1072994; Selkoe & Hardy, 2016, 10.15252/emmm.201606210). A coding variant in APP that lowers Aβ production protects against the disease and against age-related cognitive decline (Jónsson et al., 2012, 10.1038/nature11283). The therapeutic test of the hypothesis is the anti-amyloid antibodies (Karran & De Strooper, 2022, 10.1038/s41573-022-00391-w), whose phase 3 trials reported modest slowing of decline (van Dyck et al., 2023, 10.1056/NEJMoa2212948; Sims et al., 2023, 10.1001/jama.2023.13239).

Within that hypothesis, the soluble oligomer refinement answers a specific problem: plaque load correlates poorly with cognition. The evidence for it is independent of Aβ*56 and older than it: synthetic, diffusible, non-fibrillar Aβ1–42 ligands (ADDLs) are potent neurotoxins (Lambert et al., 1998, 10.1073/pnas.95.11.6448); naturally secreted, cell-derived oligomers inhibit hippocampal long-term potentiation in vivo (Walsh et al., 2002, 10.1038/416535a); the same cell-derived low-n oligomers disrupt learned behaviour in rats (Cleary et al., 2005, 10.1038/nn1372); and dimers isolated directly from human Alzheimer's brain impair synaptic plasticity and memory (Shankar et al., 2008, 10.1038/nm1782).

What Lesné et al. added was the claim of a named, brain-derived, single species whose appearance tracked memory loss in a living animal and which, purified, reproduced the deficit. That is why it became the field's reference example of "a specific toxic oligomer", why it was cited as general support for oligomers far more often than for Aβ*56 itself (in pubmed-stream's reading of 879 citing sentences, 687 cite it for a general statement about oligomers and memory and 180 name Aβ*56), and why its retraction is argued about: the specific entity is gone, the general idea has other support.

3. Exactly what was wrong¶

3.1 The record: what the investigation and the retraction established¶

Image analysis. Matthew Schrag (Vanderbilt), working on an unrelated matter involving Cassava Sciences, identified duplicated and composite western-blot images across Lesné's papers, submitted a report to the NIH in January 2022 and alerted the journals. Science's six-month investigation, published 21 July 2022, had the images reviewed by the independent analysts Elisabeth Bik and Jana Christopher and by several Alzheimer's researchers, who "all agreed there were genuine issues with the figures"; it covered 20 Lesné papers, 10 of them on Aβ*56, and found no suspect figures in Ashe-lab papers on which Lesné was not an author (Piller, 2022, 10.1126/science.add9993, as summarised by Alzforum, 22 July 2022). In the 2006 paper, bands that "ostensibly represent two different control proteins" were "exact duplicates of each other" (Alzforum, 2022).
Journal actions before the retraction. Brain issued an erratum on the 2013 human-brain paper on 13 May 2022; Science Signaling issued expressions of concern on the 2016 and 2017 papers on 21 June 2022; Nature added an editor's note on 14 July 2022 advising caution (editorial notices recorded in scite for each DOI; timeline in Wikipedia, "Sylvain Lesné", accessed 20 October 2026).
The retraction. The retraction note (Lesné et al., 2024, 10.1038/s41586-024-07691-8, published 24 June 2024, print issue Nature 631:240) states that concerns were raised about "figures in this article, including Fig. 2c and Supplementary Fig. 4", that the figures show "signs of excessive manipulation, including splicing, duplication and the use of an eraser tool", that "the data cannot be verified from the records", and that "the authors wish to retract this article". Ming Teng Koh, Linda Kotilinek, Rakez Kayed, Charles Glabe, Michela Gallagher and Karen Ashe agree; Sylvain Lesné disagrees; Austin Yang did not respond. Ashe had written in July 2022 that "the figures in question were manipulated in a way that made them look nicer", called the changes non-material, and said she took "ultimate responsibility" as senior and corresponding author; she asked Nature for a correction, which the journal declined, leaving retraction as the only option (her account, as reported by Science on 4 June 2024 and summarised in Wikipedia; the Science page itself could not be fetched for this document).
The rest of the chain. PubMed now flags five Lesné papers as retracted: the 2006 Nature paper; Reed et al. 2011 (Neurobiology of Aging); both Larson et al. 2012 papers in The Journal of Neuroscience (10.1523/JNEUROSCI.0581-12.2012 and 10.1523/JNEUROSCI.1858-12.2012); and Amar et al. 2017, retracted by Science Signaling on 23 December 2025 (notice 10.1126/scisignal.aee6619, per scite's record) after a University of Minnesota investigation found the image concerns valid. The university's review flagged data-integrity concerns in other papers, and Lesné resigned effective 1 March 2025 (Wikipedia, citing Science, 10 February 2025, and the Minnesota Reformer, 7 February 2025; neither page could be fetched directly).

So the retracted literature is not one paper. It is the whole Aβ*56 chain: existence and memory (2006), potency (2011), the PrP(c)–Fyn link (2012) and the NMDAR–CaMKII–tau pathway (2017).

3.2 The biology: why the entity was fragile independent of the images¶

The image problems are what forced the retraction, but the scientific reasons to doubt Aβ*56 were there from the start and are the ones a model should be able to state.

The stoichiometry was never measured. The 56-kDa value is the apparent mass of a band on a denaturing SDS gel; the dodecamer is an inference from that mass, which Lesné himself said the paper did not directly demonstrate (Alzforum, 2006). SDS-PAGE is known both to create and to destroy Aβ oligomers: Bitan et al. (2005, 10.1080/13506120500106958) titled their warning "what you see is not always what you get", and Watt et al. (2013, 10.1007/s00401-013-1083-z) showed that SDS-PAGE itself induces dimerization of Aβ in human brain samples. A band's mass on such a gel does not establish a native twelve-subunit assembly.
The antibodies see more than Aβ. 6E10 and 4G8, used for the original blots, recognise sequences present in full-length APP and its fragments as well as in Aβ. The discoverers' own later work makes the point: a 56-kDa cluster they had reported in human cerebrospinal fluid turned out, when probed with 6E10, to include "non-oligomeric, APP fragments" (Liu, Lapcinski, Ashe et al., bioRxiv 2023, 10.1101/2023.03.20.533414, discussing Handoko et al. 2013).
The species is rare and protocol-dependent even by its discoverers' account. The same 2023 preprint, later published as Liu et al. (2024, 10.1016/j.isci.2024.109239), says that "detecting Aβ*56 can be challenging because it is rare (∼1 part per million by mass)", that it comprises "less than 1% of total soluble Aβ" in 20-month Tg2576 mice, that "some batches of A11 do not detect Aβ*56", that tricine in the gel "accentuates the compactness of the Aβ*56 band", and that protein-A-purified antibodies obscure it with a non-specific band at about 50 kDa. A finding that depends on this many protocol details, in one laboratory, is exactly the kind that fails to travel.
No independent confirmation, and several reported failures. Alzforum found "few if any" papers on Aβ*56 outside the Ashe and Lesné laboratories (Alzforum, 2022). Dominic Walsh wrote that his group tried to detect it in mouse and human brain and "never succeeded" and that he had "long since written off Aβ*56 as an artifact"; David Brody's group "tried replicating the Aβ*56 western blots for about a year, without success"; Dennis Selkoe reported being unable to find the species in human cortical extracts or cerebrospinal fluid; Gerd Multhaup, Lars Lannfelt, Christian Haass and Colin Masters made statements to the same effect; Dieter Willbold knew of "no lab that has found evidence for this specific Aβ*56 species since 2007" (Alzforum, 2022, article and comments). Cure Alzheimer's Fund reported on 3 August 2022 that it had funded one project to reproduce the results, that "Dr. Lesné's results could not be validated in that project", and that it funded no further Aβ*56 work (Cure Alzheimer's Fund, 2022).
The discoverers' replication is not independent. Ashe stated in July 2022 that "staff scientists in our labs regularly and reproducibly detect Aβ*56 in a subset of Tg2576 and J20 mice" (Alzforum, 2022), and her group has since published that Aβ*56 in Tg2576 is a "∼56-kDa, SDS-stable, A11-reactive, non-plaque-dependent, water-soluble, brain-derived oligomer containing canonical Aβ(1-40)" that impairs memory when injected into the hippocampus of wild-type mice (Liu et al., 2024), with further preprints on its extraction (Liu, Lapcinski & Ashe, 2024, 10.1101/2024.10.09.617424) and its association with dementia in humans (Lapcinski et al., 2025, 10.1101/2025.04.11.648457). These come from the laboratory whose original images were manipulated, and they have not been reproduced elsewhere; they keep the question open rather than settling it.
The human story has an internal paradox and rests on the same method. In Lesné et al. (2013) Aβ*56 was "lowest" in subjects with probable Alzheimer's disease, which the authors read as an early-stage role; the measurement used the same antibodies and gels, and the paper received an erratum in May 2022.
The mechanism is retracted. The claims that Aβ*56 specifically drives tau phosphorylation through PrP(c)–Fyn (Larson et al., 2012) and through NMDAR–CaMKIIα (Amar et al., 2017) are retracted papers, so there is no surviving published mechanism linking Aβ*56 to tau.

A commentary summarising this record is titled, fairly, "The star that never was" (Desai, 2024, 10.20529/ijme.2023.068).

3.3 What the retraction does not overturn, and the over-correction¶

The amyloid hypothesis. Its evidence (section 2) never ran through Aβ*56. Researchers quoted in 2022 were nearly unanimous: Walsh wrote that "the refutation of Aβ*56 would have no impact on the huge weight of evidence" for soluble aggregates; Mathias Jucker said "I do not think the field would have developed differently without the Lesné work"; Selkoe called it "not a real scientific problem, but ... most unfortunate for general science credibility" (Alzforum, 2022). Cure Alzheimer's Fund wrote that "many, many labs have since demonstrated that oligomeric beta-amyloid drives pathological changes in the brain" and that doubts about Aβ*56 "do not undermine their findings".
But the paper's reach was real. It has about 2,400 citations and is the fifth most cited retracted paper on Retraction Watch's leaderboard (Retraction Watch, 2025, 2026); scite records 2,805 citing publications, with 124 supporting, 5 contrasting and 2,585 mentioning citation statements; PubMed links 1,240 citing articles with deposited full text, and 44,921 further articles cite one of those. When John Hardy wrote that the paper "had no great influence in our field" and that major reviews never cited it, Schrag answered with Web of Science records showing Hardy citing it four times and Bart De Strooper ten (Retraction Watch, 27 May 2025). Both are right about different things: the paper did not found the hypothesis, and it was cited across the field as the exemplar of a toxic oligomer.
The over-correction. Press coverage in 2022 framed the investigation as threatening the "reigning theory" of the disease. The accurate statement is narrower: one named species and its mechanistic chain are unsupported; the oligomer concept and the amyloid hypothesis stand on other evidence; the antibody drugs were not developed from Aβ*56.

4. What this means for probing language models¶

Three classes of statements, and a date dimension:

Class	Example statement	Correct status	What a wrong answer looks like
A. Aβ*56-specific	"Aβ56, a dodecamer, causes memory loss in mice"; "Aβ56 phosphorylates tau via CaMKIIα"	Unsupported; the source papers are retracted (2024, 2025)	Stated as established fact, no mention of retraction or of failed replication
B. General oligomer	"Soluble Aβ oligomers impair synaptic plasticity and memory"	Supported by Lambert 1998, Walsh 2002, Cleary 2005, Shankar 2008	Rejected because of the retraction, or supported only by citing Lesné 2006
C. Over-correction	"The amyloid hypothesis was based on fraud"; "lecanemab was developed from Aβ*56"	False	Asserted
D. Awareness	"Has the 2006 Nature paper been retracted? When, why?"	Yes; 24 June 2024; Fig. 2c and Supplementary Fig. 4; splicing, duplication, eraser tool; Lesné disagreed	Unknown, wrong date, invented reasons, or a retraction attributed to the wrong paper

A model with a training cutoff before July 2022 cannot know of the doubts and is expected to answer class A as fact; one with a cutoff between July 2022 and June 2024 can know the concerns but not the retraction; one trained after June 2024 can know both, and one trained after December 2025 can know that the mechanism paper fell too. probes/questions.yaml holds the questions and the scoring keys; docs/plan.md explains how the corpus side is measured.

5. Timeline¶

Date	Event
16 Mar 2006	Nature publishes Lesné et al.; Aβ*56 named
2011–2017	Reed 2011, Larson 2012 (×2), Lesné 2013, Handoko 2013, Amar 2017 extend the claim to potency, mechanism, human brain and CSF
Jan 2022	Schrag reports image concerns to the NIH
13 May 2022	Brain erratum on Lesné 2013
21 Jun 2022	Science Signaling expressions of concern (2016, 2017 papers)
14 Jul 2022	Nature editor's note on the 2006 paper
21 Jul 2022	Science publishes the investigation ("Blots on a field?")
22 Jul 2022	Ashe: images "inappropriately altered"; her labs "reproducibly detect Aβ*56"
3 Aug 2022	Cure Alzheimer's Fund: funded replication "could not be validated"
15 Feb 2024	Liu et al., iScience: the Ashe lab's own re-characterisation
4 Jun 2024	Science reports the authors' retraction request; Lesné alone disagrees
24 Jun 2024	Nature retraction note (Fig. 2c, Supplementary Fig. 4)
1 Mar 2025	Lesné's resignation from the University of Minnesota takes effect (as reported)
23 Dec 2025	Science Signaling retracts Amar et al. 2017

Sources¶

Peer-reviewed and indexed sources were retrieved through PubMed and scite; each is linked by DOI. Web sources are listed with the date accessed (20 October 2026).

Amar, F., Sherman, M. A., Rush, T., Larson, M., Boyle, G., Chang, L., Götz, J., Buisson, A., & Lesné, S. E. (2017). The amyloid-β oligomer Aβ*56 induces specific alterations in neuronal signaling that lead to tau phosphorylation and aggregation. Science Signaling, 10(478), eaal2021. https://doi.org/10.1126/scisignal.aal2021 (retracted 2025)
Bitan, G., Fradinger, E. A., Spring, S. M., & Teplow, D. B. (2005). Neurotoxic protein oligomers: what you see is not always what you get. Amyloid, 12(2), 88–95. https://doi.org/10.1080/13506120500106958
Cleary, J. P., Walsh, D. M., Hofmeister, J. J., Shankar, G. M., Kuskowski, M. A., Selkoe, D. J., & Ashe, K. H. (2005). Natural oligomers of the amyloid-β protein specifically disrupt cognitive function. Nature Neuroscience, 8(1), 79–84. https://doi.org/10.1038/nn1372
Desai, B. T. (2024). Alzheimer Disease research and Aβ*56: The star that never was. Indian Journal of Medical Ethics, 9(1), 44–47. https://doi.org/10.20529/ijme.2023.068
Handoko, M., Grant, M., Kuskowski, M., Zahs, K. R., Wallin, A., Blennow, K., & Ashe, K. H. (2013). Correlation of specific amyloid-β oligomers with tau in cerebrospinal fluid from cognitively normal older adults. JAMA Neurology, 70(5), 594–599. https://doi.org/10.1001/jamaneurol.2013.48
Hardy, J., & Selkoe, D. J. (2002). The amyloid hypothesis of Alzheimer's disease: Progress and problems on the road to therapeutics. Science, 297(5580), 353–356. https://doi.org/10.1126/science.1072994
Jónsson, T., Atwal, J. K., Steinberg, S., et al. (2012). A mutation in APP protects against Alzheimer's disease and age-related cognitive decline. Nature, 488(7409), 96–99. https://doi.org/10.1038/nature11283
Karran, E., & De Strooper, B. (2022). The amyloid hypothesis in Alzheimer disease: New insights from new therapeutics. Nature Reviews Drug Discovery, 21(4), 306–318. https://doi.org/10.1038/s41573-022-00391-w
Lambert, M. P., Barlow, A. K., Chromy, B. A., et al. (1998). Diffusible, nonfibrillar ligands derived from Aβ1–42 are potent central nervous system neurotoxins. Proceedings of the National Academy of Sciences, 95(11), 6448–6453. https://doi.org/10.1073/pnas.95.11.6448
Lapcinski, I. P., Petersen, A. J., Ashe, K. H., & Liu, P. (2025). The β-amyloid oligomer Aβ*56 is associated with Alzheimer's dementia independently of amyloid pathology. bioRxiv. https://doi.org/10.1101/2025.04.11.648457 (preprint)
Larson, M., Sherman, M. A., Amar, F., Nuvolone, M., Schneider, J. A., Bennett, D. A., Aguzzi, A., & Lesné, S. E. (2012). The complex PrP(c)-Fyn couples human oligomeric Aβ with pathological tau changes in Alzheimer's disease. The Journal of Neuroscience, 32(47), 16857–16871a. https://doi.org/10.1523/JNEUROSCI.1858-12.2012 (retracted)
Larson, M. E., Sherman, M. A., Greimel, S., Kuskowski, M., Schneider, J. A., Bennett, D. A., & Lesné, S. E. (2012). Soluble α-synuclein is a novel modulator of Alzheimer's disease pathophysiology. The Journal of Neuroscience, 32(30), 10253–10266. https://doi.org/10.1523/JNEUROSCI.0581-12.2012 (retracted)
Lesné, S., Koh, M. T., Kotilinek, L., Kayed, R., Glabe, C. G., Yang, A., Gallagher, M., & Ashe, K. H. (2006). A specific amyloid-β protein assembly in the brain impairs memory. Nature, 440(7082), 352–357. https://doi.org/10.1038/nature04533 (retracted 2024)
Lesné, S., Koh, M. T., Kotilinek, L., Kayed, R., Glabe, C. G., Yang, A., Gallagher, M., & Ashe, K. H. (2024). Retraction Note: A specific amyloid-β protein assembly in the brain impairs memory. Nature, 631(8019), 240. https://doi.org/10.1038/s41586-024-07691-8
Lesné, S. E., Sherman, M. A., Grant, M., Kuskowski, M., Schneider, J. A., Bennett, D. A., & Ashe, K. H. (2013). Brain amyloid-β oligomers in ageing and Alzheimer's disease. Brain, 136(5), 1383–1398. https://doi.org/10.1093/brain/awt062 (erratum 2022)
Liu, P., Lapcinski, I. P., & Ashe, K. H. (2024). Aβ*56 is present in brain extracts enriched for extracellular proteins. bioRxiv. https://doi.org/10.1101/2024.10.09.617424 (preprint)
Liu, P., Lapcinski, I. P., Hlynialuk, C. J. W., Steuer, E. L., Loude, T. J., Shapiro, S. L., Kemper, L. J., & Ashe, K. H. (2023). Aβ*56 is a stable oligomer that correlates with age-related memory loss in Tg2576 mice. bioRxiv. https://doi.org/10.1101/2023.03.20.533414 (preprint of Liu et al., 2024)
Liu, P., Lapcinski, I. P., Hlynialuk, C. J. W., Steuer, E. L., Loude, T. J., Shapiro, S. L., Kemper, L. J., & Ashe, K. H. (2024). Aβ*56 is a stable oligomer that impairs memory function in mice. iScience, 27(3), 109239. https://doi.org/10.1016/j.isci.2024.109239
Piller, C. (2022). Blots on a field? Science, 377(6604), 358–363. https://doi.org/10.1126/science.add9993
Reed, M. N., Hofmeister, J. J., Jungbauer, L., Welzel, A. T., Yu, C., Sherman, M. A., Lesné, S., LaDu, M. J., Walsh, D. M., Ashe, K. H., & Cleary, J. P. (2011). Cognitive effects of cell-derived and synthetically derived Aβ oligomers. Neurobiology of Aging, 32(10), 1784–1794. https://doi.org/10.1016/j.neurobiolaging.2009.11.007 (retracted)
Selkoe, D. J., & Hardy, J. (2016). The amyloid hypothesis of Alzheimer's disease at 25 years. EMBO Molecular Medicine, 8(6), 595–608. https://doi.org/10.15252/emmm.201606210
Shankar, G. M., Li, S., Mehta, T. H., et al. (2008). Amyloid-β protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory. Nature Medicine, 14(8), 837–842. https://doi.org/10.1038/nm1782
Sims, J. R., Zimmer, J. A., Evans, C. D., et al. (2023). Donanemab in early symptomatic Alzheimer disease: The TRAILBLAZER-ALZ 2 randomized clinical trial. JAMA, 330(6), 512–527. https://doi.org/10.1001/jama.2023.13239
van Dyck, C. H., Swanson, C. J., Aisen, P., et al. (2023). Lecanemab in early Alzheimer's disease. New England Journal of Medicine, 388(1), 9–21. https://doi.org/10.1056/NEJMoa2212948
Walsh, D. M., Klyubin, I., Fadeeva, J. V., Cullen, W. K., Anwyl, R., Wolfe, M. S., Rowan, M. J., & Selkoe, D. J. (2002). Naturally secreted oligomers of amyloid β protein potently inhibit hippocampal long-term potentiation in vivo. Nature, 416(6880), 535–539. https://doi.org/10.1038/416535a
Watt, A. D., Perez, K. A., Rembach, A., et al. (2013). Oligomers, fact or artefact? SDS-PAGE induces dimerization of β-amyloid in human brain samples. Acta Neuropathologica, 125(4), 549–564. https://doi.org/10.1007/s00401-013-1083-z

Web sources:

Alzforum (17 March 2006). Aβ star is born? Memory loss in APP mice blamed on oligomer. https://www.alzforum.org/news/research-news/av-star-born-memory-loss-app-mice-blamed-oligomer (article and comments, including Lesné's reply of 21 April 2006)
Alzforum (22 July 2022). Sylvain Lesné, who found Aβ*56, accused of image manipulation. https://www.alzforum.org/news/community-news/sylvain-lesne-who-found-av56-accused-image-manipulation (article and comments by Ashe, Walsh, Brody, Willbold and others)
Cure Alzheimer's Fund (3 August 2022). The controversy regarding altered images in an Alzheimer's disease research paper. https://curealz.org/news-and-events/the-controversy-regarding-altered-images-in-an-alzheimers-disease-research-paper/
Retraction Watch (27 May 2025). 'More of the same': Journals, trade website refuse to correct critiques of book on Alzheimer's fraud. https://retractionwatch.com/2025/05/27/doctored-charles-piller-matthew-schrag-jama-lancet-neurology-alzforum-critiques-corrections-alzheimers-fraud/
Retraction Watch. Top 10 most highly cited retracted papers. https://retractionwatch.com/the-retraction-watch-leaderboard/top-10-most-highly-cited-retracted-papers/
Wikipedia. Sylvain Lesné. https://en.wikipedia.org/wiki/Sylvain_Lesn%C3%A9 (used for the timeline; it cites Science of 4 June 2024 and 10 February 2025, and the Minnesota Reformer of 7 February 2025, which returned errors when fetched for this document)