The Algorithm That Creates Diets That Work for You

November 20, 2015 9:12 AM

Take a slice of cake and cut it in two. Eat one half, and let a friend scoff the other. Your blood-sugar levels will both spike, but to different degrees depending on your genes, the bacteria in your gut, what you recently ate, how recently or intensely you exercised, and more. The spikes, formally known as “postprandial glycemic responses” or PPGR, are hard to forecast since two people might react very differently to exactly the same food.
But Eran Elinav and Eran Segal from the Weizmann Institute of Science have developed a way of embracing that variability. By comprehensively monitoring the blood sugar, diets, and other traits of 800 people, they built an algorithm that can accurately predict how a person's blood-sugar levels will spike after eating any given meal.
posted by hippybear (36 comments total)

Cool.
posted by wemayfreeze at 9:27 AM on November 20, 2015

I tried that but my friend sneered rather than scoffed and so neither of us enjoyed the cake.
posted by ardgedee at 9:28 AM on November 20, 2015 [12 favorites]

In the future there will be a great divide in eating habits.

One group will only eat food as prescribed by this algorithm. The other will only eat Soylent. (Paging Zach Weinersmith to take this idea to it's conclusion.)
posted by matrixclown at 9:34 AM on November 20, 2015 [4 favorites]

Take a slice of cake and cut it in two.

Damn it, why are diets so HARD?
posted by mittens at 9:37 AM on November 20, 2015 [10 favorites]

I want to do cake science with friends, where can I do cake science. Where is my cake research funding.
posted by poffin boffin at 9:40 AM on November 20, 2015 [15 favorites]

also why can't we just use two separate identical slices of cake, why must the initial cake be halved
posted by poffin boffin at 9:41 AM on November 20, 2015 [5 favorites]

Guys, you're missing the point. The cake is a lie.
posted by Naberius at 9:45 AM on November 20, 2015 [6 favorites]

That must be why the friend is scoffing at it instead of scarfing it down.
posted by briank at 9:50 AM on November 20, 2015 [9 favorites]

Although people are often notoriously unreliable at documenting their meals, Segal says that his volunteers were unusually motivated. “We didn't pay them,” he says. “They joined because we explained that we'd be able to tell them which of the foods they normally eat spike their glucose levels. They came because they wanted to know and we said that if they didn't log properly, we wouldn't be able to tell them.”

See, if this means I have to disclose in any sort of detail the I-have-hot-dogs-and-mayo-but-no-buns incident from last night, I'm going to have to politely decline.
posted by mandolin conspiracy at 10:01 AM on November 20, 2015 [3 favorites]

but you just told us and now google will know forever
posted by poffin boffin at 10:08 AM on November 20, 2015 [3 favorites]

but you just told us and now google will know forever

I was intentionally vague on the finer details. Olives may also have been involved.

Oops.
posted by mandolin conspiracy at 10:13 AM on November 20, 2015 [3 favorites]

If I could know that having hot dogs and mayo with no buns (or any other foods) on any given night would be good for my blood sugar levels, I'd be totally happy to let anyone know what I'd eaten.
posted by hippybear at 10:15 AM on November 20, 2015 [3 favorites]

I was intentionally vague on the finer details. Olives may also have been involved.

Mayotini with hot dog and olive garnish
posted by jason_steakums at 10:16 AM on November 20, 2015 [8 favorites]

I-have-hot-dogs-and-mayo-but-no-buns

The other day I had that, but I skipped the hot dogs.
posted by mittens at 10:17 AM on November 20, 2015 [4 favorites]

All joking aside, this is amazing stuff:

It was remarkably accurate. When the team tested it on a fresh set of 100 volunteers, it predicted sugar spikes that matched the volunteers' actual data with a correlation of 0.7 (where 1 would be perfect). That's good: Even if the same person eats the same meal on two different days, the correlation between the two sugar spikes will be 0.77 at most. That sets a ceiling for predictability, one that the team's algorithm came very close to hitting. It certainly outperformed the crude technique of counting carbs or calories; when Zeevi and Korem tried doing that, they got correlations of just 0.38 and 0.33.
posted by mandolin conspiracy at 10:17 AM on November 20, 2015 [10 favorites]

As a type one diabetic who's half-Irish digestive system processes potatoes like they're candy I'm definitely familiar with individual variations in glycemic response.
posted by BrotherCaine at 10:17 AM on November 20, 2015 [3 favorites]

See, if this means I have to disclose in any sort of detail ...

They need a stool sample, is that too much?
posted by achrise at 10:47 AM on November 20, 2015 [2 favorites]

greg what if you deep fry a mayo cone and put hot dog chunks in it
posted by poffin boffin at 10:48 AM on November 20, 2015 [2 favorites]

This is fascinating. Whereas many people I know can have sugary protein bars or even doughnuts for breakfast with their coffee and be perfectly functional, my blood sugar and resultant mood/cognitive ability would go haywire. I've always found it strange that I can eat exactly the same thing as someone else, even so-called "healthy" foods, and suffer totally different consequences. Whole grains of pretty much any kind might as well be Milk Duds to me. And I'm weirdly sensitive to lentils, of all things.
posted by whistle pig at 10:50 AM on November 20, 2015 [3 favorites]

I love this stuff. People are so interesting and varied yet ultimately understandable via science! On the other hand, if I were organized and motivated enough to track everything I ate, I could probably figure this out for myself. I'm just not that organized and motivated, and I refuse to do things like measure portions, because that's eating disorder behavior for me.
posted by ArbitraryAndCapricious at 10:54 AM on November 20, 2015

OK, this is all fine and good but I was promised cake. Can I split it with my companion cube?
posted by nubs at 11:07 AM on November 20, 2015 [1 favorite]

Interventive Gerontology 1.0.02: First, Try to Make it to the Mean: Diet as a life extending tool, Part 1.
posted by bukvich at 11:12 AM on November 20, 2015

The ongoing vociferous arguments about which foods are and aren't "good for you" has long been known by informed people to be a red herring. This work emphasizes the degree to which an individual's response to environmental stimuli, including food, is specific to the person's genome, epigenome, and microbiome. The uninformed discussion usually proceeds with the assumption, in spite of the obvious variability in observable human traits, that we all digest and metabolize food the same way.

I hope that cartoonish view will soon wither away.
posted by Mental Wimp at 11:18 AM on November 20, 2015 [10 favorites]

Although people are often notoriously unreliable at documenting their meals, Segal says that his volunteers were unusually motivated. “We didn't pay them,” he says.

My former PhD student, now on faculty at NDSU, is collaborating on a phone app that will analyze a picture of your plate and assess it for macro and micronutrients. Coupled with the research in the FPP, this could revolutionize our understanding of how health is related to nutrition. It's kind of exciting to me to be even at the fringes of this enterprise.
posted by Mental Wimp at 11:23 AM on November 20, 2015 [5 favorites]

My former PhD student, now on faculty at NDSU, is collaborating on a phone app that will analyze a picture of your plate and assess it for macro and micronutrients.

Ooooooh, nifty.
posted by quaking fajita at 11:31 AM on November 20, 2015

Your blood-sugar levels will both spike, but to different degrees depending on your genes, the bacteria in your gut, what you recently ate, how recently or intensely you exercised, and more. The spikes, formally known as “postprandial glycemic responses” or PPGR, are hard to forecast since two people might react very differently to exactly the same food.

Vindication!!! To my mother and others: MY BODY IS NOT YOUR BODY, AND I AM NOT A BAD PERSON FOR NOT EATING YOUR WAY!

(sorry, the holidays are coming up...)
posted by sockerpup at 12:24 PM on November 20, 2015 [9 favorites]

analyze a picture of your plate and assess it for macro and micronutrients

Finally, I can learn whether plates spikes my blood sugar less or more than stoneware plates :)

Seriously though, this is awesome and I want to be part of it. I'm a type 2 diabetic and I have poor BG control no matter what sort of effort I put in (it's just a question of how poor), despite also being on metformin and two kinds of insulin. If I could have a tailored diet plan that works, doesn't consist entirely of kale, celery and wood chips, and actually satisfies me so I'm not a murderous hungerbeast all day, that would be great.
posted by Foosnark at 12:31 PM on November 20, 2015 [4 favorites]

My former PhD student, now on faculty at NDSU, is collaborating on a phone app that will analyze a picture of your plate and assess it for macro and micronutrients.

Probably a better approach than this.

The lack of depth and scale from a single image seems like it'd make it kind of hard to do reliably, though. There's been a lot of research into extracting depth from video, so maybe that could work.
posted by jedicus at 1:18 PM on November 20, 2015 [1 favorite]

analyze a picture of your plate and assess it for macro and micronutrients

I remember seeing an iPhone app that claimed to do this a few years ago. It had the user enter a descriptive label alongside the picture. Which would seem to take a bit of the work out of assessing food.
posted by ZeusHumms at 1:27 PM on November 20, 2015

I finally had time to read the article, but I may be missing something: When they said "diets that work for you"...what did they mean, exactly? What's the advantage of knowing what foods will cause a spike, if you're not actually diabetic? EXPLAIN FOOD TO ME PLEASE
posted by mittens at 1:31 PM on November 20, 2015

Blood sugar spikes and crashes can make you feel tired or hungry or angry.
posted by the agents of KAOS at 1:38 PM on November 20, 2015 [2 favorites]

I'm not diabetic or even pre-diabetic that I know of, but reading this from a previous FPP has me ordering unsweetened tea most places I go these days.
posted by hippybear at 1:39 PM on November 20, 2015 [1 favorite]

I hope this goes somewhere. As a diabetic who cannot make sense of some of my spikes (I eat to my meter, as most diabetics in control do, as it is better than blindly eat/shot/eat/shot.)

Today, I ate perfectly. Seriously, as in what they would say is pretty much a perfect day for a diabetic. Yet, my blood glucose is through the roof.

Not only have I eaten properly, I've exercised, taken my meds, and my insulin. Still ended up with a crazy number. And, I'm not on steroids, nor do I have an infection, or a cold or anything.

I don't get it. I would love to be able to do something like this.
posted by SuzySmith at 3:05 PM on November 20, 2015 [3 favorites]

SuzySmith, sorry if I'm telling you things you already know, but other things to look out for causing sugar spikes include stress, caffeine, and exercising near the anaerobic threshold or higher. Not sure if your insurance covers it, but I found having a continuous glucose monitor very helpful for pinning down exactly what causes spikes for me.
posted by BrotherCaine at 4:37 PM on November 20, 2015

mandolin conspiracy: "See, if this means I have to disclose in any sort of detail the I-have-hot-dogs-and-mayo-but-no-buns incident from last night, I'm going to have to politely decline."

Hail Eris
posted by Reverend John at 7:58 PM on November 20, 2015 [1 favorite]

Well, it seems you can't sign up for the follow up study nor can you download their food and activity app. I am not bright enough to find or derive the algorithm from the Cell article (any help with this would be appreciated).
posted by jadepearl at 4:09 AM on November 21, 2015 [2 favorites]