Hedonic g gLMS: a n new scale t that pe permits valid hedonic c - - PowerPoint PPT Presentation

Hedonic g gLMS: a n new scale t that pe permits valid hedonic c comparisons

Linda Bartoshuk Smell & Taste Center

lbartoshuk@dental.ufl.edu 352-273-5119

International Citrus & Beverage Conference (ICBC) September 14-17, 2010

Outline of talk

• 1. Introduction to supertasters
• 2. Problem: Conventional scales make

invalid comparisons across subjects (i.e., scales cannot “see” supertasters).

• 3. Solution: It’s not easy, but valid

comparisons are possible.

Do we all live in the same taste worlds?

NO NO

Super ertaster ers experience more intense oral sensations than the rest of us.

PROP DEMONSTRATION

• PROP is a medication used to suppress

thyroid function in patient’s with Grave’s

• disease. Typical dose: 200 mg/day.
• PROP paper: about 1.6 mg of PROP

This tongue has been swabbed with blue food coloring. The structures that house taste buds (fungiform papillae do not stain as well as the rest of the tongue so they appear lighter. Keep a mental picture of this tongue.

This is the tongue of a su super ertast ster er: a person born with an unusually large number of fungiform papillae.

• We count the papillae in

a circular template 6 mm in diameter just to the side of the midline.

Supertasters can have up to about 60 fungiform papillae in the template area. Most individuals have far fewer. This individual has 16 fungiform papillae in the template area. Values as low as 5 are normal.

Magn agnitud ude Es Estimat ate Tas Taste Int Intens nsity y (n (nor

• rmalized

lized t to t

• ton
• nes)

es) Magn agnitud ude E Estimat ate Bit Bitter erness o ess of . f .0032 M 0032 M P PROP (no norm rmal alized t d to t tone

• nes)

10 20 30 40

Cit itric ic A Acid id

(r=.22, p p<.05)

Suc ucros

• se

r=.46, p p<.0001 10 20 30 40 50

QHC QHCl

r=.62, p p<.0001

NaCl

(r=.49, p p<.001) 10 20 30 40 50 10 20 30 40

TA TASTE: Supertasters have the most fungiform papillae (the structures that house tastebuds) so they have the most tastebuds and thus perceive the most intense tastes.

ORAL BURN: Tastebuds are surrounded by fibers thought to mediate

• ral burn; thus supertasters perceive greater burn from oral irritants

like chilis.

300 300 200 200 100 100 r = = . .56, 56, p < p < . .0001 0001

Or Oral al b bur urn Bit Bitter erness o ess of . f .0032 M 0032 M P PROP

100 100 200 200 300 300 400 400 500 500 600 600

100 ppm 100 ppm c caps psaicin

Snyder Senior Essay, Yale University, 1996

Jalapeno peppers contain about 100 ppm capsaicin.

Duffy, Bartoshuk, Lucchina, Snyder &, Tym, 1996.

ORAL T TOUCH: Fungiform papillae are innervated by fibers mediating touch; thus supertasters perceive more intense touch sensations (e.g., creaminess) from fats

Creamines ess

hea eavy cr crea eam r = = .44, p < .44, p < .001 .001

10 100 90 90 80 80 70 70 60 60 50 40 30 20 20 10 10

NT M MT S ST

In addition, there is a connection between supertasting and flavor (flavor is taste plus RETRONASAL OLFACTION)

Food and Beverage Industry

• Has long known that adding a taste to a

beverage (e.g., adding sugar) will intensify the perception of the flavor (e.g., orange) of the beverage.

• Similarly, experiencing greater taste (as do

su supertast sters) will intensify the perception

• f flavor.

Nontasters Supertasters

Retronasal

Orthonasal Orthonasal Retronasal

• Supertasters live in a

neon eon food world.

• Those with the fewest

taste buds live in a

pa pastel el food world.

I’ve compared supertasters to others with regard to how they perceive the food world. But we cannot share each other’s experiences directly. How did I make comparisons between supertasters and others?

I did not

• t use the classic Natick

9-point scales (or any similar scales like VAS).

Natick 9-point scales

1 2 3 4 5 6 7 8 9 extremely strong extremely weak

Sensory Hedonic

1 2 3 4 5 6 7 8 9 like extremely dislike extremely neutral

The problem.

• What does “ext

extrem emel ely strong” mean?

• What does “like ext

extrem emel ely” mean?

• Let me begin with the sensory descriptor:

“ext extrem emel ely strong.”

• In ordinary conversation, we use intensity descriptors to

communicate with each other and compare experiences. “That lemonade tastes ex extrem emel ely st strong to me. Does it taste ex extrem emel ely st strong to you?”

• Labeled scales use those same descriptors to denote

perceived intensities.

• We seem to be communicating, but are we?
• No! “ex

extrem emel ely st strong ” ” may denote different perceived taste intensities to each of us.

• How do our subjects interpret “ex

extrem emel ely stro rong ng” ” on the 9-point scale?

• Do they tend to interpret “ex

extrem emel ely str strong“ as referring only to foods when they are in a food experiment?

• Let’s ask them.

We asked our subjects to rate their favorite and least favorite foods.

• The median for favorite

food was “9.”

• The median for least

favorite food was “1.”

• This tells us that the

Natick 9-point scale was interpreted as referring to food preferences only.

1 2 3 4 5 6 7 8 9 like extremely dislike extremely neutral favorite food least฀ favorite food

But consider how easily we switch contexts for ratings.

Example: elastic intensity scale

• A woman who has just experienced childbirth may

describe her pain as “ve very st y stron

• ng.

g.”

• Given a cup of tea, she may also describe the

flavor of the tea as “ve very st y stron

• ng.

g.”

Elastic Intensity Scale We understand that she does not mean to suggest that her pain and the intensity of the tea flavor are the same. She means her pain was “very strong” in the context of all pains she has experienced and the tea flavor was “very strong” in the context of all teas she has sampled.

Comparison errors caused by elasticity of our intensity scale

• What happens when we treat our elastic

scale as if it is absolute?

• We first encountered this kind of error in
• ur taste studies involving supertasters.

REVE VERAL AL AR ARTIFAC ACT (s (show

• wn b

by y red da d dashe hed l d line nes)

Ver Very y Stro rong ng Ta Tast ste RE REALI LITY TY Ot Othe hers rs ST

Qui uini nine ne Suc ucros

• se

Na NaCl

Ot Othe hers rs ST INCORRE NCORRECT A CT ASS SSUMP UMPTI TION ON th that " t "Ver ery S Str tron

• ng Ta

g Taste" ste" is th is the sa e same e absolu bsolute in te inten tensity sity for for n non

• nta

taster sters (NT) a s (NT) and d su super perta taster sters (S s (ST) T)

Qui uini nine ne Suc ucros

• se

Na NaCl

Ver Very y Stro rong ng Ta Tast ste Ver Very y Stro rong ng Ta Tast ste

We are not the first to note this error.

• Aitken, R. C. B. (1969). Measurement of feelings using visual

analogue scales. Proceedings of the Royal Society of Medicine, 62, 989-993.

• Narens, L., & Luce, R. D. (1983). How we may have been

misled into believing in the interpersonal comparability of

• utility. Theory and Decision, 15, 247-260.
• Biernat, M., & Manis, M. (1994). Shifting standards and

stereotype-based judgements. Journal of Personality and Social Psychology, 66, 5-20.

• Birnbaum, M. H. (1999). How to show that 9>221: Collect

judgements in a between-subjects design. Psychological Methods, 4, 243-249.

Illustration: PROBLEM and SOLUTION

• Select two groups

– people with many fungiform papillae: super ertaster ers – people with few fungiform papillae

• Ask both groups to describe the sweetness of a coke

– Both groups rate the sweetness to be about 2/3 of the distance from no sweet (0) to the strongest sweet they have ever tasted (100%). – Thus both groups seem to be experiencing the same sweetness.

• Now ask both groups to match the sweetness to the

loudness of a tone.

Correct conclusion: Supe upertas asters perceive twice as much sweetness as do Other ers.

Each 10 db doubles loudness subway train whistle telephone dial tone loud conversation 90 80 100 70 db

Others Supe pert rtasters rs Sweetness of a coke

Summary of Magnitude Matching

• Select a standard that is independent of the sensation we

want to compare and ask subjects to rate the sensation of interest relative to the standard.

• If the standard is roughly equal, on average, to two groups,

then we can make an absolute comparison of the sensation

• f interest across the two groups.
• This method was first used in taste (1975) and was

formalized as “magnitude matching” in 1980 (Marks & Stevens).

Let’s make this user friendly.

• Let’s take the labeled magnitude scales

we are used to (e.g., Natick 9-point scale).

• Respace the intensity descriptors so that

the scale has ratio properties (i.e., “8” denotes an intensity twice that of “4”).

• This is essentially the LMS (Labeled

Magnitude Scale) devised by Barry Green and his colleagues to be used with oral sensations (Green et al, 1993).

• Green and his colleagues empirically re-

spaced the intensity descriptors to give this scale ratio properties

• Turns out that we don’t even really need

all of those descriptors.

very strong strong moderate weak strongest imaginable

• ral sensation

no sensation

LMS

Evolution of the general Labeled Magnitude scale (gLMS)

10 20 30 40 50 60 70 80 90 100 very strong strong moderate weak barely detectable strongest imaginable sensation of any kind no sensation

gLMS

barely detectable very strong strong moderate weak strongest imaginable

• ral sensation

no sensation

LMS

no sensation strongest imaginable sensation of any kind ever฀ experienced

gVAS

฀ where 100 =฀ ฀ ฀ ฀ ฀ 0 = no sensation ฀ strongest sensation of any kind ever฀ experienced

Numbers 0 to 100

Bartoshuk, Puentes, Snyder & Sims, unpublished data

very s y stron

• ng

medi edium ver ery w y weak eak 1 2 3 4 5 6 7 8 9

Natick 9- 9-poi point categor gory s y scale Cit itric ic A Acid id

Den ensity of y of F Fungi gifor

• rm P

Papi apillae ae (# i # in t tem empl plat ate) e) 10 10 30 30 40 40 20 20 50 50

Perce ceive ved T Taste I e Inten ensity

Den ensity of y of F Fungi gifor

• rm P

Papi apillae ae (# i # in t tem empl plat ate) e)

Perce ceive ved T Taste I e Inten ensity

Citric A Acid 10 10 30 30 40 40 20 20 50 50 very s y stron

• ng

stron

• ng

mode

• derat

ate weak eak ba barel ely de y detec ectabl able stron

• nge

gest imagi aginabl able e sen ensat ation

• n
• f a
• f any k

y kind d

gL gLMS MS

Density of fungiform papillae gives us a way to check how good our scales are.

Let’s create a hedonic version

• f the gLMS

Sensory

20 40 60 80 100 most intense sensation of any kind ever experienced no sensation

Hedonic

• 100
• 80
• 60
• 40
• 20

20 40 60 80 100 most intense pleasure of any kind ever experienced neutral most intense displeasure of any kind ever experienced

1 2 3 4 5 6 7 8 9 like extremely dislike extremely neutral favorite food least฀ favorite food

• 100
• 80
• 60
• 40
• 20

20 40 60 80 100 most intense pleasure of any kind ever experienced neutral most intense displeasure of any kind ever experienced favorite฀ food least฀ favorite food

Natick 9-point scale Hedonic gLMS

Examples of what the hedonic gLMS can do

• 100
• 75
• 50
• 25

25 50 75 100 20 40 60 80 100

favorite food least favorite food

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

r = .007฀ p = .92 r = .10฀ p = .16 r = .40฀ p < .0001 r = .35฀ p < .0001

hedonic gLMS hedonic 9-pt

Supertasters Others Supertasters Others

Note that the hedonic gLMS reveals that supertasters have more extreme food likes and dislikes.

Hedonic gLMS

• 100
• 75
• 50
• 25

25 50 75 100 r = .16฀ p = .02 neutral

Dislike Like Dislike Like

• 100
• 75
• 50
• 25

25 50 75 100 20 40 60 80 100 r = .24฀ p <.0006

Supertasters Others

neutral

• The gLMS shows

that supertasters like

• range juice and

dislike grapefruit juice more than do

• thers.
• The 9-point hedonic

scale cannot reveal this.

These data were collected at lectures so we could accumulate large samples.

When to use the hedonic gLMS

• When you compare samples, this is a within

subject comparison and virtually any scales can do this.

• However, when you need to compare

groups they may be quite different with regard to food perception or liking, the older scales may provide erroneous comparisons.