Food Research International 51 (2013) 310–320

Contents lists available at SciVerse ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Textural properties of model food sauces: Correlation between simulated mastication

and sensory evaluation methods
Cheryl Chung a, Kerstin Olson b, Brian Degner b, David Julian McClements a,⁎
a
Department of Food Science, University of Massachusetts, Amherst, MA, 01003, USA
b
ConAgra Foods, Six ConAgra Drive, Omaha, NE 68102, USA

a r t i c l e i n f o a b s t r a c t

Article history: The ability of a recently developed simulated mastication method to characterize the rheological properties of
Received 22 October 2012 model food sauces was determined. This method uses integrated compression–shear–decompression motions
Accepted 5 December 2012 to mimic the motions of the tongue against the palate, and includes the addition of saliva. Model sauces
Available online 20 December 2012
contained swollen starch granules (3.75%), fat droplets (0, 5, 10, and 15%), and hydrophilic polymer (0 or 0.2%
locust bean gum). The texture of the sauces was characterized by the simulated mastication method, shear vis-
Keywords:
Rheology
cometry, and sensory methods, and the parameters derived from the instrumental analysis were correlated to
Oral simulation those derived by sensory analysis. The optical properties (lightness) and microstructure (light scattering and mi-
Mastication croscopy) of the sauces were also measured. There was a good correlation (r2 > 0.95) between the “consistency”
Reduced fat (maximum peak force) derived from the simulated mastication method and the “viscosity/thickness” derived
Starch from sensory analysis. In both cases, the consistency of the sauces increased with increasing fat content and
Sensory with locust bean gum addition. The microstructures of sauces masticated within a human mouth and within
the simulated mastication instrument were similar. Overall our results suggest that the simulated mastication
method can be used to characterize the textural properties of semi-solid foods within the mouth.
© 2012 Elsevier Ltd. All rights reserved.

1. Introduction better simulate the flow profiles experienced by foods in the oral cavity
during mastication (Le Reverend et al., 2010; Lillford, 2000).
The food industry requires instrumental methods to measure the tex- We recently developed a simulated mastication method that applies
tural characteristics of food products (Bourne, 2002). These methods are both normal and shear forces to a sample within the same instrument
used to develop a basic understanding of the major factors that contribute (Chung, Degner, & McClements, 2012a). This method integrates the
to food texture, e.g., the impact of food composition, structure, interac- principles of squeezing flow rheology and shear viscometry in a single
tions, or environmental conditions. They are also used as screening instrument. Food samples are compressed–sheared–decompressed for
tools to predict the sensory attributes of foods during the product devel- a number of cycles between two horizontal parallel plates to simulate
opment process. A wide variety of instrumental methods have been de- the movement of the tongue and palate during mastication. A commer-
veloped to measure the textural properties of foods and to correlate cial dynamic rheometer capable of measuring both shear and normal
them to specific sensory attributes (Bourne, 2002; van Aken, 2007). forces acting upon an upper plate was utilized (Fig. 1). The device con-
These instrumental methods are often categorized according to the na- sists of a measurement cell that has two stainless steel plates: an upper
ture of the forces applied to the material being tested (McClements, plate (d=40 mm) and a lower plate (d=65 mm). The lower plate is
2005). Typically, most instrumental rheological instruments apply either fixed, while the upper plate is capable of moving upwards/downwards
normal forces (e.g., compression or elongation testing) or shear forces and rotating. Initially, the two plates are moved apart and the sample
(e.g. viscometry or dynamic shear rheometry). In practice, both normal to be tested is placed on the lower plate. The upper plate is then lowered
and shear forces are applied to food materials during the mastication pro- so that there is a gap of 20 mm between the two plates. This distance
cess, and complex flow profiles are generated within the mouth (de was selected based on the jaw gape width reported for a semi-solid
Bruijne, Hendrickx, Anderliesten, & de Looff, 1993; Le Reverend, Norton, food (spread) during mastication within the human mouth (Hiiemae &
Cox, & Spyropoulos, 2010; Lillford, 2000). The successful development Palmer, 1999). The food sample is then subjected to a number of com-
of instrumental methods to provide information about the textural attri- pression–shearing–decompression cycles, with each cycle having the
butes of foods within the mouth therefore relies on test methods that following form:

Compression stage: The material is compressed from 20 to 1 mm

⁎ Corresponding author. Tel.: +1 413 545 1019; fax: +1 413 545 1262.
(down stroke) for the first cycle by bringing the upper plate down-
E-mail address: [email protected] (D.J. McClements). ward at a constant speed of 17.8 mm/s. This speed was chosen based

0963-9969/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodres.2012.12.005
 C. Chung et al. / Food Research International 51 (2013) 310–320 311

The primary objective of this study was to examine the ability of

the newly developed simulated mastication assay to characterize dif-
ferences in the textural attributes of model sauces containing differ-
ent starch, fat, and polymer ratios, as well as to establish whether
the results derived from this method could be correlated to those
obtained from sensory analysis.

2. Materials and methods

2.1. Materials

Four formulations of model turkey sauces were prepared and

provided by ConAgra Foods (Omaha, Nebraska, US). The ingredients
used in the preparation of the sauces were fat (soybean oil), surfac-
tant (Panodan 150), modified starch (Maxi-Gel 542), locust bean
gum (LBG), turkey flavor, salt, sugar and distilled water. The modified
starch (Maxi-Gel 542) is a waxy maize modified corn starch with 10
to 13% moisture and 0.6% ash contents, which was reported to have
a gelatinization peak at 69 °C, an onset temperature at 62 °C, an offset
temperature at 75 °C, and an enthalpy change of 2.27 J/g (Tate & Lyle,
Decatur, IL, U.S.). This starch is covalently cross-linked (phosphorous
oxytrichloride) and substituted (hydroxypropyl) so that it does not
disintegrate after heating, i.e., the granules swell but remain intact
so that a high viscosity is maintained. The surfactant (Panodan 150)
is a mixture of diacetyl tartaric acid esters of mono-diglycerides
(DATEM) and mono-diglycerides, which is widely utilized to prepare
emulsion-based products in the food industry.
Alpha-amylase (98.3 units/mg solid from human), mucin from
porcine stomach (type II), ammonium nitrate, potassium phosphate
(>99%), potassium chloride (>99%), potassium citrate, uric acid sodium
Fig. 1. Schematic representations and photographic images of the instrumental masti-
salt, urea (99.5%) and sodium DL-lactate (99% and 112.06 g/mol) were
cation device designed to simulate the tongue and palate in the mouth. A sample is
loaded on the lower plate. The upper plate is moved downwards from a 20 to 1 mm purchased from Sigma-Aldrich Co. (St. Louis, MO, U.S.). Sodium chloride
gap during the compression stage (then a 3 to 1 mm gap on subsequent compression was purchased from Fisher Scientific Co. (Fair Lawn, N.J., U.S.). All
stages). The upper plate is held at a fixed gap of 1 mm (with or without shearing). The chemicals used were of analytical or biological certified grade. Double
upper plate is then moved upwards from 1 to 3 mm during the decompression stage. distilled water was used as the solvent throughout the study.

2.2. Methods
on the work of another group who employed a different instrument
to simulate mastication of polysaccharide gel systems (Ishihara, 2.2.1. Starch and locust bean gum systems preparation
Nakauma, Funami, Odake, & Nishinari, 2011a, 2011b). For the subse- Starch (3.75%) suspensions or locust bean gum (0.2%) solutions
quent cycles, the material is compressed from 3 mm to 1 mm with were prepared by dispersing the weighed amount of each component
the same speed as the first compression (17.8 mm/s). Normal forces into double distilled water. These systems were then heated to 90 °C
measured on the upper plate are recorded during the compression for 5 min holding time, and then cooled in a water bath at 25 °C for at
stage. least 1 h before analysis.
Fixed gap stage: The material is subjected to a constant shear rate
(e.g., 10 s −1) for 1 s. This shear rate is widely used to measure the 2.2.2. Model sauce preparation
apparent viscosity of materials under conditions designed to simu- Four formulations of model turkey sauces were prepared by ConAgra
late the flow rate of fluids within the human mouth (Cutler, Foods (Omaha, Nebraska, US). These sauces contained the same amount
of starch and flavoring ingredients, but different amounts of fat and LBG.
Morris, & Taylor, 1983; de Wijk, Prinz, & Janssen, 2006).
The surfactant-to-oil ratio was kept constant in all systems (rather than
Decompression stage: The material is decompressed from 1 mm to
the total amount of surfactant) so as to produce physically stable sauces
3 mm (upstroke) by bringing the upper plate upwards at a constant
(Table 1). Initially, a weighed amount of surfactant was dispersed in a
speed of 17.8 mm/s, and the normal forces acting on the upper plate weighed amount of oil phase and this blend was heated to 65 °C until
are measured during this process.
The compression/decompression motions of the upper plate are
used to simulate the upward/downward movement of the tongue
against the palate, whereas the shearing motion of the upper plate Table 1
is used to simulate the sliding of the tongue against the palate. The Formulations of model sauces containing different fat and locust bean gum (LBG) con-
tents. The other ingredients included turkey flavor, salt and sugar.
number of compression–shearing–decompression cycles used de-
pends on the typical movement of the mouth during mastication for Systems Fat Surfactant Starch LBG Other ingredients (%)
the particular kind of food being analyzed, e.g., the number of times (%) (%) (%) (%)
the food is normally chewed. Food materials are mixed with saliva 5% fat–0.2% LBG 5.00 0.13 3.75 0.2 2.1
during mastication, and therefore the test setup was designed so 5% fat 5.00 0.13 3.75 – 2.1
that artificial saliva could be added to the samples prior to applying 10% fat 10.00 0.25 3.75 – 2.1
15% fat 15.00 0.38 3.75 – 2.1
the compression/decompression cycles.
312 C. Chung et al. / Food Research International 51 (2013) 310–320

the surfactant was completely dissolved. Then, coarse emulsions were 2.3.4. Simulated mastication assay
created by blending the surfactant/oil phase with 25% of the total An instrumental rheometer with a plate-plate measurement cell was
water portion using a high shear blender (LR-51B, Yamato Scientific used to measure the textural attributes of samples during simulated mas-
Co., Japan) at ~145 rpm for 30 s. The coarse emulsions formed were tication as described in the Introduction section (Kinexus Rheometer,
then passed once through a high pressure valve homogenizer at Malvern Instruments Ltd., MA, USA). An aliquot of sample (1.3 mL) was
2500 psi to reduce their droplet size (GEA Mechanical Equipment/GEA loaded onto the lower plate (Fig. 1). After loading, the mastication test se-
Niro Soavi S.p.A.; Via da Erba Edoari, 29; 43123 Parma, Italy). The dry in- quence (compression–shearing–decompression) was activated and op-
gredients (starch, locust bean gum, turkey flavor, salt, and sugar) were erated through the instrument software (Kinexus rSpace Software,
dissolved separately in the remaining 75% of the total water at 25 °C. Fi- version 1.30 Malvern Instruments Ltd., MA, USA). Experiments were car-
nally, the model sauces were prepared by mixing the emulsions with the ried out at 60 °C to mimic the temperature of a hot sauce within the
aqueous solutions containing the dissolved ingredients and then heating mouth.
to 90 °C for 5 min with continuous stirring at ~145 rpm using a high The effect of saliva on the rheological properties of the sauces was
shear blender (LR-51B, Yamato Scientific Co., Japan). also measured. For this set of experiments, 1.1 mL of sample was
loaded onto the lower plate and 0.2 mL of artificial saliva (containing
2.3. System characterization salt solution, mucin, and α-amylase) was added on top of the sample.
Mucin solution containing mucin and salt components (without
2.3.1. pH and optical properties α-amylase) was used as a control to determine the effect of dilution
The pH of the systems was measured using a pH electrode probe and α-amylase in the artificial saliva. The composition of the artificial
(827 pH LAB, Metrohm). The lightness (L*) was measured using a col- saliva and mucin used in this study has been reported elsewhere
orimeter (ColorFlez EZ, HunterLab, Reston, Virginia). (Chung et al., 2012a). The artificial saliva contained 3% porcine
mucin, a concentration that has previously been reported to simulate
2.3.2. Particle size the viscosity of human saliva (Gal, Fovet, & Adib-Yadzi, 2001; Shellis,
The particle size distribution of the systems was measured using a 1978) and α-amylase at an activity of 93 units/mL, which is similar to
laser diffraction particle size analyzer (Mastersizer 2000, Malvern Instru- that reported in other studies (Ades, Kesselman, Ungar, & Shimoni,
ments, Ltd., Worcestershire, UK). Samples were diluted by adding small 2012; Mandel, Peyrot des Gachons, Plank, Alarcon, & Breslin, 2010).
aliquots into a measurement chamber containing water until the instru-
ment gave an optimum obscuration rate of between 10 and 20%. The par-
ticle size distribution was calculated from the light scattering pattern 2.3.5. Simplified descriptive sensory tests
using Mie theory. A refractive index of 1.33 was used for the aqueous Four turkey sauces with different fat levels were presented to 25
phase, and 1.472 for the disperse phase. Particle size measurements are individuals (non-trained panel) at ConAgra Foods, Inc. The samples
reported as volume-weighted mean diameters (d43). were labeled 1 to 4; Sample 1 contained 5% fat; Sample 2 contained
10% fat; Sample 3 contained 15% fat; and, Sample 4 contained 5% fat
2.3.3. Apparent viscosity and flow behavior and 0.2% locust bean gum. Locust bean gum was used as an example
The apparent shear viscosity of the systems was measured using a dy- of a non-caloric thickening agent. The sauce samples were initially
namic shear rheometer (Kinexus Rheometer, Malvern Instruments Ltd., prepared at ~ 60 °C in 30 mL cups, but they were all cooled to around
MA, USA) with a plate-and-plate measurement cell. All measurements room temperature prior to serving to panelists, to avoid having ap-
were performed using shear rates from 0.1 to 100 s−1 at 60 °C (a temper- preciable alterations in sample temperatures between panelists. Pan-
ature at which food is normally consumed after heating/cooking). The elists were allowed to compare the samples freely, in no particular
measurement cell was set at 60 °C, and the sample was allowed to equil- order. All samples were served at the same time and panelists were
ibrate to this temperature after sample loading prior to starting the rheo- allowed to re-visit each sample as needed. Panelists were asked to
logical measurements. The time interval between measurements was rate the following attributes on a 1 to 5 scale (1 being the lowest, 5
automatically determined by the instrument software according to the being the highest): “overall liking”; “texture liking”; “texture smooth-
time required for the samples to reach steady state conditions at each ness/creaminess”; “texture viscosity/thickness”; “appearance liking”;
shear rate (Kinexus rSpace Software, version 1.30 Malvern Instruments and, “visual whiteness”. The individual 25 scores for each attribute
Ltd., MA, USA). The rheological properties of the samples were character- were tallied and the mean values are reported. The results were plot-
ized using the Herschel–Bulkley (HB) model for non-ideal plastic mate- ted as a spider diagram.
rials (Herschel & Bulkley, 1926):
2.3.6. Microstructure analysis
n
τ−τ0 ¼ K γ_ : ð1Þ A comparison of the microstructures of 15% fat systems before and
after mastication in the mouth and by the simulated mastication method
was also performed. 1.3 mL of the sauce was placed on the tongue of an
Here τ is the shear stress (Pa), τ0 is the yield stress (Pa), γ_ is the individual (adult female) and masticated for 10 s. The instrumental mas-
shear rate (s −1), and K is the consistency index (Pa.s), and n is the tication of the sample with added artificial saliva was performed as
power law index. The unknown parameters (τ0, K, n) in the above described above (Section 2.3.4). A small aliquot of each sample was
equations were found by finding the best-fit of the HB model to the placed on a microscope slide and covered with a cover slip and examined
experimental measurements of shear stress versus shear rate using under optical microscopy (40× objective lens and 10× eyepiece) (Nikon
the rheometer software (Kinexus rSpace Software, version 1.30 Malvern D-Eclipse C1 80i, Nikon, Melville, NY, U.S.). The microstructure images
Instruments Ltd., MA, USA). It should be noted that we assumed were analyzed using image analysis software (Nikon, Melville, NY, USA).
that there were no wall-slip effects in our rheological measurements.
Wall-slip effects are important in certain types of dispersed systems
(Buscall, 2010), and impact the measured rheological properties of 2.4. Statistical analysis
these systems. The reported data should therefore be seen as mainly pro-
viding information about differences in the textural characteristics of the Duplicate or triplicate analyses were performed for all measure-
different samples studied. Measurement problems associated with ments. Means, standard deviations and statistical differences between
wall-slip and how to overcome them have recently been reviewed for means (p b 0.05) were calculated using two-way analysis of variance
dispersed systems (Buscall, 2010). (ANOVA) using Microsoft Excel 2011.
 C. Chung et al. / Food Research International 51 (2013) 310–320 313

3. Results and discussion 18

Starch only
3.1. pH 16 Starch + Fat

The model sauces had pH values ranging from 4.2 to 4.9 with the 14
pH decreasing with increasing fat (and therefore surfactant) content
(Table 2). A recent study in our laboratory also found that the pH of
12
model sauces containing the same surfactant decreased as the amount
of surfactant present increased, which was attributed to the acidic na-

Volume (%)
ture of the anionic surfactant used (Chung, Degner, & McClements, 10
2012b).
8
3.2. Particle size distribution
6
The heated starch system and all the model sauces had fairly similar
mono-modal particle size distributions (Fig. 2) and mean particle diam-
4
eters (d43) falling between 39 and 41 μm (Table 2). Only a single peak
was detected by the light scattering technique despite the fact that the
sauces contained both oil droplets and starch granules. The absence of 2
a small peak for the oil droplets in the light scattering data (despite see-
ing 2 to 10 μm oil droplets in the samples using optical microscopy) can 0
be attributed to the stronger light scattering efficiency of the large swol- 10 100
len starch granules compared to the small oil droplets (McClements, Particle Diameter (mm)
2005). Consequently, the signal from the starch granules dominated
the overall signal utilized by the light scattering instrument to calculate Fig. 2. A representative particle size distribution for diluted samples: “Starch only”=3.75%
the particle size distribution. The slight difference in the particle size dis- starch suspended in water; “starch+fat”=model sauce containing 3.75% starch and 15% fat.
tributions of the samples in the absence and presence of fat droplets
(Fig. 2) may have occurred for a number of reasons: (i) the fat droplets to the 5% fat samples had little influence on their visual appearance,
may have made some contribution to the overall light scattering signal, i.e., the LBG could not mimic the light scattering properties of the fat
thereby affecting its mathematical interpretation; (ii) the fat droplets droplets.
promoted some aggregation of the starch granules; (iii) the fat droplets
adsorbed to the surfaces of the starch granules, thereby increasing their 3.4. Shear rheometry measurements
overall dimensions.
The influence of fat content and locust bean gum addition on the
3.3. Optical properties and appearance rheological properties of the model sauces was initially examined
using conventional shear rheometry. The shear stress versus shear
All the model sauces were optically opaque and had a slight brownish rate profile of each sample was measured (Fig. 3), and the apparent
appearance compared to the individual heated LBG and starch systems, shear viscosity at 10 and 50 s −1 were determined from these values
which were both whitish in color. The optical properties of the sauces (Table 3). The yield stress (τ0), consistency index (K), and flow
measured using a colorimeter indicated that all the model sauce systems index (n) of the model sauces were obtained by finding the best-fit
had high lightness, with L* values ranging from 73 to 81 (Table 2). The between the Herschel–Bulkley (HB) model and the experimentally
lightness of the model sauces increased with fat content, which can be measured shear stress versus shear rate profiles (Table 3). All sys-
attributed to the greater amount of light scattering as the concentration tems, including the starch-only and locust bean gum-only systems,
of oil droplets increased (Chantrapornchai, Clydesdale, & McClements, had similar general flow behavior patterns (Fig. 3). No flow was ob-
1999a, 1999b). A similar trend was also observed in the sensory evalua- served below a critical applied shear stress (i.e. the yield stress),
tion of product appearance, where the 15% fat system had the highest vi- and the flow rate increased non-linearly as the applied stress was in-
sual whiteness score, followed by the 10% fat system, while both the 5% creased above this critical value, which is indicative of non-ideal plas-
fat systems had similar scores (Table 2). The 5% fat model sauces in the tic materials that exhibit shear-thinning (Larson, 1999; Quemada &
absence and presence of LBG had fairly similar appearances, which indi- Berli, 2002).
cated that this gum did not scatter light strongly. A correlation coefficient In the absence of LBG, the yield stress and consistency index in-
(r2) of 0.97 was obtained between the sensory and instrumental light- creased with increasing fat content in the model sauces, while all of
ness tests. Overall, these measurements show that the addition of LBG them showed shear thinning behavior above the yield stress (n b 1).
The rheology of the samples was also highly dependent on LBG addi-
tion, with the yield stress of the model sauce (5% fat) containing LBG
Table 2 being appreciably higher than the one without it (Table 3). The model
Physical properties and sensory attributes (visual whiteness) of model sauces containing
different fat and LBG contents. The sensory evaluation scores ranged from 1 to 5, with 5
sauce containing 5% fat and 0.2% LBG had a yield stress that was
being the highest “visual whiteness”. A correlation coefficient (r2) of 0.97 was found be- between that of the 10% and 15% fat samples containing no LBG,
tween the instrumental lightness (L*) and the sensory visual whiteness. highlighting the potential of this gum for creating reduced fat prod-
ucts with textural attributes similar to those of higher fat products.
Systems pH d43 Lightness, Visual Whiteness
(μm) L* Rating Interestingly, the model sauce containing 5% fat and 0.2% LBG had
the lowest consistency index and highest flow index, suggesting
5% fat–0.2% LBG 4.89 ± 0.02c 39.70 ± 0.51a 73.88 ± 0.05b 1.68 ± 0.69a
5% fat 4.91 ± 0.01c 40.62 ± 0.25b/c 73.23 ± 0.05a 1.84 ± 0.90a that the gum altered the flow behavior of the sauces once the yield
10% fat 4.42 ± 0.01b 40.37 ± 0.09b 79.36 ± 0.05c 3.28 ± 0.69b stress was exceeded.
15% fat 4.20 ± 0.02a 41.04 ± 0.01c 81.36 ± 0.03d 4.20 ± 0.82c The increased apparent shear viscosity and yield stress of the model
Data within the same column sharing the same alphabetical superscript are not statistically sauces with increasing fat droplet and polymer concentrations can be
different (pb 0.05). attributed to the increase in effective particle concentrations that
314 C. Chung et al. / Food Research International 51 (2013) 310–320

The effective volume fraction of model sauces can be calculated as

the sum of the volume fractions of the different types of disperse
phase present:
10.0
ϕeff ¼ ϕE;S þ ϕE;F þ ϕE;P : ð3Þ

Here, ϕE,S, ϕE,F and ϕE,P are the effective volume fractions of the
Shear Stress (Pa)

starch, fat, and polymer, respectively. It is important to use the appro-

1.0 priate values of the effective volume fractions of the different disperse
phases in the above equation, since these may be appreciably differ-
ent from the initial mass fractions of starch, fat, and polymer added.
The effective volume fraction of the starch is that occupied by the
swollen starch granules after heating, whereas the effective volume
3.75% Starch
fraction of the polymers depends on their conformation due to the
0.1 0.2% LBG water trapped within their structure. The above equations show
5% Fat - 0.2% LBG that the viscosity of a system may increase steeply when more fat
5% Fat or polymer is added, particularly when the overall effective volume
10% Fat fraction is close to the ϕc value (around 60%).
15% Fat Thickening agents, such as LBG, are often incorporated into food
products as a means of fat reduction. We therefore compared the rhe-
0.0
0 20 40 60 80 100 ological properties of the model sauces containing 5% fat and 0.2%
LBG, with the sauces containing different fat levels but no LBG. The
Shear Rate (s-1)
apparent viscosity of the 5% fat sauce containing LBG was between
Fig. 3. Flow profiles (shear stress versus shear rate) of thermally processed (90 °C for that of the 10% and 15% fat sauces at 10 s −1 but was similar to the
5 min) systems: 3.75% starch; 0.2% LBG solution; and the four model sauce systems 15% fat sauce when measured at 50 s −1 shear rate (Table 3). This sug-
with different fat contents. All the model sauce systems contained 3.75% starch. One gests that the LBG altered the shear thinning behavior of the model
of the 5% fat systems contained 0.2% (LBG) in addition to 3.75% starch [5% fat–0.2%
sauces, leading to a less pronounced decrease in viscosity with in-
LBG]. All measurements were taken at 60 °C.
creasing shear rate. This may have been because the change in struc-
ture or interactions of the LBG molecules was less effected by
shearing than that of the starch granule/fat droplet mixture. Overall,
occurred in the systems (Servais, Jones, & Roberts, 2002; Velez-Ruiz, our results show that the addition of a low amount of gum (0.2%)
Gonzalez-Tomas, & Costell, 2005). To a first approximation the shear was effective at producing reduced-fat sauces with viscosities similar
viscosity of a mixed particle dispersion (such as starch granules, fat to those found in high fat products. This result is in agreement with
droplets and polymer dispersed in water) can be related to the overall other studies that have reported an increase in apparent viscosity
particle concentration using the following semi-empirical equation and shear modulus when gums are added to starch suspensions
(Brady, 1993): (BeMiller, 2011; Kim & Yoo, 2010; Sikora, Badrie, Deisingh, &
Kowalski, 2008). A number of physicochemical phenomenon were
  proposed to account for the observed increase in the viscosity of
ϕeff −2
η ¼ η0 1− : ð2Þ starch suspensions in the presence of gums: aggregation of swollen
ϕc
granules due to depletion flocculation; interaction of gum molecules
with leached starch molecules; and, thermodynamic incompatibility
Here, η0 (Pa.s) is the viscosity of the liquid surrounding the parti- (BeMiller, 2011; Kim & Yoo, 2010). However, the simple increase in
cles, ϕeff is the effective volume fraction of the particles, and ϕc is a effective particle volume fraction described by Eq. (2) may also be
critical packing parameter that is related to the volume fraction at at least partly responsible for this effect.
which the particles become close packed. For monodisperse systems We compared the apparent shear viscosities determined using
containing similar sized hard spheres ϕc is around 0.64, but for poly- this instrumental method to the “mouthfeel viscosity/thickness” at-
disperse systems containing a mixture of hard spheres with different tribute determined using sensory tests on similar samples. The senso-
sizes ϕc is higher than this value because small particles can fit into ry analysis of the model sauces was in good agreement with the
the spaces between large particles (Servais et al., 2002). Eq. (2) pre- apparent viscosity measurements, with the perceived “mouthfeel
dicts that the viscosity should increase slightly with increasing effec- viscosity/thickness” decreasing in the following order: 15% fat > 5%
tive volume fraction initially, but increase steeply once the effective fat + 0.2% LBG > 10% fat > 5% fat (Table 3). A high correlation
volume fraction approaches ϕc. coefficient (r 2 = 0.98) was calculated between the shear viscosity

Table 3
Rheological parameters derived from fitting the Herschel–Bulkley model to the shear stress versus shear rate data for model sauces (3.75% starch) containing different fat and LBG
levels. All systems fitted to the model had 0.99 correlation coefficients. The sensory rating is reported for perceived “Mouthfeel Viscosity/Thickness”.

Systems Yield stress (τ0) Consistency (K) Flow index (n) Apparent viscosity Apparent viscosity Sensory rating
(Pa) (Pa.s) (Pa.s) (Pa.s)
10 s−1 50 s−1

5% fat–0.2% LBG 1.82 ± 0.04b 1.34 ± 0.02a 0.65 ± 0.01c 0.79 ± 0.01c 0.38 ± 0.01c 3.4 ± 1.12c
5% fat 0.54 ± 0.20a 1.45 ± 0.07a 0.51 ± 0.01a 0.52 ± 0.01a 0.22 ± 0.01a 1.64 ± 0.90a
10% fat 1.81 ± 0.07b 1.68 ± 0.09b 0.52 ± 0.01a/b 0.72 ± 0.09b 0.30 ± 0.01b 2.72 ± 0.89b
15% fat 1.95 ± 0.37c 1.97 ± 0.31c 0.56 ± 0.05b 0.91 ± 0.00d 0.39 ± 0.01d 4.32 ± 0.74d

Data within the same column sharing the same alphabetical superscript are not statistically different (p b 0.05).
 C. Chung et al. / Food Research International 51 (2013) 310–320 315

(at 10 s −1) and the sensory score. Other studies that have examined to structure breakdown (Chung et al., 2012a). After shearing, the
the effects of thickener addition have also reported relatively high upper plate was moved upwards (decompression stage) and an in-
correlations between rheological properties and sensory perception creasingly negative normal force was measured until a maximum
(Kora, Latrille, Souchon, & Martin, 2002; Wendin, Solheim, Allmere, trough force was reached at the largest gap (3 mm), which can be re-
& Johansson, 1997). lated to the “adhesiveness” of the material. When the material was
held at this gap the negative normal force dissipated until a constant
3.5. Simulated mastication measurements residual trough force was reached.
In squeezing flow rheology it is important to establish the influ-
The rheological behavior of the model sauce systems was also ence of the contact between the sample and the plates on the mea-
characterized using the simulated mastication method. The rheologi- sured properties of the material, i.e., imperfect (lubricated) or
cal properties of model sauces containing similar starch levels (3.75%) perfect (frictional) conditions (Campanella & Peleg, 2002; Corradini,
but different fat and LBG levels were measured in the absence and Stern, Suwonsichon, & Peleg, 2000). This can be achieved by plotting
presence of saliva. Samples were subjected to a series of compres- log force vs. log height data obtained from a compression cycle. The
sion–shearing–decompression cycles. Squeezing flow behavior is linear part of the logarithmic plot is considered to represent the re-
presented as normal force versus mastication time profiles, while gion where squeezing flow is dominant, whereas the non-linear
shearing behavior is presented as the apparent shear viscosity mea- part represents the transient region before squeezing flow has been
sured during the fixed gap stage. The interpretation of these profiles fully established. And the data in the transient region is not taken
has been described in more detail in our recent publication (Chung into account. The slope (n) (flow index) of the linear part of the log
et al., 2012a). force-log height plot is used to determine the flow regime: n b 1 is
strong evidence for lubricated squeezing flow, whereas n > 1 is evi-
3.5.1. Behavior within the simulated mastication instrument dence for frictional flow (Corradini & Peleg, 2000; Suwonsichon &
The normal force versus mastication time profile of the material Peleg, 1999). In this study, the calculated n for the four sauce systems
containing 15% fat is shown in Fig. 4. The various textural parameters ranged from 3.4 to 4.6, which suggested that the flow was frictional
that can be derived from this profile are also highlighted on the dia- squeezing flow. No calculations were carried out for the individual lo-
gram. For clarity only the data for the first two cycles is shown in cust bean gum or starch systems as their viscosities were so low that
the figure (rather than for the full ten cycles). At the start of the test it was not possible to make accurate measurements using this
the normal force on the upper plate was close to zero because it method.
was not in contact with the sample. As the upper plate moved down-
wards the normal force acting on the upper plate increased until it 3.5.2. Influence of saliva on simulated oral behavior
reached a maximum peak force at the smallest gap between the plates During mastication food is mixed with saliva to form a bolus
(1 mm), which is taken as a measure of the “consistency” of the ma- (Foster et al., 2011; van der Bilt, 2009). Saliva contains a variety of
terial. When the material was held at a fixed gap (1 mm), the normal constituents that may influence the structural properties and rheolo-
force rapidly dissipated until it reached a relatively constant value, gy of foods within the mouth, including salts, polymers, and digestive
which is related to the “yield stress” of the material. During the enzymes.
fixed gap stage a shearing motion was applied to the samples to sim- In this series of tests, the effect of mixing samples with artificial
ulate the sliding of the tongue against the palate, and the apparent saliva on the rheological properties of the model sauce containing
shear viscosity was recorded during this time. In a previous study, 15% fat was examined (Fig. 5). Measurements were carried out in
we showed that application of this shearing force caused a reduction the absence of saliva, and for samples diluted with artificial saliva
in the normal force generated by the material, which was attributed (α-amylase, mucin and salts) or mucin solution (mucin and salts).
The samples were subjected to 10 cycles, with a 10 s −1 shear rate ap-
plied for 1 s during the shearing stage. In the absence of saliva, there
0.4
was a gradual decrease in the magnitudes of the maximum peak
C FG/S D force, maximum trough force, and apparent shear viscosity with in-
0.3 creasing number of cycles, which was attributed to progressive
Maximum breakdown in the structure of the material during the compression,
Peak Force shearing, and decompression stages. From cycle 7 onwards, almost
0.2
constant values were measured, which suggests that the microstruc-
ture and physical properties of the samples were not further altered
Normal Force (N)

0.1 by compression and shearing. In the presence of saliva, all the textural
Residual
parameters decreased more rapidly with increasing number of cycles
than in the absence of saliva (Fig. 5). This effect can be attributed to
0.0
two effects: (i) dilution of the sample by saliva; (ii) digestion of starch
0 2 4 6 granules by α-amylase. The influence of the dilution effect in isolation
-0.1 is apparent in the sample containing mucin solution (without
α-amylase), as the magnitudes of the textural parameters were
lower than observed in the system containing no saliva but higher
-0.2
than in the one with saliva (Fig. 5).
Maximum
-0.3 3.5.3. Influence of fat content and locust bean gum addition
Trough Force
In this section, the effects of fat content and addition of locust
bean gum on the textural properties of the model sauces were studied
-0.4
in the absence and presence of artificial saliva. Only the maximum
Time (s) peak forces measured during the compression stage (Fig. 6) and the
Fig. 4. Normal force vs. time plot for compression-fixed gap-decompression-fixed gap
apparent shear viscosity measured during the fixed gap stage
tests carried out on 15% fat system. Key to symbols: C= compression; FG/S = fixed gap (Fig. 7) are reported. We also measured the properties of the starch
with shearing; D = decompression. suspension (no LBG) and LBG solution (no starch) in the absence of
316 C. Chung et al. / Food Research International 51 (2013) 310–320

a fat using the same test methods. The magnitudes of the rheological
parameters of these samples were relatively low for all test cycles
0.25
No Saliva
(full data not shown): maximum peak forces b 0.05 N; apparent
shear viscosities b 0.15 Pa.s.
With Saliva
In the absence of saliva, the maximum peak force of the model
Maximum Peak Force (N)

With Mucin sauces decreased in the following order: 5% fat–0.2% LBG > 15%
0.20
fat > 10% fat > 5% fat (Fig. 6a). A slightly different trend was observed
for the apparent shear viscosity measurements, with the viscosity of
the model sauces decreasing in the following order: 15% fat > 5%
0.15
fat–0.2% LBG > 10% fat > 5% fat (Fig. 7a). The slight difference in the
trends observed for these two measurement methods may be due
to their different physical basis. The maximum peak height is mea-
0.10 sured during compression at a fixed speed (17.8 mm s −1), whereas
the shear viscosity is measured during shearing at a fixed speed
(10 s −1). Samples that undergo changes in their structural organiza-
0.05 tion after application of mechanical forces (like the ones used in this
study) will have different rheological properties when different types
and rates of applied stress are used to measure them. Overall, our re-
0.00
sults indicate that the addition of 0.2% LBG to the 5% fat systems in-
0 1 2 3 4 5 6 7 8 9 10 creased their measured textural attributes to values around those
Cycle Sequence measured for 10 to 15% fat systems containing no LBG, which high-
b -0.08 lights the potential usefulness of this gum for developing reduced
fat sauces. The various physicochemical mechanisms that may ac-
count for the observed increase in the textural attributes of the
No Saliva
model sauces upon addition of the LBG were discussed in Section 3.4.
With Saliva
There were appreciable changes in the textural profiles of all the
Maximum Trough Force (N)

-0.13
With Mucin model sauces in the presence of saliva. In particular, the magnitudes
of the maximum peak forces and the shear viscosities of the model
sauces were appreciably lower in the presence of saliva than in its ab-
-0.18 sence. In general, the maximum peak force (Fig. 6b) and apparent
shear viscosity (Fig. 7b) decreased appreciably from the first to sixth
cycle and then reached relatively low constant values. The relatively
low textural parameters measured from the sixth cycle onwards can
-0.23
be attributed to breakdown of the starch granules, since these are
known to be the main structural entities contributing to the rheological
properties of this type of system (Chung et al., 2012b; Chung, Degner &
-0.28 McClements, 2013). Comparing data obtained from the first to the sixth
cycle, the maximum peak force of the sauces decreased in the following
order: 15% fat> 5% fat–0.2% LBG> 10% fat> 5% fat (Fig. 6b), while the
apparent shear viscosity decreased in the following order: 15% fat> 5%
-0.33 fat–0.2% LBG ≈10% fat ≈ % fat (Fig. 7b). The trends observed in the
0 1 2 3 4 5 6 7 8 9 10
presence of saliva were therefore fairly similar to those measured in
Cycle Sequence
the absence of saliva. Nevertheless, the differences between the sam-
c 1.2
ples became less apparent in the presence of saliva, which can be
attributed to sample dilution and enzymatic degradation of the starch
granules. These results highlight the importance of including artificial
No Saliva
saliva within the rheological measurements in order to better mimic
Apparent Shear Viscosity (Pa.s)

1.0 With Saliva the complex processes occurring within the mouth. They also show
With Mucin that the addition of LBG can be used to mimic some of the desirable tex-
tural attributes of the fat droplets, i.e., high viscosity.
0.8
3.6. Sensory evaluations
0.6
The overall sensory attributes of model sauces containing similar
starch levels (3.75%) but different levels of fat (5 to 15%) and LBG (0 or
0.4 0.2%) were measured and presented graphically (Fig. 8). In the absence
of LBG, the overall liking score improved with increasing fat content,
which is to be expected since higher amounts of fat impart desirable
0.2 creaminess and smoothness attributes. The 5% fat and 0.2% LBG system
had the second highest overall liking rating, falling between the 10%

0.0
Fig. 5. Influence of artificial saliva and mucin solution on the change in a) maximum peak
0 1 2 3 4 5 6 7 8 9 10
forces and b) maximum trough forces during compression and decompression, respectively,
Cycle Sequence and c) shear viscosity for successive compression-fixed gap shearing–decompression cycles
on 15% fat sauce. A shear rate of 10 s−1 was applied to the samples during the fixed gap
stage.
 C. Chung et al. / Food Research International 51 (2013) 310–320 317

and 15% fat systems, indicating that LBG enhanced the desirable sensory rate to the ones without LBG, which suggested that this gum was not
perception of the low fat system. This effect can mainly be attributed to particularly effective at inhibiting starch degradation.
the ability of LBG to act as a thickening agent that increases the consis- The maximum peak force and shear viscosity measured using the
tency of the model sauces (Section 3.4). These results are in agreement simulated mastication method were correlated to related sensory pa-
with earlier studies which also showed improved sensory perception rameters. For the instrumental method, we selected the textural param-
of reduced-fat products with addition of water-soluble gums due to vis- eters measured during the first cycle as representative of the in-mouth
cosity enhancement (Akhtar, Stenzel, Murray, & Dickinson, 2005; Kora et behavior of the model sauces since this should be related to the initial
al., 2002). Another mechanism that has been proposed to account for the perception of a product when first placed in the mouth. The sensory
ability of gums to improve the sensory attributes of some food products analysis of the samples showed that there was a strong correlation be-
is their ability to retard the degradation of starch granules by amylase tween the “smoothness/creaminess” and “viscosity/thickness” rankings
within the mouth (Brennan & Tudorica, 2008). However, we found of the model sauces (Fig. 8), with a correlation coefficient, r2 = 0.993
that the maximum peak force (Fig. 6b) and apparent shear viscosity for the 4 samples tested. We therefore only compared the instrumental
(Fig. 7b) of the model sauces containing LBG decreased at a fairly similar measurements with the sensory “viscosity/thickness” rankings. For both

a a
0.30
No Saliva 5% Fat - 0.2% LBG
No Saliva 5% Fat - 0.2% LBG
5% Fat 1.2
5% Fat
0.25 10% Fat 10% Fat

Apparent Shear Viscosity (Pa.s)

15% Fat 1.0 15% Fat
Maximum Peak Force (N)

0.20
0.8

0.15
0.6

0.10
0.4

0.05 0.2

0.00 0.0
0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10
Cycle Sequence Cycle Sequence
b b
0.30
With Saliva 5% Fat - 0.2% LBG With Saliva 5% Fat - 0.2% LBG
1.2
5% Fat 5% Fat
0.25 10% Fat
10% Fat
Apparent Shear Viscosity (Pa.s)

1.0 15% Fat

15% Fat
Maximum Peak Force (N)

0.20
0.8

0.15 0.6

0.10 0.4

0.2
0.05

0.0
0.00 0 1 2 3 4 5 6 7 8 9 10
0 1 2 3 4 5 6 7 8 9 10 Cycle Sequence
Cycle Sequence
Fig. 7. Influence of artificial saliva on the change in apparent shear viscosity measured
Fig. 6. Influence of artificial saliva on the change in maximum peak forces during compres- (at 10 s−1 shear rate) during compression and decompression of model sauces with
sion and decompression for model sauces with different fat and LBG contents: (a) no saliva; different fat and LBG contents: (a) no saliva; (b) artificial saliva (amylase, mucin, and
(b) with artificial saliva (amylase, mucin, and salt). salt).
318 C. Chung et al. / Food Research International 51 (2013) 310–320

Sensory Attributes Scoring

5% Fat - 0.2% LBG
Visual 5% Fat
Whiteness
5 10% Fat
15% Fat
4

3 Appearance
Overall Liking
Liking
2

Texture
Texture Mouth
Liking Viscosity
/Thickness

Texture
Smoothness
/Creaminess

Fig. 8. Descriptive sensory evaluation of model sauce systems with varying fat content (5, 10 and 15%) and addition of locust bean gum (0.2%).

the maximum peak force (r2 >0.95) and the apparent shear viscosity that the initial textural attributes of the material before it enters the
(r2 >0.76) measured using the simulated mastication method there mouth may be the most important in determining certain textural attri-
was a reasonably good correlation between the instrumental attributes butes of this kind of sauce product.
and the sensory rankings (Fig. 9). Interestingly, we found a higher corre-
lation (r2 >0.98) between the same sensory attributes of the model 3.7. Microstructure of masticated samples
sauces and the apparent shear viscosity measured using the dynamic
shear rheometer without saliva addition (Section 3.4). This suggests The microstructure of a selected model sauce (15% fat, 3.75% starch)
was examined before and after mastication in a human mouth (in vivo)
and by the simulated mastication instrument (in vitro). Initially, the
model sauce contained a mixture of swollen starch granules and fat
Peak Force (N) droplets that were packed closely together prior to mastication
0.8 (Fig. 10). After mastication in the mouth or by the simulated mastica-
Viscosity (Pa.s)
R² = 0.7685 tion method the microstructure of the sauces was similar: the fat drop-
lets could be seen, but the starch granules were no longer visible
Instrumental Parameter

(presumably due to amylase digestion). These results strongly suggest

0.6 that the simulated mastication method was able to mimic the break-
down of model sauces within the human mouth.

4. Conclusions
0.4
This study has demonstrated the potential of a recently developed
simulated mastication method for providing valuable information about
the textural properties of model sauces during mastication. The method
R² = 0.9581 is designed to simulate certain mechanical aspects of the tongue and
0.2 palate during mastication, such as their upward/downward and shearing
motions. The method enables a number of parameters that are related to
15% Fat
5% Fat + the textural characteristics of a sample to be determined during the sim-
10% Fat
5% Fat 0.2% LBG ulated mastication process, including maximum peak force (which is re-
0.0 lated to consistency), residual force (which is related to yield stress), and
1 2 3 4 5 apparent shear viscosity.
The new simulated mastication method was used to study the in-
Sensory Ranking
fluence of fat content and locust bean gum addition on the textural
Fig. 9. Comparison of instrumental parameters measured using the simulated mastica- properties of model sauces thickened by swollen starch granules. As
tion device with a sensory attribute (“viscosity/thickness”). expected, the consistency of the sauces increased as the fat content
 C. Chung et al. / Food Research International 51 (2013) 310–320 319

Model Sauce (As Is) Oil Droplet

Starch

Oil Droplet

50 m
Starch

Oil Droplet
Mouth Masticated Sauce

Oil Droplet

50 m

Instrument Masticated Sauce Oil Droplet

Oil Droplet

50 m

Fig. 10. Micrographs (400× magnification) of model sauce (15% fat, 3.75% starch) without saliva (top) and sauce samples masticated in the mouth (for 10 s) (middle) and by the
instrument (bottom) with compression-fixed gap shearing-decompression test sequence in the presence of artificial saliva.

increased from 5 to 15%. The addition of 0.2% LBG to the 5% fat sauce References
gave it a consistency that was similar to higher fat products (10 to
Ades, H., Kesselman, E., Ungar, Y., & Shimoni, E. (2012). Complexation with starch for
15% fat). These studies showed that LBG addition was a useful strategy encapsulation and controlled release of menthone and menthol. LWT- Food Science
for creating low fat products with similar textural attributes as high fat and Technology, 45, 277–288.
products. However, measurements of the optical properties of the sam- Akhtar, M., Stenzel, J., Murray, B. S., & Dickinson, E. (2005). Factors affecting the perception
of creaminess of oil-in-water emulsions. Food Hydrocolloids, 19, 521–526.
ples showed that LBG addition could not be used to mimic the desirable BeMiller, J. N. (2011). Pasting, paste, and gel properties of starch-hydrocolloid combinations.
“lightness” associated with the scattering of light waves by fat droplets. Carbohydrate Polymers, 86, 386–423.
Finally, we have shown that the textural parameters determined Bourne, M. C. (2002). Food texture and viscosity: Concept and measurement (2nd ed.).
London, UK: Academic Press.
from the instrumental simulated mastication method correlated Brady, J. F. (1993). The rheological behavior of concentrated colloidal dispersions. Journal
well with textural parameters obtained from sensory analysis. This of Chemical Physics, 99(1), 567–581.
new instrumental method may therefore be a valuable tool for inves- Brennan, C. S., & Tudorica, C. M. (2008). Evaluation of potential mechanisms by which die-
tary fibre additions reduce the predicted glycaemic index of fresh pastas. International
tigating the factors that influence the textural properties of semi-solid
Journal of Food Science and Technology, 43, 2151–2162.
food materials within the mouth during mastication. Buscall, R. (2010). Wall slip in dispersion rheometry. Journal of Rheology, 54(6), 1177–1183.
Campanella, O. H., & Peleg, M. (2002). Squeezing flow viscometry for nonelastic semiliquid
foods — Theory and applications. Critical Reviews in Food Science and Nutrition, 42(3),
241–264.
Acknowledgments Chantrapornchai, W., Clydesdale, F., & McClements, D. J. (1999a). Influence of droplet charac-
teristics on the optical properties of colored oil-in-water emulsions. Colloids and Surfaces
We thank ConAgra foods for financially supporting this work, and A: Physicochemical and Engineering Aspects, 155(2–3), 373–382.
Chantrapornchai, W., Clydesdale, F., & McClements, D. J. (1999b). Theoretical and experi-
Bicheng Wu, Gordon Smith and Ware Flora for valuable scientific and mental study of spectral reflectance and color of concentrated oil-in-water emulsions.
technical discussions about the work. Journal of Colloid and Interface Science, 218(1), 324–330.
320 C. Chung et al. / Food Research International 51 (2013) 310–320

Chung, C., Degner, B., & McClements, D. J. (2012a). Instrumental mastication assay for Kora, E. P., Latrille, E., Souchon, I., & Martin, N. (2002). Texture-flavor interactions in
texture assessment of semi-solid foods: Combined cyclic squeezing flow and shear low fat stirred yogurt: How mechanical treatment, thickener concentration and
viscometry. Food Research International, 49, 161–169. aroma concentration affect perceived texture and flavor. Journal of Sensory Studies,
Chung, C., Degner, B., & McClements, D. J. (2012b). Rheology and microstructure of bimodal 18, 367–390.
particulate dispersions: Model for foods containing fat droplets and starch granules. Larson, R. G. (1999). The structure and rheology of complex fluids. Oxford, UK: Oxford
Food Research International, 48, 641–649. University Press.
Chung, C., Degner, B., & McClements, D. J. (2013). Physicochemical characteristics of mixed Le Reverend, B. J. D., Norton, I. T., Cox, P. W., & Spyropoulos, F. (2010). Colloidal aspects
colloidal dispersions: Models for foods containing fat and starch. Food Hydrocolloids, 30, of eating. Current Opinion in Colloid and Interface Science, 15, 84–89.
281–291. Lillford, P. (2000). The materials science of eating and food breakdown. MRS Bulletin,
Corradini, M. G., & Peleg, M. (2000). Lubricated squeezing flow viscometry for dulce de 25, 38–43.
leche (milk sweet). Food Science and Technology International, 6(4), 339–344. Mandel, A. L., Peyrot des Gachons, C., Plank, K. L., Alarcon, S., & Breslin, P. A. S. (2010).
Corradini, M. G., Stern, V., Suwonsichon, T., & Peleg, M. (2000). Squeezing flow of semi Individual differences in AMY1 gene copy number, salivary α-amylase levels, and
liquid foods between parallel Teflon coated plates. Rheologica Acta, 39, 452–460. the perception of oral starch. PLoS One, 5(10), e13352.
Cutler, A. N., Morris, E. R., & Taylor, L. J. (1983). Oral perception of viscosity in fluid McClements, D. J. (2005). Food emulsions: Principles, practices, and techniques (2nd ed.).
foods and model systems. Journal of Texture Studies, 14, 377–395. Boca Raton, Fl: CRC Press.
de Bruijne, D. W., Hendrickx, H. A. C. M., Anderliesten, L., & de Looff, J. (1993). Mouthfeel Quemada, D., & Berli, C. (2002). Energy of interaction in colloids and its implications in
of foods. In E. Dickinson, & P. Walstra (Eds.), Food colloids and polymers: Stability and rheological modelling. Advances in Colloid and Interface Science, 98(1), 51–85.
mechanical properties (pp. 204–213). Cambridge, UK: Royal Society of Chemistry. Servais, C., Jones, R., & Roberts, I. (2002). The influence of particle size distribution on
de Wijk, R. A., Prinz, J. F., & Janssen, A. M. (2006). Explaining perceived oral texture of the processing of food. Journal of Food Engineering, 51(3), 201–208.
starch-based custard desserts from standard and novel instrumental tests. Food Shellis, R. P. (1978). A synthetic saliva for cultural studies of dental plaque. Archives of
Hydrocolloids, 20, 24–34. Oral Biology, 23(6), 485–489.
Foster, K. D., Grigor, J. M. V., Cheong, J. N., Yoo, M. J. Y., Bronlund, J. E., & Morgenstern, M. Sikora, M., Badrie, N., Deisingh, A. K., & Kowalski, S. (2008). Sauces and dressings: A review
P. (2011). The role of oral processing in dynamic sensory perception. Journal of of properties and applications. Critical Reviews in Food Science and Nutrition, 48, 50–77.
Food Science, 76(2), R49–R61. Suwonsichon, T., & Peleg, M. (1999). Rheological characterization of almost intact and stirred
Gal, J. Y., Fovet, Y., & Adib-Yadzi, M. (2001). About a synthetic saliva for in vitro studies. yoghurt by imperfect flow viscometry. Journal of the Science of Food and Agriculture, 79,
Talanta, 53(6), 1103–1115. 911–921.
Herschel, W. H., & Bulkley, R. (1926). Konsistenzmessungen von gummibenzollosunge. van Aken, G. A. (2007). Relating food microstructure to sensory quality. In D. J. McClements
Kolloid Zeitschrift, 34, 291–300. (Ed.), Understanding and controlling the microstructure of complex foods (pp. 449–482).
Hiiemae, K., & Palmer, J. B. (1999). Food transport and bolus formation during complete Cambridge, U. K.: Woodhead Publishing Limited.
feeding sequences on foods of different initial consistency. Dysphagia, 14, 31–42. van der Bilt, A. (2009). Oral physiology, mastication and food perception. In D. J.
Ishihara, S., Nakauma, M., Funami, T., Odake, S., & Nishinari, K. (2011a). Swallowing pro- McClements, & E. A. Decker (Eds.), Designing functional foods: Measuring and con-
files of food polysaccharide gels in relation to bolus rheology. Food Hydrocolloids, 25, trolling food structure breakdown and nutrient absorption (pp. 3–37). Cambridge:
1016–1023. Woodhead Publishing Limited.
Ishihara, S., Nakauma, M., Funami, T., Odake, S., & Nishinari, K. (2011b). Viscoelastic and Velez-Ruiz, J. F., Gonzalez-Tomas, L., & Costell, E. (2005). Rheology of dairy custard model
fragmentation characters of model bolus from polysaccharide gels after instrumental systems: Influence of milk-fat and hydrocolloid type. European Food Research and
mastication. Food Hydrocolloids, 25, 1210–1218. Technology, 221, 342–347.
Kim, D. D., & Yoo, B. (2010). Rheological behaviors of hydroxypropylated sweet potato Wendin, K., Solheim, R., Allmere, T., & Johansson, L. (1997). Flavor and texture in sourmilk
starches influenced by guar gum, locust bean and xanthan gums. Starch-Starke, 62, affected by thickeners and fat content. Food Quality and Preference, 8(4), 281–291.
584–591.