How Food Choices Shape Memory, Cognition & Cravings | Dr. Scott Kanoski

Why does it matter? Because hunger two hours from now is being programmed right now — by what your hippocampus is encoding, not what's in your stomach

Dr. Scott Kanoski runs a USC lab devoted to the gut-brain-hippocampus axis — a circuit that simultaneously controls memory formation and appetite regulation, and that turns out to be shockingly vulnerable to the foods most common in the modern diet. The central, counterintuitive discovery: the organ that forms episodic memories is also the organ that interprets hunger signals. Break it with sugar and fat early in life, and the damage outlasts the diet.

• The memory of a meal — not stomach stretch or caloric content — governs how hungry you are two hours later
• A staircase ramp of acetylcholine in the hippocampus encodes every real meal; artificial sweeteners cannot trigger this circuit because it requires post-oral nutrient sensing, not just taste
• Early-life sugar at soft-drink concentrations leaves hippocampal memory deficits that persist into adulthood even after the diet is corrected, but exercise can rescue them
• Eating with familiar companions reliably increases intake via hippocampal oxytocin — a mechanism Kanoski's lab confirmed at the genetic level

Your hunger two hours after a meal isn't driven by digestion — it's controlled by what your hippocampus encoded

The stomach is not running the show two hours after you eat. The hippocampus is.

Suzanne Higgs and colleagues had participants eat a pizza lunch, then primed one group to vividly recall that meal before a subsequent cookie taste test. Those who were reminded of lunch consumed fewer cookies. The memory of recent eating, Kanoski explains, contributes directly to how much you eat at the next occasion — a mechanism operating independently of caloric content or stomach fullness.

Jeff Brunstrom's group made the architecture more explicit with a magic bowl of tomato soup fitted with a peristaltic pump that secretly added or removed soup while participants ate. Right after the meal, hunger tracked actual consumption — the wisdom of the body, as Kanoski calls it. A couple hours later, that flipped: hunger ratings were now driven by what participants thought they'd consumed, not what they had.

The hippocampus is needed to interpret interoceptive signals, not merely receive them. The most striking demonstration: densely amnesic patients with bilateral medial temporal lobe damage including the hippocampus will eat three full meals in a row when presented sequentially, because they cannot encode having just eaten. Rate their hunger after the third meal, and they score a 7 out of 10. The gut signal arrives. There is simply no hippocampus to translate it into behavior.

Rodent lesion studies confirm the same: animals without a hippocampus cannot use internal hunger or satiety state as a predictive cue, even when that state is extreme.

Eating while distracted — phone, laptop, ambient stress — likely degrades hippocampal encoding of the meal. That degraded memory is not just a missed experience. It may genuinely increase intake at the next meal, independent of calories consumed.

Real food triggers a staircase of acetylcholine in the hippocampus that stevia cannot replicate

Acetylcholine is the molecular currency running from gut to hippocampus every time you eat — and it does not activate for sweet taste alone.

Using fiber photometry to track acetylcholine release in freely behaving animals, Kanoski's lab recorded what happens during a meal. As the animal holds food, chews, and swallows, hippocampal acetylcholine ramps up. During pauses — a glance around, a brief rest — it ramps down. Across the full meal, the pattern describes a staircase: up-down, up-down, with each baseline landing slightly higher than the last. By the end, postprandial acetylcholine is elevated above where it started.

Kanoski interprets that elevated baseline as the signal that allows the animal to consolidate what just happened: what was eaten, where, how to return. The Portland croissant you still remember years later is the staircase doing exactly what it evolved to do.

The disqualifying result for artificial sweeteners: when animals consume stevia instead of caloric food, the acetylcholine staircase does not appear. Taste alone — even convincing, pleasurable sweet taste — fails to trigger the circuit. The full pattern requires what Kanoski calls post-oral nutrient sensing: actual caloric consequence downstream from the mouth. His lab has not yet infused nutrients without taste, but the working hypothesis is that both taste and caloric consequence are needed to drive the vagal-acetylcholine-hippocampus axis completely.

The gut-to-hippocampus encoding circuit is not triggered by sweetness. It is triggered by nutrients. There is a hard biological ceiling on how far non-caloric sweeteners can substitute for real food in sustaining hippocampal function, and Kanoski's data locate it clearly.

Soft-drink levels of sugar during childhood cause hippocampal memory deficits that outlast diet correction — but exercise can reverse them

Juvenile diet damage to the hippocampus does not disappear when the diet improves. That is Kanoski's central developmental finding, and it should recalibrate how early-life food choices are weighted.

Young rats given ad libitum access to a sugar solution — matching the caloric percentage of a standard soft drink, otherwise eating healthy chow — developed hippocampal-dependent memory deficits as adults, even after the sugar was removed. A full junk food model produced the same persistence: animals fed chocolate peanut butter cups, ruffled potato chips, and a high-fat, high-sugar rodent diet showed memory impairments that lasted even after switching to healthy food.

The gut microbiome appears to be part of the mechanism. Sugar-exposed animals show dramatic shifts in bacterial populations, with specific changes correlating with the memory deficits — pointing toward a gut-brain pathway that stays dysregulated even after the dietary insult stops.

Exercise is the notable exception. There is evidence in rats that physical activity reverses hippocampal memory deficits from early junk food exposure, and part of that rescue happens through direct changes in the brain — not only through improved metabolic function.

Diet correction alone is not enough once the developmental window has closed. For children and adolescents, early food quality and sustained physical activity are the two levers most likely to protect the cognitive architecture that will regulate memory and hunger for the rest of their lives.

A single gut bacterium elevated twofold by childhood sugar is causally sufficient to produce hippocampal memory impairments

One gut bacterium, elevated twofold by early sugar exposure, is enough to cause hippocampal memory deficits in animals that never tasted sugar. That is Kanoski's causal demonstration — not a correlation.

Parabacteroides was elevated roughly twofold in sugar-exposed rats and negatively correlated with memory performance. Kanoski's lab then used antibiotic treatment to partially deplete the gut bacterial population and enriched Parabacteroides into that depleted environment — in animals with no prior sugar exposure. Memory impairments appeared anyway. The enrichment used was approximately eightfold above what sugar naturally produced, and the deficits were correspondingly more severe, but the principle holds: a single bacterial shift is causally sufficient.

From the heat-treated diet model — currently a preprint — the complementary picture: a lactic acid-producing bacterium was completely eliminated by high-heat cooking of otherwise healthy food. Its abundance was positively correlated with memory performance. Reintroducing it to impaired animals reversed the deficits.

The cooking mechanism matters independently of macronutrients. The Maillard reaction — the browning chemistry that makes roasted food taste better — produces advanced glycation endproducts, or AGEs, as a byproduct. An AGE-enriched diet during development caused memory impairments in Kanoski's model, with the gut bacteria as intermediary: high-heat preparation eliminated a protective population, and the cognitive consequences followed. Cooking method is not a flavor variable. It is a microbiome variable with downstream effects on cognition.

The hippocampus is loaded with appetite hormone receptors because foraging decisions require memory, not reflexes

Foraging is not a reflex. That single reframe is the foundation of Kanoski's case for why the hippocampus belongs at the center of feeding science alongside the hypothalamus.

The hippocampus expresses dense receptor populations for insulin, ghrelin, leptin, and GLP-1 — the same hormonal suite that fills the hypothalamus, the brain's classic feeding center. Kanoski argues this overlap is not incidental. Deciding where to forage, what to eat, whether to compete, how to return — these are cognitively demanding behaviors that simple homeostatic switches cannot execute alone.

The gut communicates with the hippocampus through a three-stage pathway: vagus nerve to brainstem, brainstem to medial septum, cholinergic medial septum neurons projecting into the hippocampus. That is the fast neural channel. Circulating hormones — insulin, leptin, ghrelin — provide the slower humoral update on longer-term energy reserves.

The self-reinforcing damage loop follows directly: a high-fat, high-sugar diet impairs hippocampal function, which degrades interoceptive awareness of fullness, which removes the very corrective signal that would otherwise limit intake. Kanoski is explicit: interoceptive awareness of hunger versus fullness is plastic and susceptible to what we eat.

Protecting the hippocampus is not a separate goal from protecting appetite regulation. They are the same system.

Sever the sensory vagus nerve and native GLP-1 loses its power to reduce food intake — the primary satiety mechanism is peripheral, not central

Cut the sensory vagus nerve as it rises from the GI tract and inject native short-acting GLP-1. Food intake does not fall. The satiety signal disappears entirely.

This is Wolf Gang Longhans's work, cited by Kanoski as the cleanest evidence for natural GLP-1's primary mechanism. Knock down GLP-1 receptor expression on the vagus nerve using RNA interference, and the animals eat larger meals. The native molecule acts through a paracrine mechanism: GLP-1 released from intestinal L cells acts on receptors on the nearby vagus nerve terminal, which sends the satiation signal to the brainstem.

Whether the newer-generation drug agents — semaglutide, tirzepatide — enter the brain at all remains controversial. Fluorescent tagging studies show accumulation in circumventricular organs, which lack blood-brain barrier protection. Radiolabeling work from Elizabeth Rhea and Bill Banks at the University of Washington suggests some earlier-generation GLP-1 drugs cross readily; newer long-acting agents are less clear.

GLP-1 receptors are expressed densely throughout the brain — including hippocampus, VTA, striatum, and nucleus accumbens — which makes effects on memory, food noise, and reward circuitry biologically plausible. But plausible is not confirmed. The people taking newer GLP-1 agents specifically for cognitive benefit are operating ahead of the evidence.

The same oxytocin that suppresses meal size in the brainstem drives overeating in the hippocampus when you eat with friends

Oxytocin in the brainstem makes meals smaller. Oxytocin in the hippocampus makes meals bigger when you are with familiar people. Same molecule, opposite behavioral outcomes — brain region determines direction.

The brainstem mechanism is well-established: Harvey Grill's work identified oxytocin signaling there as primarily reducing meal size by increasing satiation. Historically, Kanoski notes, oxytocin's role in food intake was considered unidirectional — appetite suppressing. Data from his lab and others are now complicating that picture.

A well-documented phenomenon called social facilitation of eating holds that people eat more in groups, and more still when those groups are familiar. Kanoski's lab showed this requires hippocampal oxytocin signaling. Genetically knocking down oxytocin receptor expression in the hippocampus eliminated the effect entirely: animals ate the same amount regardless of whether they were alone or surrounded by familiar companions.

The mechanism by which social familiarity amplifies food intake runs through a known receptor in a mapped brain region. If caloric intake matters to you, dining alone or with strangers is not circumstance — it is a real biological lever.

The vagus nerve may be the next therapeutic target for dementia — not just a conduit for satiety

Kanoski's lab is currently implanting vagal nerve stimulation devices in a transgenic Alzheimer's rat model. The devices fire only when the animals eat, because natural stomach distension triggers the mechanical sensor. Early results already show memory rescue in diet-impaired animals. Whether that extends to reducing amyloid or tau pathology in Alzheimer's remains unanswered.

But the framing it implies is genuinely new: the gut is not background noise for the brain. It is upstream of it, continually feeding the hippocampus signals that either sustain or erode its function across a lifetime.

Every meal is a dose.

---

Topics: nutrition, memory, hippocampus, gut-brain axis, GLP-1, microbiome, acetylcholine, oxytocin, dopamine, food cravings, Alzheimer's, developmental nutrition, vagus nerve, interoception, advanced glycation end products, Maillard reaction, ultra-processed food