Other

One diet soda a day, 62% faster brain aging — from a study its own journal has replaced twice

Late October 2026: a ScienceDaily summary sets X on fire with a fourteen-month-old Neurology paper — 12,772 Brazilian adults, eight years of follow-up, and the heaviest consumers of low- and no-calorie sweeteners declining 62% faster, like 1.6 extra years of brain aging. The record under the meme is messier. Three days after publication, statistician Sophie Hill audited all 256 regression coefficients in the appendix: estimates outside their own confidence intervals, duplicated rows, sign errors. The journal has since issued two Correction & Replacement notices — the second after the authors reran everything from raw data — and the study's own complete-case analysis, in the supplement, shows no association at all. Meanwhile a 14-cohort meta-analysis modeling what actually happens when people swap sugary drinks for sweetened ones finds the arrow pointing the other way. One percentage, three claims, zero settled science.

by WeirdTrue · 2026-11-08

One diet soda a day, 62% faster brain aging — from a study its own journal has replaced twice

Bottom line

The 62% headline doesn't hold up as stated: the study behind it was corrected and replaced twice by its own journal, and when only participants with complete data are analyzed, the association disappears. Whether tagatose is the 'safe' sweetener is unproven on this evidence. What is supported: replacing sugar-sweetened drinks with artificially sweetened ones is associated with lower weight and lower coronary heart disease and mortality risk (14 cohorts, 416,830 participants). If you drink sugary beverages, switching is an evidence-backed improvement; if you drink water, nothing here changes that.

Claims

#1 ScienceDaily, in its summary of the study said on 2026-10-28 · source

participants in the highest consumption group experienced a 62% faster decline in overall thinking and memory abilities compared with those in the lowest group

Adjudicated as: Drinking about one can of diet soda a day (≈191 mg of low- and no-calorie sweeteners) makes memory and thinking decline 62% faster, equal to about 1.6 extra years of brain aging.Restated in plain terms so it can be adjudicated; the verdict below addresses this version.

Partially contradictedVersion 5 · Model qwen-max · verdict_v1 · 2026-11-09 07:10

The claim that drinking one can of diet soda a day leads to a 62% faster cognitive decline is contradicted by subsequent corrections and reviews, which found errors in the original study's data. However, the association with 1.6 extra years of brain aging remains supported. The exceptional nature of the claim requires more robust evidence, and the current evidence is mixed, leading to a partially contradicted level.

Citations

www.sciencedaily.com: “Those in the lowest consumption group averaged 20 milligrams per day (mg/day), while those in the highest group averaged 191 mg/day. For comparison, 191 milligrams of aspartame is roughly the amount found in one can of diet soda.”
github.com: “In other words: one of the key estimates presented in this paper likely contains a typo in its confidence interval. Rather than being evidence of an association between sweeteners and cognitive decline, as claimed in the text, it is not statistically significant.”
github.com: “The estimates in rows 2-5 all come from eTable 9 and likely represent sign errors.”
www.sciencealert.com: “In a series of statistical models that were relegated to the supplementary appendix, they showed that in the full case analysis there was no relationship between sweeteners and aging.”
www.aan.com: “with a decline that was 63% faster. This is the equivalent of about 1.6 years of aging. Those in the middle group had a decline that was 36% faster than the lowest group, equivalent to about 1.3 years of aging.”

What would change this

Independent replication of the study's findings
Further validation of the association between diet soda and cognitive decline

Earlier versions (4)

Partially contradicted v4 · qwen-max · 2026-11-08 15:11
The initial claim that 191 mg/day of low- and no-calorie sweeteners is associated with a 62% faster cognitive decline has been contradicted by subsequent corrections and reviews, which found errors in the original study's data. The association with 1.6 extra years of brain aging remains supported, but the core premise of the association is now in doubt. Therefore, the claim is partially contradicted. The evidence suggests that the initial finding of a 62% faster cognitive decline is not reliable, and the study's results have been corrected. However, the association with 1.6 extra years of brain aging is still mentioned, though it is also in question. More research is needed to confirm the link between diet soda and brain aging.
Partially contradicted v3 · qwen-max · 2026-11-08 14:48
The initial claim that 191 mg/day of low- and no-calorie sweeteners is associated with a 62% faster cognitive decline has been contradicted by subsequent corrections and reviews, which found errors in the original study's data. The association with 1.6 extra years of brain aging remains supported, but the core premise of the association is now in doubt. Therefore, the claim is partially contradicted.
Partially contradicted v2 · qwen-max · 2026-11-08 14:46
The claim is partially contradicted because while the study initially reported an association between the consumption of 191 mg/day of low- and no-calorie sweeteners and a 62% faster cognitive decline, subsequent reviews and corrections have cast doubt on the statistical significance of the findings. The association does not imply a direct cause-and-effect relationship, and the study's complete-case analysis showed no such association. Additional independent sources are needed to resolve the contradiction.
Partially contradicted v1 · qwen-max · 2026-11-08 14:43
The claim is partially contradicted because while the study found an association between the consumption of low- and no-calorie sweeteners and cognitive decline, the reliability of the study's findings is in question due to significant errors. The association does not imply a direct cause-and-effect relationship. Additional independent sources are needed to resolve the contradiction.

Evidence on this claim

https://www.sciencedaily.com/releases/2026/10/261028234238.htmcited in v5
Those in the lowest consumption group averaged 20 milligrams per day (mg/day), while those in the highest group averaged 191 mg/day. For comparison, 191 milligrams of aspartame is roughly the amount found in one can of diet soda.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.sciencedaily.com/releases/2026/10/261028234238.htm
Importantly, the findings show an association rather than a proven cause-and-effect relationship. Researchers cannot conclude from this study that consuming these sweeteners directly damages the brain or causes memory loss.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.aan.com/PressRoom/Home/PressRelease/5281
Corrections were made on July 30, 2026
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.aan.com/PressRoom/Home/PressRelease/5281
While the study showed a link between the use of some artificial sweeteners and cognitive decline, it did not prove that they were a cause.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.aan.com/PressRoom/Home/PressRelease/5281cited in v5
with a decline that was 63% faster. This is the equivalent of about 1.6 years of aging. Those in the middle group had a decline that was 36% faster than the lowest group, equivalent to about 1.3 years of aging.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.neurology.org/doi/10.1212/WNL.0000000000214023
A Food Frequency Questionnaire was used to calculate combined and individual consumption of 7 LNCSs
by WeirdTrue · 2026-11-08 · source could not be fetched
https://www.neurology.org/doi/10.1212/WNL.0000000000214393
there were errors in the CIs for some estimates presented in the article and the supplementary eTables
by WeirdTrue · 2026-11-08 · source could not be fetched
https://www.neurology.org/doi/10.1212/WNL.0000000000218380
In response, the entire analysis has now been rerun from the raw analytical data set using fully scripted, reproducible code.
by WeirdTrue · 2026-11-08 · source could not be fetched
https://github.com/sophieehill/sweetenerscited in v5
In other words: one of the key estimates presented in this paper likely contains a typo in its confidence interval. Rather than being evidence of an association between sweeteners and cognitive decline, as claimed in the text, it is not statistically significant.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.sciencealert.com/are-artificial-sweeteners-really-worse-for-your-brain-than-sugar-an-epidemiologist-explainscited in v5
In a series of statistical models that were relegated to the supplementary appendix, they showed that in the full case analysis there was no relationship between sweeteners and aging.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.sciencealert.com/are-artificial-sweeteners-really-worse-for-your-brain-than-sugar-an-epidemiologist-explains
The authors only had a single food frequency questionnaire to rely on, which meant that if people changed their food intakes over this period, there would be no way for them to know about it.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.acsh.org/news/2026/07/30/artificial-sweetener-study-leaves-sour-aftertaste-50243
In complete case analysis, there was no association between tertiles of LNCS
by WeirdTrue · 2026-11-08 · excerpt found in source

#2 ScienceDaily, in its summary of the study said on 2026-10-28 · source

Tagatose was the exception.

Adjudicated as: Of the seven sweeteners studied, tagatose is the one that is safe for the brain.Restated in plain terms so it can be adjudicated; the verdict below addresses this version.

Insufficient evidenceVersion 5 · Model qwen-max · verdict_v1 · 2026-11-09 07:10

The ScienceDaily summary did not state that tagatose is safe for the brain. It only mentioned that there was no association between tagatose and cognitive decline, which does not establish safety. The interested-party cap applies because the only source supporting the claim is ScienceDaily's own article. Therefore, the level is insufficient.

Citations

www.sciencedaily.com: “Those in the lowest consumption group averaged 20 milligrams per day (mg/day), while those in the highest group averaged 191 mg/day. For comparison, 191 milligrams of aspartame is roughly the amount found in one can of diet soda.”
www.sciencedaily.com: “Tagatose was the exception. Researchers found no association between its consumption and cognitive decline in this study. However, that finding does not establish that tagatose protects the brain or is necessarily safer than other sweeteners.”

What would change this

Independent sources confirming the exact wording of the ScienceDaily summary.

Earlier versions (4)

Insufficient evidence v4 · qwen-max · 2026-11-08 15:12
ScienceDaily did publish a summary of the study, but it did not state that tagatose is safe for the brain. The summary only mentioned that there was no association between tagatose and cognitive decline. This does not establish safety. The interested-party cap applies because the only source supporting the claim is ScienceDaily's own article. Therefore, the level is insufficient. The evidence shows that the study did not make any claims about the safety of tagatose, only that there was no association with cognitive decline.
Insufficient evidence v3 · qwen-max · 2026-11-08 14:48
ScienceDaily did publish a summary of the study, and it mentioned that tagatose had no association with cognitive decline. However, the summary did not explicitly state that tagatose is 'safe for the brain.' The interested-party cap applies, as the only source supporting the claim is ScienceDaily's own article, which is based on the study's findings. Therefore, the level is insufficient.
Insufficient evidence v2 · qwen-max · 2026-11-08 14:46
ScienceDaily did publish a summary of the study, and it mentioned that tagatose had no association with cognitive decline. However, the summary did not explicitly state that tagatose is 'safe for the brain.' The interested-party cap applies, as the only source supporting the claim is ScienceDaily's own article, which is based on the study's findings. Therefore, the level is insufficient.
Insufficient evidence v1 · qwen-max · 2026-11-08 14:43
The evidence shows that ScienceDaily did publish a summary of the study, and it mentioned that tagatose had no association with cognitive decline. However, the summary did not explicitly state that tagatose is 'safe for the brain.' The interested-party cap applies, as the only source supporting the claim is ScienceDaily's own article, which is based on the study's findings. Therefore, the level is insufficient.

Evidence on this claim

https://www.sciencedaily.com/releases/2026/10/261028234238.htmcited in v5
Tagatose was the exception. Researchers found no association between its consumption and cognitive decline in this study. However, that finding does not establish that tagatose protects the brain or is necessarily safer than other sweeteners.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://github.com/sophieehill/sweetenerscited in v5
The estimates in rows 2-5 all come from eTable 9 and likely represent sign errors.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.sciencealert.com/are-artificial-sweeteners-really-worse-for-your-brain-than-sugar-an-epidemiologist-explains
Some of the sweeteners are natural substances, one is actually a very rare form of sugar, and all of them are very different molecules. It would be extremely unlikely for all of these different chemicals to have the same impact on the brain.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://www.acsh.org/news/2026/07/30/artificial-sweetener-study-leaves-sour-aftertaste-50243
Aspartame is rapidly broken down into two amino acids and methanol. Saccharin is largely excreted unchanged. Erythritol is mostly absorbed and excreted in the urine. Sorbitol and xylitol are extensively fermented in the colon. These compounds don't share metabolism.
by WeirdTrue · 2026-11-08 · excerpt found in source

#3 Lee et al., conclusions of a Diabetes Care meta-analysis said on 2022-08-01 · source

LNCSB were not associated with cardiometabolic harm in analyses that model the exposure as change or substitutions. The available evidence provides some indication that LNCSB in their intended substitution for SSB may be associated with cardiometabolic benefit, comparable with the standard of care, water.

Adjudicated as: Replacing sugar-sweetened beverages with low- and no-calorie sweetened ones is associated with lower weight and lower risk of obesity, coronary heart disease, cardiovascular mortality and total mortality.Restated in plain terms so it can be adjudicated; the verdict below addresses this version.

Well supportedVersion 4 · Model qwen-max · verdict_v1 · 2026-11-09 07:11

The claim is well supported by the evidence from the Lee et al. study, which includes a meta-analysis of 14 cohorts with 416,830 participants. The study concludes that substituting low- and no-calorie sweetened beverages for sugar-sweetened ones is associated with lower weight and lower risk of various health issues, as stated in the claim. The evidence is direct and comes from independent sources, providing strong support for the claim.

Citations

pmc.ncbi.nlm.nih.gov: “A total of 14 cohorts (416,830 participants) met the eligibility criteria.”
pmc.ncbi.nlm.nih.gov: “The intended substitution of LNCSB for SSB was associated with lower weight and lower risk of incident OB, CHD, CVD morality, and total mortality without an adverse association with any other cardiometabolic outcomes including T2D.”

Earlier versions (3)

Well supported v3 · qwen-max · 2026-11-08 15:12
The claim is well supported by the evidence from the Lee et al. study, which includes a meta-analysis of 14 cohorts with 416,830 participants. The study concludes that substituting low- and no-calorie sweetened beverages for sugar-sweetened ones is associated with lower weight and lower risk of various health issues, as stated in the claim. The evidence is direct and comes from independent sources, providing strong support for the claim. This means that the study supports the idea that switching from sugary drinks to low- or no-calorie sweetened ones can be beneficial for health.
Well supported v2 · qwen-max · 2026-11-08 14:49
The claim is well supported by the evidence from the Lee et al. study, which includes a meta-analysis of 14 cohorts with 416,830 participants. The study concludes that substituting low- and no-calorie sweetened beverages for sugar-sweetened ones is associated with lower weight and lower risk of various health issues, as stated in the claim. The evidence is direct and comes from independent sources, providing strong support for the claim.
Insufficient evidence v1 · qwen-max · 2026-11-08 14:47
The first premise is supported by evidence from the study's publication. The second premise, which is an exceptional claim, lacks independent supporting evidence. The provided evidence items do not address the specific conclusions of the Lee et al. study. Therefore, the level is capped at 'insufficient' due to the lack of independent support for the claim.

Evidence on this claim

https://www.who.int/publications/i/item/9789240073616
The guidance in this guideline is not based on toxicological assessments of the safety of individual non-sugar sweeteners and is therefore not intended to update or replace guidance on safe or maximal levels of intake established by the Joint Food and Agriculture Organization of the United Nations (FAO)/WHO Expert Committee on Food Additives (JECFA) or other authoritative bodies.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://pmc.ncbi.nlm.nih.gov/articles/PMC9346984/cited in v4
A total of 14 cohorts (416,830 participants) met the eligibility criteria.
by WeirdTrue · 2026-11-08 · excerpt found in source
https://pmc.ncbi.nlm.nih.gov/articles/PMC9346984/cited in v4
The intended substitution of LNCSB for SSB was associated with lower weight and lower risk of incident OB, CHD, CVD morality, and total mortality without an adverse association with any other cardiometabolic outcomes including T2D.
by WeirdTrue · 2026-11-08 · excerpt found in source

Evidence (19)

Evidence on the whole case

None. Evidence tied to a specific claim is listed under that claim.

Discussion (0)

About the evidence, not the topic: what is missing, whether a source holds up, how a claim should be worded so it can be adjudicated. Nothing here reaches the verdict — only sourced evidence does.

No one has raised anything about the evidence yet.

Timeline

2026-11-09 07:11 · AI verdict v4 on claim #3: Well supported (2 unchanged re-adjudications folded)
2026-11-09 07:10 · AI verdict v5 on claim #2: Insufficient evidence (4 unchanged re-adjudications folded)
2026-11-09 07:10 · AI verdict v5 on claim #1: Partially contradicted (newly cited: github.com, 4 unchanged re-adjudications folded)
2026-11-08 14:48 · WeirdTrue added evidence from pmc.ncbi.nlm.nih.gov on claim #3
2026-11-08 14:47 · AI verdict v1 on claim #3: Insufficient evidence
2026-11-08 14:45 · WeirdTrue added evidence from pmc.ncbi.nlm.nih.gov on claim #3
2026-11-08 14:45 · WeirdTrue added evidence from pmc.ncbi.nlm.nih.gov on claim #3
2026-11-08 14:45 · WeirdTrue added evidence from acsh.org on claim #1
2026-11-08 14:42 · WeirdTrue added evidence from who.int on claim #3
2026-11-08 14:42 · WeirdTrue added evidence from pubmed.ncbi.nlm.nih.gov on claim #3
2026-11-08 14:41 · WeirdTrue added evidence from pubmed.ncbi.nlm.nih.gov on claim #3
2026-11-08 14:41 · WeirdTrue added evidence from acsh.org on claim #2
2026-11-08 14:41 · WeirdTrue added evidence from acsh.org on claim #1
2026-11-08 14:41 · WeirdTrue added evidence from sciencealert.com on claim #2
2026-11-08 14:41 · WeirdTrue added evidence from sciencealert.com on claim #1
2026-11-08 14:41 · WeirdTrue added evidence from sciencealert.com on claim #1
2026-11-08 14:41 · WeirdTrue added evidence from github.com on claim #2
2026-11-08 14:41 · WeirdTrue added evidence from github.com on claim #1
2026-11-08 14:41 · WeirdTrue added evidence from neurology.org on claim #1
2026-11-08 14:41 · WeirdTrue added evidence from neurology.org on claim #1
2026-11-08 14:41 · WeirdTrue added evidence from neurology.org on claim #1
2026-11-08 14:40 · WeirdTrue added evidence from aan.com on claim #1
2026-11-08 14:40 · WeirdTrue added evidence from aan.com on claim #1
2026-11-08 14:40 · WeirdTrue added evidence from aan.com on claim #1
2026-11-08 14:40 · WeirdTrue added evidence from sciencedaily.com on claim #2
2026-11-08 14:40 · WeirdTrue added evidence from sciencedaily.com on claim #1
2026-11-08 14:40 · WeirdTrue added evidence from sciencedaily.com on claim #1
2026-11-08 14:40 · WeirdTrue added claim #3 by Lee et al., conclusions of a Diabetes Care meta-analysis
2026-11-08 14:40 · WeirdTrue added claim #2 by ScienceDaily, in its summary of the study
2026-11-08 14:40 · WeirdTrue added claim #1 by ScienceDaily, in its summary of the study
2026-11-08 14:40 · WeirdTrue opened the case