Date: February 4th, 2010

Speaker's Name and Affiliation:

• Margriet Westerterp
• Maastricht University Medical Centre

Seminar Title: "Representation in the brain of stress induced food choice in the absence of hunger"

Chair: Harry R. Kissileff, Ph.D.

Rapporteur: Kathleen L. Keller, Ph.D.

Attendees and their Affiliation:

Kathleen Keller	Columbia/Obesity Research Center
Harry Kissileff	Obesity Research Center
Karen Ackroff	Brooklyn College
Richard Bodnar	Queens College
Susan Carnell	Columbia
Virginia Chang	Columbia
Swana de Gysel
Chungwon Kim	Columbia
Roni Friedman	Columbia
Laurence Nolan	Wagner College
Miranda Johnson	NJMS
Annemarie Olsen	NYORC
Barry Levin	NJMS
Gerry Smith	WMC
JA Grinker	U of Michigan
John Glendenning Barnard
Tony Sclafani	Brooklyn College
Marcia Vitolo	UFCSPA-Brazil
Paula Campagnolo	UFCSPA-Brazil
Majella Okeeffe	Columbia
Carol Maggio	NYORC
Joe Vasselli	NYORC
Allan Geliebter	NYORC
Muhammad Mansour	IHN-CUMC
BT Walsh	CUMC-PI

Summary: (Prepared by the Rapporteur)

Human eating behavior may be influenced non-homeostatically by the rewarding value of foods, i.e. 'liking' (pleasure/palatability) and 'wanting' (incentive motivation). Stress may diminish the rewarding value of food, i.e. 'liking' and 'wanting', possibly leading to eating in the absence of hunger. The first study assessed effects of acute stress on the rewarding value of food in normal weight (NW) and visceral overweight (VO) subjects. Per test-session, conducted randomly in a stress or rest condition, 'liking' and 'wanting' for 72 items divided in six categories (bread, filling, drinks, dessert, snacks, stationery (control)) were measured twice, each time followed by a wanted meal. Appetite profile, heart rate, mood state and level of anxiety were measured. Mean 'wanting' for items decreased pre- to post-meal in rest in NW and VO, and in stress in NW (p<0.001). VO showed higher fat and carbohydrate intake during the second meal (in the absence of hunger) under stress vs. rest (p=0.02). The difference (stress-rest) in 'wanting' post-meal and the difference (stress-rest) in energy-intake for the 2nd meal was higher in VO than in NW (p<0.05). To conclude, in VO stress prevented a decreased 'wanting' and promoted additional energy intake in the absence of hunger, in contrast to NW.

With respect to assessment of liking of foods during stress compared to rest as a function of body-weight, thirty-nine normal weight and visceral overweight subjects tasted 67 food items randomly during rest and stress, and rated pleasantness of taste and characterized food perception using visual analog scales. Pleasantness of taste was decreased similarly during stress in both groups. Also food perception was decreased during stress compared to rest. The correlation (p<0.01) between pleasantness of taste and characterization of food perception was not different between visceral overweight and normal weight, neither during rest or stress. Pleasantness of taste scored continuously higher in normal weight compared to visceral overweight (p<0.01). Characterization of food perception was more pronounced in normal weight than in visceral overweight (p<0.01); normal weight rated sweetness higher (p<0.01). Conclusion: Stress reduced pleasantness of taste similarly for visceral overweight and normal weight subjects. Pleasantness of taste is always lower in visceral overweight; the decrease in characteristics of food perception during stress is smaller. With respect to related brain activity, the effect of acute stress on food choice and food choice reward was determined using a similar set-up as in the first study. Energy intake in the absence of hunger, under stress, implied more food intake characterized as crispy and rich in taste (p<.05). Then brain activation was reduced in reward-areas: Amygdala, hippocampus, cingulate cortex, and putamen. Conclusion: Reward signaling and reward sensitivity was significantly lower under stress, coinciding with increased energy intake from food choice for more crispiness and richness in taste. The changes in putamen activation may reflect specifically decreased reward prediction sensitivity.

Question & Answers:

Q. What do you mean by the term 'visceral obesity'?
A. The visceral obese have higher waist to hip ratios (above the recommended cutoffs of waist circumference for men and women).

Q. What were the stressors you used?
A. A math test. One of the tests was with no distractions while the other test was very hard and did not have a solution. The latter was the stressful situation.

Q. What was the interval between the 1st and 2nd meal?
A. 2 hours

Q. What size meals did you use?
A. The first was a 'normal' breakfast. The second meal was extremely small. When the subjects were served this second meal, they were not hungry.

Q. Was the food the subjects selected then served to them as their meal?
A. Yes, immediately after the test, subjects got the foods they chose and they needed to consume all of them.

Q. What kind of units did you use to measure work?
A. We used points earned for the memory games and the amount of correct responses.

Q. What did you do if people didn't eat all of their foods?
A. Everyone finished their foods, so this was not an issue.

Q. Are you going to talk about stress differences in the visceral obese vs. the nonvisceral obese?
A. I will show the data on cortical later. The effects of the stressor (the math test) were not very strong.

Q. What does the 'wanting score' mean?
A. This is the points they earned with the memory test.

Q. What was the placebo condition?
A. The pictures of stationery (pens, pencils).

Q. Do you have a slide that shows the composition of the meals?
A. They consisted of 67 food items and the meals for the subjects depended on which of those food items they selected.

Q. The stress is only affecting the second meal. How do you explain that?
A. They weren't hungry in the second meal, only in the first. This may be the explanation.

Q. The types of carbohydrates you offered to the subjects differed. You have a sugar rich stimulus versus a complex carbohydrate rich stimulus. How might this have affected the results.
A. Most of the subjects ate the breads. They hardly selected the sweets. I will show you what the subjects chose in a bit.

Q. What does 'fullness of taste' mean?
A. Usually these are items such as cheese, meaty, tomato. They are not always savory items. Fullness of taste means 'richness of taste.'

Q. If you would have switched the protocol to later in the day, do you think the macronutrient choices would have changed?
A. Yes, probably.

Q. Did the subjects expect to eat the foods after they selected them?
A. In the wanting test, subjects had to eat the wanted food items. In the liking test, we used a 'sip and spit' procedure.

Q. What happened during the fMRI?
A. The stimuli used were the same pictures of foods that subjects were presented with in the other protocol. As pictures were presented to subjects, we asked questions like 'What food do you like best' or 'What do you want to eat just now?' Subjects were shown a randomized sample of the entire set of pictures.

Q. What was your fMRI protocol?
A. We did both region of interest testing and general statistical mapping.

Q. How did you assess wanting and liking?
A. By asking subjects which brand they liked better, one vs. the other, and by asking subjects how much they wanted a particular item.

Q. These are all aligned structurally?
A. Yes.

Q. Is this the same procedure that Finlayson & Blundell use?
A. It is different. He asks for 2 types of liking.

Q. It seems like a lot of your findings are in the white matter.
A. (Answer not given)

Q. How many Regions of Interest did you look for?
A. 5-6, I think

Q. I thought in the first study you only found a relationship in the visceral obese but in the second study your findings are in normal weight subjects. Do you know why?
A. Yes, that's true. The second experiment was more about what subjects ate, as opposed to what they wanted to eat. There could have also been more stress related to the fMRI procedure.

Q. What does 'time' in this case mean?
A. Time means before and after the meal.

Q. Did you have any way of evaluating 'chronic stress?'
A. We did physical stress tests and we looked at more vs. less stressed individuals, but we did not find any effects.

Q. If you are making a choice in advance, how can you predict the rewarding value of a food?
A. These foods were all highly familiar.

Q. Was there any change in the value of non-food items as a function of stress?
A. Yes, the ratings of the non-food items go up, but only in relation to the ratings of the food items, which go down.

Q. What is your conception of the difference between wanting and liking?
A. In most cases, I believe that they go together. But, for highly addictive foods, wanting can increase but not liking. In highly restrained individuals, liking can be high without wanting being high.