Brain, livin' on ketones - a molecular neuroscience look at the ketogenic diet

Brain, livin’ on ketones – a molecular neuroscience look at the ketogenic diet

Edited October 3, 2013: A 2.0 version of this post can be found at Scientific American MIND Guest blogs, here. And here’s me talking about it. Feel free to check it out!

Remember when your high school biology teacher said that the brain absolutely NEEDS glucose to function? Well, that’s not entirely true. Under severe carbohydrate restriction, the brain can adapt and start burning ketones as fuel.

Originally devised as a therapy for drug-resistant epilepsy in children, the ketogenic diet (keto) has been gaining popularity lately. It’s a high fat, moderate protein and low carbohydrate diet (LCHF) designed to force the body to go into a state called metabolic ketosis. With the advent of books like “Good Calories, Bad Calories” and “Why we get fat”, LCHF diets are increasingly touted as the magic bullet to weight loss. While there is considerable interest in the medical community in using the ketogenic diet to manage metabolic syndrome or prevent cardiovascular disease, more attention has focused on its role in drug-resistant seizure management and (potentially) neuroprotective effects in brain damage. In the last decade, keto has been shown to improve memory in patients at risk for Alzheimer’s disease, stabilize mood in type II bipolar disorder, reduce symptoms in Parkinson’s disease and even ameliorate some behavioral and social deficits in autism. Keto also seems to decrease brain cancer progression. ALL without observable side effects. Although most of these studies were unblinded (hence placebo can’t be ruled out), the effect is still amazing.

What is going on in the brain? And why aren’t pharmaceutical companies racing to package keto into a convenient treat-all 3-a-day pill?

How does the body go into ketosis?



Simple speaking, strict carbohydrate restriction depletes liver glycogen and forces the body to turn to other macronutrients for energy. Proteins are metabolically costly to utilize (not to mention dangerous – heart is also a muscle), and is often used as a last resort. By providing adequate amounts of fat, the liver uses dietary and body fat as fuel and produces ketones. “Ketones”, or “ketone bodies” is actually an umbrella term for 3 different molecules, β-hydroxybutyrate (BHB), acetoacetate (ACA) and acetone. All three can be delivered into the brain and metabolically converted into ATP in both neurons and glia. The three are interrelated: BHB and ACA can convert into each other, while ACA can turn into acetone. Extra ketones are eliminated through urine, or in the case of acetone, breath.

What are ketones doing in the brain? Answer: it’s complicated

There’s a reason neuroscientists and neurologists still haven’t figured out why keto is so effective in treating epilepsy. Holistically, the body is now running in a different metabolic state with changes in hormone levels (to say the least) which influences the nervous system. Locally, the brain is running on 3 types of semi interchangeable ketone bodies, the effects of which often can’t be teased apart. Hence it’s hard to say whether specific molecular and cellular alterations observed clinically or experimentally in animal models are a direct effect of ketosis, or simply a secondary phenomenon. Nevertheless, several hypotheses have been put forward to explain keto’s neuroprotective effects.

Source: Melo TM et al. Neuronal–glial interactions in rats fed a ketogenic diet. Neurochemistry InternationalVolume 48, Issues 6–7, May–June 2006, Pages 498–507

Source: Melo TM et al. Neuronal–glial interactions in rats fed a ketogenic diet. Neurochemistry International
Volume 48, Issues 6–7, May–June 2006, Pages 498–507

An oldie turned newbie: local changes in pH

Ketone metabolism generate pH-lowering metabolites, hence a change in pH was proposed early on as a way keto influences brain function. However there’s no evidence that keto significantly lowers brain pH, although mild decreases in pH may be possible in local microdomains. This hypothesis is attractive as many receptors are modulated by pH, such as acid-sensing ion channel (involved in stroke) and NMDA receptors (involved in learning, memory and excitotoxicity in stroke/neurodegenerative diseases), which may explain keto’s possible effect in stroke protection or cognitive improvement. Discarded a while back, this hypothesis recently resurfaced as the mechanism behind keto’s positive effect on Type II bipolar disorder management, which relies on blood acidification.

A favorite: Bioenergetics?



Ketones can be turned into energy effectively by the brain. In fact, BHB may provide a more efficient source of energy for brain per unit oxygen than glucose. A microarray study showed that keto induced a coordinated upregulation of genes encoding energy metabolism and mitochondrial enzymes, increasing the number of mitochondria in the hippocampus, a brain area associated with learning and memory. This increased energy capacity has been shown to enable hippocampal neurons to better withstand low glucose exposure, which happens in stroke. Better bioenergetics is proposed to limit seizure activity by stabilizing neuron resting membrane potential (so they’re not as excitable) or activate KATP channels through adenosine release. There are no studies directly addressing if energy efficiency is the reason for cognitive improvement in neurodegenerative diseases under keto, or if it contributes to enhanced cognitive performance in healthy individuals.

Cute as it is, there's not a hell lot of evidence.

Cute as it is, there’s not a hell lot of evidence. Source:

Another favorite: Antioxidative & anti-inflammatory effects?

Mitochondrial respiration, while generating ATP, also produces many reactive oxygen species (ROS). An acute increase in ROS is associated with stroke damage, while accumulation of ROS is one of the major hallmarks of aging and age-related neurodegenerative diseases. Keto can induce upregulation of mitochondrial uncoupling proteins (UCPs) in rat, which correlates with decreased ROS generation and increased resistance towards chemically induced seizure. Keto also increases the body’s own antioxidant defense system, namely glutathione levels in the hippocampus and protects mitochondrial DNA from ROS damage. If, and how, these antioxidative effect relate to neuroprotection is not yet clear.

Poly unsaturated fatty acids (PUFA) in the keto diet, such as DHA and EPA have garnered a lot of attention. Some evidence links them to decreasing neuronal excitability in hippocampus, which may contribute to decreasing seizure generation. PUFAs can also directly act on receptors called peroxisome proliferator-activated receptors (PPARs), the latter of which translocates to the nucleus and shuts down expression of pro-inflammatory factors. As inflammation increasingly recognized as a contributor to seizures, Alzheimer’s and metabolic syndrome, this mechanism may be a crucial one in keto’s favorable effects.

Maybe: direct drug-like actions of ketone bodies

Direct injection of ACA and acetone into animal models of epilepsy prevented seizures, hinting that ketone bodies may directly suppress seizure activity. However, other studies show that ketone levels may not correlate with seizure control. BHB and ACA have also been proposed to directly influence excitatory/inhibitor neural transmission. However, direct application of the two ketones had no effect on (1) excitatory responses in hippocampal neurons after stimulation (2) spontaneous epilepsy-like activity in a brain slice model of epilepsy (3) whole-cell currents evoked by glutamate, kainate, and GABA in cultured hippocampal neurons. Looks like a nail in the coffin for that theory. As far as I know, direct actions of ketone has not been linked to neuroprotection.

 Directly inhibiting cell death?

Keto seems to suppress the expression of pro-cell death proteins such as caspase-3 and clusterin (both of which mediates cell death in Huntington’s disease, among others), which correlates with enhanced recovery from seizure episode or stroke in patients. How much of a role this mechanism plays in disease states is still unknown.

Putting it all together: no pill yet!

The effects of keto are multifactorial and complicated. At the moment it’s impossible to tease apart which mechanisms are the driving forces behind keto’s powers, and which ones are secondary manifestations. Or they may operate equally, who knows? Hence packing everything up into a neat little keto pill is going to require a lot of effort (… …don’t even mention raspberry ketones!).

In the end, what do all these data tell us? A ketogenic diet with calorie restriction is most likely beneficial to weight loss. It is effective in controlling seizures in children and adults. It may improve cognition in patients with neurodegeneration or enhance mood stability in patients with Type II bipolar disorder. And that’s a BIG “may”. Without randomized controlled trials, it’s really difficult to say.

What about keto as a potential therapy (or adjunct therapy) for neurodegeneration? As increased ROS accumulation and mitochondrial dysfunction are common threads in age-related neurodegenerative diseases, it is conceivable that keto could be beneficial with its antioxidative and anti-inflammatory actions. To what degree is anyone’s guess.

Pass the bacon, maybe?

The ketogenic diet (along with other low-carb diets) is gaining popularity as a weight-loss measure. Some go on the keto diet because they experience “fewer sugar crashes, enhanced energy levels, mental clarity and decreased hunger”. A quick browse through progress photos on is probably enough to convince most people with a few extra pounds to give the diet a shot. Based on the above studies, keto is touted as a safe, effective tool for weight loss, with the added bonus of improved cognition. A few studies in epileptic children and healthy volunteers demonstrate that keto exterts a biphasic effect on cognition, with initial lethargy and subsequent heightened vitality, physical functioning, and alertness. Whether this translates into overall cognitive enhancement remains to be seen.

What are the potential side effects? A small study involving 21 obese women showed impaired higher cognitive function after 28-days on the keto diet. On the contrary, another study involving 83 obese patients showed improved blood lipid profile after being on a 24-week ketogenic diet without significant side effects. A caveat of most of these studies is the lack of an adequate control group, making results hard to interpret.

Some have also raised concerned regarding metabolic and long-term complications. Children following ketogenic diets have higher rates of dehydration, constipation, and kidney stones. Other reported adverse effects include hyperlipidemia, impaired neutrophil function, optic neuropathy, osteoporosis, and protein deficiency. However, ketogenic diet for the management of seizure is different than that for weight-loss. The former generally requires 80-90% fat calories (although this has decreased slightly in the Modified Atkins Diet) while the latter proposes ~60% fat calories with sufficient protein for muscle maintenance.

Whether keto promotes cognitive improvement and neuroprotection in the general public remains to be seen. While waiting for science to catch up though, I’m going for that bacon, spinach & cheese omelet. For science!

Edited Sep 2,2013: For those interested in exploring more, here is an accessible journal review that looks at potential therapeutic uses of nutritional ketosis in many other diseases. Note I am not promoting using keto as a sole treatment option – if there are efficient pharmaceuticals available please do not forgo them in favour of ketosis.


Masino and Jong. 2012. Mechanisms of Ketogenic Diet Action. Jasper’s basic mechanisms of the epilepsies. 4th ed.

Hallböök T, Ji S, Maudsley S, & Martin B (2012). The effects of the ketogenic diet on behavior and cognition. Epilepsy research, 100 (3), 304-9 PMID: 21872440

Dashti HM et al. 2004. Long-term effects of a ketogenic diet in obeses patients. Exp Clin Cardiol. 9(3): 200-205.

Rho, J., & Sankar, R. (2008). The ketogenic diet in a pill: Is this possible? Epilepsia, 49, 127-133 DOI: 10.1111/j.1528-1167.2008.01857.x

Phelps, J., Siemers, S., & El-Mallakh, R. (2012). The ketogenic diet for type II bipolar disorder Neurocase, 1-4 DOI: 10.1080/13554794.2012.690421

Ruskin DN, Ross JL, Kawamura M Jr, Ruiz TL, Geiger JD, & Masino SA (2011). A ketogenic diet delays weight loss and does not impair working memory or motor function in the R6/2 1J mouse model of Huntington’s disease. Physiology & behavior, 103 (5), 501-7 PMID: 21501628

Krikorian R, Shidler MD, Dangelo K, Couch SC, Benoit SC, & Clegg DJ (2012). Dietary ketosis enhances memory in mild cognitive impairment. Neurobiology of aging, 33 (2), 2147483647-27 PMID: 21130529

Denke, M. (2001). Metabolic effects of high-protein, low-carbohydrate diets The American Journal of Cardiology, 88 (1), 59-61 DOI: 10.1016/S0002-9149(01)01586-7