Chem. Senses 28: 527-536,
2003
© Oxford University Press 2003
Understanding Creaminess Perception of Dairy Products Using Free-Choice Profiling and Genetic Responsivity to 6-n-Propylthiouracil
1 Department of Food Science, Cook College, Rutgers University, New Brunswick, NJ 2 International Flavors and Fragrances, Inc., Dayton, NJ, USA
Correspondence to be sent to: Beverly J. Tepper, Rutgers University, Department of Food Science, Cook College, 65 Dudley Road, New Brunswick, NJ 08901-8520, USA. e-mail: tepper{at}aesop.rutgers.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Flavor and texture contribute to the perception of creaminess in dairy products, but the nature of this interaction is not well understood. Previous studies on the genetic ability to perceive the bitter compound 6-n-propylthiouracil (PROP) reveal the existence of individual differences in creaminess perception. The objective of the present study was to use PROP-classified subjects to gain insight into this individual variation to better understand the cues for creaminess in dairy products, and to ascertain the contributions of flavor and texture to the integrated perception of creaminess. Ten nontasters and 10 supertasters of PROP participated in the study. Subjects evaluated nine commercial dairy products using Free-Choice Profiling (FCP), a type of descriptive analysis that allows subjects to rate products on individual lists of descriptors. Generalized Procrustes Analysis was used to develop separate consensus spaces for nontasters and supertasters. The models for both groups accounted for
54% of the variance in the data
and were resolved in two dimensions (a dairy flavor/texture axis and a
sweet–sour continuum). The products were arranged in a similar pattern
along the dimensions in both models. However, nontasters used a limited number
of simple terms (sour, sweet, milky and mouthcoating) to describe the
products, whereas supertasters used a more complex vocabulary (rich, buttery,
creamy, light, grainy, gritty and sandy). The model for nontasters gave equal
weight to the sweet–sour and dairy flavor/texture dimensions (28 and 26%
variance, respectively); whereas, the model for supertasters relied more
heavily on the dairy flavor/texture dimension (34% variance), and less so on
the sweet–sour dimension (20% variance). These data suggest that the
overall impression of creaminess was similar for nontasters and supertasters,
but the cues the two groups used to judge creaminess differed.
Key words: creaminess perception, 6-n-propylthiouracil, Free-choice Profiling
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Fat content contributes to the perception of creaminess in dairy products. Creaminess is a highly integrated and complex perception that encompasses both flavor and texture sensations (Mela, 1988
). A fundamental understanding of creaminess perception could
be obtained by decomposing creaminess into its underlying sensory components.
Accomplishing this task has been difficult since changes in stimulus fat
content simultaneously affect the texture of dairy products
(Mela, 1988;
Li et al., 1997;
Richardson-Harman et al.,
2000), as well as the flavor release
(Mela, 1988;
Li et al., 1997;
Richardson-Harman et al.,
2000).
The texture of dairy products can be ascertained from tactile sensations
produced in the mouth (Mela,
1988). These attributes include slipperiness, greasy mouthfeel
(Tuorila, 1986), creaminess
and residual mouthfeel (Bom Frost et
al., 2001). Kokini and Cussler demonstrated that texture
perception in the mouth could be modeled as a mathematical function of
thickness and smoothness (Kokini and
Cussler, 1983). In fluid dairy products, the presence of small,
even-sized fat globules coupled with adequate viscosity enhances the
perception of creaminess (Richardson
et al., 1993).
The contribution of milk fat to perceived fat content and creaminess is
less well understood. Mela demonstrated equivalent results when the perceived
fat content of fluid dairy products was assessed with and without nose clips
(Mela, 1988), suggesting that
flavor cues made little or no contribution to the overall assessment of
creaminess. However, Li et al. showed that altering the fat content
of ice creams modulated the pattern of release of flavor compounds
(Li et al., 1997).
The latter findings are consistent with a large body of work suggesting that
fat content affects the intensity, duration and balance of flavors present in
many foods (Lucca and Tepper,
1995), including dairy products
(Elmore et al.,
1999).
Other experiments have shown that supplementing dairy
products with added flavor enhances creaminess perception. For example,
Lawless and Clark showed that when vanilla flavor was added to 1% fat milk,
the perception of fat and creaminess increased
(Lawless and Clark, 1992). In
another study, adding dairy flavor to milk model systems ranging from 0 to 10%
fat resulted in greater perceived fat content and creaminess in the higher fat
samples (Tepper and Kuang,
1996). Therefore, flavor may make a more significant contribution
to the perception of creaminess in foods than some data seem to suggest.
Creaminess is a critical sensory attribute for consumer acceptance of
products (Daget et al.,
1987; Richardson-Harman et
al., 2000). However, consumers use the term creaminess
interchangeably to describe flavor and textural perceptions in dairy products,
most often not distinguishing between them. Thus, creaminess may be a
difficult precept to quantify if individuals are using the term as a flavor
descriptor, a texture descriptor or as an integrative perception of both
characteristics. It is possible that individuals may vary in their sensory
acuity for specific flavor and/or texture attributes that impart creaminess in
dairy products. As a consequence, individuals might utilize proportionately
greater or fewer flavor cues, relative to texture, to judge creaminess in
these foods. Understanding these sources of variation may provide insight into
individual differences in creaminess perception as well as the underlying
dimensions of creaminess perception.
Individual differences in perception have been observed among individuals
classified by PROP taster status (Tepper
and Nurse, 1997). PROP (6-n-propylthiouracil) is a bitter
tasting compound, the perception of which is genetically determined
(Bartoshuk et al.,
1994). Individuals can be grouped as nontasters, medium tasters
and supertasters based upon their sensitivity to PROP
(Bartoshuk et al.,
1994). Tasters (medium and supertasters) are more sensitive than
nontasters to the bitterness of caffeine, and to the sweetness of sucrose and
some artificial sweeteners (Bartoshuk
et al., 1994;
Lucchina et al.,
1998). Differences among taster groups have also been observed for
trigeminal sensations, including irritation from alcohol and capsaicin, the
burn of chili peppers (Karrer and
Bartoshuk, 1991; Bartoshuk
et al., 1994; Tepper
and Nurse, 1997; Lucchina
et al., 1998).
Some studies have also shown that PROP tasters are also more sensitive to
fat (Tepper, 1998). Tepper and
Nurse (Tepper and Nurse, 1997)
found that medium tasters and supertasters were able to discriminate between
10 and 40% fat salad dressings, whereas nontasters assessed the fat content of
the samples as equivalent. Duffy et al. investigated creaminess
perception in fluid dairy products and found that PROP supertasters gave the
highest ratings of creaminess to milk products with fat contents ranging from
11.5% to 54% (Duffy et al.,
1996). It has been hypothesized that PROP tasters are more
responsive to oral texture because they perceive more tactile sensations on
the tongue (Duffy et al.,
1996). Other studies have not reported taster-group differences in
creaminess perception with sweetened milk mixtures
(Drewnowski et al.,
1998) or flavored puddings and dairy drinks
(Yackinous and Guinard, 2001)
formulated in the laboratory. These data raise doubts about the role of PROP
status in the perception of more complex dairy foods. Further investigation of
this issue was a primary aim of the present study.
Conventional descriptive analysis has been used to characterize the texture
attributes of fluid dairy products
(Tuorila, 1986;
Bom Frost et al.,
2001). However, descriptive analysis forces all subjects to
utilize the same terminology and define each of these terms in the same way,
ignoring individual differences in perception. Free-Choice Profiling (FCP)
differs from conventional descriptive analysis in that each subject describes
the perceived qualities of samples using his or her own individual list of
terms, rather than the group lexicon
(Oreskovich et al.,
1991; Heymann,
1994). Few limitations are put upon subjects in FCP, allowing them
freedom and flexibility to describe individual perceptions of products. Li
et al. used FCP to study vanilla ice cream varying in fat content and
found that subjects used a variety of both flavor and textural attributes to
describe the samples (Li et al.,
1997). If individuals use different criteria to define and assess
creaminess, the use of FCP might help to expose these differences.
The present study utilizes FCP with individuals classified by PROP taster
status. This approach was intended to gain insight into individual differences
in perception to better understand the cues for creaminess and assess the
contribution of flavor and texture to the overall perception of creaminess. It
was hypothesized that supertasters will perceive more creaminess in dairy
products, as compared with nontasters, and that the perception of creaminess
would relate specifically to the textural components of these foods. Previous
research on creaminess perception has utilized model systems, which are not
fully representative of actual food products
(Daget et al., 1987;
Tepper and Kuang, 1996;
Richardson-Harman et al.,
2000). A range of commercially available dairy products was used
in the present study to better understand these relationships in real
foods.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Subject screening and selection
PROP screening
Seventy-six adults of the International Flavors and Fragrances, Inc.
Creative Center, South Brunswick, NJ were previously screened for their PROP
sensitivity. These individuals comprised a subject pool from which
participants in the current study were drawn. The pool consisted of 37 females
and 39 males with a mean age of 38.8 ± 8.0 (mean ± SD) (range
21–60 years). Participants were healthy individuals from various
technical backgrounds including
creation and application of flavors and were free of oral or nasal illness at the time of the study.
Subjects were classified for PROP taster status utilizing the one-solution
method (Tepper et al.,
2001). The method involves obtaining suprathreshold ratings for
0.32 mmol/l PROP (Aldrich Chemical, Milwaukee, WI) and 0.1 mol/l NaCl (Fischer
Scientific, Fairlawn, NJ). Both solutions were prepared with room temperature
spring water, however the PROP solution required mild heat to dissolve into
solution.
To screen for PROP taster status, two sessions were conducted on separate
days. On each day, subjects were presented with 10 ml of 0.1 mol/l NaCl
solution followed by 10 ml of 0.32 mmol/l PROP solution in three-digit coded
sample cups. Evaluations were conducted in individual testing booths equipped
with white lighting. Using Compusense version 4.0 direct data entry system
(Guelph, Ontario), subjects rated the intensity of the solutions using the
Labeled Magnitude Scale (LMS scale). The LMS scale is considered
quasilogarithmic with label descriptors similar to magnitude estimation (Green
et al., 1993,
1996). The end descriptors
included barely detectable and strongest imaginable.
Subjects were instructed to taste and rate the samples relative to oral
stimuli experienced in everyday life. Subjects rinsed their mouths with water
prior to tasting each sample. Then, they put the entire 10 ml sample in their
mouth, rated the intensity of the stimulus and expectorated. There was a 1 min
break between the NaCl and PROP evaluations. Replicate evaluations for the
solutions were obtained, separated by at least 24 h but no more than 5 days.
All participants signed an informed consent form prior to participating in the
screening.
For each subject, the average of the two PROP intensity ratings was
calculated. Three taster groups were identified using K-means cluster analysis
(JMP version 4.05, SAS Institute, Cary, NC). These grouping were used to
establish numerical cutoffs for PROP. The cutoffs were confirmed by
calculating the 95% confidence interval around the mean for each of the
groups, according to the classification method developed by Tepper et
al. (Tepper et al.,
2001). Mean NaCl ratings for each group were also determined.
Among the 76 subjects screened, individuals whose ratings for PROP
intensity were <11.5 on the LMS were classified as nontasters, those whose
ratings were between 11.5 and 61 were classified as medium tasters, and those
whose ratings were >61 were classified as supertasters (see
Figure 1). Occasionally, when a
borderline rating for PROP was given by a subject, the rating was compared
with the rating for NaCl to help clarify group assignment
(Tepper et al.,
2001). For example, if a subject gave PROP a borderline rating of
61, but he/she gave NaCl a much lower rating, this individual would be
classified as a supertaster. Conversely, if the subject gave both PROP and
NaCl a borderline rating of 61, the individual would be classified as a medium
taster. Twenty-one nontasters, 28 medium tasters and 27 supertasters were
identified. Follow-up analyses also revealed that the NaCl mean intensity
ratings were not significantly different among the three taster groups
(P = 0.20), which has been observed in previous studies
(Tepper and Nurse, 1997;
Tepper et al., 2001;
Tepper and Ullrich, 2002;
Zhao et al.,
2003).
|
FCP subjects
Ten nontasters and 10 supertasters were selected from this subject pool to
participate in the FCP study. There were five females and five males in each
PROP taster group. Across the groups, subjects were matched as closely as
possible for age, height and weight. All participants were classified as
unrestrained eaters using a 10 question abbreviated version of the
Three-Factor Eating Questionnaire (Tepper
et al., 1997). All subjects were considered semi-trained,
as they had frequently participated in other sensory evaluations with a
variety of food products, and were familiar with tasting and evaluating
samples. Previously, Heymann suggested that FCP should be used with subjects
with prior experience in sensory methods, as `sensory naive' subjects did not
produce consistent results (Heymann,
1994).
Taste stimuli
The nine foods used for the FCP study represented a range of market dairy
products. They were chosen based upon their fat content, viscosity, sweetness
and flavor (see Table 1). Many
of the same foods were used in a previous study by Kokini and Cussler, which
investigated the texture of dairy products
(Kokini and Cussler, 1983).
All products were obtained from local New Jersey grocery stores and served at
their customary refrigerated temperature, except the vanilla ice cream, which
was served frozen, and the sweetened condensed milk, which was served at room
temperature.
|
Procedure
The FCP study started with two orientation sessions. Each subject completed the series individually. In the orientation sessions, each subject tasted the sample set and developed his or her own list of terms to describe the samples. Subjects were prompted to generate words to fully describe the appearance, taste/flavor and mouthfeel/texture of the products. The name of the attribute could be anything that the subject desired; however the subject had to be able to use the attribute consistently across all products. Subjects were informed that they would use their personal list of attributes in future tasting sessions.
There were four evaluation sessions; five samples were served in the first session followed by the remaining four samples in the second session. The final two sessions included a replicate evaluation of all products. Approximately 2 oz. samples were served in a counterbalanced order in 4 oz. soufflé cups under white lighting. Subjects rated the terms on a 15 cm line scale with the anchors of none at all to extremely strong using Compusense version 4.0 direct data entry software (Guelph, Ontario, Canada).
Analysis
The total and average number of terms used by nontasters and supertasters to describe the samples were separately tallied by appearance, flavor and texture attributes. Group differences were analyzed by Analysis of Variance (ANOVA).
Generalized Procrustes Analysis (GPA) was then used to condense the
individual evaluations into a consensus space
(Oreskovich et al.,
1991). Individual attribute means cannot be calculated, because
the terms are not standardized across subjects. Interpretation of the GPA data
was obtained by examining the position of the foods and attributes in the
n-dimensional space. Separate n-dimensional spaces were
calculated for nontasters and supertasters.
The GPA was performed with the statistical analysis software Senstools (OP&V Research, Guelph, Ontario, Canada). Eigenvalues and variance for each dimension were computed using the Procrustes Analysis of Variance (PANOVA), which demonstrates the relative importance of each dimension to the model. Then, a permutation test was conducted on the total variance accounted for from the PANOVA, to indicate if a true consensus space among subjects was obtained.
The final output of the GPA analysis is a visualization of the foods and attributes in an n-dimensional space. To interpret the spaces, three pieces of information are assimilated including (i) the product orientations in n-dimensional space, (ii) the dimension loadings (how much each dimension captures the total variance), and (iii) how the attributes relate to the dimensions. Significant relationships between attributes and dimensions were identified by correlations >0.5 on the main axis and <0.3 on all other axes.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Lexicon
The total number of terms used by individual subjects to describe the products ranged from 14 to 53. PROP supertasters and nontasters did not differ in the average number of terms used (P = 0.46). The terms were separately tallied by appearance, flavor/taste and mouthfeel/texture characteristics (see Table 2). The terms sweet, sour, vanilla, creamy, smooth and thick were used by most of the subjects in both the nontaster and supertaster groups, however supertasters tended to use different mouthfeel/texture terms than nontasters.
|
Generalized Procrustes Analysis
GPA analysis was conducted for appearance, flavor and texture attributes. Separate analyses were performed for the nontaster and supertaster groups. The analyses revealed that both groups used appearance attributes similarly in their consensus spaces. Therefore, to better understand the contribution of oral sensory perception to these products, the appearance attributes were removed from subsequent analyses. The following section describes models based on the flavor and texture evaluations the samples.
GPA revealed that the data from each taster group was resolved in three
major dimensions, each with eigenvalues greater than one. The total variance
accounted for in three dimensions was 77.1% for the nontasters and 75.2% for
the supertasters. These values were significantly different from chance as
determined by the permutation test, indicating that a true consensus space was
achieved (P 
0.05). Dimensions 1 and 2 accounted for the majority
of the variance in the models, and together were also significantly different
from chance by the permutation test (P 0.05). Dimension 3 was
difficult to interpret due to a limited number of terms loading on the
dimension. Therefore, the final analyses focused on the solution in two
dimensions.
For both the nontasters and supertasters, the total variance accounted for
in the two-dimensional models was 54%. For nontasters, the percent of
variance accounted for in Dimensions 1 and 2 were approximately equal (28 and
26%, respectively). For supertasters, Dimension 1 accounted for a higher
percent of variance in the model (34%) as compared to Dimension 2 (20%).
The two-dimensional space for nontasters is described in Figure 2. The subjects' attribute labels are shown in italics. For nontasters, Dimension 1 was described as a continuum from sweet taste/flavor to sour and buttermilk taste/flavor. Dimension 2 was described by the terms milky and bland taste/flavor on one end, to sweet taste/flavor and mouthcoating, thick mouthfeel/texture on the other end.
|
For supertasters (Figure 3), the two dimensions are rotated. Dimension 1 reflected a gradation of dairy flavor and texture terms. The terms rich, buttery, creamy taste/flavor and creamy, thick, coating, heavy mouthfeel/texture described one end of the continuum. The other end of the continuum included watery, light mouthfeel/texture and bland, watery, milky/dairy taste/flavor. Dimension 2 consisted of basic tastes plus textural terms, such that sour and salty taste/flavor were contrasted with grainy, gritty, sandy mouthfeel/texture. The latter textural terms were absent from the nontasters basic taste dimension.
|
The figures also show the relationship between the products in the two-dimensional spaces. The relative positions of the foods on the dimensions were similar for nontasters and supertasters. The foods that loaded on the sweet–sour continuum (Dimension 1 for nontasters and Dimension 2 for supertasters) included sour cream, cream cheese and vanilla ice cream. Sweetened condensed milk, whole milk and skim milk loaded high on the dairy flavor/texture dimension (Dimension 2 for nontasters and Dimension 1 for supertasters). For supertasters and nontasters, the foods maintained the same position relative to the major axis, but the order of the dimensions was reversed.
For both taster groups, the dimensions of the model explain the major characteristics of dairy products. However, supertasters used more terms overall to describe their perceptions, and also used a greater number of attributes related to texture.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The objectives of this study were to gain insight into individual differences in perception to better understand the cues for creaminess and to assess the contribution of flavor and texture to the overall perception of creaminess. These goals were accomplished by investigating individual descriptions and consensus spaces of dairy products from FCP and by studying the group effects of PROP-classified subjects.
Overall, there were a number of similarities in the perceptual spaces of
nontasters and supertasters. The models for both groups were resolved in two
dimensions and were described by a dairy flavor/texture axis and a
sweet–sour axis. Both models accounted for 54% of the variance
in the data. Also, the products were arranged in a similar pattern along the dimensions in the models for both groups. Thus, both groups described their perceptions of the products using a combination of dairy flavor and texture terms plus basic tastes.
There were also striking differences in certain features of the models of nontasters and supertasters. Nontasters used a limited number of simple terms such as sour, sweet, milky and mouthcoating to describe the products. In contrast, supertasters used a more complex vocabulary, including terms such as rich, buttery, creamy, light, grainy, gritty and sandy. The model for nontasters gave equal weight to the sweet–sour and dairy flavor/texture dimensions (28% and 26% variance, respectively); whereas, the model for supertasters relied more heavily on the dairy flavor/texture dimension (34% variance), and less so on the sweet–sour dimension (21% variance). Thus, although the perceptual spaces for creaminess were similar for nontasters and supertasters, the cues used to judge creaminess differed for the two groups.
A fundamental question raised by this research is whether supertasters
perceive more creaminess intensity in dairy products. Studies have
investigated this question but have failed to come to come to a consensus on
the issue. Duffy et al. (Duffy
et al., 1996) reported that PROP tasters perceived more
creaminess intensity in high-fat milk products (11.5–54% fat). Other
research found no differences in creaminess and/or fattiness perception among
PROP taster groups for sweet-fat milk mixtures
(Drewnowski et al.,
1998) or chocolate dairy drinks and vanilla puddings varying in
flavor and fat content (Yackinous and
Guinard, 2001). In contrast, Tepper et al.
(Tepper et al., 2002)
recently reported a moderate effect of PROP status on judgements of creaminess
intensity in sweetened and unsweetened milks. Creaminess ratings of milks with
increasing fat content rose more rapidly for supertasters than for medium
tasters and nontasters. Similar results were obtained for sweetened and
unsweetened samples. Although the present study did not directly compare
intensity judgments across taster groups, the GPA models revealed qualitative
differences in the use of descriptive terms for creaminess by nontasters and
supertasters. Thus, current evidence supports a role for PROP status in the
perception creaminess in dairy products, but the magnitude of this effect
remains controversial.
Whatever advantage PROP tasters might have in perceiving creaminess may be
due to greater acuity of oral texture sensations arising from variation in
tongue anatomy (Bartoshuk et al.,
1994; Duffy et al.,
1996; Tepper and Nurse,
1997). Supertasters have more taste papillae innervated by
trigeminal and other nerve fibers, which may produce a greater somatosensory
sensation on the tongue. As expected, supertasters in the present study
utilized more textural terms to describe the products including those related
to creaminess, heaviness and particulates. An unexpected finding was that
supertasters also employed more flavor terms to describe the products. This
finding raises the intriguing possibility that flavor perception could also
vary with PROP taster status. Support for this idea comes from a recent study
showing that PROP supertasters have a lower olfactory threshold than
nontasters for diacetyl, a butter-type flavor found in many dairy products
(Yackinous and Guinard, 2001).
Specific anosmia to this compound has been previously reported, but the nature
of the olfactory blindness is poorly understood
(Lawless et al.,
1994). Diacetyl may also elicit pungency or nasal irritation
(Arctander, 1994), contributing
to the reported differences in olfactory thresholds. Thus, it is possible that
PROP taste responsivity plays a role in other sensory systems such as
olfaction and nasal pungency. Further investigation of these areas is
warranted.
It was initially expected that using FCP would help to deconstruct
creaminess into its constituent parts, projecting flavor and texture terms
onto separate dimensions of the models. However, the results of the present
study did not support this hypothesis. Our results agree with the findings
from an earlier study by Li et al. on FCP of ice creams
(Li et al., 1997). In
that study, both flavor and texture terms also loaded on the same dimension.
Tepper and Kuang (Tepper and Kuang,
1996) used multidimensional scaling to investigate the perception
of milk model systems. They found that flavor and texture loaded on separate
dimensions of the model. However, two features of the Tepper and Kuang study
design might have facilitated this separation. First, fat content and dairy
flavor were separately manipulated in the samples. Second, subjects were
instructed to rate the degree of difference among sample pairs for specified
attributes; they were not free to choose their own descriptors. Thus, the
present data support the notion that flavor and texture may be so highly
integrated in dairy products that humans do not easily separate these
sensations. The exception is descriptive training, which is designed to
deconstruct complex sensory stimuli into their component parts
(Meilgaard et al.,
1991).
Theories of flavor interactions may provide insight into the integration of
the senses in creaminess perception. Prescott
(Prescott, 1999) recently
discussed the concept of flavor in terms of a psychological construct produced
by odor and taste integration. Implicit in the concept is the idea that
odor/taste mixtures are sensed as unique flavor entities rather than as the
sum of their individual components. Learning may play a significant role in
this interaction, as demonstrated by experiments showing that taste
enhancement is more pronounced for flavor–odor pairs that are familiar
and typically appear together in foods. A common example is the pairing of
strawberry odor with sucrose, which enhances the perceived sweetness of the
mixture. The same pairing with peanut butter does not produce this effect
(Frank and Byram, 1988). This
concept might also apply to creaminess perception, whereby the pairing of an
appropriate flavor with a dairy base enhances the overall
creaminess impression. Experiments by Lawless and Clark
(Lawless and Clark, 1992) and
Tepper and Kuang (Tepper and Kuang,
1996) showed that adding vanilla or dairy flavor to milk systems
enhanced the creaminess perception.
The milk fat present in dairy products has unique properties as a source of stimulus fat and volatile flavor compounds. Increasing the concentration of fat in a dairy food also enhances the release of other volatile compounds that may already be present, further enhancing the overall creaminess impact. Thus, `creamy flavor' and `creamy texture' may be inexorably linked in the experience of the human assessor because they are physically linked in real dairy products. The neural mechanisms that give rise to these individual sensations and the psychophysical processes that integrate them are poorly understood. Elucidation of these mechanisms would improve our understanding of the effects of physical properties on sensory perceptions.
The GPA models developed in the present study captured 54% of the
variance in two dimensions, which is similar to the results reported by Raats
and Shepherd (Raats and Shepherd,
1992). In their study, a two-dimensional model was obtained for
milks differing in fat content, which accounted for 57% of the variance
in the data. Li et al. (Li et
al., 1997) captured over 86% of the variance in three
dimensions, however their solution was primarily uni-dimensional, as the first
dimension describing the fat content of ice creams contributed over 81% of the
variance. In the current study, three-dimensional models of these data were
studied initially; however, the third dimension was difficult to interpret
because few terms loaded on this dimension. Also, the appearance terms were
removed from the analyses in an effort to understand how sensations in the
mouth rather than visual characteristics contributed to creaminess. Including
the appearance evaluations in the analyses increased the percent of variance
captured in the models by 25% for both nontasters and supertasters.
However, Gonzalez Vinas et al.
(Gonzalez Vinas et al.,
2000) have argued that in real food products, appearance can be an
overwhelming driver of the model minimizing the contributions of flavor and
textural assessments of the products. Thus, eliminating the appearance
characteristics seemed justified in light of these data.
The present study combined existing sensory techniques in a novel approach
to investigate creaminess perception in dairy products. The rationale for
selecting these methods bears mentioning, given the current debate surrounding
the use of scaling methods (Green et
al., 1993; Bartoshuk, 2000;
Bartoshuk et al.,
2002), particularly as they relate to PROP screening and
classification (Tepper et al.,
2001; Bartoshuk et
al., 2002; Rankin et
al., 2003; Zhao et
al., 2003). PROP screening in the present study followed the
methodology of Tepper et al.
(Tepper et al.,
2001), which utilizes the LMS. This method is valid as compared
with an accepted procedure (Bartoshuk
et al., 1994) and has high test–retest reliability
as reported by Tepper et al. and Rankin et al.
(Tepper et al., 2001;
Rankin et al., 2003).
Group assignments in this method depend primarily on numerical cutoff scores
for PROP taste intensity; NaCl ratings are used to clarify group assignment
when borderline ratings are given [see Tepper et al.
(Tepper et al., 2001)
for a full explanation of these procedures]. The validity of using NaCl as a
reference standard has been called into question on the basis of data reported
by Pruntkin et al. (Pruntkin
et al., 2000) who found that intensity ratings for NaCl
varied with PROP status. Also, Prescott et al.
(Prescott et al.,
2001) studied binary mixtures of tastants and found a small but
statistically significant increase in saltiness intensity for NaCl/quinine
hydrochloride mixtures among supertasters. These data contrast with numerous
observations from this laboratory, which show no systematic differences in the
intensity of NaCl solutions as a function of taster status
(Tepper and Nurse, 1997;
Tepper et al., 2001;
Tepper and Ullrich, 2002;
Zhao et al., 2003).
The reasons for this disparity are presently unknown and require further
study.
A standard, 15 cm line scale was used to collect ratings for the dairy
products in the present study. This scale has a long history of use in
descriptive analysis and free-choice profiling
(Meilgaard et al.,
1991) and permitted the models obtained in this study to be
directly compared with data from other FCP studies on dairy products. Research
by Bartoshuk and colleagues (Bartoshuk, 2000;
Bartoshuk et al.,
2002) has emphasized the importance of preserving valid
across-group comparisons with the use of labeled scales. The generalized LMS
(gLMS), a variant of the LMS, seeks to address this issue by placing all
stimuli (including gustatory, visual, auditory, etc.) on a common scale with
the upper anchor labeled as `strongest imaginable sensation of any kind'. A
universal scale would be desirable on theoretical grounds because it permits
subjects the maximum freedom to rate stimuli and would remain independent of
context. At present, published data comparing the gLMS to other scales are
limited. However, recent work by Horne et al.
(Horne et al., 2002)
found no advantage of the gLMS as compared with the LMS in a study examining
sweetness perception in PROP classified individuals. Additional studies should
be conducted on the gLMS to establish its true capabilities.
In conclusion, the results of this study showed that the term `creaminess' may have different meanings to individuals, a finding that could have important implications for language development and utilization of terms in sensory testing. Our findings also showed that characterizing subjects by PROP status was an effective tool for systematically studying individual differences in perception that are biologically mediated. Previously, this variation has been attributed to experimental `noise' and differences in methodology. Finally, it is important to determine if the results obtained here are relevant to consumer perception and acceptance of products. Future studies will address this topic.
|
Acknowledgments |
|---|
This research was funded by International Flavors and Fragrances, Inc., Dayton, NJ and was conducted in partial fulfillment of the Ph.D. degree by S.V.K. Portions of this work were presented at the 2002 Association for Chemoreception Sciences Annual Meeting and the 2002 European Chemoreception Research Organization Satellite Symposium on `Sensitivity to PROP (6-n-propylthiouracil): Its Measurement, Significance and Implications'.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Arctander, S. (1994) Perfume and Flavor Chemicals (Aroma Chemicals). Carol Stream, Allured Publishing Corporation.
Bartoshuk, L.M. (2000a) Comparing sensory
experiences across individuals: recent psychophysical advanced illuminate
genetic variation in taste perception. Chem. Senses,25
, 447-460.
Bartoshuk, L.M. (2000b) Psychophysical advances aid the study of genetic variation in taste.Appetite , 34,105 .[CrossRef][Web of Science][Medline]
Bartoshuk, L.M., Duffy, V.B., Fast, K., Green, B.G., Pruntkin, J.M. and Snyder, D.J. (2002) Labeled scales (e.g., category, Likert, VAS) and invalid across-group comparisons: what we have learned from genetic variation in taste. Food Qual. Pref.,14 , 125-138.
Bartoshuk, L.M., Duffy, V.B. and Miller, I.J. (1994) PTC/PROP tasting: anatomy, psychophysics, and sex effects. Physiol. Behav., 56,1165 -1171.[CrossRef][Medline]
Bom Frost, M., Dijksterhuis, G. and Martens, M. (2001) Sensory perception of fat in milk. Food Qual. Pref., 12,327 -336.[CrossRef]
Daget, N., Joerg, M. and Bourne, M. (1987) Creamy perception I: in model dessert creams.J. Texture Stud. , 18,367 -388.
Drewnowski, A., Henderson, S.A. and Barratt-Fornell, A. (1998) Genetic sensitivity to 6-n-Propylthiouracil and sensory responses to sugar and fat mixtures. Physiol. Behav., 63,771 -777.[CrossRef][Medline]
Duffy, V.B., Bartoshuk, L.M., Lucchina, L.A., Snyder, D.J. and Tym, A. (1996) Supertasters of PROP (6-n-propylthiouracil) rate the highest creaminess to high-fat milk products (abstract). Chem. Senses, 21,598 .
Elmore, J.R., Heymann, H., Johnson, J. and Hewett, J.E. (1999) Preference mapping: relating acceptance of `creaminess' to a descriptive sensory map of a semi-solid. Food Qual. Pref., 10,465 -475.[CrossRef]
Frank, R.A. and Byram, J. (1988)
Taste–smell interactions are tastant and odor dependent.Chem. Senses
, 13,445
-455.
Gonzalez Vinas, M.A., Garrido, N. and Wittig de Penna, E. (2000) Free choice profiling of Chilean goat cheese. J. Sens. Stud., 16,239 -248.
Green, B.G., Shaffer, G.S. and Gilmore, M.M.
(1993) Derivation and evaluation of a semantic scale of oral
sensation magnitude with apparent ratio properties. Chem.
Senses, 18,683
-702.
Green, B.G., Dalton, P., Cowart, B., Shaffer, G., Rankin, K.
and Higgins, J. (1996) Evaluating the `Labeled
Magnitude Scale' for measuring sensations of taste and smell.Chem. Senses
, 21,323
-334.
Heymann, H. (1994) A comparison of free choice profiling and multidimensional scaling of vanilla samples.J. Sens. Stud. , 9,445 -453.
Horne, J., Lawless, H.T., Speirs, W. and Sposato,
D.J. (2002) Bitter taste of saccharin and
acesulfame-K. Chem. Senses, 27,31
-38.
Karrer, T. and Bartoshuk, L.M. (1991) Capsaicin desensitization and recovery on the human tongue.Physiol. Behav , 49,757 -764.[CrossRef][Medline]
Kokini, J.L. and Cussler, E.L. (1983) Predicting the texture of liquid and melting semi-solid foods.J. Food Sci. , 48,1221 -1225.[CrossRef]
Lawless, H.L. and Clark, C.C. (1992) Psychological biases in time-intensity scaling. Food Technol., 11,81 -90.
Lawless, H.L., Antinone, M.J., Ledford, R.A. and Johnston, M. (1994) Olfactory responsiveness to diacetyl. J. Sens. Stud., 9,47 -56.
Li, Z., Marshall, R., Heymann, H. and Fernando, L. (1997) Effect of milk fat content on flavor perception of vanilla ice cream. J. Dairy Sci.,80 , 3133-3141.[Abstract]
Lucca, P.A. and Tepper, B.J. (1995) Fat replacers and the functionality of fat in foods. Trends Food Sci. Technol., 5,12 -18.
Lucchina, L.A., Curtis, O.F., Putnam, P., Drewnowski, A., Pruntkin, J.M. and Bartoshuk, L.M. (1998) Psychophysical measurement of 6-n-propylthiouracil (PROP) taste perception. Ann. N. Y. Acad. Sci.,855 , 816-819.[CrossRef][Web of Science][Medline]
Meilgaard, M., Civille, G.V. and Carr, B.T. (1991) Sensory Evaluation Techniques. CRC Press, New York.
Mela, D.J. (1988) Sensory assessment of fat content in fluid dairy products. Appetite,10 , 37-44.[Web of Science][Medline]
Oreskovich, D.C., Klein, B.P. and Sutherland, J.W. (1991) Procrustes analysis and its applications to free-choice and other sensory profiling. In H.T. Lawless and B.P. Klein (eds), Sensory Science Theory and Applications in Foods. New York, Marcel Dekker, pp. 353-393.
Prescott, J. (1999) Flavour as a psychological construct: implications for perceiving and measuring the sensory qualities of food. Food Qual. Pref.,10 , 349-356.[CrossRef]
Prescott, J., Ripandelli, N. and Wakeling, I. (2001) Binary taste mixture interactions in PROP non-tasters, medium-tasters and super-tasters (abstract). Chem. Senses,16 , 993-1003.
Pruntkin, J.M., Duffy, V.B., Etter, L., Fast, K., Gardner, E., Lucchina, L.A., Snyder, D.J., Tie, K., Weiffenbach, J. and Bartoshuk, L.M. (2000) Genetic variation and inferences about perceived taste intensity in mice and men. Physiol. Behav., 60,161 -173.
Raats, M.M. and Shepherd, R. (1992) Free-choice profiling of milks and other products prepared with milks at different fat contents. J. Sens. Stud,,7 , 179-203.
Rankin, K.M., Godinot, N., Tepper, B.J., Kirkmeyer, S.V. and Christensen, C.M. (2003) Assessment of different methods for PROP status classification. In J. Prescott and B.J. Tepper (eds), Genetic Variation in Taste Sensitivity. New York, Marcel Dekker, in press.
Richardson, N.J., Booth, D.A. and Stanley, N.L. (1993) Effect of homogenization and fat content on oral perception of low and high viscosity model creams. J. Sens. Stud., 8,133 -143.
Richardson-Harman, N.J., Stevens, R., Walker, S., Gamble, J., Miller, M., Wong, M. and McPhearson, A. (2000) Mapping consumer perceptions of creaminess and liking for liquid dairy products. Food Qual. Pref., 11,239 -246.[CrossRef]
Tepper, B.J. (1998) 6-n-Propylthiouracil: a genetic marker for taste, with implications for food preference and dietary habits. Am. J. Hum. Genet., 63,1271 -1276.[CrossRef][Web of Science][Medline]
Tepper, B.J. and Kuang, T. (1996) Perception of fat in a milk model system using multidimensional scaling. J. Sens. Stud., 11,175 -190.[Web of Science]
Tepper, B.J. and Nurse, R.J. (1997) Fat perception is related to PROP taster status. Physiol. Behav., 61,949 -954.[CrossRef][Medline]
Tepper, B.J. and Ullrich, N.V. (2002a) Influence of genetic taste sensitivity to 6-n-propylthiouracil (PROP) dietary restraint and disinhibition on body mass index in middle-aged women. Physiol. Behav., 75,305 -312.[CrossRef][Medline]
Tepper, B.J. and Ullrich, N.V. (2002b) Taste, smell and the genetics of food preferences. Topics Clin. Nutr., 17,1 -14.
Tepper, B.J., Choi, Y.S. and Nayga Jr, R.M. (1997) Understanding food choice in adult men: influence of nutrition knowledge, food beliefs and dietary restraint. Food Qual. Pref., 8,307 -317.[CrossRef]
Tepper, B.J., Christensen, C.M. and Cao, J. (2001) Development of brief methods to classify individuals by PROP taster status. Physiol. Behav.,73 , 571-577.[CrossRef][Medline]
Tepper, B.J., Goldstein, G.L. and Ullrich, N.J. (2002) PROP taster status and sensory responses to sweetened and unsweetened milk mixtures. Society for the Study of Ingestive Behavior: Annual Meeting 2002, Santa Cruz, CA.
Tuorila, H. (1986) Sensory profile of milks with varying fat contents. Lebensmittel-Wisenechaft Technol., 19,344 -345.
Yackinous, C. and Guinard, J.-X. (2001) Relation between PROP taster status and fat perception, touch and olfaction. Physiol. Behav., 72,427 -437.[CrossRef][Medline]
Zhao, L., Kirkmeyer, S.V. and Tepper, B.J. (2003) A paper screening test to assess genetic taste sensitivity to 6-n-propylthiouracil. Physiol. Behav.,78 , 625-633.[CrossRef][Medline]
Accepted June 2, 2003
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Lim, L. Urban, and B. G. Green Measures of Individual Differences in Taste and Creaminess Perception Chem Senses, July 1, 2008; 33(6): 493 - 501. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. D. Mattes Effects of linoleic acid on sweet, sour, salty, and bitter taste thresholds and intensity ratings of adults Am J Physiol Gastrointest Liver Physiol, May 1, 2007; 292(5): G1243 - G1248. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Hayes and V. B. Duffy Revisiting Sugar-Fat Mixtures: Sweetness and Creaminess Vary with Phenotypic Markers of Oral Sensation Chem Senses, March 1, 2007; 32(3): 225 - 236. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||