I didn't come up with the idea that excessive food reward increases calorie intake and can lead to obesity, far from it. The idea has been floating around the scientific literature for decades. In 1976, after conducting an interesting diet study in humans, Dr. Michel Cabanac stated that the "palatability of the diet influences the set point of the ponderostat [system that regulates body fatness]" (1).
Currently there is a growing consensus that food reward/palatability is a major contributor to obesity. This is reflected by the proliferation of review articles appearing in high-profile journals. For the scientists in the audience who want more detail than I provide on my blog, here are some of the reviews I've read and enjoyed. These were written by some of the leading scientists in the study of food reward and hedonics:
Palatability of food and the ponderostat. Michel Cabanac, 1989.
Food reward, hyperphagia and obesity. Hans-Rudolf Berthoud et al., 2011.
Reward mechanisms in obesity: new insights and future directions. Paul J. Kenny, 2011.
Relation of obesity to consummatory and anticipatory food reward. Eric Stice, 2009.
Hedonic and incentive signals for body weight control. Emil Egecioglu et al., 2011.
Homeostatic and hedonic signals interact in the control of food intake. Michael Lutter and Eric J. Nestler, 2009.
Opioids as agents of reward-related feeding: a consideration of the evidence. Allen S. Levine and Charles J. Billington, 2004.
Central opioids and consumption of sweet tastants: when reward outweighs homeostasis. Pawel K. Olszewski and Allen S. Levine, 2007.
Oral and postoral determinants of food reward. Anthony Sclafani, 2004.
Reduced dopaminergic tone in hypothalamic neural circuits: expression of a "thrifty" genotype underlying the metabolic syndrome? Hanno Pijl, 2003.
If you can read all these papers and still not believe in the food reward hypothesis... you deserve some kind of award.
Hello Stephan. Thanks for your blog. I wonder if you have any comment on the product Sensa.
This series has been great. Have you done any study on causes of accumulation of large amount of Visceral fats? I have seen very active thick people (even some dancers are fat) I assume they might be having mostly subcutaneous fat. I assume this because pigs are very active in-spite they have large amount of subcutaneous fat.
It's good to have references. So far I could only briefly glance at them but it seems to me that the literature tends to be resolutely behaviorist, making no distinction between mental state and behavior. This might significantly cloud the issues.
How about defining (the degree of) food reward (of food F for organism O at time T) in terms of behavior and/or strictly behavior oriented mental state: (the degree to which) O eats more/wants to eat more of F at T).
Palatability (of F for O at T) could be reserved for the notion of hedonistic value, ie.enjoyment of food.
It's obvious I think that these are linked but distinct notions. (Take the chips manufacturer's explanation in Pretlow's presentation that mem linked earlier: chips are designed to peak quickly and fade quickly, to make people (want to) eat more. Chips are purposely designed to be unsatisfying, ie. low palatability high reward food in my terms.)
Couldn't a certain amount of confusion be cleared up along these lines? In particular what you are wisely recommending is not low palatability food but a low reward high palatability diet.
Great series. It's funny it comes up at this time. Just in the last 5-6 weeks I had identified that I was experiencing some severe cravings after not having any real cravings for a long time. The only thing that changed for me was I was chewing sugar-free gum all the time.
That caused me to experiment and research a bit, and I read The Shangri-La Diet and it made more sense. Where that book makes the point of food reward driving up setpoint and calories with no food driving it down, I saw it a little differently.
I see it more as a matrix. Calories on one side, taste on the other. So the combinations are: No calories/no taste (fasting), No calories/taste (artificial ingredients like in my gum), calories/no taste (Seth's ELOO), and calories/taste (Real food).
I think that anything with taste drives up your fat regulation set point. So chewing all that sugar-free gum created calorieless taste events. I craved more calories.
Snacking between meals would produce a similar effect if I'm correct. So I set out to minimize what I called "taste events". I was pretty comfortable eating once or twice a day, but the occasional snack, sugar-free drink, etc. would end up in there.
I simply decided to eat once or twice a day and the meal(s) would be the only taste events I would have daily. No snacks. No flavors. No other tastes.
Within two days the cravings were gone and in another day or two it was difficult to eat all the food I planned to eat for protein goals.
I think this in part explains why people drinking diet drinks all day tend to gain weight or find it difficult to lose weight. The taste events cause enough craving for additional calories that the desire gets satisfied somehow, and more calories are consumed because of the stimulation from the diet drinks.
About the time I had figured this out for myself, your Food Reward series started! It sure reinforced my thinking and provided more solid evidence for my conclusions.
I really like the way that Bill Strahan just restated/interpreted it!
The interesting about this is that I have similar "rewards"/responses to Woo. Now I've never had a weight problem, but I know that I can eat heavily flavored/spiced meat or nuts and stop without problem. But, even with a plain dry baked potato or rice, I have a strong desire to keep eating, and I want more food [all kinds]. Fermented dairy and unsweetened chocolate/cacao satisfy my appetite in small amounts.
In previous posts you suggested that people with either weight problems or known blood sugar problems be carefull with dietary carbohydrate.
Have you fully changed your opinion in this regard and if so could you expand a bit on that specifically as it might relate to your current theory?
Unpalatable = lower calorie intake = blood sugar/overweight reversal and as such no need to restrict carbs specifically?
I still want someone to do a trial of low food reward diet vs other diets. The effect will probably show itself to be rather minor.
And while they're at it, they could also include zero added fat/sugar, zero trans fat, zero soy, zero BPA, sane omega balance, otherwise normal diet and check how much difference that makes. I'm not expecting instant miracles here, but it might work pretty well long term.
I also like Bill Strahan's comment. I found that eating more natural and soaking grains is affecting our children and myself a great deal. Our baby can't get enough of the soaked whole wheat bread we make every week. The toddlers eat my whole wheat pancakes (no sugar added) without syrup all the time and they get full. The better we eat, the less cravings we seem to have.
I think it also has to connect to the toxins in our body. I've eaten poorly most my life (not fast food poor, just refined sugars and processed grains) so my body doesn't absorb as many nutrients from the good foods we are now eating. Working on detoxifying the junk that has been built up for years so I don't have as many cravings and can actually get energy from a normal serving of food. Thank you for your post.
The theoretical basis it is based on is plausible enough that I don't dismiss it offhand, but I don't find it that convincing either. I'd like to see its efficacy demonstrated in a convincing manner.
I know there has been research done on what causes visceral fat accumulation vs. fat accumulation in other places, but I'm not that familiar with it. Two factors that come to mind that promote visceral fat accumulation are high-heat cooking (mice) and large doses of fructose (humans).
I think I see your point. My definition of reward is mostly behavior-based. One exception is lateral hypothalamus self-stimulation threshold. That has been used as a measure of internal reward sensitivity in rodents, although there are skeptics out there.
My definition of palatability is strictly the hedonic value of food-- distinct from reward. Measurements of palatability include both behavior and "internal state" factors. Behavioral measures of palatability can be taken in humans and animals because there are stereotyped orofacial responses to palatable food. As far as I know, "internal state" measures of palatability can only be taken in humans, i.e. by asking them how much they like the taste of a food.
One of the problems with all this is that reward and hedonic systems are tightly intertwined and the two factors typically travel together. So it's hard to keep them separate and people are often sloppy about it. I'm sure there are some parts of my series where I didn't use the best terminology, and you will see that to some extent in the scientific lit as well.
I'm not sure my recommendations are for a high palatability diet. I think the food will be simple, and possibly (although not necessarily) lower palatabiltiy than the typical diet. But I think palatability will increase somewhat as people become accustomed to it.
That is a nice theory. It does handily explain how calorie-free foods could increase body fatness. It would also concur pretty well with Seth Roberts' findings. I'll keep your idea in mind.
Interesting. Are you sure it isn't just because potatoes and rice have a lower calorie density than meat and nuts, so you have to eat more to get the same number of calories?
Yes, I have changed my mind on that. At some point, I'm going to do housekeeping and take down/modify the old posts that I've diverged from.
Hi Stephen -
I'm a longtime follower of your blog and have found your food reward series very interesting. However, I find your explanations somewhat unsatisfying - in my opinion you have not really explained the actual mechanism, only given a compelling story to interpret the experiments you've looked at. Granted, it IS a compelling story, but even if your 'explanation' fits the evidence, you are in much the same position as others who have given a consistent interpretation to the data they have looked at. Maybe you have a better theory that fits more data, but where is the mechanism?
Basically, I would like an explanation that reaches down to the biochemical level. Does 'food reward', say, decrease leptin sensitivity (a hormone whose action can be understood at the biochemical level) and thus decrease appetite? If so, how does it do that, chemically? Or is there some other explanation for how food reward works? And what exactly is food reward? Do, say, spices really contribute to food reward, and if so, what is the actual mechanism? Again, biochemical explanations could answer these questions and suggest new experiments.
Or is your explanation purely in terms of neurobiology, dopamine receptors, etc? If so, it needs to be established by what (biochemical) mechanism these receptors in the brain are affected. And by what mechanism does the brain then influence other cells in the body, causing increase in appetite for certain foods, or signaling to cells to store or release more fat, etc?
I'm not really sure, but I bet calories play a role. If I don't eat lots of fat I lose weight. I just know that fruit and starches like sweet potatoes are impossible to stop eating until I'm uncomfortable, and I want more as soon as I feel okay again. Salt, fat, and MSG don't give me a strong reward I suppose (though the only time I recently ate MSG was with duck heart & tongue because I didn't know "delicious food" meant MSG). But, if we're talking about reward and weight gain, then there's no problem, since they don't cause it, at least for me. Maybe a potato is much less rewarding for someone else, as I haven't had refined sugars or vegetable oils in about a year. I find it hard to accept though that using butter with starch is what's causing people to be fat.
I hope you will agree that a mechanism is not required to demonstrate that a phenomenon exists. In my opinion, the mechanism is interesting but secondary.
The papers I linked to in this post contain a lot of mechanism. It's beyond the scope of this blog for me to get into all the details because the blog is for a general audience. However, I'm currently preparing a review article for a major endocrine journal that goes into more detail (although they limited us to 4,000 words and 120 refs so it will still be a fairly brief overview). I don't know how much of it is going to survive the editorial process as we're still in early draft stage.
In brief, the mechanism hinges on the effects of highly palatable food on CNS reward and hedonic systems. Reward circuits, including but not limited to dopaminergic cells in the VTA, synapse in the hypothalamus and exert a considerable influence on hypothalamic regions that regulate food intake and body fatness. Dopamine action in the lateral hypothalamus suppresses food intake. My hypothesis is that dopamine signaling in certain key regions, particularly through the D2 receptor, becomes progressively desensitized due to overexposure to dopamine and perhaps the effects of leptin and insulin on reward signaling.
Opioidergic systems that participate in hedonic and reward processes antagonize anorexigenic signaling in the hypothalamus, so I think that is also likely to play a role.
The combination of naloxone (opioid antagonist) and bupropion (dopamine-NE specific reuptake inhibitor) synergistically cause fat loss in overweight humans, consistent with the idea that obesity results from the combination of excessive opioidergic signaling and insufficient dopaminergic signaling. The proposed mechanism of the drug combination is a synergistic effect on hypothalamic neurons that regulate body fatness.
Are you on a low-carb diet? If so, that may be the reason carb foods are so rewarding to you. Adding fat to a high-fat diet doesn't increase reward, but adding starch does. Similarly, adding starch to a high-starch diet doesn't increase reward, but adding fat does. Just a possibility.
Wow! That is quite a turnaround.
I'm not really so suprised with regard to only general nutrition or even weight loss but advocating noncontrolled carb consumption [and uncontrolled fat as well if I understand your theory] for people with blood sugar issues?
That's a tough sell for me and I don't even have blood sugar issues.
I've been following along pretty closely in my "not even amatuer" manner and I'm having a hard time making that leap with you.
What prompted the turn around? I feel like I missed something.
Also, I watched some footage of a kitavan village that was shot recently and although the men do seem on the lean side generally the younger to middle aged women could have passed for americans in any mall parking lot without batting an eye.
The older women that I saw were indeed exactly as described, very very lean.
Any thoughts on that?
That's funny - I have been having cravings lately, but they are solely for spicey food. (I'm not on a low-carb diet anymore - maybe that's why I'm not craving carbs.) I had lost a lot of weight over recent months while having GI pain triggered by lots of different kinds of foods, and so was on a bland diet by default.
Now that that's resolved (don't know why, but thank goodness), I can eat spicy food. It *has* triggered a bigger appetite and I've gained a couple of pounds. I find this reassuring: I was getting to the lower end of my "ideal" BMI and didn't seem to be able to stop losing weight. It was weird. I was a little worried I might have cancer. Having an appetite and a bit of weight gain tells me it was this food reward thing. (I had an upper endoscopy and several blood tests that ruled out any serious condition).
But now I can't stop eating spicy food. If I had a gallon of gazpacho in front of me, it would be gone in minutes. At least now I know what I can do to lose weight again. But it's hard to say goodbye to flavor. I think this might be the sticking point for a lot of people with this plan!
I also like Bill's concept of the "taste event."
@ Stephan - Thinking about my experience on a bland diet - is it possible for people to go below their ideal weight this way? Or will the body find its best weight and settle there? I'm wondering if I've just been conditioned by having a 10-pound higher set point most of my adult life into thinking that a BMI of 18 or 19 is "too low."
It wasn't that much of a turnaround actually. There was a relatively brief period of time (6-8 months) when I thought that starch per se was part of the cause of obesity and metabolic dysfunction in the general population, after I read GCBC. There are some old posts on the blog that still reflect that viewpoint. Then I snapped out of it. Now, I find the idea totally untenable. The only way I was able to believe that hypothesis at the time was my relative lack of knowledge about human physiology, controlled diet studies and traditional cultures, which all consistently indicate that a starch-based diet can support very good health and in some cases can even be therapeutic. "Noncontrolled carb consumption" is how most of the world eats, and as long as the diet is natural/traditional, they avoid the diseases of civilization.
Each person has to decide how much carbohydrate they can handle given their own glycemic control. I encourage people to buy a blood glucose meter and use it. Limiting carbohydrate makes sense to me in the case of diabetes. However, the average person has good enough glycemic control to eat a well planned whole foods high carb diet if they want to, and their glycemic control will typically improve over time on such a diet. My own glycemic control has improved since increasing my starch intake. I used to get spikes of ~130 mg/dL sometimes when I was restricting carb; now I rarely go over 120 mg/dL even though I eat 2-3 times more carb.
About Kitava, I don't know much about how things have changed there since Lindeberg did his study in the early 90s, but I do know that industrial food has been slowly creeping in. His study indicated that both men and women in all age groups were relatively lean. There were a few overweight women in that video you mentioned, but none I saw were clearly obese. The majority appeared fairly lean to me, although it's hard to tell because their clothes were so baggy.
@Stephen - thanks for your response. I agree you don't need a mechanism to show that a phenomenon exists. But would you agree that the mechanism provides the how and the why? Without the mechanism I feel there's no *real* understanding, just fitting a curve to a set of points...
So I think it would be very interesting for you to expand on your explanation of the mechanisms - perhaps you could do it in a way that would be readable to a general audience. I actually would enjoy learning whatever biology were necessary to understand the actual mechanisms. Would the paper you are working on now be accessible to a motivated layperson with a general science background, Wikipedia access, and some patience? :) If not what would be a good starting point?
Reduced dopaminergic tone in hypothalamic neural circuits: expression of a "thrifty" genotype underlying the metabolic syndrome?
^^^ Does not support food reward hypothesis.
IT DOES support a central dopaminergic control of metabolism, however.
The piece missing here is that "hypothalamic dopamine receptors" are in turn CONTROLLED by the nutrients in the diet, as well as season and other factors. When obese prone humans see a lot of dietary glucose, POOF GOES THE DOPAMIEN RECEPTORS, just like a hibernating animal.
Dopamine receptors magically reappear when glucose is restricted.
Doesn't sound so much like food reward and more like an evolutionarily conserved thrifty gene, where in which carbohydrate represents a seasonal zeigeiber triggering a hibernation-preparation like set of behaviors.
Calorie-unrestricted low-fat high-carb diets reliably cause fat loss in controlled weight loss trials. In most cases, lean non-industrial cultures that have become overweight/obese due to a diet/lifestyle change have gone from very high-carb to lower-carb higher-fat. The US diet 120 years ago was higher carb, lower fat than today and there was very little obesity. Starch is clearly not the culprit.
It's not worth looking for a mechanism for a hypothesis if it doesn't fit the basic empirical evidence.
I have been reading your posts regularly. One thing I have not seen much of is a discussion about fruit.
Clearly there can be downsides to fruit since they have a lot of sugar and are high in palatability (except for "fruit/vegetables" such as tomatoes and avocados). In the native diet, there is generally only a short time period where fruit is available. For Pacific Northwest Indians, I doubt they had more than two month's worth of available fruit, and it was never a real major component of their diet.
I have seen menus of feasts served to George Washington, and there was very little sweetness evident. Also, people have argued that our fruits are much more perfect and sweet compared to what was historically available. Also, I have noticed some obese people (generally women) who snack on fruits all day. Perhaps they think it is healthy, but they ought to know better.
So here is the flip side:
Color vision is believed to have developed to be able to find fruit. Our evolutionary relatives eat fruit, as I recall, chimpanzees are big fruit eaters, while gorillas have their share of fruit in an omnivorous diet. Also, workers in the cane fields, who ate sugar cane prodigiously, were clearly not obese.
Fruit flavor does not derive from industrial processes. It is not a new addition to the diet, such as industrial seed oils. Many cultures have had healthy diets which included large amounts of wine, which generally has a substantial residual sugar content.
Could fruit be eaten like starch in the low palatability diets you propose?
PS -- In terms of macronutrients in the diet, you indicated that fat consumption is up as a portion of the diet over the past century. Reading Gary Taubes and others, I strongly recollect that the post-1970 diet held flat or negative in fat while increasing carbohydrates. Also, remember the Pima Indians, with their massive carbohydrate consumption. I guess I don't see eye to eye with you when you try to link increase fat consumption to modern obesity.
It's important to keep sight of the main problem: industrially processed food. Fruit, although somewhat rewarding, is something we've been dealing with for a very long time, probably millions of years. Fruit is not a necessary part of the diet but it can be part of a healthy diet. Some people do well with a lot of fruit, others with less. I prefer to eat one piece of fruit per day and minimal sugars otherwise.
Regarding diet macronutrients, I'm not saying we got fatter in the US because we started eating more fat. As you said, that doesn't square with the diet changes that occurred in the last 40 years in the US. All I'm saying is that you can't pin the obesity epidemic on carbohydrate, either. If you look at diet and obesity trends in the US over the last 100 years, neither starch nor fat consumption correlate with obesity. In fact, when we were at our leanest in the late 1800s, we ate more starch and less fat than today as a percentage of calories. That doesn't mean the increase in dietary fat is what made us gain weight, but it does argue against starch as a dominant cause of obesity.
so stephan what is your best guess as to what drove up the obesity rates, the industrial seed oils mostly?
"In fact, when we were at our leanest in the late 1800s, we ate more starch and less fat than today as a percentage of calories."
But were total calories the same? Was average daily calorie expenditure the same? (After all, there was a lot more manual labor in the 1800s than today...) What about stress levels, sun exposure, or any number of other variables? And how exactly was leanness measured... does it take into account that average height has changed over time, say? Etc, etc.
I think looking at historical trends or observing different cultures is at best hypothesis generating (and downright misleading in many cases) - you cannot use it to dismiss or prove hypotheses about causes of obesity. To do that requires meticulously run RCTs, lots of them, and even then there are huge difficulties because it is generally impossible to measure only one variable (e.g., reduce fat and you must add either protein or carbs or you change total calories, etc).
Stephen, btw, I read GCBC too and found it very compelling at the time, but didn't feel like it was the full picture, which is why I've found your blog and your food reward posts very interesting. But I think you are a little hasty in completely dismissing starch as a culprit (or perhaps that is not what you are saying) of obesity. My current opinion is that in healthy, insulin-sensitive individuals, starch is fine in some quantities, but for others who (for whatever reason, maybe due to inflammatory modern industrial foods, excess fructose, or something else??) have more limited insulin/glucose processing capacity, starch actually does pose a problem.
I'm still quite confused about the whole food "reward" thing.
it seems one can be explained in both ways:
monotonicity = boredom = negative feed back => discourage over-eating?
monotonicity = positive feedback loop (habitual forming) => over-eating?
for me, i tend to get bored & under-eat on a monotonic diet (less variety).
none of the "high reward" food you've listed would make me binge due to boredom factor.
A comment from my own experience plus what I have read, is that cravings for carbohydrates stem from glucose in the blood stream. If you go on a low carb diet, your body stops sensing glucose in the blood stream and stops asking for it. Since I've been on a low carb diet, I have seen a very clear drop in my cravings for sweets. It's still there - I'd like a milk shake for instance. But it's not strong enough to drive me to action - to go out and get one. So, you can consider avoiding the whole reward issue by just dropping the carbs altogether from your diet.
Regarding fruit, it seems to me that the fruit we have readily available today may be far different than pre-agriculture fruit, thanks to selective breeding. How should we judge our level of adaptation?
Hi Stephan, thank you for writing a great blog, this series has been fascinating. I've been doing some crawling on PubMed, and while my hypothesis is still premature I've been thinking that highly rewarding food may actually work to induce leptin resistance. Any thoughts on this?
I'm not so sure. You might want to check out Denise Minger's latest post on her blog "Raw Food SOS".
Hi Stephen S,
I think that's highly plausible, but I'm not aware of any direct evidence to support it. There is plenty of mechanism in the dopamine and opioid literature suggesting it's possible. I'd love to see the idea tested. It would make perfect sense and it would tie homeostasis and reward/hedonic ideas together nicely.
Stephen: Can you comment on the effects of estrogen on the "food reward" theory? As with many women, for the first two weeks of my cycle, with estrogen high, there seems to be a very stable food reward relationship physically and mentally. However, as estrogen drops in the second two weeks, despite years of dietary trial, error and monitoring, nothing seems to shut off the intense need for food reward. Any thoughts?
Very useful. Keep it updated.
"The combination of naloxone (opioid antagonist) and bupropion (dopamine-NE specific reuptake inhibitor) synergistically cause fat loss in overweight humans,"
Stephan, do you have a citation for that? Naloxone is a relatively short acting drug, was it some sort of time release formulation?
Have you ever read anything from Tim Ferriss? In his "the 4 hour body" he outlines a diet for weight loss that is somewhat similar to your own guidelines. IIRC he calls it The Slow Carb Diet.
From the top of my head:
- Don't eat most carbohydrates (anything that is or can be made white, potatos, rice, pasta, etc.);
- Eat lots of legumes (beans, lentils, etc.);
- Reduce fat intake;
- Drink lots of water.
And the two most important points:
- Eat the SAME few meals every day;
- One free day per week in which you can eat everything you want ad nauseam.
So, low variety, somewhat low carb and one day to look forward to and keep cravings in check.
Berridge KC, Ho CY, Richard JM, DiFeliceantonio AG. The tempted brain eats: pleasure and desire circuits in obesity and eating disorders. Brain Res. 2010 Sep 2;1350:43-64. Epub 2010 Apr 11. Review. Full text here: http://bit.ly/pboUvG
"Your Brain on Food: Food Reward and Obesity", Univ. of Washington - http://depts.washington.edu/uwcphn/news/summits/food_reward_0206.shtml
What can be the potential mechanism of leptin and rewards signalling?
From my understanding leptin enters the brain matter two ways:
1. blood brain barrier endothelial cells active transport
2. choroid plexus cells active transport into cerebrospinal fluid
Once inside brain maybe the hormone enters the neuron circuits within the reward parts of the brain cells or secondary messengers instead transactivate dna/rna expression of a given dopamine gene and increasing dopamine receptor density synthesis Maybe this process is also happening in the hypothalamus as well. Anyway, as a result more dopamine receptors available, increases sensitivity of dopamine-receptor interaction, resulting in less potential of wanting food, even after being fed.
Let me know your thoughts
Post a Comment