Ultra-Processed Food Is Quietly Taxing Your Attention. The Mediterranean Salad Next to It Won't Pay the Bill.

The researchers tried to wash the effect out. It would not wash out.

In a new cross-sectional study of 2,192 Australian adults aged 40 to 70, every 10 percent rise in ultra-processed food intake came with a small but measurable drop in attention on standardized cognitive tests, and a meaningful jump on a modifiable-dementia-risk score. Then the team added a Mediterranean-diet adjustment, the move that usually buries this kind of signal, and the attention association held. The risk-score association held too. The bag of chips on the desk is an attention bill, and the salmon-and-olive-oil plate next to it doesn’t pay it down.

That is the line I keep coming back to, because it inverts the deal a lot of us think we have struck with our diets. Eat the Mediterranean things. Earn back the convenience things. The data say the ledger isn’t balancing the way we assumed.

Dr. Barbara Cardoso, the senior author at Monash University’s Victorian Heart Institute, put the dose in a form anyone can picture: a 10 percent increase in ultra-processed food is roughly a standard bag of chips added to your daily intake. That’s the unit attached to the attention drop her team reported.

And this isn’t some outlier cohort of junk-food enthusiasts. These were middle-aged Australians enrolled in the Healthy Brain Project, three quarters of them women, most with at least 15 years of education. Their average pull from ultra-processed foods was 40.6 percent of daily calories. That’s the baseline in an educated, middle-aged Australian cohort, which is sobering enough on its own. Soft drinks, packaged snacks, ready meals, breakfast bars, the flavored yogurts marketed as virtuous. The NOVA classification used in the study lumps them all into Category 4, the ultra-processed bucket, and the literature has been building on this bucket for a decade. The Lancet Healthy Longevity ran a direct call last December for governments to start treating ultra-processed foods as a public-health category in their own right, not just a proxy for sugar or salt.

Here is where my biology brain woke up. Attention and processing speed slid with intake. In this battery, the memory composite did not show the same association.

The team used the Cogstate Brief Battery, four short subtests for processing speed, visual attention, visual memory, and working memory. The attention composite slid by a β of about negative 0.21 between the top and bottom quintiles of ultra-processed intake, with a clean trend across the middle quintiles. The memory composite didn’t budge in the same way. That split is the interesting clue, because attention and processing speed are the kind of cognition that lean on the brain’s cardiometabolic plumbing, while episodic memory is hippocampal and tends to crash later and more abruptly. A diet pattern that hits attention first is, plausibly, running through a cardiometabolic route. I want to be careful here: the Cardoso paper itself doesn’t dissect the white-matter biology, so treat that pathway as the working hypothesis, not the finding.

The mechanism end of that hypothesis is where a small but elegant 2016 Food & Function paper earns its keep. Ninety-eight ready-to-eat foods, compared head to head: the ultra-processed ones produced sharper blood-glucose spikes and weaker satiety than minimally processed foods of the same calorie load. Do that meal after meal, year after year, and you are building the kind of metabolic profile that, in larger cohorts, shows up as worse attention long before it touches memory.

What about the dementia score? The team used the CAIDE risk score, a validated tool that bundles age, education, sex, cholesterol, blood pressure, activity level, and BMI into a single number. They built a “modifiable” version that strips out what you can’t change. Every 10 percent rise in ultra-processed intake bumped the modifiable score by 0.24 points, p less than 0.001, and that one held its significance after the Mediterranean-diet adjustment. In plain language: ultra-processing tracks with a cardiometabolic risk profile that adding kale on the side does not cancel out. The food-processing axis is doing something on its own, separate from how Mediterranean your plate looks otherwise.

I came into this paper expecting the usual story, an observational signal that mostly tracks calories and sugar. That’s not what this study found. Mediterranean adherence is essentially a quality knob, and turning the quality knob did not rescue attention. That moves the question from “eat fewer chips” to “the processing itself is contributing something my macros don’t capture,” which is a much more uncomfortable framing for the industry that designs these foods. Big Food has spent years arguing that NOVA is too crude, that “ultra-processed” sweeps in fortified breakfast cereals alongside soda, and that nutrients are what matter while processing is incidental. This dataset makes that defense harder to sustain.

The limits are worth naming honestly, in voice, not as a disclaimer dump. Causality first: it’s cross-sectional, so the arrow could in principle run the other way, with already-distracted people drifting toward convenient food. Measurement next: the food-frequency questionnaire was the EPIC-Norfolk instrument, validated for British eating patterns and not formally validated for Australians, and the cohort skews female, white, and educated. And the effect sizes are small in absolute terms, a Cohen’s f² of 0.03 for the attention finding, which is what you’d expect for one diet variable inside a 56-year-old’s cognition. None of that erases the signal. It frames where it goes next: a longitudinal follow-up from this cohort, and ideally a randomized feeding study that can put a causal hook into the processing axis specifically.

Here’s what I’m taking from it, plainly. If you eat well by the usual scorecards and still pull 40 percent of your calories from packaged products, the data now say your attention may be paying a tax your scorecard isn’t registering. The cleanest move isn’t a new supplement and it isn’t a new diet name. It’s to walk through a week of your meals and count how much of what you ate came out of a factory wrapper. For a lot of us that number will be larger than we want to admit. The bag of chips on the desk is the bill. The salad next to it isn’t covering it.

Sources

1. Cardoso et al., Alzheimer’s & Dementia: DADM, 2026 – full text on PMC
2. Cardoso et al., journal landing page (Wiley)
3. Monash University press release via EurekAlert (23 April 2026)
4. Fardet, Food & Function (2016) – minimally vs ultra-processed foods, satiety and glycemic response
5. The Lancet Healthy Longevity, “Addressing the issue of ultra-processed foods” (December 2025)