Science

Chocolate eating styles, as defined by SCIENCE

 
I recently read Salt, Sugar, Fat by Michael Moss (a New York Times journalist who previously worked in Pakistan, but when his colleague David Rohde was kidnapped, the NYT transferred Moss to the safer beat of reporting on the food industry). The book provides fascinating (and sometimes horrifying) insight into the research and testing that goes into manipulating the various sensory aspects of food, and into developing food products with just the right amounts of salt, sugar and/or fat to make them incredibly alluring. Every variable in the eating experience can be fine-tuned to make food that people feel compelled not only to eat, but to eat lots of. For example, food scientists have worked out the exact amount of pressure a corn chip should break at in order to produce chips of an ideal crispness and crunchiness to maximise the enjoyment the average consumer experiences while eating (and I assume that if you buy some Doritos and eat them, you’re experiencing the outcome of that research).

Honing foods to optimise their appeal to consumers obviously requires the characterisation of many aspects of both the food and the consumers. Which is presumably how this paper, ‘Characterisation of chocolate eating behaviour’, came about, and which asks the question:

So, like, how do people eat chocolate?

You know, apart from, like, putting it in their mouth and somehow getting it down their gullet.

Turns out people fall into three distinct groups when it comes to how they eat chocolate:

1. the fast chewers
2. the thorough chewers
3. the suckers

The researchers found this out by giving the 40 study participants each two samples of milk chocolate. The samples were the same in ingredient composition and were similar in how viscous they would be when melted, but differed slightly in some textural aspects. I love this bit of the paper where the researchers say “For confidentiality reasons no further detail concerning sample composition can be provided.” So manipulation of the texture of the chocolate was a key variable being investigated in the study, but you’re being coy about how the texture differed between the two chocolate samples being tested? Well, one researcher does work for Mars and the study was funded by Mars, so omg u guyz this is a Mars chocolate formulation secret!

Anyway, the researchers have deigned to go into detail about how people ate the chocolate samples. Eating behaviours were characterised using surface electromyography (electrodes placed on the skin surface that record muscle activity) and electroglottography, which uses electrodes to record the activity of the glottis, i.e. the vocal cords. Electroglottography was originally developed for use in speech therapy, but has been re-purposed here to learn more about how and when people swallow.

The results were, firstly, that Chocolate A (whatever that was) was significantly preferred to Chocolate B (whatever that was). Participants considered Chocolate A to coat the mouth more and be softer at first bite (so hey, maybe you’re eating Chocolate A these days whenever you eat a Mars chocolate product). In terms of how the chocolate was eaten, each participant apparently had a unique way of eating the chocolate – no two participants were identical in terms of chew rate, total chewing time, number of swallows, how long it took them to get to the point of their last swallow, etc.

Cluster analysis was then performed on this information (this grouped people so that their eating behaviour was more similar to other people in the same group than people in other groups) which gives us the three styles of chocolate eating: fast chewing, thorough chewing, sucking.

If you do the cluster analysis for the data obtained for the two chocolate types separately, the same people still fall into the same eating behaviour groups, suggesting that people’s chocolate eating style is consistent across different types of chocolate. We can’t know this for sure from this study, since we don’t know how similar or different the two chocolate samples were (just that they were both milk chocolate), but maybe chocolates that are more obviously different would be eaten differently (I think my eating behaviour for a piece of milk chocolate versus a piece of, say, 80% dark chocolate is different, but maybe I’m just some weird anomaly).

And to finish, an excerpt from the paper:

“From a wider perspective, it has been shown that variation in bite size and oral processing time impacts on food intake. ‘Slow’ eating has also been shown to decrease food intake and result in increased satiety. If variation in chocolate eating behaviour exists then this could also impact on chocolate intake levels which in turn may have health implications.”

If we make those assumptions, this presents rather a quandary to Mars: Chocolate A was much preferred by participants compared to Chocolate B, but participants also took longer to eat Chocolate A, which might mean that people would eat less of it. What do you do, Mars? What do you do? (Presumably some more testing and then whatever option seems most likely to give them the best profits.)

The influence of shape on taste, or, what the hell Cadbury this chocolate is, like, über-sweet, how dare you

Nancy Drew Charles Spence is on the case of The Chocolate That Suddenly Became Too Sweet. In short, Cadbury changed the shape of their Dairy Milk bars, and consumers are complaining that Cadbury must have also reformulated the chocolate because now it tastes much sweeter. Spence, however, postulates that the shape of the bar might be influencing people’s perceptions of sweetness. Other studies have found that people seem to implicitly associate shapes with particular gustatory sensations – previous research has found that “sweetness has been paired with roundness in a range of different foodstuffs ranging from fruit juices to yoghurts and from milk chocolate to fruit juices. By contrast, angularity has been matched with bitterness, carbonation, and sourness in beers, sparkling waters, fruits, fruit juices, dark chocolates, salt and vinegar crisps, and so on”. If sweetness is associated with roundness, maybe making something round rather than angled will make people, at some level, experience the thing as more sweet.

However, implicit associations are a very different thing from actually changing the sensory experience. Conceptually, sure, round/sweet and sharp/bitter or whatever make sense – humans are very good at generalising lots of concepts and ideas across the senses and across all other sorts of categories. That doesn’t demonstrate any cause-and-effect – just an association. As the ancient words of wisdom go, correlation does not equal causation. I need to look more into the research of how manipulating a thing’s shape actually affects people’s sensory experience of the thing. (I’m assuming it exists; in fact, I’m reasonably sure I’ve read some of it, ages ago).

In the meantime, though, the whole Dairy Milk ~fiasco~ kind of makes me want to hunt around some newsagents/convenience stores/vending machines to see if I can find some of the old rectangular bars and do a direct taste comparison between them and the new rounded bars. Figure 1 in Spence’s paper has the old and new bar side by side – I hope someone who was around at the time cut the bars up and did some taste-testing with people (I reckon you could’ve broken each bar into its 6 pieces and done a rather underpowered randomised, within-subjects experiment on 6 people, hey?).

Actually, testing this whole thing properly wouldn’t be too hard, and doesn’t require the stomache-achingly sweet experience of eating Cadbury Dairy Milk chocolate. Get some chocolate, melt it, form it into different shapes as it cools and sets (e.g. pour some so it forms little round pools; just break off some straight-edged bits to get some angular pieces), then ask people to give you a hand. You can’t really test this yourself because you know what you’re testing and that could influence your perception – the mere knowledge that you could possibly expect the rounder shape to seem sweeter can bias your evaluation of the sweetness. So tell some friends that you just want to compare two brands of chocolate to find out which one is sweeter, and get each person to sample both chocolate shapes (making sure not to draw their attention to the shapes or to say anything about your hypothesis that shape might be important). There are a bunch of other things that would need to be considered in order for it to be a properly sound experiment – like counterbalancing the order in which the chocolate shapes are tasted (half the people should have the round one first, half should have the sharp one first, to make sure that the order in which people consume the shapes doesn’t have an effect on the evaluation), or maybe getting people to rate the sweetness of each piece on a scale rather than just saying which one is more or less sweet than the other, or making sure the two differently shaped pieces of chocolate are matched in terms of weight and temperature and other such factors – but even just as a quick and messy experiment you might be able to get an idea of whether shape might just be having an influence on the experience of sweetness.

And then you can spend hours thinking about how millions of different factors influence millions of other different factors in the context of the sensory experience of eating, because complex, multisensory experiences like that are ridiculously deep rabbit holes. Or you can just whinge that your precious Dairy Milk has been changed forever and vent your consumer outrage spleen in Cadbury’s general direction.

Bowls made of depleted uranium, what could possibly go wrong?

Just a relatively quick post, as things are have been and are continuing to insist upon being busy – I’ve had a round of assessment for my PhD (my mid-candidature review), participated in a speaking competition about my PhD, done a talk about “seducing with neuroscience” (maybe I’ll post something about that at some stage – it’s not food-related but it is very interesting, and despite the evocative name it’s actually about how neuroscience information influences people’s judgments about the credibility of explanations), and now I’m madly analysing data to get an abstract submitted for a conference in November prior to leaving for a conference coming up in September (yep, it’s overseas-trip-time for me again, for the fourth time in 18 months – more on that later too, I guess).

SO! Another one to file away in the “how to influence people’s perception of food” category – can you influence people’s perception of food by manipulating something as simple as the weight of the dish it’s served in?

YES! It seems…

In a study by Piqueras-Fiszma and colleagues, participants were asked to rate 3 yoghurt samples in terms of flavour intensity, density, price expectation and liking. The yoghurt samples were served in three bowls that were all identical except for their weight: they were white ceramic bowls to which a hidden weight could be attached, so that the lightest bowl was 375g, the intermediate-weight bowl was 675g, and the heaviest bowl was 975g. Each bowl was used to serve a 150g sample of plain old Greek yoghurt (important information for study replication: it was purchased from Tesco!).

Participants had to hold the bowl in one hand while they sampled the yoghurt, and each bowl was taken away before the next one was given, so that there was never a chance to directly compare the weights.

Of course, it was the exact same yoghurt in each of the three samples that each participant tried – the only difference was the weight of the bowl the yoghurt was served in. But the participants didn’t know that. They probably assumed they were taste-testing different yoghurts.

Results showed that the heavier the bowl, the higher people rated the perceived density of the yoghurt when they sampled it, and the more money they expected to pay for it. Also, the heavier the bowl, the more participants liked the yoghurt. Perceived flavour intensity was not significantly affected by the weight of the bowls.

So everyone go out immediately and buy really heavy bowls and plates! Serve your things in heavy dishes all the time! Force your dinner guests to hold the crockery in their hands as they eat! Make people think your food is more fancy, more expensive, more lovely!

Or not. These results probably can’t be generalised too much – maybe you only get this effect for particular foods, and yoghurt just happens to be one of them. Maybe the effect changes depending on other properties of the food, e.g. maybe the volume of yoghurt is difficult to visually gauge because it’s just a big amorphous lump, so people’s perceptions are only influenced when the amount of food is difficult to get an idea of. Maybe putting something a bit more discrete, like an apple, into the different bowls would get a different result.

And it would be interesting to do a study investigating whether the weight of the vessel influences the amount of food eaten – maybe if the bowl is heavier, the food seems denser, and perhaps the brain is tricked into thinking the food is more calorific and so satiety occurs sooner. Who knows? I mean, there are many cues that the brain uses to determine satiety, so maybe it would be difficult to detect an effect of serving dish weight, but it is intriguing…

Anyway, it’s brownie time. Or blondie, as the case may be, I don’t know what defines one versus the other. Do blondies simply lack the predominance of cocoa or milk/dark chocolate in the batter? Anyway, the photos above are of some ridiculously delicious blondies (we’ll go with blondies for the name), featuring one of the greatest things known to humankind: peanut butter. They are peanut butter and waffle blondies (they have waffle crumbs throughout them) with dark chocolate chips. I recently made a variation of the recipe, photos of which are below: peanut butter, waffle and malt biscuit blondies with white chocolate chips and strawberries. I made those for a friend who recently endured an incredibly difficult experience (understatement of the century) and has just gotten out of hospital.

Recipe for peanut butter blondies (and variations thereof)…

Panko panko panko

This might be the most difficult post I’ve ever had to write. Not because of the hours involved in devising recipes, cooking them, photographing them, going through countless databases and articles in order to find some interesting food-related scientific research to write about, actually writing the post, etc. The difficulty is mainly because of the title of the paper I’m writing about:

Overcoming the urge to splurge: Influencing eating behaviour by manipulating inhibitory control.

The 5th word in the title: absolutely the worst word in the English language.

I hate it so much. Who would have thought that a single word, by virtue of the mere sound of its phonemes, could conjure up such grotesque imagery and such a visceral revulsion in a person? I don’t understand how that word isn’t onomatopoeic for the sound of vomiting. It makes me want to vomit, it sounds like the noise people make when vomiting – it would be put to good use if it were employed in that manner.

I also hate the meaning that it stands for: some sort of unjustified failure of self-control, a lapse in good judgement, a lack of consideration and deliberation. People use the word in and of itself to try to attempt to justify their unjustified behaviour, as if the fact that the behaviour has a name means that it’s legitimate. Oh I just spent $900 on an antique Royal Doulton “Bunnykins” tea set from 1954, someone (I don’t know who) might say, but it’s OK – it was a… splurge. OH MY GOD. I would prefer if people just owned up to it and said they made an impulsive, spontaneous, poorly informed decision, but hey, it felt good at the time. It would be even better if they just made an informed, thoughtful, deliberate decision after a long and critical analysis of the pros and cons, but hey – humans as a species are nothing if not impulsive and irrational.

And I should know. Well, I should know something about it, to some limited extent. One of the things I study in my PhD is the ability to control behaviour, albeit in a bit of a simplified way: I use simple computer-based tasks that test people’s ability to withhold a response that they’re used to making. So my research participants press a button as quickly as possible to indicate whether they’ve seen an X or an O flash up on the screen, but sometimes, infrequently, a red box flashes up around the X or the O and that means the participant has to withhold the response they were going to make. This is a measure of a person’s inhibitory control. The greater your inhibitory control, the less trouble you have inhibiting your response when you see that red box, and this is something that naturally varies from person to person. A pretty straight-forward task, seemingly, but incredibly important in daily life – you have to constantly evaluate your environment and change your behaviour accordingly. You don’t want to be driving through a red light because you have trouble inhibiting your foot from pressing the accelerator.

But wow – I never anticipated the potential power of this simple computer task to actually improve people’s inhibitory control in a way that might improve their eating habits. All I needed to do was chuck in some photos of things like cupcakes or chocolate and I could have been doing all my participants a huge favour (and simultaneously wrecking my actual PhD project).

This is what my task looks like:

On trials like the one in the top stream, an X or O pops up and the participant has to respond as quickly as possible to identify the letter. But on trials like the one in the bottom stream, an X or an O pops up followed a fraction of a second later by a red box, the stop-signal, which indicates to participants that they need to withhold the response they were going to make.

If I adapted my task to be like the one used in the urge to spl*rge study by Houben, this is what it would look like:

Sometimes participants might see the top stream – M&Ms flash up and they press a button to indicate that they’ve seen a food item (the alternative is that they might see something like a chair flash up and they press a button to indicate that they’ve seen a non-food item). Other times, participants might see the bottom stream – chips flash up and they go to press a button to indicate that they’ve seen a food item, but then they hear an auditory tone that tells them that they have to withhold that response that they were going to make (again, the stop-signal).

The manipulation here is that one of the 3 different food items that are displayed (chips, nuts or M&Ms) is always paired with the stop-signal, one is never paired with the stop-signal, and one is only paired with the stop-signal 50% of the time. This is to see if participants might be pairing their response inhibition with a particular type of food and perhaps through associative learning they might become better at inhibiting responses to that food in general – maybe if they’re used to inhibiting their response every time they see M&Ms, and you give them some real M&Ms, they won’t be so impulsive and eat so many of them. Maybe.

And yeah, you kind of get that. If you take a subset of participants who have particularly bad inhibitory control (they’re more than 1 standard deviation below average for inhibition speed), you can actually decrease their consumption of the food that was always paired with the stop signal, compared to the food that was paired with the stop-signal 50% of the time (the control condition). So if they were always inhibiting their response to M&Ms but not to chips, after the experiment, they’ll eat fewer M&Ms than they perhaps otherwise would.

However, you have to be careful with this one – it has the power for both good and evil. If you take a subset of participants who have particularly good inhibitory control (more than 1 standard deviation above average for inhibition speed), it seems there is a trend towards them eating more of the food that was never paired with a stop-signal compared to the food that was paired with the stop-signal 50% of the time. So they get used to responding quickly and maybe impulsively to a particular food, and when you actually give them that food after the task, they eat more of it than they perhaps otherwise would.

There you go – if for some reason you desperately want to improve your ability to resist a particular food (have you had to take out a personal loan to finance your Valrhona habit?), all you have to do is spend a few thousands dollars on a computer and some task-programming software like E-Prime or Presentation, spend days or possibly weeks learning how to program a response inhibition task containing photos of the food you want to resist and voilà – if you’ve got poor inhibitory control to start with, you might be better at resisting that food, at least immediately after doing that task. Who knows how long-lasting the effects are? Actually, that would be an interesting study to do.

In the meantime, here are some maple syrup and panko biscuits. For those who don’ t know, panko are Japanese breadcrumbs. Panko’s potential for use in sweet recipes was first brought to my attention by the Chuao Chocolatier panko chocolate bar I tried in San Francisco – I found the bar kind of disappointing because try as I might I could not detect the panko flavour, but I believed it could still work well if done differently, which is why I came up with these biscuits. They’re delightfully crunchy and the panko gives them a nice wheaty flavour and trust me, you do not want to inhibit your eating of them.

Recipe for maple syrup & panko biscuits…

Chocolatuuuunnnnggghhfffffff.

Chocolate! Oh my god, chocolate. Chocolate! Master/mistress of our hearts! Emperor/empress of all sweet things! Ruler of our minds! Don’t you know, it has chemicals that make us happy! It has compounds that are the ones that also flood your brain when you’re in love! Or during moments of passion! Chocolate contains chemicals of passion! And love! And happiness! And joy! And all of that! Right? Right?

(Can you just imagine my withering gaze right now? You should get a chill up your spine.)

It’s not right, actually. Of course it isn’t. Of course. It’s never that simple. But saying that chocolate contains happy/lovey-dovey/alluring chemicals is a nifty thing to say, even if it’s not necessarily true, or not true to the extent we think it is. It plays into our beliefs, seems to confirm our existing expectations, and makes us all feel warm and fuzzy and comforted – chocolate will always be there for us, to hold us and nurture us and to spur us to neglect urgent work in favour of going on an epic mission to find chocolate, to console ourselves in its melty embrace, to search desperately through the pantry until we are totally convinced there’s not any chocolate in there, and then to search again just to make sure, and then to walk or drive or ride in a horse-drawn carriage to an appropriate vendor of chocolate-related delights.

But what say you, Science? (Science should be the true master/mistress of everyone’s hearts, OK?)

Science says things just aren’t that clear. Unsurprisingly. Oh, Science. Oh, oh, Science.

Firstly, the most widespread inaccuracy: that a chemical in chocolate – phenylethylamine – is the chemical that also responsible for the feeling of love in the brain, or the feeling of passion, or whatever the latest churnalism article in the Daily Mail (yeah, I’m not even in the UK and I’m going to point my finger at it) or whatever hobbling excuse for a newspaper has regurgitated.

While it is true that chocolate contains phenylethylamine, and that phenylethylamine may have effects on cognition, mood and emotion, very little of the chemical could ever get to the brain, since when it is consumed, most of it is converted into a different chemical by the protein monoamine oxidase-B (Suzuki et al. 1981). Barely any phenylethylamine makes it into circulation in the blood and from there into the brain.

Oh and high levels of phenylethylamine are associated with psychosis, particularly in certain individuals with paranoid schizophrenia (Janssen et al. 1999). Just out of interest.

So everyone can quit jabbering about phenylethylamine. It is really not the magical aphrodisiac love-drug that people make it out to be.

Right, moving on. Theobromine! That’s the next chemical to which people ascribe chocolate’s beautiful, bewitching powers.

There’s actually a little bit more validity in that one, although it’s still not totally resolved and is open to debate.

In one study (Smit et al. 2004), participants were given either:
a) pills containing an amount of cocoa power that would have the same theobromine and caffeine amounts as a 50g bar of chocolate
b) pills containing the same amount of theobromine and caffeine as a 50g bar of chocolate (plus some inactive cellulose as filler)
c) pills containing just inactive cellulose as filler (the placebo condition, naturellement)

Participants in conditions (a) and (b) had faster simple reaction times and reported feeling more energetic than those in the (c) condition. That certainly suggests that theobromine and/or caffeine could be having psycho-active effects: they’re doing something to the brain. But it remains unclear whether it’s just theobromine that’s responsible for the effect, just caffeine, or a combination of both.

To make the conditions of the study a little more relevant to the real-world, doses of theobromine/caffeine were also administered in a second experiment using actual pieces of chocolate. The chocolate was made so that it either had no theobromine/caffeine in it (like white chocolate), a small amount of theobromine/caffeine (the same as milk chocolate), or a larger amount of theobromine/caffeine (the same as dark chocolate). When participants ate the chocolate with the large amount of theobromine/caffeine, it greatly improved reaction time. Chocolate with the small and large amounts of theobromine/caffeine improved working memory (tested by getting participants to press a button when they saw 3 odd or even numbers flash up in a row in a constant stream of numbers on a computer screen). The large and small doses of theobromine/caffeine did not result in participants feeling more energetic compared to the chocolate with no theobromine/caffeine or compared to placebo.

Both of the experiments also found, or rather, didn’t find, something rather interesting: there was no strong, clear effect of theobromine/caffeine on “hedonic tone”, a measure of the participants’ levels of contentment and pleasure. So it seems a bit tenuous that theobromine is this magical chemical that so many people assume must be in chocolate in order for chocolate to exert its effects upon us. But people want there to be a chemical in chocolate that explains why so many people love it – I think people get a kick out of imagining that they’re messing with their brain chemistry in order to elevate their mood or to experience enjoyment, thereby associating chocolate consumption (rather loosely) with illicit drug-taking and getting a kick out of the fact that they’re doing something vaguely taboo like that, albeit on a very diluted scale.

Maybe we should, then, look at what other possible ways chocolate could have established its lauded position as reliable and beloved comforter to many. Is it the chemicals? Or is it something else? Or is it the chemicals and something else?

But maybe I’ll get into that next time.

(See, if I leave you with this tremendously epic cliff-hanger, I’ll feel bad for not following it up again soon, so that will give me the impetus to blog again. It’s like training a puppy! But not really like that at all.)

The photos in this post are of the birthday cake I made for Dr Tash PhD last year, and I tried to make said cake as sour as possible (per her predilection for sourness). I can’t actually remember how I made this cake – I think it was a lemon drizzle cake sandwiched in between two layers of white chocolate mud cake, but I could be wrong. I also have no idea what the purple stars are, except that they were some sort of sour agar gel I made and are not, contrary to appearances, slices of beetroot. Just so you know.

References
Suzuki et al. 1981. Oxidation of beta-phenylethylamine by both types of monoamine oxidase: examination of enzymes in brain and liver mitochondria of eight species. Journal of Neurochemistry, 36(3), 1298-1301.
Janssen et al. 1999. Does phenylethylamine act as an endogenous amphetamine in some patients? International Journal of Neuropsychopharmacology, 2(3), 229-240.
Smit et al. 2004. Methylxanthines are the psycho-pharmacologically active constituents of chocolate. Psychopharmacology, 176(3-4), 412-419.