“Eat well—it’s good for your brain” sounds like sensible advice. But scientifically, it hides a difficult question: What, exactly, connects food with thinking?
Diet may influence the brain. But the relationship can also run in other directions. Young people with stronger cognitive skills may make different food choices. Family income, sleep, puberty, body composition, and many other factors may affect both diet and cognition. Even when diet and thinking are related, identifying the brain structures involved is far from simple.
In a recent study, my colleagues and I examined these relationships in 3,308 adolescents from the Adolescent Brain Cognitive Development (ABCD) Study. We combined information about their diets, performance on cognitive tests, and structural MRI scans.
Our central question was not whether one particular “brain food” makes adolescents smarter. It was whether broader patterns of eating are associated with cognition—and whether part of that association is also reflected in the structure of the developing brain.
Looking Beyond Individual Foods
People do not consume nutrients in isolation. They eat combinations of fruits, grains, dairy products, legumes, sugary drinks, and many other foods. These choices tend to cluster together.
We therefore used a statistical method to identify five broad dietary patterns from 10 food categories. Together, these patterns captured 77 percent of the dietary variation in our sample.
We then created cognition-aligned dietary scores: combinations of those eating patterns that were most closely associated with either crystallized or fluid cognition.
Crystallized cognition refers roughly to knowledge and skills acquired through learning and experience, including vocabulary. Fluid cognition is the ability to reason, solve new problems, and adapt to unfamiliar situations.
To avoid tailoring each adolescent’s score to that same adolescent’s cognitive results, we used a procedure called family-level cross-fitting. In simple terms, each participant’s dietary score was constructed using relationships estimated from other families – not from that participant or their siblings.
The cognition-aligned dietary patterns were associated with both forms of cognition. A one-standard-deviation increase in the dietary score corresponded to approximately a 0.10-standard-deviation increase in crystallized cognition and a 0.06-standard-deviation increase in fluid cognition.
These are modest associations, not dramatic effects. But they remained after accounting for age, sex, puberty, socioeconomic conditions, income relative to family size, sleep, and body mass index. The results were also similar when ethnicity and waist circumference were included in sensitivity analyses.
A Small Bridge Deep in the Brain
The most consistent brain finding involved the fornix.
The fornix is a relatively small bundle of white-matter fibers that connects the hippocampus – central to memory – with other regions involved in learning, motivation, and cognitive control. You might think of it as part of the brain’s communication infrastructure for memory.
We measured the organization of white matter using fractional anisotropy, or FA. Higher FA can reflect more coherent organization of fibers, although it should not automatically be interpreted as a “better brain.”
FA in the fornix was related to cognitive performance. More strikingly, it statistically accounted for a small portion of the association between dietary patterns and both crystallized and fluid cognition. The left and right fornix each accounted for about 6 percent of the diet–crystallized-cognition relationship and approximately 8 to 10 percent of the diet–fluid-cognition relationship.
Other brain regions were associated with cognition, especially areas involved in memory and language. However, after correcting for the large number of comparisons, no gray-matter region reliably accounted for the relationship between diet and cognition. The fornix finding was the one that proved most consistent.
What This Study Does—and Does Not—Show
These results do not demonstrate that changing an adolescent’s diet will alter the fornix or improve intelligence. Diet, cognition, and brain structure were measured during the same general period. We therefore cannot establish which came first – or whether an unmeasured influence contributed to all three.
Nor did we discover a menu guaranteed to produce better grades. Our dietary scores were derived from this particular sample and need to be tested in other populations.
What the study does provide is a more detailed map of an intriguing relationship. Diet and cognition were linked even after accounting for body composition and socioeconomic and behavioral factors. Moreover, that relationship was not distributed randomly throughout the brain: it converged most consistently on a memory-related white-matter pathway.
Adolescence is a period of continuing brain remodeling. White-matter pathways are still developing, cognitive demands are growing, and young people are gaining greater control over what they eat. Nutrition may be one piece of this developmental puzzle – but it operates alongside sleep, stress, education, family resources, and many other influences.
The practical message is therefore both promising and restrained. Healthy eating may matter for more than physical health, and the fornix offers a plausible neural clue about how diet and cognition travel together. But a clue is not yet a causal explanation. The next step is to follow adolescents over time and test whether changes in diet precede changes in brain structure and cognitive performance.
For now, we can say that the teenage brain and the teenage diet appear to be connected—and that a small white-matter bridge involved in memory may be part of the story.