Eating about 2 ounces of peanuts a day improved brain insulin response in older adults

Most research connecting food to brain aging arrives as an observational link: people who eat more of something tend to score better on a memory test years later, and the two get connected with a line that stops short of proof. A trial published this month is a different shape: a real intervention, in a real group of older adults, testing one specific, measurable thing: how a brain region tied to memory responds to insulin.

The study

Researchers at Maastricht University Medical Center's NUTRIM institute, led by PhD candidate Lucia Kerkhof with senior author Peter Joris, ran a randomized, controlled, crossover trial (registered as NCT05724654) in 31 healthy adults aged 60 to 75. Each participant ate 60 grams of unsalted, skin-roasted peanuts a day, about 2 ounces or two servings, for 16 weeks, and also went through a control period without peanuts, with at least 8 weeks of washout between the two. Because every participant did both the peanut period and the control period, each person acts as their own comparison, which helps separate the effect of the peanuts from ordinary differences between people.

To measure brain insulin responsiveness, researchers gave participants insulin through a nasal spray and used MRI to track the resulting change in regional cerebral blood flow, a method that isolates the brain's response to insulin without the swings in blood sugar a meal or an injection would cause elsewhere in the body. The results, published in the journal Food & Function, found that after the peanut period, one cluster in the right hemisphere of the brain, covering parts of the lingual gyrus, hippocampus, parahippocampal gyrus, thalamus and posterior cingulate gyrus, areas with known roles in memory, showed a more favorable blood flow response to insulin than after the control period.

A companion paper from the same trial and the same 31 participants, published in 2025 in the journal Clinical Nutrition, reported two further findings from the same peanut period: overall cerebral blood flow increased by about 3.6%, and scores on a delayed verbal recognition memory test improved by about 5.8%, roughly 1.4 more words recalled. Other measures from that same test battery, including executive function and processing speed, did not change significantly.

The caveat that has to come first

The newer trial was funded by The Peanut Institute Foundation, an organization that promotes peanut consumption. That does not make the measurements wrong, brain imaging and a nasal insulin spray are not the kind of outcome easy to nudge with funding, but it is the same kind of disclosure this site flags whenever the group paying for a study has a direct stake in a favorable result, and it is a reason to want an independently funded replication before treating the finding as settled.

Independent of funding, the trial is small. Thirty-one participants is enough to detect a real effect if the effect is large and the measurement is precise, which MRI-based cerebral blood flow is, but it is not enough to rule out the finding being a fluke, and a single study, however well designed, is rarely the last word on anything. The result is also a single cluster within a region-by-region brain scan, not a change across the whole brain, and a surrogate measure of how the brain responds to insulin, not a measurement of dementia risk, diagnosed cognitive impairment, or any clinical outcome. Nobody in this trial was tracked long enough, or assessed for anything close to a dementia diagnosis, to say whether this translates into less cognitive decline later in life.

Why brain insulin response might matter at all

Insulin does more in the body than regulate blood sugar. Separately from the insulin resistance that drives type 2 diabetes, researchers have spent over a decade studying a brain-specific version of insulin resistance as one of several candidate contributors to age-related cognitive decline and Alzheimer's disease, concentrated in some of the same regions, including the hippocampus, that showed the effect in this trial. That is the broader context that makes a result like this interesting to the researchers who ran it. It is also exactly the gap this trial does not close: it measured a brain region's blood flow response to an insulin nasal spray, not whether that response change led to any measurable difference in disease risk.

The study's authors point to a few candidate explanations for why peanuts specifically might produce this effect: the unsaturated fat content, the fiber, and L-arginine, an amino acid the body converts into nitric oxide, which has been studied separately for a role in learning and memory. These are proposed mechanisms, not something this trial tested directly, and the paper does not claim to have proven which, if any, of the three is responsible.

How much is 60 grams, and what a serving looks like

Precise peanut content varies by product, especially once peanuts are mixed with other ingredients, so the numbers below combine well-established whole-food values with a qualitative ranking where a single confident figure does not exist:

• Peanuts, dry-roasted, unsalted, 1 oz (about 28 kernels): roughly 7 g protein, 14 g fat, 2 g fiber, around 165 calories. Two of these is close to the trial's daily dose.
• Peanuts, raw, 1 oz: similar protein and fat, slightly more fiber.
• Peanuts, oil-roasted, 1 oz: similar, with slightly more fat from the roasting oil.
• Peanuts, dry-roasted, salted, 1 oz: the same numbers as the unsalted version, plus roughly 230 mg of sodium, worth noticing if you are also watching the sodium cap.
• Peanut butter, natural, no added oil or sugar, 2 tbsp (about 32 g): close to the same peanut content as whole peanuts of the same weight, since nothing else has been added.
• Peanut butter, standard commercial, 2 tbsp: usually includes added oil, salt and sometimes sugar, so a somewhat smaller share of peanut by weight than the natural version, and this varies a great deal by brand.
• Peanut butter, reduced-fat, 2 tbsp: typically replaces some of the removed fat with added sugar or starch, a smaller peanut share again.
• Boiled peanuts (in shell, shelled yield), ½ cup: lower fat and calorie density than roasted, since boiling leaches some fat out and the kernel absorbs water.
• Defatted peanut flour, 2 tbsp: concentrated protein with most of the fat removed.
• Trail mix with peanuts, ¼ cup: peanut share depends entirely on the blend, commonly a third to half of the mix by weight.
• Thai or Indonesian-style peanut sauce, 2 tbsp: built from peanut butter diluted with liquid, so meaningfully less actual peanut per spoonful than the jar it started from.
• Peanut brittle or peanut candy, 1 oz: sugar carries most of the weight here; peanut is a minority ingredient.
• Mixed nuts containing peanuts, 1 oz: peanut share varies by brand, often the largest single nut by count since it is the cheapest to include.

Only whole, unsalted, skin-roasted peanuts were tested in this trial. None of the mixed or processed forms above were, and they were not designed to be read as equivalent.

Is there a recommended amount?

There is no RDA, AI or official recommended intake for peanuts specifically, because peanuts are a food, not an isolated nutrient. US dietary guidance groups nuts and seeds under the protein foods category and treats them as one of several interchangeable protein sources rather than setting a target amount. The 60 grams a day used in this trial is the amount this specific study tested, not a recommendation from any health authority, and no outcome from this trial establishes that a different amount would or would not produce the same result.

Too little, and too much

Too little is not a meaningful category here. Peanuts are a food, not a required nutrient, and there is no deficiency that results from eating none.

Too much has two real, well-documented angles, neither of them about the amount used in this trial specifically. The first is allergy: peanut is one of the most common foods behind severe allergic reactions, and this trial excluded anyone with a nut allergy, so its results say nothing about whether this amount would be safe for someone who has one. The second is aflatoxin, a toxin produced by mold that can grow on peanuts in storage. It is well enough documented that the FDA sets an action level of 20 parts per billion for aflatoxin in peanuts and peanut products, and US Department of Agriculture rules set an even stricter 15 parts per billion limit on domestic raw peanuts before they can be sold for human food. Neither of those numbers has anything to do with this trial; they reflect routine food-safety monitoring that applies to peanuts generally, the same kind of documented, category-level risk this site flags for rice and arsenic or cocoa and cadmium, not a number anyone can read off a photograph of a snack.

Tracking it in AIM

AIM does not track peanuts, or any single food, as one of its fourteen standard nutrients, so if you want to watch how close you get to the amount this trial tested, add "peanuts" as a custom metric in grams or ounces, with 60 g (2 oz) as a target if that is the number you want to track against. Any custom nutrient works the same way.

What a photograph cannot do is tell you anything about this study's actual outcome: brain insulin responsiveness is not something a camera, or any food log, can measure, and neither is aflatoxin contamination in a particular bag of peanuts. A log can honestly tell you how many grams of peanuts you ate this week. It cannot tell you what that did to your brain, and neither, yet, can this one trial.