The Brain's Fuel Gap: Ketones, Glucose and Cognitive Aging
As we age, the brain gradually loses its ability to take up glucose, its main fuel — but its ability to take up ketones, the backup fuel, stays intact. Dr Stephen Cunnane explains how his team measured this gap with PET imaging and tested whether medium-chain triglycerides could narrow it. His conclusion: ketones are no magic bullet, but pairing them with lower carbohydrate intake, movement and sleep addresses the energy side of brain aging that drug trials have largely ignored.
Overview
Dr Stephen Cunnane, a professor at the University of Sherbrooke, has spent his career on a deceptively simple question: what fuels an aging brain? The prevailing assumption was that reduced fuel uptake in Alzheimer's simply reflected dying cells — but if that were true, uptake of every fuel should fall together. Using PET imaging with both a glucose tracer and a ketone tracer in the same people, his team found something different: glucose uptake was reduced, while ketone uptake remained normal, and rose further when ketones were supplied.
He calls the shortfall the brain energy gap. Crucially, glucose uptake by the brain is driven by demand from within the brain, whereas ketone uptake is driven simply by how much circulates in the blood — which makes ketones a practical lever. Insulin is the switch between the two fuels, and insulin resistance puts the brain in double jeopardy: it blocks glucose entry while also keeping fatty acids locked in storage so ketones never get made.
The BENEFIT trial gave people with mild cognitive impairment 30 g of medium-chain triglyceride a day for six months and narrowed the gap at peak ketone levels, with matching improvements in memory, language and executive function. A separate care-home feasibility study cut carbohydrates modestly and saw glucose fall around 11% within days — while residents still got their two daily desserts. Cunnane's take-home is a combination, not a single fix.
Key quotes
5The brain is a hybrid car. So it's running on two fuels that are complementary.
We repeated what had been shown umpteen dozen times with the glucose tracer, but the ketone uptake was normal.
You can have a bowl of ice cream now and eat 150 grams of glucose. It's not going to change the availability of glucose in the brain by one molecule.
The brain ends up in double jeopardy, because the insulin resistance is blocking the uptake of glucose, but the insulin is also blocking the production of the fatty acids that will become the replacement fuel.
There's no magic in ketones. We still need to reduce insulin resistance, and ketones don't do that by themselves.
Key ideas
9The brain runs on two fuels
Glucose is the main fuel, but ketones are a complementary reserve that can become the dominant fuel during fasting or prolonged food restriction. Cunnane calls this the hybrid brain — and the reserve fuel is the part most people never think about.
PET imaging made the gap visible
Positron emission tomography with a glucose tracer showed, more than 40 years ago, a problem getting glucose into the parietal cortex as Alzheimer's develops. Cunnane's team added a ketone tracer so both fuels could be compared in the same person.
Ketone uptake stays normal
If reduced uptake simply reflected dead cells, ketone uptake should fall too. It did not — and when a ketone source was given, uptake in the affected regions actually increased, suggesting those cells are underfuelled rather than gone.
Healthy aging and MCI are not the same picture
In healthy older adults, glucose uptake sits roughly 5–6% below younger adults, mainly in the frontal cortex. In mild cognitive impairment the shortfall is larger and affects different regions — a difference of both degree and location.
The brain still takes up most of its glucose
Even in Alzheimer's, the brain takes up roughly 70–75% of the glucose it should. The good news is that the machinery is not switched off; the concerning news is that the missing quarter tracks closely with cognitive decline.
Ketones use a different, shorter route
Ketones have their own transporter and travel a short metabolic path to the mitochondria, while glucose passes through roughly ten steps that can each become a bottleneck. The enzymes that process ketones appear to stay intact.
Ketone levels in blood drive brain uptake
Brain glucose use responds to demand inside the brain, not to what you just ate. Ketone uptake, by contrast, follows plasma levels — which is exactly why a ketone source can be used to top up the brain's energy supply.
The BENEFIT trial: 30 g of MCT a day
Around 40 people with mild cognitive impairment took 30 g of medium-chain triglyceride daily, emulsified in milk and split between breakfast and supper, with brain PET imaging before and after. A later extension doubled the sample size for the cognitive outcomes.
How much of the gap actually closed
At peak blood ketone levels the gap narrowed by roughly 40%, but ketones fall back within two to three hours. Averaged over 24 hours, Cunnane estimates the real improvement at closer to 10% — an honest limit on what twice-daily dosing can do.
Practical takeaways
6- 1
Modest carb reduction works fast 45:30
In a care-home feasibility study, cutting carbohydrates about 30% brought fasting and post-meal glucose down within days in residents in their eighties — no ketogenic diet required.
- 2
You can keep the dessert 49:00
Participants insisted on two desserts a day and still lowered their glucose by around 11%. Smaller portions and better-quality choices left the plan palatable enough that nobody dropped out.
- 3
Choose a ketone source you'll actually use 51:40
MCT is the least expensive and most flexible option — in a drink, coffee, soup or salad dressing. Start low and build up over about ten days so the gut adapts.
- 4
Movement amplifies the effect 50:10
In Cunnane's imaging work, participants who exercised on the same day got more ketones into the brain for the same blood level — an unexpected result that makes activity part of the recipe.
- 5
Sleep and social life belong in the plan 53:00
Cunnane flags sleep as underrated for the brain's repair processes, and notes how loss, illness and isolation in later life compound poor eating and poor rest.
- 6
Cut the liquid and table sugar first 52:00
His simplest starting point: soft drinks, fruit juice, candy and sugar added in recipes and coffee. Reducing these is the most direct route to better insulin sensitivity.
Topics & chapters
15Double jeopardy: the opening thesis
Cunnane sets up the core problem — insulin resistance blocks glucose entry while also preventing ketone production.
Meet Dr Stephen Cunnane
An introduction to his work at the University of Sherbrooke on brain energy metabolism and cognitive decline.
What is the brain energy gap?
The hybrid car metaphor, and why the shortfall was long assumed to be a consequence of dying cells.
PET imaging explained
How tracers work, and what the glucose tracer revealed about the parietal cortex four decades ago.
Ketones as the reserve fuel
Evidence from prolonged medically supervised fasting that ketones can become the brain's major fuel.
The surprising result
Ketone uptake stays normal, and rises further when a ketone source is provided — replicated by other methods.
Aging brain versus MCI
The trajectory of declining glucose use, and why the affected regions differ between healthy aging and impairment.
The 'type 3 diabetes' question
Why the parietal cortex and precuneus are affected, and what share of glucose uptake actually remains.
Why ketones still get in
Separate transporters, a shorter metabolic path, and enzymes that appear to remain functional.
Demand versus supply
Brain glucose use is set by internal demand; ketone uptake follows blood levels — the strategic opening for supplementation.
Insulin as the fuel switch
How insulin decides between the two fuels, and why insulin resistance disables both at once.
The BENEFIT trial design
Funding, participant criteria, the milk emulsion of 30 g MCT daily, and why the primary outcome was metabolic.
Cognitive and imaging outcomes
Memory, language and executive function improved; connectivity and white matter imaging extended the picture to all five domains.
The care-home carbohydrate study
A feasibility study using continuous glucose monitoring, testing whether a modest carb reduction is safe and workable.
Practical takeaways and what's next
Carbs, MCT, exercise, sleep and social life — plus the CogniKet trial and a planned multimodal study.