- Email: [email protected]
Facial expressions and genetic sensitivity to 6-n-propylthiouracil predict hedonic response to sweet
Physiology & Behavior, Vol. 52, pp. 75-82, 1992
0031-9384/92 $5.00 + .00 Copyright © 1992PergamonPress Ltd.
Printed in the USA.
Facial Expressions and Genetic Sensitivity to 6-n-Propylthiouracil Predict Hedonic Response to Sweet HEATHER
LOOY 1 AND
HARVEY
P. W E I N G A R T E N z
Department o f Psychology, M c M a s t e r University, Hamilton, Ontario, Canada, L 8 S 4K1 R e c e i v e d 11 J u n e 1991 LOOY, H. AND H. P. WEINGARTEN. Facial expressions and genetic sensitivity to 6-n-propylthiouracil predict hedonic response to sweet. PHYSIOL BEHAV 52(1) 75-82, 1992.--Individuals differ in their hedonic response to sweet: sweet likers show increasing liking with increasing sucrose concentration, while sweet dislikers show increasing dislike with increasing concentration. Our results indicated that naive raters can correctly classify sweet likers and dislikers by observing subjects' facial responses to the taste of sucrose. Also, for both adults and children, the sweet liker/disliker distinction correlated strongly with the genetically determined ability to taste 6-n-propylthiouracil (PROP): PROP nontasters were almost always sweet likers, whereas sweet dislikers were almost always PROP tasters. Sweet dislikers also reported a purer sweet sensation than likers, who perceived nonsweet components in pure sucrose solutions. These results suggest that the sweet liker/disliker distinction is robust and valid, and that sensitivity to PROP may influence preference for sweet by altering the quality of sweet. Taste perception
Genetics
PROP
PTC
Sucrose
N U T R I T I O N research has repeatedly demonstrated a strong positive relationship between sweetness and food palatability, preference, and acceptance (4,18,33,35,36). However, while it is speculated that a pleasurable response to sweet is innate (8,35), and, therefore, that sweet-tasting foods are by definition preferred, adults show significant variation in hedonic (pleasure) response to sweet-tasting substances (5,7,12,23,30,31). The majority of adults fall into two distinct populations: sweet likers, who show increasing liking with increasing concentration of sweet, and sweet dislikers, who show increasing dislike with increasing concentration (7,22,31). The distinction between likers and dislikers is important in understanding individual differences in taste reactivity and the effects of some manipulations on taste. For example, Cabanac's alliesthesia effect (6,7) is more easily observed in sweet dislikers than likers (22). Specifically, sweet dislikers show an attenuation of their dislike for concentrated sweet tastes when food deprived; sweet likers show no change in sweet reactivity whether sated or deprived (22). Data such as these suggest that individual differences in hedonic response, such as the sweet liker/disliker distinction, may be important for understanding the relationship between taste psychophysics and food preferences or intake. The experiments reported in this paper examine factors that may underlie individual differences in hedonic response to sweet. Differing hedonic response patterns to sweet may reflect quantitative or qualitative differences in the perception of sweet stim-
Nutrition
Human
Hedonic
Sweet
uli, or differing hedonic responses may represent differences in cognitions or attitudes regarding sweet (potentially fattening) foods. To explore these possible explanations of the sweet liker/ disliker dichotomy, we first evaluated the validity of hedonic self-report measures by assessing whether these reports (upon which the sweet liker/disliker distinction is based) predict another index of hedonic response to taste, facial expressions (18,35). Such predictive validity would support the hypothesis that the sweet liker/disliker distinction reflects a genuine difference in taste percept, rather than a purely cognitive, culturally determined response. Second, we asked whether hedonic response patterns to sucrose were related to an established difference in taste perception, the genetically determined ability to taste the bitter substance 6-n-propylthiouracil (PROP). Third, we examined directly the possibility that sweet likers and dislikers differ in their perceptions of the taste quality of pure sucrose solutions. GENERAL METHOD
Subjects Subjects were male and female university undergraduates registered in an introductory psychology course. Subjects gave informed written consent prior to testing and received course credit for their participation.
t Present address: Department of Psychology, The King's College, Edmonton, Alberta, Canada. 2 Requests for reprints should be addressed to Dr. Harvey P. Weingarten.
75
76
l,{)()'~' A N D V v E I N ( i A R l t N
METHOD FOR DETERMINING SWFIf-.T LIKER/I)ISLIKER S I A I U S
Stimuli Sweet solutions were made by dissolving sucrose in room temperature (22°C) distilled water. Sucrose concentrations used were 0.05, 0.10, 0.21, 0.42, and 0.83 M. The exact number of concentrations used varied in different experiments. In each experiment, stimuli were presented in blocks. Each block included all concentrations used. The order of stimuli was randomized in a block and there were five or six replications of each block.
Rating Scales' Subjects rated each sucrose solution on two scales, one for sweetness intensity, and one for hedonic value. The rating scales were 200 m m visual analogue scales, with anchor points at 30 and 170 mm. The intensity scale was preceded by the question: " H o w sweet was the taste?" and the anchors were labelled not sweet at all and extremely sweet. The hedonic scale was preceded by the question: " H o w much did you like the taste?" and the anchors were labelled disliked very much and liked very much, with a neutral point at 100 m m labelled neither liked nor disliked. Subjects were instructed to make a mark on the line which represented their response to each question. Ifa given solution tasted twice as sweet, or was liked twice as much, as another solution, they were to make a mark twice as far along the scale. The intensity and hedonic value scales were reproduced on one sheet of paper, but a separate sheet was provided for each solution.
Procedure Subjects were requested to come to each session after having eaten within I hour prior to testing. Each subject was tested alone, seated before a tray of plastic medicine cups, each containing about 10 ml of sucrose solution. The cups were numbered in order. Subjects were instructed to sip solution #1, swish it around their mouth, and spit it out (the sip and spit procedure). A funnel attached to a large jug was provided for spitting. They were then to rate the intensity and hedonic value of the taste. rinse their mouth with distilled water, and proceed to the next solution. They were not permitted to swallow, nor to taste any solution more than once.
Data Analysis The intensity ratings were scored by measuring the distance to the mark on the scale, from 0 to 200 mm, beginning before the not sweet at all end. Hedonic ratings were measured from the neutral point to the extremes, liking ranging from 0 to + 100 mm, disliking from 0 to - 100 mm. Each subject's intensity and hedonic ratings were averaged over the replications of each sucrose concentration. These averaged data were analyzed using analyses of variance (ANOVAs), as described for each experiment. Classification of a subject as a sweet liker or disliker was based on an assessment of whether the hedonic ratings of the sucrose solutions by that individual increased (sweet liker) or decreased (sweet disliker) with ascending sucrose concentration. METHOD FOR DETERMINING PROP TASTER/NONTASTERSTATUS
PROP Solutions 6-n-Propylthiouracil (PROP) solutions were made by dissolving P R O P in room temperature (22°C) distilled water. Fifteen concentrations were prepared, ranging from 3.2 × 10 -3 M to 1.0 × 10 -6 M in 0.25 log steps.
A standard tbrced-choice detection procedure was used to determine the threshold concentration for detection of the taste of P R O P for each subject (2,28,37). Subjects were tested individually. They were seated behind a barrier which prevented them from seeing the solutions which the experimenter supplied. They were instructed "to sip, swish, and spit out each solution provided. Prior to tasting any solution, the}' were to rinse their mouth with distilled water. 1 o obtain an initial approximation of each subject's detection threshold, the experimenter began with the weakest concentration. She gave each subject about 10 ml of this solution in a plastic medicine cup and asked subjects to indicate whether the solution had a taste. This was repeated with e v e u second increasing concentration of P R O P until the subject reported the presence of a taste. Subjects were then given two cups of solution. One contained distilled water, the other the PROP solution for which the subject had first reported a taste. They were asked to taste both solutions, and to indicate which one of the two had a taste. They were required to make this decision even when they could not perceive a difference. If the subject chose correctly, the test was repeated with the same concentration. If they were correct again, the next lower concentration was tested. If the subject was incorrect (on either the first or second try), the next higher concentration was tested. Subjects were not given feedback regarding the accuracy of their responses. This procedure was continued until seven turn-around points were obtained. The first was discarded, and the geometric mean of the concentration at the other six provided the precise concentration which represented the subject's P R O P detection threshold. Subjects with a threshold greater than or equal to 2.0 x 10 4 31 were classified as nontasters; those with a threshold at or below 1.0 X I0 4 :ll were classified as tasters; and those who fell between these two cutpoints were labelled borderline (14). E X P E R I M E N T 1: FACIAL EXPRESSIONS Because of current sociocultural pressures toward thinness ( 13,15), it may be that a dislike for sweet represents a culturally determined response rather than a true difference in taste perception. Specifically, subjects' self-reported hedonic evaluations may not necessarily reflect their genuine taste perception, but rather a more cognitive decision based on associations between the taste and the social acceptability of liking sweet foods. Both human and animal studies indicate that there are distinct facial responses that reflect hedonic evaluations of taste stimuli (18,32,35). To evaluate the predictive validity of the sweet liker/ disliker distinction, Experiment I correlated subjects" facial expressions to sucrose with their hedonic ratings of these same solutions. In this study, subjects were videotaped while they tasted and rated sucrose solutions. Then, raters blind to the subjects' hedonic ratings viewed the videotapes and classified the subjects as sweet likers, dislikers, or neutral on the basis of facial expressions alone. Consistency between facial responses and hedonic ratings would support the conclusion that subjects report their genuine hedonic reactions on the rating scale. Alternatively, inconsistency between the two would suggest that subjects may be basing their hedonic evaluations on more cognitively assessed, socially determined associations of the taste.
77
PREDICTORS OF H E D O N I C RESPONSE TO SWEET METHOD Z
Subjects Sixty-six undergraduate student volunteers participated in the study. There were 38 females and 28 males, ranging in age from 18 to 44 years (mean = 20.3, SD = 3.5).
Stimuli The test stimuli consisted of 0.05, 0.10, 0.21, 0.42, and 0.83 M sucrose. Eight replications of each concentration were used, for a total of 40 solutions.
I-
Z IJJ I-Z --
0 Z
Procedure Subjects evaluated the intensity and hedonic value of sucrose solutions according to the procedure described in the General Method Section. Directly facing the subject, across the table at which they were seated, was a video camera. Subjects were told that the camera was recording the test session, and asked to disregard its presence as much as possible. When all subjects had been tested, a videotape was prepared containing sequences in which subjects were tasting the highest sucrose concentration (0.83 M). Only sweet likers and sweet dislikers were included on the tape. A few subjects did their tasting in such a m a n n e r that their facial expressions could not be seen clearly on the tapes and, due to this technical difficulty, had to be excluded. In total, the videotape included the facial responses of 27 females and 21 males. The videotape was rated by 20 volunteer raters. The raters viewed the tape individually. They were instructed to decide whether the subjects liked, were neutral toward, or disliked what they were tasting, on the basis of what they saw on the tape. After classifying a subject, they indicated how confident they felt about that classification by circling a n u m b e r from 1 to 5 (not confident at all to completely confident). Raters were reminded that neutral was a genuine category, and was not to be used when they felt uncertain of a subject's classification. The raters were permitted to watch the tasting sequences for any subject as often as required to make a classification. RESULTS Of the 66 subjects tested, 55 could be classified clearly as sweet likers or dislikers, based on their self-reported hedonic evaluations. The 11 whose data were not used included subjects with neutral hedonic responses across concentrations (n = 7), and those with erratic responses across concentrations (n = 4). Of the 55 subjects classified, 7 were excluded from the videotape for technical reasons (see Procedure Section). Figure 1 shows the mean ratings of sweetness intensity across concentrations for the 55 subjects classified as sweet likers (n = 22) and dislikers (n = 33). While intensity ratings across concentration were very similar in the two groups, ANOVA revealed that dislikers reported a small, but significant, greater sweet intensity over all concentrations than likers [group, F(1, 53) = 4.32, p < 0.05. Figure 1 also shows the mean hedonic ratings for likers and dislikers across concentrations. By definition, sweet likers showed an increasingly positive response with increasing sucrose concentrations, while dislikers reported an increasingly negative reaction. These differences were confirmed by ANOVA, with an expected group × concentration interaction, F(4, 216) = 143.93, p < 0.0001. Naive, independent raters found the classification of sweet dislikers quite clear. On the basis of facial expressions, raters accurately identified sweet dislikers 86% of the time. In contrast,
~ ~ ~ a LU "1"
200 175
r i ~
75
~
~
,,~7~
,s
50 ~
25" 0 . . . . . . . . . . . . . . . 0 . 0 0.1 0.2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9
1.0
100 75 50
25 0 -25 -50 -75 -100 0 . 0 0.1 0 . 2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9 SUCROSE
1.0
CONCENTRATION
FIG. 1. Top: mean (_SEM) intensity ratings of sucrose concentrations for sweet likers and dislikers. Bottom: mean (_+SEM) hedonic ratings of sucrose concentrations for sweet likers and dislikers. (SEMs not seen fall within the symbol.)
only 40% of sweet likers were correctly identified on the basis of facial expressions. However, when errors in the classification of sweet likers were made, they were classified as neutral, never as dislikers. In fact, in 87% of the cases, sweet likers were identified as either liking sweet or neutral to it. Raters were also significantly more confident in their assessment of sweet disliking than liking, t(25) = 3.90, p < 0.001. After they had evaluated the videotape, half of the raters were asked to indicate the cues they used to make their classifications. (Many other raters spontaneously mentioned the cues they had used.) For classification as a disliker, raters mentioned: frowning, tongue thrusting, whole-face grimacing, closing, widening, or rolling of the eyes, nostrils flaring, head rearing back, and a downturned or open mouth. All raters reported that dislike was very easy to detect. Representative photographs of individuals classified as sweet dislikers are shown in Fig. 2. For likers, the cues mentioned were: raised eyebrows, licking of the lips, nodding, and smiling. However, these cues were considered subtle, inconsistent, and difficult to interpret. The neutral category was used when raters could see no evidence of affect in either direction. DISCUSSION Consistent with the results of our (22) and other (30,31 ) studies, two main hedonic response patterns to sucrose are evident. These hedonic profiles allow classification of subjects as sweet likers or dislikers. Only 11 of the 66 subjects could not be classifted into one of these two groups. Independent raters, blind to subjects' hedonic evaluations, were remarkably accurate and confident in classifying sweet dislikers on the basis of facial expressions. Consistent with previous studies (16,35), they were somewhat less accurate and confident in classifying liking. However, sweet likers not classified as likers
I.()()'~ A N D W E I N ( i A R q E N
78
FIG. 2. Representative photographs (taken from videotapes) showing facial responses of sweet dislikers to 0.83 M sucrose. were classified as neutral, not as dislikers. Although, as Gilbert et al. (16) have suggested, facial expressions may not be reflexive indices of hedonic response, the cues raters used to make their classifications, especially those related to disliking, are consistent with those identified by others (18,35) as reflecting hedonic response to tastes. E X P E R I M E N T 2: P R O P SENSITIVITY A N D H E D O N I C R E S P O N S E T O SWEET The previous study suggests that differing hedonic response patterns to sweet reflect differences in the actual taste percept. The genetically determined sensitivity to the bitter taste of 6-npropylthiouracil (PROP), and the structurally related compound phenylthiocarbamide (PTC), is one of the best identified markers of taste difference among humans (38). Thus, we investigated whether sensitivity to P R O P correlates with sweet liker/disliker status.
High sensitivity to the bitter taste of P R O P and related compounds is encoded on a single dominant gene (38). Those most sensitive to P R O P (tasters) are either homozygous or heterozygous for the dominant allele, while those least sensitive (nontasters) are homozygous for the recessive allele. PROP sensitivity predicts how an individual will perceive a number of taste stimuli. For example, P R O P tasters find other bitter-tasting compounds more intensely bitter than do nontasters (3,26). In addition, sucrose and saccharin taste more intensely sweet to tasters than to nontasters (14,26). While the studies cited above have focussed on the relationship between P R O P sensitivity and taste intensity, no one has investigated whether there is any link between P R O P sensitivity and taste hedonics. However, indirect evidence for a relationship between taster/nontaster status and hedonics comes from research suggesting that P R O P sensitivity predicts food likes and dislikes, particularly for those foods containing bitter components (11,17,21).
P R E D I C T O R S O F H E D O N I C R E S P O N S E T O SWEET
79
NONTASTERS
TASTERS
TABLE 1 RELATIONSHIP BETWEEN PROP SENSITIVITY AND
24
SWEET STATUSAMONG ADULTS
Sweet Likers n=60
20
Sweet
PROP taster PROP nontaster
Liker
Disliker
Totals
>(D Z bJ
29 31 60
81 3 84
110 34 144
0 bJ rr b_
16 12 84-
J -6.0
Sweet liker/disliker and P R O P taster/nontaster status were assessed for all subjects who participated in the experiments reported in this paper. We present a summary of this accumulated data for all the adult subjects tested over this series of studies, as well as a similar assessment for a group of 7- to 10-year-old children• METHOD
Subjects
Procedure For adults, sweet liker/disliker status and P R O P sensitivity were determined as described previously. The procedures were almost identical, but simplified, for the children. Children were tested individually during the school day. Six replications of three sucrose concentrations, 0.05, 0.21, and 0.83 M, were used, for a total of 18 solutions• The wording of the visual analogue rating scales was simplified. The intensity question read: " H o w sweet was the taste?" and the anchors labelled not sweet at all and very sweet. The hedonic value question read: " H o w much did you like the taste?" and the anchors were labelled did not like at all, beneath which was a sketch of a grimacing face with its tongue protruding, and liked a lot, beneath which was a sketch of a happy face licking its lips. The P R O P procedure was followed as described in the General Method section, except the children were given feedback regarding the accuracy of their responses, and the process was stopped when each child's taster or nontaster classification was clear. Precise determination of the detection thresholds would have been too time consuming for the children.
-5.0
-4.5
-4.0
-5.0
-4.s
-4.0
I
-3.5
-3.0
-2.5
-i.0
-~.s
24- SweetDislikers n=84 20>L)
,,=
A total of 144 university undergraduate volunteers were tested and classified as sweet likers or dislikers, P R O P tasters or nontasters. There were 51 males and 93 females, ranging in age from 18 to 45 years (mean = 20.7 years). Also tested was a sample of 72 children, 31 females and 29 males, ranging in age from 7 to 10 years (mean = 9.0 years). The children were volunteers, and informed consent for their participation was given by their parents, as well as the principal and teachers of their schools. Children were rewarded with a sticker for their participation.
I -5.5
16-
84 O -6.0
-5.5
II
-3.5
LOG(X) OFPROPTHRESHOLD
FIG. 3. Frequency distribution of PROP detection thresholds for sweet likers (top) and dislikers (bottom). Subjects with thresholds greater than or equal to 2 × 10 4 M were classified as nontasters; thresholds at or below 1 >( 10-4 M were classified as tasters.
sweet (×2 = 44.89, p < 0.0001). Sweet dislikers were almost always P R O P tasters, while P R O P nontasters were almost always sweet likers. The reverse was not true. The same relationship was observed among the children (see Table 2). O f the 60 subjects who could be classified clearly as sweet likers or dislikers, 43 (71.7%) were P R O P tasters, while 17 (28.3%) were P R O P nontasters. Again, chi-square analysis revealed a significant relationship: sweet dislikers were P R O P tasters, while P R O P nontasters were sweet likers (X2 = 13.73, p < 0.0005). DISCUSSION
The proportions of tasters and nontasters in both adult and child samples falls within the expected range (39), and is con-
RESULTS
TABLE 2
O f the 144 adult subjects, 34 (23.6%) were P R O P nontasters, while 110 (76.4%) were P R O P tasters. Table 1 summarizes the relationship between P R O P sensitivity and sweet liker/disliker status for the adults, using the categorical classification of taster versus nontaster. Figure 3 shows the relationship between actual P R O P detection threshold and sweet liker/disliker classification in adults. There was a highly significant, asymmetrical relationship between P R O P taster/nontaster status and hedonic response to
RELATIONSHIP OF PROP SENSITIVITY TO SWEET STATUS AMONG CHILDREN Sweet
Taster Nontaster
Liker
Disliker
Totals
21 17 38
22 0 22
43 17 60
80
l()()x( AND WEIN(iAR I [ N
sistent with the predicted frequency based on a single-gene model of PROP sensitivity (38). The data reveal a dramatic correlation between PROP sensitivity and hedonic response to sweet for both adults and children. The PROP nontasters are almost always sweet likers, while sweet dislikers are almost all PROP tasters. There were proportionately more sweet likers among children than among adults, a finding consistent with demonstrations that children prefer higher concentrations of sucrose than adults (8-10). The increased number of sweet likers among children also suggests that some who are sweet likers in childhood become sweet dislikers by adulthood. Since, in both the adult and child samples, PROP nontasters are almost always sweet likers, the implication is that only childhood sweet likers who are also PROP tasters make the shift to sweet disliker status in adulthood. The relationship between PROP taster/nontaster status and hedonic response to sucrose may depend on the observation that PROP tasters and nontasters differ in their perceptions of sweet intensity (14,26). However, PROP tasters and nontasters may also differ in the quality of the taste sensation arising from sucrose. It is not known what structural or functional effects the PROP-related gene has on the gustatory, system. The PROP tasters may have a type of bitter receptor that is absent in nontasters (19) and an increased taste bud density relative to nontasters (29). These effects could underlie differences in taste quality that, in turn, could contribute to different assessments of hedonic value. EXPERIMENT 3: TASTE QUALITY The hypothesis that sweet likers and dislikers perceive different taste qualities in sucrose solutions is explored in this experiment. METHOD
Subjects Subjects were 48 undergraduate volunteers. There were 34 females and 14 males, with a mean age of 20.9 years.
Stimuli Five sucrose concentrations, 0.05, 0.10, 0.21, 0.42, and 0.83 M, were used. There were five replications of each concentration, for a total of 25 solutions. The same solutions, in a different random order within blocks, were presented in both experimental sessions. The PROP solutions described previously were used to determine taster/nontaster status.
Rating Scales The rating scales to measure perceived intensity and hedonic value have been described in the General Method section. A procedure developed by McBurney (27,34) was employed to assess the taste quality of the solutions. Subjects were asked to provide a numerical magnitude estimate of the total taste intensity (disregarding taste quality) of each solution. Subjects were then instructed to indicate what proportion of the total taste intensity was due to a bitter, sweet, sour, and/or salty quality. They could do this either by dividing their numerical magnitude estimate appropriately across qualities, or by giving the percent of the taste that was bitter, sweet, sour, or salty. (This latter strategy was preferred by most subjects; the experimenter simply converted these percentages into proportions of the magnitude estimate.) A separate rating sheet was provided for each solution.
])r(~('('dllre
Subjects participated in two sessions, each 2(I to 30 nlinutes long, with no more than 7 days between sessions. In one session, the procedures for evaluating the intensity and hedonic value of the sucrose solutions and determining PROP taster/nontaster status were followed (described in the General Method section). In the other session, subjects were instructed to sip-and-spit the solutions, and follow the method tbr assessing taste quality as described above. They were given two practice trials to ensure that they understood the magnitude estimation/quality assessment procedure. The experimenter was careful to give no indication of the actual contents of the solutions, to avoid biasing subjects' expectations about what the solutions ought to taste like. To reduce further the demand characteristics of the procedure, subjects were assured that there was no correct taste, that they were to simply report what they really perceived. The order of the two sessions was counterbalanced across subjects. l)ala. Inaly.sLs Numerical magnitude estimates of total taste intensity were equated across subjects using the following procedure (2,27,24): each subject's responses to the five sucrose concentrations were averaged. The averages for all subjects were then equated to a value of t0 by multiplying by the appropriate constant. The constant used for each subject was then applied to all the magnitude estimates for every concentration for that subject. These adjusted data were averaged arithmetically across subjects, as usual. This procedure reduces variability introduced by the absolute size of the magnitude estimates without affecting the ratios between the estimates (2). RESU 1. I'S
Of the 48 subjects tested, 42 could be unambiguously classified as sweet likers or dislikers. The mean intensity ratings obtained from the visual analogue scale for sweet likers and dislikers across sucrose concentrations are shown in Fig. 4. The patterns for each group were remarkably similar, although dislikers reported a small, but significantly greater [group, F( 1, 40) - 4.73, p < 0.05] sweet intensity. Figure 4 also shows the mean hedonic ratings across concentrations for sweet likers and dislikers. As expected, ANOVA revealed a significant group × concentration interaction, F(4, 168) = 129.04, p < 0.0001. Figure 5 shows the average intensity of the sweet component of the taste of the sucrose solutions for sweet likers and dislikers. ANOVA revealed a significant difference between likers and dislikers, F(I, 40) = 19.03, p < 0.0005, the expected change across concentrations, F(4, 156) = 256.02, p < 0.0001, and a group × concentration interaction, F(4, 156) = 2.99, p < 0.05. Inspection of Fig. 5 indicates that these findings result from a significantly greater perceived sweet component in the sucrose solutions, particularly at higher concentrations, in sweet dislikers. Because the contribution of any one of the other three quality components (bitter, sour, and salty) to the taste was quite small, these were combined into a total nonsweet component. The average intensities of this nonsweet component for sweet likers and dislikers are shown in Fig. 5. ANOVA revealed a significant group effect, F(I, 40) = 21.03, p < 0.0001, but no change across concentrations and no interaction: sweet likers consistently reported the presence of a small nonsweet component to the taste of the sucrose solutions, while sweet dislikers perceived virtually nothing apart from sweet.
PREDICTORS OF HEDONIC RESPONSE TO SWEET
0 Z
200
I...< tr
150
:
/ o
> I--
z uJ I-z 0O
O z
no
5o
0 . 0 0.1 0 . 2 0.3 0 . 4 0.5 0.6 0.7 0 . 8 0.9
1.0
lOO
7s i 50
o
-2s
0
-50 -75
~
°
/
~'s I , _
~
t~ LU
,¢~
25
: "
"1" - 1 0 0 .
.
.
.
.
.
.
.
.
.
.
0 . 0 0.1 0 . 2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9
SUCROSE
1.0
CONCENTRATION
FIG. 4. Top: mean (_+SEM)intensity ratings of sucrose concentrations for sweet likers and dislikers. Bottom: mean (_+SEM)hedonic ratings for sweet likers and dislikers. (SEMs not seen fall within the symbol.)
DISCUSSION
81 relative to the intensity of those components tasted separately (1,20). Because sweet likers perceive other qualities apart from sweet in sucrose solutions (see Experiment 3), the intensity of the sweet component may have been reduced relative to sweet dislikers who perceive no accompanying additional nonsweet taste component. Several observations suggest that the sweet liker/disliker distinction represents individual differences in the actual perceptual experience of sweet. First, raters can reliably classify subjects as sweet likers or dislikers based solely upon the subject's facial responses. Since it is highly unlikely that subjects consistently modified their facial expressions to be consistent with their cognitions or attitudes to sweet, we suggest that the sweet liker/ disliker distinction is a valid index of individual differences in the hedonic response to sweet. Second, there is a dramatic correlation between PROP sensitivity and sweet liking/disliking in both adults and children. We believe that this is the first report of a link between a marker of genetic taste difference and hedonic response to simple taste stimuli. This correlation is consistent with the view that the presence or absence of the PROP trait reflects structural or functional differences in the gustatory system that affects the perception of simple tastes. Third, and most directly, likers and dislikers report different taste qualities when tasting pure sweet solutions. Sweet dislikers tasted a purer sweet taste in sucrose solutions than sweet likers. This difference suggests that the hedonic response to sucrose may reflect, in part, a difference between sweet likers and dislikers in the perceived quality of the taste stimuli. In summary, these data imply that the sweet liker/disliker distinction is an important and valid difference in hedonic response to sweet, and may reflect underlying sensory differences.
Taste quality assessments indicate that sweet likers perceive a more complex taste in sucrose solutions than do dislikers, who report a very pure sweet sensation. Since hedonic response to sucrose is strongly related to PROP taster/nontaster status, the difference in the taste quality of sucrose may be the result of whatever structural or functional modifications in the taste system underlie PROP sensitivity. What those modifications are, and how they mediate the more complex taste experience of the sweet likers (who comprise almost all of the PROP nontasters), are issues for further research.
z
wluJ
4
GENERAL DISCUSSION
~
o
I-Z
UJ
24
0
20 i
n
16
8
~
j,~ ~
0 . 0 0.1
The series of experiments reported here were designed to explore the factors underlying the sweet liker and disliker distinction in taste hedonics. The results suggest that most individuals, 82% of adults (124/152) and 83% of children (60/72), can be classified easily as either a sweet liker or disliker based simply upon their self-reports of the hedonic value of sucrose within the ranges used. It is unclear whether a difference in perceived sweet intensity accompanies the sweet liker/disliker distinction. In some experiments (Experiments 1 and 3), but not others (22), dislikers report a small, but significantlygreater, sweet intensity in sucrose solutions than do likers. It is unlikely that this intensity difference accounts for the fundamentally different profiles in hedonic response that characterize the sweet liker/disliker distinction. The results of the taste quality experiment (Experiment 3) provide a possible explanation for the difference in perceived sweet intensity when it is apparent. Differences in intensity in likers and dislikers may be due to mixture suppression, which refers to the weakening of the perceived intensity of components of a mixture,
I-Z LU Z
24-
n
20-
O
•
,
L,kers
0 . 2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9
1.0
O 16; I--tO 1 2 [ LLI
8=
GO
0
Z O
L
5
o-e
•
•
0.0 0.1 0.2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0 . 9 1.0
Z SUCROSE
CONCENTRATION
FIG. 5. Top: average size (-+SEM)of the sweet component of the total intensity of sucrose solutions for sweet likers and dislikers. Bottom: average size (_+SEM) of the nonsweet (bitter, salty, and sour, summed) component for sweet likers and dislikers. (SEMs not seen fall within the symbol.)
82
I,OOY ANI) W k l I N G A R I 1 . N
These conclusions are relevant b o t h for extending our understanding of the m e c h a n i s m s for encoding gustatory, stimuli a n d for exploring the impact o f individual differences in taste reactivity on food preferences and intake. F u r t h e r experiments are needed to establish the generality of the sweet liker/disliker distinction to other gustatory stimuli, a n d to identify the specific sensory, differences that underlie these hedonic patterns.
A('KNOWLkD(iI!MEN IS This research was funded by the Natural Sciences and Engineering Research Council (NSERC) of Canada. A prelimina~ report of these data was presented at the 13th Annual Meeting of the Association lot Chemoreception Sciences, Sarasota. FL, April 1991. These data fi)rm part of a thesis submitted by H.L. to McMaster University in partial fulfillment of the requirements for the Ph.D. degree.
REFERENCES t. Bartoshuk, L. M. Taste mixtures: Is mixture suppression related to compression? Physiol. Behav. 14:643-649; 1975. 2. Bartoshuk, L. M. The psychophysics of taste. Am. J. Clin. Nutr. 31: 1068-1077; 1978. 3. Bartoshuk, L. M. Bitter taste of saccharin related to the genetic ability to taste the bitter substance 6-n-propylthiouracil. Science 205:934935; 1979. 4. Blundell, J. E.; Rogers, P. J.; Hill, A. J. Uncoupling sweetness and calories: Methodological aspects of laboratory studies on appetite control. Appetite 11:54-61 ; 1988. 5. Brala, P. M.; Hagen, R. L. Effects of sweetness perception and caloric value of a preload on short term intake. Physiol. Behav. 30:1-9; 1983. 6. Cabanac, M. Physiological role of pleasure. Science 173:1103-1107; 1971. 7. Cabanac, M. Sensory pleasure. Q. Rev. Biol. 54:1-22; 1979. 8. Cowart, B. J. Development of taste perception in humans: Sensitivity and preference throughout the life span. Psychol. Bull. 90:43-73; 1981. 9. Desor, J. A.; Beauchamp, G. K. Longitudinal changes in sweet preferences in humans. Physiol. Behav. 39:639-641 ; 1987. 10. Desor, J. A.; Greene, L. S.; Mailer, O. Preferences for sweet and salty in 9- to 15-year-old and adult humans. Science 190:686-687: 1975. 11. Drewnowski, A. Genetics of taste and smell. World Rev. Nutr. Diet. 63:194-208: 1990. 12. Enns, M. P.; Vanltallie, T. B.; Grinker, J. A. Contributions of age, sex and degree of fatness on preferences and magnitude estimations for sucrose in humans. Physiol. Behav. 22:999-1003; 1979. 13. Garner, D. M.; Garfinkel, P. E.; Schwartz, D.; Thompson, M. Cultural expectations of thinness in women. Psychol. Rep. 47:483-49 t ; 1980. 14. Gent, J. F.; Bartoshuk, L. M. Sweetness of sucrose, neohesperidin dihydrochalcone, and saccharin is related to genetic ability to taste the bitter substance 6-n-propylthiouracil. Chem. Senses 7:265-272; 1983. 15. George, R. S.; Krondl, M. Perceptions and food use of adolescent boys and girls. Nutr. Behav. a:115-125: 1983. 16. Gilbert, A. N.; Fridlund, A. J.; Sabini, J. Hedonic and social determinants of facial displays to odors. Chem. Senses 12:355363; 1987. 17. Glanville, E. V.; Kaplan, A. R. Food preference and sensitivity of taste for bitter compounds. Nature 205:851-852: 1965. 18. Grill, H. J.; Norgren, R. The taste reactivity test. I, Mimetic responses to gustatory stimuli in neurologically normal rats. Brain Res. 143: 263-279; 1978. 19. Hall, M. J.; Bartoshuk, L. M.; Cain, W. S.: Stevens, J. C. PTC taste blindness and the taste of caffeine. Nature 253:442-443; 1974.
2(1. Kroeze, J. H. A. Masking and adaptation of sugar sweetness intensity. Physiol. Behav. 22:347-351; 1979. 2i. Krondl, M.; Coleman, P.: Wade, J.; Milner, J. A twin study examining the genetic influence on food selection. Hum. Nutr.: Appl. Nutr. 37A:189-198; 1983. 22. Looy, H.; Weingarten, H. P. Effects of metabolic state on sweet taste reactivity in humans depend on underlying hedonic response profile. Chem. Senses 16:123-130; 1991. 23. Lucas, F.: Bellisle, F. The measurement of food preferences in humans: Do taste-and-spit tests predict consumption? Physiol. Behav. 39:739-743; 1987. 24. Marks, L. E. Sensory processes: The new psychophysics. New York: Academic Press; 1974. 25. Marks, L. E. On the cross-modal perception of intensity. J. Exp. Psychol. [Hum. Percept.] 12:517-534:1986. 26. Marks, L. E.; Stevens, J. C.; Bartoshuk, L. M.; Gent, J. F.; Rifkin, B.: Stone. V. K. Magnitude-matching: The measurement of taste and smell. Chem. Senses 13:63-87; 1988. 27. McBurney, D. H.: Bartoshuk, L. M. Interactions between stimuli with different taste qualities. Physiol. Behav. 10:1101-1106; 1973. 28. McBurney, D. H.; Collings, V. B. Introduction to sensation/perception, 2nd ed. Englewood Cliffs: Prentice Hall; 1984. 29. Miller. I. J.: Reedy, F. E. Variations in human taste bud density and taste intensity perception. Physiol. Behav. 47:1213-1219; 1990. 30. Moskowitz, H. R.: Kumraiah, V.; Sharma, K. N.: Jacobs, H. L.: Sharma, S. D. Effects of hunger, satiety and glucose load upon taste intensity and taste hedonics. Physiol. Behav. 16:47 t-475; 1976. 31. Pangborn, R. M. Individual variation in affective responses to taste stimuli. Psychonom. Sci. 2 I:125-126; 1970. 32. Pfaffman, C.: Norgren, R.; Grill, H. J. Sensory affect and motivation. Ann. NY Acad. Sci. 290:18-34; 1977. 33. Sclafani, A. Carbohydrate taste, appetite, and obesity: An overview. Ncurosci. Biobehav. Rev. 11 : 131 - 153; 1987. 34. Smith, D. V.; McBurney, D. H. Gustatory, cross-adaptation: Does a single mechanism code the salty taste? J. Exp. Psychol. 80:101105; 1969. 35. Steiner, J. E. Human facial expressions in response to taste and smell stimulation. Adv. Child Dev. Behav. 13:257-295; 1979. 36. Weingarten, H. P.; Brdard, M. Diet palatability: Its definition, measurement, and experimental analysis. In: Cooper, S. J.; Liebman. J. M., eds. The neuropharmacology of appetite. Oxford: Oxford University Press; 199 I. 37. Wetherill, G. B.; Levitt, H. Sequential estimation of points on a psychometric function. Br. J. Math. Stat. Psychol. 18:1-10; 1965. 38. WhisselI-Buechy, D. Genetic basis of the phenylthiocarbamide polymorphism. Chem. Senses 15:27-37; 1990. 39. Whissell-Buechy, D. Effects of age and sex on taste sensitivity to phenylthiocarbamide (PTC) in the Berkeley Guidance sample. Chem. Senses 15:39-57: 1990.
0031-9384/92 $5.00 + .00 Copyright © 1992PergamonPress Ltd.
Printed in the USA.
Facial Expressions and Genetic Sensitivity to 6-n-Propylthiouracil Predict Hedonic Response to Sweet HEATHER
LOOY 1 AND
HARVEY
P. W E I N G A R T E N z
Department o f Psychology, M c M a s t e r University, Hamilton, Ontario, Canada, L 8 S 4K1 R e c e i v e d 11 J u n e 1991 LOOY, H. AND H. P. WEINGARTEN. Facial expressions and genetic sensitivity to 6-n-propylthiouracil predict hedonic response to sweet. PHYSIOL BEHAV 52(1) 75-82, 1992.--Individuals differ in their hedonic response to sweet: sweet likers show increasing liking with increasing sucrose concentration, while sweet dislikers show increasing dislike with increasing concentration. Our results indicated that naive raters can correctly classify sweet likers and dislikers by observing subjects' facial responses to the taste of sucrose. Also, for both adults and children, the sweet liker/disliker distinction correlated strongly with the genetically determined ability to taste 6-n-propylthiouracil (PROP): PROP nontasters were almost always sweet likers, whereas sweet dislikers were almost always PROP tasters. Sweet dislikers also reported a purer sweet sensation than likers, who perceived nonsweet components in pure sucrose solutions. These results suggest that the sweet liker/disliker distinction is robust and valid, and that sensitivity to PROP may influence preference for sweet by altering the quality of sweet. Taste perception
Genetics
PROP
PTC
Sucrose
N U T R I T I O N research has repeatedly demonstrated a strong positive relationship between sweetness and food palatability, preference, and acceptance (4,18,33,35,36). However, while it is speculated that a pleasurable response to sweet is innate (8,35), and, therefore, that sweet-tasting foods are by definition preferred, adults show significant variation in hedonic (pleasure) response to sweet-tasting substances (5,7,12,23,30,31). The majority of adults fall into two distinct populations: sweet likers, who show increasing liking with increasing concentration of sweet, and sweet dislikers, who show increasing dislike with increasing concentration (7,22,31). The distinction between likers and dislikers is important in understanding individual differences in taste reactivity and the effects of some manipulations on taste. For example, Cabanac's alliesthesia effect (6,7) is more easily observed in sweet dislikers than likers (22). Specifically, sweet dislikers show an attenuation of their dislike for concentrated sweet tastes when food deprived; sweet likers show no change in sweet reactivity whether sated or deprived (22). Data such as these suggest that individual differences in hedonic response, such as the sweet liker/disliker distinction, may be important for understanding the relationship between taste psychophysics and food preferences or intake. The experiments reported in this paper examine factors that may underlie individual differences in hedonic response to sweet. Differing hedonic response patterns to sweet may reflect quantitative or qualitative differences in the perception of sweet stim-
Nutrition
Human
Hedonic
Sweet
uli, or differing hedonic responses may represent differences in cognitions or attitudes regarding sweet (potentially fattening) foods. To explore these possible explanations of the sweet liker/ disliker dichotomy, we first evaluated the validity of hedonic self-report measures by assessing whether these reports (upon which the sweet liker/disliker distinction is based) predict another index of hedonic response to taste, facial expressions (18,35). Such predictive validity would support the hypothesis that the sweet liker/disliker distinction reflects a genuine difference in taste percept, rather than a purely cognitive, culturally determined response. Second, we asked whether hedonic response patterns to sucrose were related to an established difference in taste perception, the genetically determined ability to taste the bitter substance 6-n-propylthiouracil (PROP). Third, we examined directly the possibility that sweet likers and dislikers differ in their perceptions of the taste quality of pure sucrose solutions. GENERAL METHOD
Subjects Subjects were male and female university undergraduates registered in an introductory psychology course. Subjects gave informed written consent prior to testing and received course credit for their participation.
t Present address: Department of Psychology, The King's College, Edmonton, Alberta, Canada. 2 Requests for reprints should be addressed to Dr. Harvey P. Weingarten.
75
76
l,{)()'~' A N D V v E I N ( i A R l t N
METHOD FOR DETERMINING SWFIf-.T LIKER/I)ISLIKER S I A I U S
Stimuli Sweet solutions were made by dissolving sucrose in room temperature (22°C) distilled water. Sucrose concentrations used were 0.05, 0.10, 0.21, 0.42, and 0.83 M. The exact number of concentrations used varied in different experiments. In each experiment, stimuli were presented in blocks. Each block included all concentrations used. The order of stimuli was randomized in a block and there were five or six replications of each block.
Rating Scales' Subjects rated each sucrose solution on two scales, one for sweetness intensity, and one for hedonic value. The rating scales were 200 m m visual analogue scales, with anchor points at 30 and 170 mm. The intensity scale was preceded by the question: " H o w sweet was the taste?" and the anchors were labelled not sweet at all and extremely sweet. The hedonic scale was preceded by the question: " H o w much did you like the taste?" and the anchors were labelled disliked very much and liked very much, with a neutral point at 100 m m labelled neither liked nor disliked. Subjects were instructed to make a mark on the line which represented their response to each question. Ifa given solution tasted twice as sweet, or was liked twice as much, as another solution, they were to make a mark twice as far along the scale. The intensity and hedonic value scales were reproduced on one sheet of paper, but a separate sheet was provided for each solution.
Procedure Subjects were requested to come to each session after having eaten within I hour prior to testing. Each subject was tested alone, seated before a tray of plastic medicine cups, each containing about 10 ml of sucrose solution. The cups were numbered in order. Subjects were instructed to sip solution #1, swish it around their mouth, and spit it out (the sip and spit procedure). A funnel attached to a large jug was provided for spitting. They were then to rate the intensity and hedonic value of the taste. rinse their mouth with distilled water, and proceed to the next solution. They were not permitted to swallow, nor to taste any solution more than once.
Data Analysis The intensity ratings were scored by measuring the distance to the mark on the scale, from 0 to 200 mm, beginning before the not sweet at all end. Hedonic ratings were measured from the neutral point to the extremes, liking ranging from 0 to + 100 mm, disliking from 0 to - 100 mm. Each subject's intensity and hedonic ratings were averaged over the replications of each sucrose concentration. These averaged data were analyzed using analyses of variance (ANOVAs), as described for each experiment. Classification of a subject as a sweet liker or disliker was based on an assessment of whether the hedonic ratings of the sucrose solutions by that individual increased (sweet liker) or decreased (sweet disliker) with ascending sucrose concentration. METHOD FOR DETERMINING PROP TASTER/NONTASTERSTATUS
PROP Solutions 6-n-Propylthiouracil (PROP) solutions were made by dissolving P R O P in room temperature (22°C) distilled water. Fifteen concentrations were prepared, ranging from 3.2 × 10 -3 M to 1.0 × 10 -6 M in 0.25 log steps.
A standard tbrced-choice detection procedure was used to determine the threshold concentration for detection of the taste of P R O P for each subject (2,28,37). Subjects were tested individually. They were seated behind a barrier which prevented them from seeing the solutions which the experimenter supplied. They were instructed "to sip, swish, and spit out each solution provided. Prior to tasting any solution, the}' were to rinse their mouth with distilled water. 1 o obtain an initial approximation of each subject's detection threshold, the experimenter began with the weakest concentration. She gave each subject about 10 ml of this solution in a plastic medicine cup and asked subjects to indicate whether the solution had a taste. This was repeated with e v e u second increasing concentration of P R O P until the subject reported the presence of a taste. Subjects were then given two cups of solution. One contained distilled water, the other the PROP solution for which the subject had first reported a taste. They were asked to taste both solutions, and to indicate which one of the two had a taste. They were required to make this decision even when they could not perceive a difference. If the subject chose correctly, the test was repeated with the same concentration. If they were correct again, the next lower concentration was tested. If the subject was incorrect (on either the first or second try), the next higher concentration was tested. Subjects were not given feedback regarding the accuracy of their responses. This procedure was continued until seven turn-around points were obtained. The first was discarded, and the geometric mean of the concentration at the other six provided the precise concentration which represented the subject's P R O P detection threshold. Subjects with a threshold greater than or equal to 2.0 x 10 4 31 were classified as nontasters; those with a threshold at or below 1.0 X I0 4 :ll were classified as tasters; and those who fell between these two cutpoints were labelled borderline (14). E X P E R I M E N T 1: FACIAL EXPRESSIONS Because of current sociocultural pressures toward thinness ( 13,15), it may be that a dislike for sweet represents a culturally determined response rather than a true difference in taste perception. Specifically, subjects' self-reported hedonic evaluations may not necessarily reflect their genuine taste perception, but rather a more cognitive decision based on associations between the taste and the social acceptability of liking sweet foods. Both human and animal studies indicate that there are distinct facial responses that reflect hedonic evaluations of taste stimuli (18,32,35). To evaluate the predictive validity of the sweet liker/ disliker distinction, Experiment I correlated subjects" facial expressions to sucrose with their hedonic ratings of these same solutions. In this study, subjects were videotaped while they tasted and rated sucrose solutions. Then, raters blind to the subjects' hedonic ratings viewed the videotapes and classified the subjects as sweet likers, dislikers, or neutral on the basis of facial expressions alone. Consistency between facial responses and hedonic ratings would support the conclusion that subjects report their genuine hedonic reactions on the rating scale. Alternatively, inconsistency between the two would suggest that subjects may be basing their hedonic evaluations on more cognitively assessed, socially determined associations of the taste.
77
PREDICTORS OF H E D O N I C RESPONSE TO SWEET METHOD Z
Subjects Sixty-six undergraduate student volunteers participated in the study. There were 38 females and 28 males, ranging in age from 18 to 44 years (mean = 20.3, SD = 3.5).
Stimuli The test stimuli consisted of 0.05, 0.10, 0.21, 0.42, and 0.83 M sucrose. Eight replications of each concentration were used, for a total of 40 solutions.
I-
Z IJJ I-Z --
0 Z
Procedure Subjects evaluated the intensity and hedonic value of sucrose solutions according to the procedure described in the General Method Section. Directly facing the subject, across the table at which they were seated, was a video camera. Subjects were told that the camera was recording the test session, and asked to disregard its presence as much as possible. When all subjects had been tested, a videotape was prepared containing sequences in which subjects were tasting the highest sucrose concentration (0.83 M). Only sweet likers and sweet dislikers were included on the tape. A few subjects did their tasting in such a m a n n e r that their facial expressions could not be seen clearly on the tapes and, due to this technical difficulty, had to be excluded. In total, the videotape included the facial responses of 27 females and 21 males. The videotape was rated by 20 volunteer raters. The raters viewed the tape individually. They were instructed to decide whether the subjects liked, were neutral toward, or disliked what they were tasting, on the basis of what they saw on the tape. After classifying a subject, they indicated how confident they felt about that classification by circling a n u m b e r from 1 to 5 (not confident at all to completely confident). Raters were reminded that neutral was a genuine category, and was not to be used when they felt uncertain of a subject's classification. The raters were permitted to watch the tasting sequences for any subject as often as required to make a classification. RESULTS Of the 66 subjects tested, 55 could be classified clearly as sweet likers or dislikers, based on their self-reported hedonic evaluations. The 11 whose data were not used included subjects with neutral hedonic responses across concentrations (n = 7), and those with erratic responses across concentrations (n = 4). Of the 55 subjects classified, 7 were excluded from the videotape for technical reasons (see Procedure Section). Figure 1 shows the mean ratings of sweetness intensity across concentrations for the 55 subjects classified as sweet likers (n = 22) and dislikers (n = 33). While intensity ratings across concentration were very similar in the two groups, ANOVA revealed that dislikers reported a small, but significant, greater sweet intensity over all concentrations than likers [group, F(1, 53) = 4.32, p < 0.05. Figure 1 also shows the mean hedonic ratings for likers and dislikers across concentrations. By definition, sweet likers showed an increasingly positive response with increasing sucrose concentrations, while dislikers reported an increasingly negative reaction. These differences were confirmed by ANOVA, with an expected group × concentration interaction, F(4, 216) = 143.93, p < 0.0001. Naive, independent raters found the classification of sweet dislikers quite clear. On the basis of facial expressions, raters accurately identified sweet dislikers 86% of the time. In contrast,
~ ~ ~ a LU "1"
200 175
r i ~
75
~
~
,,~7~
,s
50 ~
25" 0 . . . . . . . . . . . . . . . 0 . 0 0.1 0.2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9
1.0
100 75 50
25 0 -25 -50 -75 -100 0 . 0 0.1 0 . 2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9 SUCROSE
1.0
CONCENTRATION
FIG. 1. Top: mean (_SEM) intensity ratings of sucrose concentrations for sweet likers and dislikers. Bottom: mean (_+SEM) hedonic ratings of sucrose concentrations for sweet likers and dislikers. (SEMs not seen fall within the symbol.)
only 40% of sweet likers were correctly identified on the basis of facial expressions. However, when errors in the classification of sweet likers were made, they were classified as neutral, never as dislikers. In fact, in 87% of the cases, sweet likers were identified as either liking sweet or neutral to it. Raters were also significantly more confident in their assessment of sweet disliking than liking, t(25) = 3.90, p < 0.001. After they had evaluated the videotape, half of the raters were asked to indicate the cues they used to make their classifications. (Many other raters spontaneously mentioned the cues they had used.) For classification as a disliker, raters mentioned: frowning, tongue thrusting, whole-face grimacing, closing, widening, or rolling of the eyes, nostrils flaring, head rearing back, and a downturned or open mouth. All raters reported that dislike was very easy to detect. Representative photographs of individuals classified as sweet dislikers are shown in Fig. 2. For likers, the cues mentioned were: raised eyebrows, licking of the lips, nodding, and smiling. However, these cues were considered subtle, inconsistent, and difficult to interpret. The neutral category was used when raters could see no evidence of affect in either direction. DISCUSSION Consistent with the results of our (22) and other (30,31 ) studies, two main hedonic response patterns to sucrose are evident. These hedonic profiles allow classification of subjects as sweet likers or dislikers. Only 11 of the 66 subjects could not be classifted into one of these two groups. Independent raters, blind to subjects' hedonic evaluations, were remarkably accurate and confident in classifying sweet dislikers on the basis of facial expressions. Consistent with previous studies (16,35), they were somewhat less accurate and confident in classifying liking. However, sweet likers not classified as likers
I.()()'~ A N D W E I N ( i A R q E N
78
FIG. 2. Representative photographs (taken from videotapes) showing facial responses of sweet dislikers to 0.83 M sucrose. were classified as neutral, not as dislikers. Although, as Gilbert et al. (16) have suggested, facial expressions may not be reflexive indices of hedonic response, the cues raters used to make their classifications, especially those related to disliking, are consistent with those identified by others (18,35) as reflecting hedonic response to tastes. E X P E R I M E N T 2: P R O P SENSITIVITY A N D H E D O N I C R E S P O N S E T O SWEET The previous study suggests that differing hedonic response patterns to sweet reflect differences in the actual taste percept. The genetically determined sensitivity to the bitter taste of 6-npropylthiouracil (PROP), and the structurally related compound phenylthiocarbamide (PTC), is one of the best identified markers of taste difference among humans (38). Thus, we investigated whether sensitivity to P R O P correlates with sweet liker/disliker status.
High sensitivity to the bitter taste of P R O P and related compounds is encoded on a single dominant gene (38). Those most sensitive to P R O P (tasters) are either homozygous or heterozygous for the dominant allele, while those least sensitive (nontasters) are homozygous for the recessive allele. PROP sensitivity predicts how an individual will perceive a number of taste stimuli. For example, P R O P tasters find other bitter-tasting compounds more intensely bitter than do nontasters (3,26). In addition, sucrose and saccharin taste more intensely sweet to tasters than to nontasters (14,26). While the studies cited above have focussed on the relationship between P R O P sensitivity and taste intensity, no one has investigated whether there is any link between P R O P sensitivity and taste hedonics. However, indirect evidence for a relationship between taster/nontaster status and hedonics comes from research suggesting that P R O P sensitivity predicts food likes and dislikes, particularly for those foods containing bitter components (11,17,21).
P R E D I C T O R S O F H E D O N I C R E S P O N S E T O SWEET
79
NONTASTERS
TASTERS
TABLE 1 RELATIONSHIP BETWEEN PROP SENSITIVITY AND
24
SWEET STATUSAMONG ADULTS
Sweet Likers n=60
20
Sweet
PROP taster PROP nontaster
Liker
Disliker
Totals
>(D Z bJ
29 31 60
81 3 84
110 34 144
0 bJ rr b_
16 12 84-
J -6.0
Sweet liker/disliker and P R O P taster/nontaster status were assessed for all subjects who participated in the experiments reported in this paper. We present a summary of this accumulated data for all the adult subjects tested over this series of studies, as well as a similar assessment for a group of 7- to 10-year-old children• METHOD
Subjects
Procedure For adults, sweet liker/disliker status and P R O P sensitivity were determined as described previously. The procedures were almost identical, but simplified, for the children. Children were tested individually during the school day. Six replications of three sucrose concentrations, 0.05, 0.21, and 0.83 M, were used, for a total of 18 solutions• The wording of the visual analogue rating scales was simplified. The intensity question read: " H o w sweet was the taste?" and the anchors labelled not sweet at all and very sweet. The hedonic value question read: " H o w much did you like the taste?" and the anchors were labelled did not like at all, beneath which was a sketch of a grimacing face with its tongue protruding, and liked a lot, beneath which was a sketch of a happy face licking its lips. The P R O P procedure was followed as described in the General Method section, except the children were given feedback regarding the accuracy of their responses, and the process was stopped when each child's taster or nontaster classification was clear. Precise determination of the detection thresholds would have been too time consuming for the children.
-5.0
-4.5
-4.0
-5.0
-4.s
-4.0
I
-3.5
-3.0
-2.5
-i.0
-~.s
24- SweetDislikers n=84 20>L)
,,=
A total of 144 university undergraduate volunteers were tested and classified as sweet likers or dislikers, P R O P tasters or nontasters. There were 51 males and 93 females, ranging in age from 18 to 45 years (mean = 20.7 years). Also tested was a sample of 72 children, 31 females and 29 males, ranging in age from 7 to 10 years (mean = 9.0 years). The children were volunteers, and informed consent for their participation was given by their parents, as well as the principal and teachers of their schools. Children were rewarded with a sticker for their participation.
I -5.5
16-
84 O -6.0
-5.5
II
-3.5
LOG(X) OFPROPTHRESHOLD
FIG. 3. Frequency distribution of PROP detection thresholds for sweet likers (top) and dislikers (bottom). Subjects with thresholds greater than or equal to 2 × 10 4 M were classified as nontasters; thresholds at or below 1 >( 10-4 M were classified as tasters.
sweet (×2 = 44.89, p < 0.0001). Sweet dislikers were almost always P R O P tasters, while P R O P nontasters were almost always sweet likers. The reverse was not true. The same relationship was observed among the children (see Table 2). O f the 60 subjects who could be classified clearly as sweet likers or dislikers, 43 (71.7%) were P R O P tasters, while 17 (28.3%) were P R O P nontasters. Again, chi-square analysis revealed a significant relationship: sweet dislikers were P R O P tasters, while P R O P nontasters were sweet likers (X2 = 13.73, p < 0.0005). DISCUSSION
The proportions of tasters and nontasters in both adult and child samples falls within the expected range (39), and is con-
RESULTS
TABLE 2
O f the 144 adult subjects, 34 (23.6%) were P R O P nontasters, while 110 (76.4%) were P R O P tasters. Table 1 summarizes the relationship between P R O P sensitivity and sweet liker/disliker status for the adults, using the categorical classification of taster versus nontaster. Figure 3 shows the relationship between actual P R O P detection threshold and sweet liker/disliker classification in adults. There was a highly significant, asymmetrical relationship between P R O P taster/nontaster status and hedonic response to
RELATIONSHIP OF PROP SENSITIVITY TO SWEET STATUS AMONG CHILDREN Sweet
Taster Nontaster
Liker
Disliker
Totals
21 17 38
22 0 22
43 17 60
80
l()()x( AND WEIN(iAR I [ N
sistent with the predicted frequency based on a single-gene model of PROP sensitivity (38). The data reveal a dramatic correlation between PROP sensitivity and hedonic response to sweet for both adults and children. The PROP nontasters are almost always sweet likers, while sweet dislikers are almost all PROP tasters. There were proportionately more sweet likers among children than among adults, a finding consistent with demonstrations that children prefer higher concentrations of sucrose than adults (8-10). The increased number of sweet likers among children also suggests that some who are sweet likers in childhood become sweet dislikers by adulthood. Since, in both the adult and child samples, PROP nontasters are almost always sweet likers, the implication is that only childhood sweet likers who are also PROP tasters make the shift to sweet disliker status in adulthood. The relationship between PROP taster/nontaster status and hedonic response to sucrose may depend on the observation that PROP tasters and nontasters differ in their perceptions of sweet intensity (14,26). However, PROP tasters and nontasters may also differ in the quality of the taste sensation arising from sucrose. It is not known what structural or functional effects the PROP-related gene has on the gustatory, system. The PROP tasters may have a type of bitter receptor that is absent in nontasters (19) and an increased taste bud density relative to nontasters (29). These effects could underlie differences in taste quality that, in turn, could contribute to different assessments of hedonic value. EXPERIMENT 3: TASTE QUALITY The hypothesis that sweet likers and dislikers perceive different taste qualities in sucrose solutions is explored in this experiment. METHOD
Subjects Subjects were 48 undergraduate volunteers. There were 34 females and 14 males, with a mean age of 20.9 years.
Stimuli Five sucrose concentrations, 0.05, 0.10, 0.21, 0.42, and 0.83 M, were used. There were five replications of each concentration, for a total of 25 solutions. The same solutions, in a different random order within blocks, were presented in both experimental sessions. The PROP solutions described previously were used to determine taster/nontaster status.
Rating Scales The rating scales to measure perceived intensity and hedonic value have been described in the General Method section. A procedure developed by McBurney (27,34) was employed to assess the taste quality of the solutions. Subjects were asked to provide a numerical magnitude estimate of the total taste intensity (disregarding taste quality) of each solution. Subjects were then instructed to indicate what proportion of the total taste intensity was due to a bitter, sweet, sour, and/or salty quality. They could do this either by dividing their numerical magnitude estimate appropriately across qualities, or by giving the percent of the taste that was bitter, sweet, sour, or salty. (This latter strategy was preferred by most subjects; the experimenter simply converted these percentages into proportions of the magnitude estimate.) A separate rating sheet was provided for each solution.
])r(~('('dllre
Subjects participated in two sessions, each 2(I to 30 nlinutes long, with no more than 7 days between sessions. In one session, the procedures for evaluating the intensity and hedonic value of the sucrose solutions and determining PROP taster/nontaster status were followed (described in the General Method section). In the other session, subjects were instructed to sip-and-spit the solutions, and follow the method tbr assessing taste quality as described above. They were given two practice trials to ensure that they understood the magnitude estimation/quality assessment procedure. The experimenter was careful to give no indication of the actual contents of the solutions, to avoid biasing subjects' expectations about what the solutions ought to taste like. To reduce further the demand characteristics of the procedure, subjects were assured that there was no correct taste, that they were to simply report what they really perceived. The order of the two sessions was counterbalanced across subjects. l)ala. Inaly.sLs Numerical magnitude estimates of total taste intensity were equated across subjects using the following procedure (2,27,24): each subject's responses to the five sucrose concentrations were averaged. The averages for all subjects were then equated to a value of t0 by multiplying by the appropriate constant. The constant used for each subject was then applied to all the magnitude estimates for every concentration for that subject. These adjusted data were averaged arithmetically across subjects, as usual. This procedure reduces variability introduced by the absolute size of the magnitude estimates without affecting the ratios between the estimates (2). RESU 1. I'S
Of the 48 subjects tested, 42 could be unambiguously classified as sweet likers or dislikers. The mean intensity ratings obtained from the visual analogue scale for sweet likers and dislikers across sucrose concentrations are shown in Fig. 4. The patterns for each group were remarkably similar, although dislikers reported a small, but significantly greater [group, F( 1, 40) - 4.73, p < 0.05] sweet intensity. Figure 4 also shows the mean hedonic ratings across concentrations for sweet likers and dislikers. As expected, ANOVA revealed a significant group × concentration interaction, F(4, 168) = 129.04, p < 0.0001. Figure 5 shows the average intensity of the sweet component of the taste of the sucrose solutions for sweet likers and dislikers. ANOVA revealed a significant difference between likers and dislikers, F(I, 40) = 19.03, p < 0.0005, the expected change across concentrations, F(4, 156) = 256.02, p < 0.0001, and a group × concentration interaction, F(4, 156) = 2.99, p < 0.05. Inspection of Fig. 5 indicates that these findings result from a significantly greater perceived sweet component in the sucrose solutions, particularly at higher concentrations, in sweet dislikers. Because the contribution of any one of the other three quality components (bitter, sour, and salty) to the taste was quite small, these were combined into a total nonsweet component. The average intensities of this nonsweet component for sweet likers and dislikers are shown in Fig. 5. ANOVA revealed a significant group effect, F(I, 40) = 21.03, p < 0.0001, but no change across concentrations and no interaction: sweet likers consistently reported the presence of a small nonsweet component to the taste of the sucrose solutions, while sweet dislikers perceived virtually nothing apart from sweet.
PREDICTORS OF HEDONIC RESPONSE TO SWEET
0 Z
200
I...< tr
150
:
/ o
> I--
z uJ I-z 0O
O z
no
5o
0 . 0 0.1 0 . 2 0.3 0 . 4 0.5 0.6 0.7 0 . 8 0.9
1.0
lOO
7s i 50
o
-2s
0
-50 -75
~
°
/
~'s I , _
~
t~ LU
,¢~
25
: "
"1" - 1 0 0 .
.
.
.
.
.
.
.
.
.
.
0 . 0 0.1 0 . 2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9
SUCROSE
1.0
CONCENTRATION
FIG. 4. Top: mean (_+SEM)intensity ratings of sucrose concentrations for sweet likers and dislikers. Bottom: mean (_+SEM)hedonic ratings for sweet likers and dislikers. (SEMs not seen fall within the symbol.)
DISCUSSION
81 relative to the intensity of those components tasted separately (1,20). Because sweet likers perceive other qualities apart from sweet in sucrose solutions (see Experiment 3), the intensity of the sweet component may have been reduced relative to sweet dislikers who perceive no accompanying additional nonsweet taste component. Several observations suggest that the sweet liker/disliker distinction represents individual differences in the actual perceptual experience of sweet. First, raters can reliably classify subjects as sweet likers or dislikers based solely upon the subject's facial responses. Since it is highly unlikely that subjects consistently modified their facial expressions to be consistent with their cognitions or attitudes to sweet, we suggest that the sweet liker/ disliker distinction is a valid index of individual differences in the hedonic response to sweet. Second, there is a dramatic correlation between PROP sensitivity and sweet liking/disliking in both adults and children. We believe that this is the first report of a link between a marker of genetic taste difference and hedonic response to simple taste stimuli. This correlation is consistent with the view that the presence or absence of the PROP trait reflects structural or functional differences in the gustatory system that affects the perception of simple tastes. Third, and most directly, likers and dislikers report different taste qualities when tasting pure sweet solutions. Sweet dislikers tasted a purer sweet taste in sucrose solutions than sweet likers. This difference suggests that the hedonic response to sucrose may reflect, in part, a difference between sweet likers and dislikers in the perceived quality of the taste stimuli. In summary, these data imply that the sweet liker/disliker distinction is an important and valid difference in hedonic response to sweet, and may reflect underlying sensory differences.
Taste quality assessments indicate that sweet likers perceive a more complex taste in sucrose solutions than do dislikers, who report a very pure sweet sensation. Since hedonic response to sucrose is strongly related to PROP taster/nontaster status, the difference in the taste quality of sucrose may be the result of whatever structural or functional modifications in the taste system underlie PROP sensitivity. What those modifications are, and how they mediate the more complex taste experience of the sweet likers (who comprise almost all of the PROP nontasters), are issues for further research.
z
wluJ
4
GENERAL DISCUSSION
~
o
I-Z
UJ
24
0
20 i
n
16
8
~
j,~ ~
0 . 0 0.1
The series of experiments reported here were designed to explore the factors underlying the sweet liker and disliker distinction in taste hedonics. The results suggest that most individuals, 82% of adults (124/152) and 83% of children (60/72), can be classified easily as either a sweet liker or disliker based simply upon their self-reports of the hedonic value of sucrose within the ranges used. It is unclear whether a difference in perceived sweet intensity accompanies the sweet liker/disliker distinction. In some experiments (Experiments 1 and 3), but not others (22), dislikers report a small, but significantlygreater, sweet intensity in sucrose solutions than do likers. It is unlikely that this intensity difference accounts for the fundamentally different profiles in hedonic response that characterize the sweet liker/disliker distinction. The results of the taste quality experiment (Experiment 3) provide a possible explanation for the difference in perceived sweet intensity when it is apparent. Differences in intensity in likers and dislikers may be due to mixture suppression, which refers to the weakening of the perceived intensity of components of a mixture,
I-Z LU Z
24-
n
20-
O
•
,
L,kers
0 . 2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0.9
1.0
O 16; I--tO 1 2 [ LLI
8=
GO
0
Z O
L
5
o-e
•
•
0.0 0.1 0.2 0.3 0 . 4 0 . 5 0 . 6 0.7 0 . 8 0 . 9 1.0
Z SUCROSE
CONCENTRATION
FIG. 5. Top: average size (-+SEM)of the sweet component of the total intensity of sucrose solutions for sweet likers and dislikers. Bottom: average size (_+SEM) of the nonsweet (bitter, salty, and sour, summed) component for sweet likers and dislikers. (SEMs not seen fall within the symbol.)
82
I,OOY ANI) W k l I N G A R I 1 . N
These conclusions are relevant b o t h for extending our understanding of the m e c h a n i s m s for encoding gustatory, stimuli a n d for exploring the impact o f individual differences in taste reactivity on food preferences and intake. F u r t h e r experiments are needed to establish the generality of the sweet liker/disliker distinction to other gustatory stimuli, a n d to identify the specific sensory, differences that underlie these hedonic patterns.
A('KNOWLkD(iI!MEN IS This research was funded by the Natural Sciences and Engineering Research Council (NSERC) of Canada. A prelimina~ report of these data was presented at the 13th Annual Meeting of the Association lot Chemoreception Sciences, Sarasota. FL, April 1991. These data fi)rm part of a thesis submitted by H.L. to McMaster University in partial fulfillment of the requirements for the Ph.D. degree.
REFERENCES t. Bartoshuk, L. M. Taste mixtures: Is mixture suppression related to compression? Physiol. Behav. 14:643-649; 1975. 2. Bartoshuk, L. M. The psychophysics of taste. Am. J. Clin. Nutr. 31: 1068-1077; 1978. 3. Bartoshuk, L. M. Bitter taste of saccharin related to the genetic ability to taste the bitter substance 6-n-propylthiouracil. Science 205:934935; 1979. 4. Blundell, J. E.; Rogers, P. J.; Hill, A. J. Uncoupling sweetness and calories: Methodological aspects of laboratory studies on appetite control. Appetite 11:54-61 ; 1988. 5. Brala, P. M.; Hagen, R. L. Effects of sweetness perception and caloric value of a preload on short term intake. Physiol. Behav. 30:1-9; 1983. 6. Cabanac, M. Physiological role of pleasure. Science 173:1103-1107; 1971. 7. Cabanac, M. Sensory pleasure. Q. Rev. Biol. 54:1-22; 1979. 8. Cowart, B. J. Development of taste perception in humans: Sensitivity and preference throughout the life span. Psychol. Bull. 90:43-73; 1981. 9. Desor, J. A.; Beauchamp, G. K. Longitudinal changes in sweet preferences in humans. Physiol. Behav. 39:639-641 ; 1987. 10. Desor, J. A.; Greene, L. S.; Mailer, O. Preferences for sweet and salty in 9- to 15-year-old and adult humans. Science 190:686-687: 1975. 11. Drewnowski, A. Genetics of taste and smell. World Rev. Nutr. Diet. 63:194-208: 1990. 12. Enns, M. P.; Vanltallie, T. B.; Grinker, J. A. Contributions of age, sex and degree of fatness on preferences and magnitude estimations for sucrose in humans. Physiol. Behav. 22:999-1003; 1979. 13. Garner, D. M.; Garfinkel, P. E.; Schwartz, D.; Thompson, M. Cultural expectations of thinness in women. Psychol. Rep. 47:483-49 t ; 1980. 14. Gent, J. F.; Bartoshuk, L. M. Sweetness of sucrose, neohesperidin dihydrochalcone, and saccharin is related to genetic ability to taste the bitter substance 6-n-propylthiouracil. Chem. Senses 7:265-272; 1983. 15. George, R. S.; Krondl, M. Perceptions and food use of adolescent boys and girls. Nutr. Behav. a:115-125: 1983. 16. Gilbert, A. N.; Fridlund, A. J.; Sabini, J. Hedonic and social determinants of facial displays to odors. Chem. Senses 12:355363; 1987. 17. Glanville, E. V.; Kaplan, A. R. Food preference and sensitivity of taste for bitter compounds. Nature 205:851-852: 1965. 18. Grill, H. J.; Norgren, R. The taste reactivity test. I, Mimetic responses to gustatory stimuli in neurologically normal rats. Brain Res. 143: 263-279; 1978. 19. Hall, M. J.; Bartoshuk, L. M.; Cain, W. S.: Stevens, J. C. PTC taste blindness and the taste of caffeine. Nature 253:442-443; 1974.
2(1. Kroeze, J. H. A. Masking and adaptation of sugar sweetness intensity. Physiol. Behav. 22:347-351; 1979. 2i. Krondl, M.; Coleman, P.: Wade, J.; Milner, J. A twin study examining the genetic influence on food selection. Hum. Nutr.: Appl. Nutr. 37A:189-198; 1983. 22. Looy, H.; Weingarten, H. P. Effects of metabolic state on sweet taste reactivity in humans depend on underlying hedonic response profile. Chem. Senses 16:123-130; 1991. 23. Lucas, F.: Bellisle, F. The measurement of food preferences in humans: Do taste-and-spit tests predict consumption? Physiol. Behav. 39:739-743; 1987. 24. Marks, L. E. Sensory processes: The new psychophysics. New York: Academic Press; 1974. 25. Marks, L. E. On the cross-modal perception of intensity. J. Exp. Psychol. [Hum. Percept.] 12:517-534:1986. 26. Marks, L. E.; Stevens, J. C.; Bartoshuk, L. M.; Gent, J. F.; Rifkin, B.: Stone. V. K. Magnitude-matching: The measurement of taste and smell. Chem. Senses 13:63-87; 1988. 27. McBurney, D. H.: Bartoshuk, L. M. Interactions between stimuli with different taste qualities. Physiol. Behav. 10:1101-1106; 1973. 28. McBurney, D. H.; Collings, V. B. Introduction to sensation/perception, 2nd ed. Englewood Cliffs: Prentice Hall; 1984. 29. Miller. I. J.: Reedy, F. E. Variations in human taste bud density and taste intensity perception. Physiol. Behav. 47:1213-1219; 1990. 30. Moskowitz, H. R.: Kumraiah, V.; Sharma, K. N.: Jacobs, H. L.: Sharma, S. D. Effects of hunger, satiety and glucose load upon taste intensity and taste hedonics. Physiol. Behav. 16:47 t-475; 1976. 31. Pangborn, R. M. Individual variation in affective responses to taste stimuli. Psychonom. Sci. 2 I:125-126; 1970. 32. Pfaffman, C.: Norgren, R.; Grill, H. J. Sensory affect and motivation. Ann. NY Acad. Sci. 290:18-34; 1977. 33. Sclafani, A. Carbohydrate taste, appetite, and obesity: An overview. Ncurosci. Biobehav. Rev. 11 : 131 - 153; 1987. 34. Smith, D. V.; McBurney, D. H. Gustatory, cross-adaptation: Does a single mechanism code the salty taste? J. Exp. Psychol. 80:101105; 1969. 35. Steiner, J. E. Human facial expressions in response to taste and smell stimulation. Adv. Child Dev. Behav. 13:257-295; 1979. 36. Weingarten, H. P.; Brdard, M. Diet palatability: Its definition, measurement, and experimental analysis. In: Cooper, S. J.; Liebman. J. M., eds. The neuropharmacology of appetite. Oxford: Oxford University Press; 199 I. 37. Wetherill, G. B.; Levitt, H. Sequential estimation of points on a psychometric function. Br. J. Math. Stat. Psychol. 18:1-10; 1965. 38. WhisselI-Buechy, D. Genetic basis of the phenylthiocarbamide polymorphism. Chem. Senses 15:27-37; 1990. 39. Whissell-Buechy, D. Effects of age and sex on taste sensitivity to phenylthiocarbamide (PTC) in the Berkeley Guidance sample. Chem. Senses 15:39-57: 1990.