How Diet Shapes Memory and Cognition: Dr. Kanoski's Neuroscience Discoveries

How High Sugar and Junk Food Can Permanently Damage Memory

In an episode of the Huberman Lab podcast, Dr. Scott Kanoski, a professor at the University of Southern California, shared an alarming finding: consuming large amounts of sugar or junk food during adolescence may lead to hippocampus-dependent memory deficits — and this damage persists even after the intake stops.

The first model in Kanoski's lab was the "high-sugar model" — animals had free access to a sugar solution comparable in sugar content to soft drinks, while the rest of their diet remained healthy. The results showed that these animals exhibited hippocampus-dependent memory deficits later in life. Even after they stopped consuming sugar, the damage continued into late life. In another "junk food model," animals consumed high-fat, high-sugar foods such as chocolate peanut butter cups and wavy potato chips, and similarly showed memory impairment that was not fully reversed after switching to a healthy diet.

Kanoski pointed out that the hippocampus is like a "canary in the coal mine" — it is far more sensitive to dietary damage than other brain regions. In maze experiments, control animals could remember food locations, while animals in the diet-manipulated group performed "essentially at chance levels," showing no evidence of memory.

The Gut-Brain Axis: The Key Roles of the Vagus Nerve and Acetylcholine

Kanoski's research revealed a three-stage pathway from the gut to the hippocampus: gut signals travel via the vagus nerve to the brainstem, then project from the brainstem to the medial septum, and finally reach the hippocampus through cholinergic neurons.

Using fiber photometry recording techniques, his team observed an elegant pattern: when animals eat, acetylcholine levels in the hippocampus show a sawtooth-like fluctuation — rising during chewing and swallowing, falling during pauses and looking around. More critically, baseline acetylcholine levels after eating are higher than before eating, forming a "staircase" rise. Kanoski believes this post-meal elevation of acetylcholine is the neural basis for animals remembering "what they ate and where" — from an evolutionary perspective, this is crucial for survival.

One detail: sweetness alone is not sufficient to trigger this response. When animals consumed stevia (a non-nutritive sweetener), the acetylcholine release pattern in the hippocampus did not occur. Kanoski speculates that a combination of taste and post-meal nutrient sensing is required to drive this signal.

The vagus nerve is the tenth cranial nerve, extending from the brainstem to the abdominal cavity, innervating the heart, lungs, gastrointestinal tract, and other visceral organs. About 80% of its fibers are afferent fibers, meaning they transmit sensory information from the gut to the brain in a one-way direction, rather than the traditionally assumed descending control from brain to gut. This makes the vagus nerve the primary channel through which the gut "reports" in real time to the brain about feeding status, microbial metabolites, and inflammatory signals. The three-stage pathway in Kanoski's research — gut → vagus nerve → nucleus tractus solitarius in the brainstem → medial septum → hippocampus — relies precisely on this afferent pathway. Cutting or stimulating the vagus nerve can respectively block or enhance the post-meal acetylcholine response in the hippocampus, providing direct anatomical evidence for the "gut-brain axis."

Acetylcholine is an important excitatory neurotransmitter in the central nervous system, released in the hippocampus by cholinergic neurons projecting from the medial septum. It is closely associated with the encoding of learning and memory: when acetylcholine levels rise, neurons respond more strongly to input signals, synaptic plasticity (such as long-term potentiation, LTP) is more likely to occur, thereby promoting memory formation. The post-meal "staircase" rise in acetylcholine observed by Kanoski means that the act of eating itself creates a neurochemical window in the hippocampus conducive to memory encoding. This also explains why blocking acetylcholine receptors (such as with scopolamine) causes anterograde amnesia.

Artificial Sweeteners: Three Camps of Controversy and Mouse Evidence

Regarding artificial sweeteners, Kanoski admits he does not belong to any camp, because "there is currently no decisive evidence supporting any of the three camps."

He distinguished three camps: one that believes artificial sweeteners are harmful to health, one that believes they are completely inert at reasonable doses, and one that is skeptical based on an "unnatural" intuition. In clinical randomized trials, there is no reliable trend in the effects of artificial sweeteners on body weight — some studies show benefits, others the opposite.

Kanoski's lab published a study in mice from adolescence to adulthood, giving animals three sweeteners (stevia, acesulfame potassium, saccharin) at doses within the FDA acceptable daily intake. When the animals reached adulthood (after sweetener intake had stopped), researchers found memory impairment and changes in glucose regulation — including a larger insulin peak. But Kanoski emphasized that this is a preclinical animal model, the effects are not "bombshell level," and human studies are still insufficient. He himself admits: "I drink Diet Coke regularly. It tastes good."

How Memory in Turn Affects Eating

Kanoski's research is not limited to the effects of diet on memory, but also includes how memory shapes eating behavior in reverse.

In one study, participants were divided into two groups several hours after eating a pizza lunch: one group was guided to recall the contents of their lunch, while the other was not. Subsequently, in a cookie taste test, participants who were guided to recall their lunch ate fewer cookies. This suggests that memory of recent eating suppresses subsequent food intake.

Even more revealing is Brunstrom et al.'s "magic soup bowl" experiment: by using a hidden pump to increase or decrease the amount of soup without participants' knowledge, researchers found that immediate post-meal hunger ratings depended on actual intake (the body's wisdom), but hunger several hours later was more driven by "how much one believed they had eaten."

Kanoski also mentioned that patients with hippocampal damage, even after eating three consecutive meals, still report a fullness rating of 7/10 — they feel the bodily signals but cannot interpret their meaning. This is a form of "interoceptive agnosia."

Interoception refers to the body's ability to perceive internal physiological states (such as hunger, fullness, heartbeat, and visceral sensations). The insula is the primary cortical center for interoception, receiving visceral signals from the vagus nerve and spinal cord and integrating them into a "sense of bodily state." Interoceptive agnosia refers to an individual's inability to accurately identify or interpret these bodily signals — for example, patients with hippocampal damage can feel gastric fullness (the bodily signal is present) but cannot interpret it as the meaning of "being full." This agnosia is not only seen in brain-injured patients but is also closely associated with eating disorders (such as binge eating disorder and anorexia nervosa). Kanoski's findings suggest that the hippocampus may be a key node connecting interoceptive signals with eating memories: it both receives interoceptive input and stores contextual memories of "what was eaten last time and how much," thereby calling upon this information in subsequent eating decisions.

Brunstrom's "magic soup bowl" experiment is a classic paradigm in the psychology of eating. In the experiment, participants used a soup bowl connected to a hidden pump that increased or decreased the amount of soup without their knowledge. The results showed that immediate fullness was primarily determined by actual intake (body signals dominate), but delayed fullness (several hours later) was more driven by how much participants "thought" they had eaten (memory and belief dominate). This dissociation indicates that eating regulation operates on two timescales: short-term reliance on physiological feedback, and long-term reliance on cognitive memory. It also explains why eating while watching TV tends to lead to overeating — distracted attention weakens the encoding of eating memories, leading to reduced subsequent intake suppression.

Ultra-Processed Foods, the Sugar Industry Lobby, and Reshaping Eating Behavior

When the discussion turned to the food environment, the topic touched on the reality that ultra-processed foods account for over 60% of calorie intake in the United States, as well as the sugar industry's interference with scientific research.

Kanoski revealed that early in his career, after his team published early research on sugar intake, the sugar industry issued public statements attempting to downplay the study. Huberman also mentioned that a representative from the American Heart Association, when asked whether "soda is healthy," gave a "frustratingly evasive answer."

But Kanoski is optimistic about the plasticity of dietary preferences. He pointed out that after reducing sugar intake, people recalibrate their sweetness threshold — foods they once enjoyed may become "too sweet." He shared his own experience: after taking a sip of a milkshake, he found it "incredibly sweet," whereas he once enjoyed vanilla milkshakes. This plasticity of neural circuits means that a healthy diet not only protects the brain but also reshapes our cravings for food.

The Microbiome: Preliminary Exploration from Correlation to Causation

Kanoski is not a microbiologist, but curiosity drove him into this field. In an early study, the gut microbiota composition of mice that consumed sugar was "completely different" from that of controls. They found that the genus Parabacteroides was elevated about twofold in the sugar group and was negatively correlated with memory performance. In subsequent experiments, through antibiotic treatment and microbiota enrichment, they replicated memory impairment in animals that had never consumed sugar — providing causal evidence rather than mere correlation.

In another study on heat-processed foods (advanced glycation end products produced by the Maillard reaction), the genus Lacticoccus was completely eliminated after heat treatment, and it was positively correlated with memory performance. Reintroducing this bacterium was able to rescue memory impairment.

Kanoski emphasized that microbiome research is still in its "infancy": when the Human Microbiome Project concluded in 2012, there were about 7,000 related papers on PubMed; today there are over 250,000. But the question of "which bacteria are good and which are bad" still has no clear answer.

Parabacteroides and Lacticoccus (more commonly spelled Lactococcus) are two different genera in the gut microbiota. Parabacteroides belongs to the phylum Bacteroidetes and is a common Gram-negative anaerobic bacterium in the gut; some species are involved in polysaccharide metabolism and short-chain fatty acid production. Its elevated abundance was associated with decreased memory performance in Kanoski's research, but the specific mechanism remains unclear and may be related to low-grade inflammation triggered by endotoxin (LPS) release. Lacticoccus belongs to the phylum Firmicutes and is a type of lactic acid bacteria, commonly used in fermented foods and generally considered beneficial to the host. After Maillard reaction products eliminated it, memory was impaired; reintroduction rescued memory, suggesting it may influence brain function by producing neuroactive metabolites (such as gamma-aminobutyric acid or short-chain fatty acids).

Advanced glycation end products (AGEs) are irreversible compounds formed during the thermal processing of food when reducing sugars react with the amino groups of proteins or lipids in the Maillard reaction. Foods typically rich in AGEs include grilled meats, fried foods, and baked pastries. AGEs not only trigger oxidative stress and inflammation through their receptor (RAGE) but can also alter gut microbiota composition. In Kanoski's research, heat-processed foods led to the complete elimination of Lacticoccus, which is precisely achieved by AGEs through antibacterial effects or by altering the gut environment. Notably, AGEs also accumulate naturally in the body and are associated with aging and neurodegenerative diseases, so dietary AGEs may accelerate this process.

The Evolutionary Logic of Eating and Memory

During the conversation, Huberman proposed an integrative perspective: the memory system evolved first for survival, and nothing is more fundamental than maintaining nutrition and hydration. Kanoski's research supports this view — the hippocampus not only contains GLP-1 receptors but also receptors for insulin, leptin, ghrelin, and other metabolic hormones. In addition to receiving neural signals from the vagus nerve, the hippocampus also learns about the body's energy state through circulating hormones.

Kanoski pointed out that eating is never a reflexive behavior: "Animals decide whether to forage, where to forage, what to eat — these are not simple behaviors." Therefore, beyond the brainstem and hypothalamus traditionally considered to control basic homeostasis, higher cognitive structures such as the hippocampus are also deeply involved.

This also means that after long-term poor diet weakens hippocampal function through the gut-brain axis, it not only impairs memory but may further worsen eating decisions — creating a vicious cycle. The key to breaking this cycle may lie in the choices made at every meal.

GLP-1 receptors (glucagon-like peptide-1 receptors) are the targets through which GLP-1, released by enteroendocrine L cells after eating, acts. GLP-1 not only promotes insulin secretion and inhibits glucagon release but also acts on the brain through the vagus nerve and bloodstream to produce satiety. The presence of GLP-1 receptors in the hippocampus means that GLP-1 released from the gut after a meal can directly regulate hippocampal synaptic plasticity and memory function. Similarly, receptors for insulin, leptin, and ghrelin are also expressed in the hippocampus: insulin promotes glucose uptake and synapse formation, leptin enhances LTP, and ghrelin promotes synaptic density and memory formation. These metabolic hormones "inform" the hippocampus of the body's energy state, allowing memory encoding to be adjusted according to nutrient availability. From an evolutionary perspective, this ensures that animals prioritize forming food-related memories when energy is sufficient, while when hungry they focus more on foraging behavior rather than memory consolidation. Long-term high-sugar, high-fat diets can induce resistance to these hormones, thereby weakening the hippocampus's metabolic signal sensing and memory function.

One-line Summary

Adolescent high sugar or junk food can permanently damage hippocampal memory function through the gut-brain axis, with bidirectional causal links to microbiome changes.

Paragraph Summary

In the Huberman Lab podcast, Dr. Scott Kanoski of USC revealed that high sugar or junk food intake during adolescence leads to hippocampus-dependent memory deficits, with damage persisting into late life even after intake stops. The mechanism involves a three-stage pathway from the gut via the vagus nerve, brainstem, and medial septum to the hippocampus, with acetylcholine as a key transmitter for post-meal memory formation. Although evidence on artificial sweeteners is mixed, mouse studies show memory impairment and changes in glucose regulation. Memory also suppresses subsequent food intake in reverse, and patients with hippocampal damage cannot interpret satiety signals. Microbiome research has preliminarily demonstrated causal relationships between specific bacteria and memory. Kanoski is optimistic about the plasticity of dietary preferences, believing that a healthy diet can reshape cravings and break the vicious cycle.