Oral nicotinamide riboside has not demonstrated improved cognition in these three randomized trials of older adults with cognitive concerns. Blood NAD⁺ rose in studies that measured it, and some exploratory brain or blood biomarkers changed. The main cognitive comparisons did not show a significant benefit. That is the answer to retain before interpreting a headline about brain blood flow or an Alzheimer-related blood test. [1][2][3]
“Brain fog” is a symptom description, not one diagnosis. Mild cognitive impairment, subjective cognitive decline and cognitive problems following an infection are different populations. This article examines the older-adult cognitive trials; the long-COVID evidence review addresses a separate clinical setting.
What the randomized comparisons found
| Study and participants | Design and outcome | Finding | Main limit |
|---|---|---|---|
| Orr et al.; 20 adults with MCI [1] | Ten-week randomized placebo-controlled pilot; cognition and safety | Blood NAD⁺ rose 2.6-fold; cognition did not improve. | Small sample; brain-flow signals did not survive multiple-comparison correction. |
| Wu et al.; 46 randomized, 37 completed [2] | Crossover trial; eight-week NR phases; RBANS primary outcome | No significant cognitive benefit. pTau217 changed as a secondary marker. | Short study; biomarker findings do not establish disease modification. |
| Martens et al.; 42 completers with amnestic MCI [3] | Twelve-week randomized pilot; cognition primary outcome | Blood NAD⁺ doubled; cognition and total brain blood flow did not improve. | Completer analysis; exploratory regional findings need replication. |
The first pilot: a strong blood response, no cognitive change
The Orr study followed 20 older adults with mild cognitive impairment over ten weeks. Participants received NR or placebo, and the NR arm reached a 2.6-fold increase in blood NAD⁺. The Montreal Cognitive Assessment and other cognitive and psychological measures remained stable. Adverse-event reporting did not differ between groups. [1]
Brain blood flow decreased in some regions in the NR group. The authors noted that the statistical significance would not have survived correction for multiple comparisons. When researchers examine many regions or biomarkers, some apparently positive findings can arise by chance. Such observations can help design a later study, but they are weaker than a prespecified, confirmed result.
The paper also explored DNA methylation and epigenetic-age measures. A shift in those laboratory calculations does not establish that memory improved or that dementia was prevented. This is the same distinction discussed in our guide to judging whether NAD is working: measure the outcome you actually care about.
A pTau217 finding needs its cognitive result beside it
Wu and colleagues randomized 46 adults with subjective cognitive decline or mild cognitive impairment; 37 completed the crossover trial. Each active-treatment phase lasted eight weeks. The primary cognitive measure was the Repeatable Battery for the Assessment of Neuropsychological Status, or RBANS. It did not improve significantly compared with placebo. [2]
One secondary blood marker, phosphorylated tau 217, fell by 7% with NR and rose by 18% with placebo in the between-group comparison (p = 0.02). Other measured plasma biomarkers did not show significant differences. The pTau217 finding is a reason to investigate the biology further, not evidence that NR treats Alzheimer’s disease. Participants were selected for cognitive concerns, and a marker change cannot substitute for demonstrated preservation of everyday function.
Results also differed depending on the analysis: step counts improved in the within-person comparison, but not significantly between the randomized groups. Quoting only the favorable analysis would conceal that distinction. The paper disclosed commercial relationships for some investigators; disclosure helps readers assess the work, but the methods and complete results remain the basis for interpretation. [2]
The 2026 pilot did not reverse that pattern
A twelve-week phase II pilot reported 42 completers with amnestic mild cognitive impairment, a condition in which memory impairment is prominent. NR doubled blood NAD⁺ but did not significantly improve cognitive function, total cerebral blood flow or blood pressure. Exploratory regional blood-flow findings, especially in the hippocampus, warrant follow-up. [3]
This trial analyzed participants who completed the protocol. Its authors identified the need for future intention-to-treat analyses and biomarker-based enrollment. Those details matter because people who complete a study can differ from people who leave it, and similar memory symptoms may arise from different underlying disease processes.
What these studies do and do not tell a reader
The repeated null cognitive findings argue against presenting NR as an established memory supplement. They do not prove that every dose, duration or subgroup will be ineffective. Trials were small and relatively short, so a future adequately powered study could answer a different question. That possibility is a research rationale, not a present-day benefit claim.
They also studied oral NR, not commercial NAD⁺ injections. Our NR explainer describes the molecule, while the broader evidence guide separates blood NAD⁺ responses from outcomes people can notice. Persistent or changing memory symptoms deserve assessment of their cause rather than an assumption that NAD⁺ is low.