Wednesday, 6 July 2016

Just my imagination

Every first year psychology student learns the difference between sensation and perception, although whether or not this distinction has any long-term impact is open to debate. Psychology students are probably like the rest of us. We tend to imagine that our perceptions are somehow a true picture of the outside world, and that our cognitions - memories, thoughts, and imaginings - belong to another part of the brain which doesn't really interfere much with what we perceive.

And yet it is our cognitions (“top-down processes” as the psychologists say) that give the meaning to what we perceive. If we taste something, for example, the tasting experience will be examined in the light of a mental model of that taste that we already have – essentially a prediction or expectation of the taste based on our memory of similar experiences or other sources of information (perhaps from advertising or packaging, and so on). This tells us two things. The first of these is that we can’t really understand what we experience and enjoy (or not) without knowing something about these top-down processes

More specifically, it tells us that we carry around mental models of our experiences, and this raises the question of just how “real” these internal representations of the stuff that’s out there really are. This, it must be said, is not a novel question, having been regularly asked by John Locke and others, as they shared pints at their local pub (The Empiricist & Bishop) in 18th century Britain. But more recently, and particularly in relation to tastes and smells, we have had some actual evidence about “imagined” sensory experiences.

In the early to mid 1990s, there were a number of studies published in what was known as memory psychophysics. The idea was to ask what would happen if you held mental image of sensations (memories), combined them mentally with actual physically-present sensations, and then rated the intensity of the combination. While, on the surface, this sounds like the ultimate waste of time, in fact it was one way of asking whether or not people explicitly or implicitly knew how different sensations interact. Dick Stevenson and I explored this idea in both mixtures of sweet and sour tastes and the hot sensations of capsaicin (from chilli) with these tastes [1]. Mixing an imagined taste (e.g., sourness) with a physically present sweet taste showed the same pattern of suppression (less sourness, less sweetness) as found when the two tastes are actually mixed together and tasted. But then, we consciously know this anyway - how do you make lemon juice less sour? More interesting was the fact that when participants imagined adding a burning sensation to sweet or sour tastes, only sweetness was rated as less intense. This is exactly what happens in actual mixtures – capsaicin only reliably reduces sweetness. But most importantly, the participants, when debriefed about their knowledge of such interactions, felt that any taste would be suppressed by the burn. In other words, they carried implicit information about the interactions, of which they were not consciously unaware.

A few years later, in an interesting exploration of the interactions of tastes and congruent odours – e.g., sweetness and vanilla – a group at McGill University in Canada [2] found that imagined odours could enhance the detectability of a sweet taste, in the same way that a physically present odour could. And in exactly the same way as a physically present odour, the imagined odour had to be congruent. So, imagined strawberry odour was effective in increasing sweet taste detectability (because strawberry odour carries with it the quality of sweetness, learned from previous experiences). An imagined ham odour, which is incongruent with sweetness, was not.

So, it appears that we can conjure effectively real sensory stimuli from imagination. Most fascinating of all though is how these cognitively-based stimuli can influence behaviour, including potentially our food preferences and choices. Again, to some extent, we know how influential thoughts can be.  Imagine sucking a lemon and saliva will start to flow; similarly, think about food close to dinner time to get the gastric juices flowing. This latter effect is consistent with data showing that thinking about food is associated with craving and, especially if you are restricting intake by dieting, succumbing to the temptation to eat. It seems that exposure to food as a mental image acts just like an initial bite, which tends to elicit an appetiser effect (“you can’t eat just one!”) [3].

A demonstration of the potential impact of mental imagery on actual food consumption was published in the prestigious journal Science in 2010 [4]. Noting that both sensory input and mental images can elicit similar responses across a range of different behaviours, these researchers examined whether repeatedly thinking about a food actually decreased desire to consume the food. In other words, can thinking act just like an actual bite or sip, in producing a type of sensory-specific satiety (SSS) effect? We know that the first bite or first sip is always the best, with pleasure decreasing after that [5]. In fact, losing pleasure in eating or drinking is just as much a reason that people stop eating as being full. A series of experiments using those much-loved research foods M&Ms and cheese cubes showed exactly this effect. For either food, multiple repeats (x 30) of thinking about eating led to less actual later ad-lib consumption than did fewer (x 3) thinking episodes. Important to the interpretation of these results was the fact they controlled for simple exposure to the image of the food – thus, imagining placing M&Ms into a bowl, as opposed to imaging eating them, had no impact on subsequent consumption.

How did these researchers reconcile the two potential effects of imagining food – the appetiser effect (increased desire to eat) and the SSS (reduced pleasure in eating) effect? A final study in their paper measured two separate reasons for eating, namely liking and wanting (see: Learning to want). Repeatedly imagining a food was found not to produce a decrease in liking for the food. So, in fact, the effect is unlike SSS which is based on reduced liking with repeat exposure. In addition, the researchers asked participants to perform a reinforcement computer game in which points in the game could be exchanged for consumption of cheese cubes. Exactly as would be predicted if the imagining effects were due to reduced motivation to consume (reduced wanting), those previously asked to imagine eating for the high multiple occasions (30 vs. 3) were less motivated to respond to obtain the cheesy reward.

The practical implications of research such as this for reducing overeating are pretty obvious. But, more generally, the studies tell us that the more we know about cognitive “top-down” processes that feed into our perceptions, emotions and motivations, the more we will understand food choices. What’s in the brain is just as important as what’s in the mouth.

1. Stevenson, R.J. and J. Prescott, Judgments of chemosensory mixtures in memory. Acta Psychologica, 1997. 95(2): p. 195-214.

2. Djordjevic, J., R.J. Zatorre, and M. Jones-Gotman, Effects of Perceived and Imagined Odors on Taste Detection. Chem. Senses, 2004. 29: p. 199-208.

3. Yeomans, M.R., Palatability and the micro-structure of feeding in humans: the appetizer effect. Appetite, 1996. 27(2): p. 119-33.

4. Morewedge, C.K., Y.E. Huh, and J. Vosgerau, Thought for Food: Imagined Consumption Reduces Actual Consumption. Science, 2010. 330: p. 1530-1533.

5.    Hetherington, M., B.J. Rolls, and V.J. Burley, The time course of sensory-specific satiety. Appetite, 1989. 12: p. 57-68.

Tuesday, 22 March 2016

Cueing for food

When was the last time you were overcome with a desire to eat a particular food? Did you say lunchtime, dinnertime, when I was last hungry and so on? Thank you for playing - next contestant, please. We may want to eat at these times, but it’s rare for such occasions to prompt a desire for something specific. On the other hand, walking past a bakery or pizzeria or curry restaurant can easily produce a desire to consume the source of the aromas we pick up. You’ll perhaps notice that I didn’t include vegetable market or health food shop among these examples, and there’s a good reason for that.

Food sensory cues - sights, sounds (sizzling, for example), smells - are signaling not just the presence of food, but also the food’s palatability. And they are everywhere. The multimedia environment in which we are all immersed in affluent societies means that cues signaling food palatability are present for pretty much the majority of the time we are awake, through television, advertising, and our urban ‘smellscape’.

This is not to say that ‘internal’ cues such as hunger are irrelevant to motivating us to eat. Indeed, hunger makes us much more vulnerable to ‘external’ food sensory cues which mainly come about through learning. That is, our bodies associate the ingested fats and sugars of foods with their flavours and aromas, which not only becomes highly liked as a consequence but also become a powerful trigger for the desire to consume, or wanting. This desire or drive is at least as much about hedonic hunger as it is about an empty stomach (see: Learning to Want for further detail).

It is important not to underestimate the role that cue-elicited wanting plays in overeating. Obesity is much less common in societies where regular mealtimes are observed and snacking is low. It is ‘non-essential’ eating, driven by a desire for food pleasure (read: fat, sugar, salt) that provides the greatest challenge to a health weight. We are all influenced to greater or lesser degrees by external food cues, and we will all tend to eat more of a palatable food than one that is not palatable. But the fact that cues can have such a powerful impact on eating has raised the question of whether those who have difficulty controlling their food intake are somehow more aware or more reactive or perhaps more vulnerable to external cues.

Even if you find chocolate or desserts highly appealing, you may or may not give in to the temptation to indulge when either is available. The idea of eating restraint - habitual monitoring and control of food intake – seems to be important here. Food cues (the sight or smells of a pizza, or even thoughts about food) exert their greatest influence on desire to eat if you are a restrained eater – this isn’t very surprising if you consider that restraint means denying yourself food pleasures.

It is easy enough to think of restrained eaters as wildeyed and salivating, forever battling with urges to eat. But of course restrained eaters as a group must include both those who are successful at resisting these urges – successful dieters – as well as those who succumb. Moreover, even if you are not a restrained eater, you may still be someone who tends to give in to the temptations posed by the sight or smell of your favourite foods. Thus, another characteristic – disinhibition – has been implicated in determining responses to food cues. Many of us have tried to lose weight at some stage. This will often involve a considerable effort to inhibit your desire to consume the palatable foods that you will encounter. If you are effective at doing this, then you can be said to be low in disinhibition (the failure of inhibition); conversely, if resistance is futile, then you can be seen as high in disinhibition.

Irrespective of its origin, when measured on the Three Factor Eating Questionnaire (the other factors are restraint and hunger) (1), up to 50 per cent of women in the USA show high disinhibition, and this percentage is highest in young women. Estimated rates in men tend to be much lower, perhaps around 25 per cent, depending on age and other demographic factors.

Disinhibition is not just about reacting to all foods at all times. Those who show high levels of disinhibition are especially responsive to cues that signal the calorie content of foods and thus prone to ‘wanting’ highly palatable foods: sweet foods and beverages and food high in fat, such as butter, cream and ice cream. And because of the nature of the trait, they tend to act on these desires when palatable food is available and consume more. Those high in disinhibition are therefore not simply gluttons who want to continue to eat, but individuals whose responses to sensory cues are exaggerated.

Needless to say, degree of disinhibition is thought to be crucial in determining successful dieting. It is highest among those dieters who have most trouble maintaining their ideal weight and in dieters who are most bothered by food images that lead to craving. It is, for example, associated with increased chocolate intake by those who identify themselves as ‘chocolate addicts’.

To be able to understand the impact of food cues on eating, disinhibition and restraint need to be considered together. If you are low in disinhibition, it is more likely that you will be a successful restrained eater. If you are not trying to restrain your eating, then your level of disinhibition might seem to be of less concern. But if you do not restrain your eating and you are high in disinhibition, then you will still be highly responsive to palatable food cues. 

How these interactions might influence reactions to palatable foods is illustrated by a study (1) in which a group of women were offered their most tempting snack food – including chocolate, crisps, candy, nuts and cookies. The women were firstly asked to taste a small amount of the snack, and then were given the remainder to take home and asked to come back a day later. Half of the women were instructed not to consume any of the snack until they returned the next day (the temptation condition), while the other half of the group were given no instructions (the control group). On returning to the laboratory the next day, all women were offered their snack to eat and the amount they consumed was measured. Regardless of the group they were in, those who measured low in restraint or those high in restraint but low in disinhibition ate approximately the same amount as they had on the previous day. By contrast, those high in both disinhibition and restraint ate significantly more if they were in the temptation group.

The combination of restraint and a strong tendency to disinhibition therefore made these women particularly vulnerable to the temptation of a palatable food that had been restricted. This is, of course, exactly the set of circumstances that those who experience cravings report during periods of dieting. It is not completely clear as yet whether eating restraint is a risk for weight gain if you are disinhibited or disinhibition is a risk if you diet a lot. Increasingly, though, studies are teasing apart the relative contribution of these factors.

Most recently, Martin Yeomans at the University of Sussex (2) has been characterizing those high in food disinhibition, finding that their responses to food may be in fact a reflection of a more general, underlying personality dimension. He pre-exposed participants varying in both restraint and disinhibition to pictures of ‘unhealthy’ foods and measured their responses on a number of tests of general impulsivity (included a test of willingness to delay reward: $2 now or $10 next month?) and risk-taking. Although none of the tests were specifically related to food or eating, after exposure to food cues, high disinhibition women were more impulsive and higher in risk-taking. In contrast, those high in restrained eating showed no such effects, irrespective of exposure to prior visual cues.

Thus, the trend that is emerging is that disinhibition is the important personality factor that predisposes to weight gain, and this is due to how disinhibition alters your response to food cues. Consistent with this, disinhibited eaters consume far more of an available snack following a high carbohydrate meal than do those low in disinhibition. Moreover, this overeating was not driven by hunger but rather by the hedonic properties of the snack. While it might seem that asking such individuals to restrain themselves is a sensible approach, it is clearly asking for trouble, with the risk of failure being high.


1. Stunkard AJ, Messick S (1985) The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research 29: 71–83.

2. Soetens, B., Braet, C., Van Vlierberghe, L., & Roets, A. (2008). Resisting temptation: Effects of exposure to a forbidden food on eating behaviour. Appetite, 51, 202-205.

3. Yeomans, M. R., & Brace, A. (2015). Cued to Act on Impulse: More Impulsive Choice and Risky Decision Making by Women Susceptible to Overeating after Exposure to Food Stimuli. Plos One, 10, e0137626.

4. Chambers, L. & Yeomans, M.R. (2011) Individual Differences in Satiety Response to Carbohydrate and Fat. Predictions from the Three Factor Eating Questionnaire (tfeq)’, Appetite, 41, 316–23.

Wednesday, 13 January 2016

Yule tired

So, Christmas was fun, right? It’s always filled with light hearts and good humour! And the food – turkey is such a surprise and so creative and delicious! If, at this stage, you are nodding, then you may very well be in a minority. We know that family tensions increase at Christmas, and so do suicides. For whatever reasons, many of those who celebrate Christmas day without the company of their parents or extended family view themselves as having a lucky escape. It’s not about being in the depths of winter either. Those in the USA sensibly organize their Christmas meals in early November to coincide with the date when local indigenous tribes took pity on the local settlers from England, recognizing that they were Puritans and hence unlikely ever to organize their own celebrations. If you are raised in an Anglo-Saxon culture in the southern hemisphere, it is even worse of course. A huge winter meal during a 35oC day is not culinary experience one dreams about. And yet it is commonplace, and produces similar emotional consequence.

Of course, children love Christmas – there’s lots of colour and movement, there are presents to break, and lots of cakes, puddings and pies to ramp up blood sugars to maximum. Sometime around puberty, though, the presents start to become less exciting (more Lego …. great!) and there comes a belated realization that the food is most likely exactly what it was last year and at every previous Christmas. In Victorian times and in northern hemisphere winters there was probably nothing nicer than tucking into a goose (where’s the meat?) or turkey (is it really meant to be this dry?). But does anybody really now look forward to such foods that we could eat at any time of the year, but interestingly choose not to? And let’s not forget the groans that commence after the meal. Who’s more stuffed – you or the turkey?

This brings me to the point that popped into my own head during a recent mince pie and Madeira binge. We have a pretty good idea of how food preferences are formed (if you don’t, may I recommend [1]), namely through different forms of associative learning following a small number of innate preferences and dislikes present at birth. It is pretty easy to see how children would learn to love the Christmas feast. But let’s look at the context for the meal, since we know that this is an important factor in preference development [2].  Given the unreasonable expectations of the day, the family feuds, the obnoxious drunk uncle, the disappointing presents, why do we clearly want to eat the same foods again and again?

Perhaps it’s comfort food? (see: Tastes Like Home) Certainly, that might be part of it, but why isn’t this undermined by the often uncomfortable context and the questionable gastronomic qualities of the food itself? If emotions are so important in determining food preferences [3], why don’t Christmas family feuds produce food dislikes?

In fact, we know far less about acquiring food dislikes than we do about how our preferences are formed in the first place. And one good reason for this is that … perhaps surprisingly … acquired food dislikes are not very common. Of course, pairing flavours with illness produces long-lasting ‘taste’ aversions, but this accounts for only a small fraction of the foods people say they don’t like. The effects of dietary monotony or boredom with particular foods are well known [4], but this seems to be about regularity of consumption and not the development of an actual dislike. Something similar occurs during a meal (sensory-specific satiety) when the first sip or bite of a food tastes best.

When we consider most food dislikes we are really talking about those foods that either haven’t been tried more than once or that are disliked for purely moral or cognitive reasons such as meat for a vegetarian (or tofu for everyone else). It becomes quite difficult to acquire a food dislike once it has been regularly eaten and a preference has formed because the body has learned through these eating occasions both that the food will not poison us and also that it carries calories or some other valued nutrient. Even ‘empty calories’ are still calories.

The learning processes through which food preferences are formed are known as evaluative conditioning (EC). EC is much like classical (aka, Pavlovian) conditioning (CC) in that it involves pairing of something neutral with something to which we already have a like or dislike. The effect is to make the neutral thing into something that evokes a response. So, the sound of a bell paired with an electric shock repeatedly leads to the bell eliciting a response as though a shock were about to be delivered, initially even when the shock is removed. Similarly, a novel flavour paired with sugar produces a liked flavour. Pairing flavour with bitter tastes might be expected to produce dislikes, but if we do this often enough then the food’s energy will offset this effect and increase liking for the flavour (known as post-ingestive or flavour-calorie learning).

One of the few ways in which EC and CC differ may account for the relative lack of learned dislikes. When the pairing is stopped, a process known as extinction (in CC) occurs. Having established that the bell is a useful signal for an upcoming shock, repeatedly presenting the bell without the shock will eventually lead to a recognition that it has stopped being a good signal and there’s no point in responding as though a shock were about to happen. In contrast, once a flavour has been paired with a liked sweet taste it remains liked, even if it is never (as far as we know) again paired with sugar [5]. Your first thought will be that it’s all down to the relative importance of these processes to survival. Unfortunately, this doesn’t work. EC occurs not just with food flavours but even in cases where there is no survival advantage (e.g., pairing a liked picture with one that is neutral); conversely, predicting that red lights signal oncoming traffic clearly helps avoid an untimely demise.

There is some very limited evidence that post-ingestive learning – that is, a flavour paired with calories – can extinguish.  In one study with humans, considerably reducing the energy content of a familiar, preferred meal led to a pronounced decline in liking over repeated eating occasions, as compared to the same meal without the energy reduction [6]. This seem s to suggest that, as with CC, the participants in the study were learning something new about these food flavours and their ability to signal upcoming energy.

But in the absence of low-cal turkey and pudding, it appears we are stuck with our Christmas preferences and food habits. Which suggests very strongly the power of early learning against anything that the adult yuletide experience can throw at it.

1.         Prescott, J., Taste Matters. Why we like the foods we do. 2012, London: Reaktion Books.
2.         Meiselman, H.L., et al., Demonstrations of the influence of the eating environment on food acceptance. Appetite, 2000. 35: p. 231-237.
3.         Spinelli, S., et al., Emotional responses to branded and unbranded foods. Food Quality and Preference, 2015. 42: p. 1-11.
4.         Meiselman, H.L., C. deGraaf, and L.L. Lesher, The effects of variety and monotony on food acceptance and intake at  a midday meal. Physiol. Behav., 2000. 70: p. 119-125.
5.         Baeyens, F., et al., Once in contact always in contact:  Evaluative conditioning is resistant to extinction. Advances in Behavioural Research and Therapy, 1988. 10: p. 179-199.
6.         O’Sullivan, H.L., et al., Effects of repeated exposure on liking for a reduced-energy-dense food. Am J Clin Nutr, 2010. 91: p. 1584-1589.