Journal of Food Engineering 109 (2012) 736744

Contents lists available at SciVerse ScienceDirect

Journal of Food Engineering

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

Particle breakdown dynamics of heterogeneous foods during mastication: Peanuts embedded inside different food matrices
Scott C. Hutchings a,, Kylie D. Foster a, John E. Bronlund b,d, Roger G. Lentle a,d, Jim R. Jones b, Marco P. Morgenstern c
a

Institute of Food, Nutrition and Human Health, Massey University, New Zealand School of Engineering and Advanced Technology, Massey University, New Zealand The New Zealand Institute for Plant & Food Research Ltd., New Zealand d The Riddet Institute, Massey University, New Zealand
b c

a r t i c l e

i n f o

a b s t r a c t
Heterogeneous foods are widely consumed but little is known about the dynamics of their breakdown during mastication. To investigate these dynamics the particle size and bolus mass at regular stages of the chewing sequence were investigated using a single subject, who was carefully selected according to strict mastication and dental criteria. A standardized volume of both a gelatine gel and chocolate matrix containing roasted peanuts was chewed and expectorated after a specied number of chews or at the point of swallowing. The mass of various components of the bolus was determined along with peanut particle parameters (d50, broadness (b), and specic surface area) at each interval. Results showed that for this subject the type of matrix inuenced the pathway that the peanut particles (embedded inside each matrix) were broken down. The d50 of peanut particles within the bolus of the gelatine gel were greater than within the bolus of the chocolate throughout the chewing sequence until the point of swallowing. At the point of swallowing, the d50 of peanut particles inside the bolus of the gelatine gel and chocolate were the same. The broadness value (b) and specic surface area of the peanut particles in the bolus of the chocolate were greater than those in the bolus of the gelatine gel throughout the chewing sequence. The matrices also induced different dynamic trends in the mass of bolus components during the chewing sequence. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 4 August 2011 Received in revised form 11 November 2011 Accepted 12 November 2011 Available online 25 November 2011 Keywords: Mastication Size reduction Particle size distribution Food bolus Chewing dynamics

1. Introduction Chewing is a dynamic process during which the physical parameters of the food bolus are continually modied until it is suitable for swallowing (Lucas, 2004). Dynamic changes depend on food characteristics, as the type of food product affects both mastication (Kohyama and Mioche, 2004) and the size of particles in the bolus (Peyron et al., 2004). Changes in food properties will consequently causes changes in jaw movement (Mioche et al., 2002; Peyron et al., 2002; Kohyama and Mioche, 2004), and associated muscle activity (Gonzalez et al., 2001; Kohyama et al., 2002; Peyron et al., 2002). Studies that have investigated dynamic changes during mastication have explored a wide range of foods, including meat, breakfast cereals, vegetables, fruit, gels, and articial test foods (Lucas and Luke, 1983; Olthoff et al., 1984; Mioche et al., 2002; Peyron et al., 2002; Kohyama and Mioche, 2004; Lenfant et al., 2009). These studies have shown that during the chew Corresponding author. Tel.: +64 9 414 0800x41260; fax: +64 9 443 9640.
E-mail address: s.hutchings@massey.ac.nz (S.C. Hutchings). 0260-8774/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2011.11.011

ing sequence the size of particles is reduced (Lucas and Luke, 1983; Olthoff et al., 1984), the mass of the bolus decreases (Peyron et al., 2004), stickiness, springiness, cohesion, and moisture content of the bolus increase, while hardness decreases (Hutchings and Lillford, 1988; Lenfant et al., 2009; Peyron et al., 2009). The food industry has a particular interest in chewing dynamics because mastication and the particle size distribution of the food bolus has been related to the glycaemic response (Read et al., 1986; Suzuki et al., 2005; Ranawana et al., 2010), texture perception (Brown et al., 1994, 2000), and the extent of avor release (Taylor, 1996; Alfonso et al., 2002). An enhanced understanding of this process may aid manufacturers in food design, with the objective of improving the sensory and/or nutritional benets of foods. Interestingly, previous research into changes in the physical properties of the bolus during chewing has been conducted largely on homogenous foods (Lucas and Luke, 1983; Lenfant et al., 2009). However, the majority of commercially available foods are heterogeneous, and a number of different foods are frequently masticated simultaneously during the consumption of a meal. The physical

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

737

changes in the food bolus of these heterogeneous foods during chewing have not been well described and the dynamics of their mastication are not well understood (Hutchings et al., 2011). Hutchings et al. (2011) asked eight subjects to chew both gelatine gel and chocolate matrices in which peanut quarters were embedded. Results showed that the size distribution of the masticated peanut particles at the point of swallowing, measured as the d50, did not differ between the two matrices. This was despite signicant differences in the initial textural properties of the matrices, in the manner in which they were chewed, and in the broadness (spread) of their peanut particle size distributions. It was therefore hypothesized that the rate of breakdown of particles embedded inside another medium are inuenced by the properties of that medium since the nal d50 was the same. Hence, by using different food mediums, it may be possible to control the dynamics of particle breakdown (from the initial chews until the product is ready for swallowing). Therefore, an investigation into the dynamic changes that occur during the production of the bolus of foods used by Hutchings et al. (2011) was required. The aim of this particular study was to compare changes in the size distribution of peanut particles and bolus mass of two heterogeneous food systems (peanut pieces embedded inside a gelatine gel matrix and a chocolate matrix) during the chewing sequence until the point of swallowing.

ing the bar before and after the bite, and the number of chews was counted by the researcher. The degree to which the data of each applicant approached those of a normal healthy population was assessed by comparing bite and chewing data of the Fruit and Nut bar with that from the study of Hutchings et al. (2009), in which 45 subjects took part. The consistency of the data for each applicant was determined by the magnitude of the standard deviation. Hence applicants with a low standard deviation value were considered more consistent. The general health and dental history of each volunteer was assessed by the administration of a screening questionnaire, and through inspection of the teeth, gums, mouth, and jaw by a qualied dentist. All applicants gave informed consent, and the study was approved by the Massey University Ethics Committee (Southern A Application 08/17). The selected subject who had mean data for each parameter that were the closest to those of a prior healthy normal population and were most consistent was a 29 year-old male (Table 1). The subject had class 1 occlusion, no signicant tooth crowding, no obvious tooth decay, and healthy periodontal condition. He did not suffer from any pain or clicking during mastication, nor did he have any other oral or health issues that could inuence oral processing. 2.2. Experimental procedure The subject was asked to refrain from eating or drinking for 2 h prior to each test session. The experimental procedure involved serving two test foods, previously used in Hutchings et al. (2011): a gelatine gel matrix (250 bloom) and a chocolate matrix, both containing embedded peanut quarters. The peanuts made up 11.3% of the test food by volume, where the remaining 88.7% of the volume was the matrix. Each test food was placed between the subjects molars on his preferred chewing side (his right side) before mastication began. The subject was told to begin masticating at the same time the researcher started to time the chewing process. The subject was asked to expectorate the food bolus into a container after 5, 10, 15, 20, or 25 chews, or at the point at which he wanted to swallow. The subject did not know the moment at which he had to expectorate the bolus, but was told that he had to follow the researchers instructions when asked to stop. The order of test foods and the order of chewing duration were randomized. Six sessions were conducted, and each session contained 12 samples (six gelatine gel samples chewed to 5, 10, 15, 20, 25 chews or the swallowing point, and six chocolate samples chewed to 5, 10, 15, 20, 25 chews or the swallowing point). Following the expectoration of each bolus, the subject rinsed his mouth with 25 mL distilled water and expectorated the washings (debris + water) into a separate container. Both the expectorated bolus and the washings were weighed and frozen (18 C). Each experimental session was conducted in a room at 20 C, with test foods equilibrated at 20 C at least 1 h before serving. In the rst four sessions boluses for the analysis of peanut particle size and peanut retention (the percentage dry weight of peanuts remaining in the bolus that was inside the original matrix) were obtained. In the nal two sessions boluses were obtained to assess the total dry weight of the bolus (matrix and peanuts) and to determine moisture content of the expectorated bolus. Separate sessions were required to determine total bolus dry weight and expectorated bolus moisture content, as drying caused shrinkage of peanut particles which could not be used for particle size analysis. The wet weight of the total bolus, the wet weight of the expectorated bolus, and the wet weight of the debris collected by washing the mouth with water, were determined in every session. For this study the wet or dry weight of the total bolus can be dened as the sum of the expectorated bolus and the debris. All pea-

2. Materials and methods 2.1. Subject screening and selection Mastication studies involving multiple subjects report wide variability in the number of chews, chewing time, and mastication frequency between subjects (Lassauzay et al., 2000; Peyron et al., 2002), and also some variability in the food bolus between subjects (Mishellany et al., 2006; Jalabert-Malbos et al., 2007). Therefore, a considerable number of participants and replicates are required to obtain signicant results. This poses a problem in studies analyzing the bolus at various points during the chewing sequence, as large amounts of resource are needed, which constrains the number of variables that can be explored. For this particular study, the purpose of the work was to evaluate the effect of food structure on breakdown pathways during oral processing, rather than to explore trends in chewing behavior among a population. In such a scenario, a standardized chewing test can be used. A number of masticatory robots have been developed for this purpose (Salles et al., 2007; Xu et al., 2008; Woda et al., 2010), but have only shown some effectiveness in the simulation of the biological mouth (Mishellany-Dutour et al., 2011; Xu and Bronlund, 2010). Consequently, a single subject was used as a chewing device for this study as an alternative to a robot. The variability within subject replicates is much smaller than between different subjects in terms of mastication parameters (Lassauzay et al., 2000) and the particle size distribution in the food bolus (Mishellany et al., 2006). Single subject studies have been used previously in mastication literature (Howel and Brudevold, 1950; Yven et al., 2010), and are used in research investigating the glycaemic index (Parada and Aguilera, 2009). A screening process was specically designed to select a subject who did not possess unusual oral processing characteristics, and was consistent in their oral processing. Eight applicants were asked to bite, chew and swallow 5 bar-shaped samples of a standard Fruit and Nut bar that was used in a previous study (Hutchings et al., 2009). Bite weight, and the corresponding number of chews each applicant took to process what they had acquired, were recorded for each applicant. Bite weight was determined by weigh-

738

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

Table 1 The bite weight and number of chews of a standard Fruit and Nut bar of the selected subject and those of a previous experimental population (Hutchings et al., 2009) (mean SD). Bite weight (g) Previous population Selected subject 6.02 2.15 6.73 0.88 Number of chews 34.8 16.2 32.0 1.9 Number of subjects 45 1 Number of replicates per subject 1 5

nut particle size and peanut retention measurements were undertaken on the total bolus, where the expectorated bolus and the debris were combined. 2.3. Assessment of natural bite size and selection of serving size

piece, with three inserted into each side of the test piece. Peanuts were inserted by manually piercing them through the surface of the matrix. Fig. 1 illustrates the test foods and sample dimensions once peanuts were inserted into the matrix. 2.5. Analysis of the physical properties of the matrices

Serving size was determined by assessing the natural bite length of the subject (and hence natural bite volume). The subject was asked to take four natural bites from the gelatine gel and chocolate matrices prepared as bars (containing 11.3% peanut quarters (v/v)) with a standardized and uniform cross-section (20 mm height, 30 mm width, 100 mm length). After the subject took a bite, he masticated and swallowed the sample. The mean bite length of the four bites was determined by measuring the average change in length of the bar after each bite. The mean bite length of the four bites was 15 3 mm (SD). This value was used as a basis for standardizing the volume of the serving. Therefore, the serving volume of each test piece was 9000 mm3 (20 30 15 mm), and each test piece contained 11.3% peanut quarters (v/v). The average weight of peanuts embedded inside each matrix was 1.13 0.02 g (mean SD). The average serving weight of the chocolate matrix (excluding peanuts) was 10.62 0.16 g, and the average serving weight of the gelatine gel matrix (excluding peanuts) was 11.59 0.13 g (mean SD). Hutchings et al. (2009) introduced the concept of standardizing the volume of food chewed by using a common bite length, after observing regulatory in bite length between six different food bars in a study with 45 subjects. 2.4. Preparation of test foods The preparation of test foods, and the analysis of initial test food properties, followed the procedure used in Hutchings et al. (2011), where matrices were prepared by setting the chocolate and gelatine gel in separate molds at 4 C and 20 C, respectively (without peanuts). In this study six roasted, unsalted peanut quarters (moisture content = 1.99 g/100 g) were embedded into each matrix test

Texture prole analysis (TPA) was used to assess the physical properties of the matrices. This employed a Texture Analyzer TAXT2 (Stable Microsystems, Surrey, UK), and involved two successive uni-axial compressions of the matrices (without peanuts) using a at cylindrical probe (diameter: 50 mm) to 80% strain. The sample dimensions before compression were 13 20 15 mm, where the 20 15 mm side was facing the compression probe (automatic trigger force of 0.1 N, test and post test speed: 0.8 mm/s, 24 replicates). Dimensions were smaller than the samples served to the subject for the mastication sessions to avoid overloading the 50 kg load cell during TPA. The TPA parameters of hardness, cohesiveness, springiness, and chewiness were calculated according to standard methods (Bourne, 2002), and all tests were undertaken at 20 C. Hardness can be dened as the resistance to force, cohesiveness as the strength of internal bonds within a food, chewiness as the measurement of the work required to masticate a food, and springiness as the measurement of the elasticity of a food (Szczesniak, 1963; Pons and Fiszman, 1996; Bourne, 2002). 2.6. Analysis of the food bolus Boluses from the rst four sessions were analyzed according to the methods described in Hutchings et al. (2011) to measure the particle size distribution and peanut retention. To undertake this analysis, the expectorated bolus and debris were combined (following the measurement of wet weight). Determining the particle size distribution (in terms of projected 2D particle area) involved sieving the masticated bolus across a 355 lm sieve (where the smallest particles were lost) to isolate peanut particles from the matrix, taking a digital photograph of the remaining peanut particles using a at bed scanner, and counting the particles and assessing size using Image J (1.37a, National Institute of Health, USA). The particle size output from Image J was collected as a cumulative size distribution for each bolus, where cumulative area (mm2) was plotted against particle diameter (mm). The particle diameter was determined for each particle by assuming each particle was circular. To derive particle size parameters the cumulative distribution of each bolus was then tted to a RosinRammler function as in Hutchings et al. (2011) and Olthoff et al. (1984) , as shown in Eq. (1):

Matrix Embedded peanuts

20 mm

"  # b x 0:354 Q 1 exp ln 2 d50 0:354

15 mm

30 mm
Fig. 1. Illustration of the test food, with peanut quarters embedded inside the food matrix.

where x is the sieve class of any given particle (mm), and Q is the area fraction of particles that have a smaller diameter than x. d50 is the median diameter, where the sum of individual particles (in terms of area) that were larger than this diameter contribute 50% of the total 2D area, and the sum of individual particles (in terms of area) that were below this diameter make up the other 50% of

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

739

the total 2D area. b is the slope of the cumulative area distribution curve, known as the broadness value. A higher broadness value means the presence of particles with a smaller spread in size, and a lower broadness value means the presence of particles with a higher spread in size. R2 values were calculated to assess accuracy of the t of the cumulative size distribution to the RosinRammler function for each bolus. R2 was greater than 0.98 in every case, apart from distributions after 5 and 10 chews in the gelatine gel, which were above 0.91 and 0.94, respectively. The specic surface area of peanut particles in each bolus was calculated by measuring the total projected 2D area of peanut particles in each bolus, and dividing by the dry weight of peanut particles. Peanut retention was calculated as the percentage dry weight retained (across the sum of the expectorated bolus and debris). This was determined by measuring the dry weight of peanuts remaining in the total bolus (by placing the same particles which had been used for particle size analysis in an air dry oven at 105 C for 24 h), and by calculating the initial dry weight of peanuts in each test piece (as the initial total weight of peanuts in each test piece was known, and the initial moisture content of the peanuts was also known). In the case of boluses obtained in the nal two sessions, total dry weight of the bolus (matrix and particles) was determined by placing the expectorated bolus and washings (which contained the debris) in an air dry oven at 105 C for 24 h. The moisture content of the expectorated bolus was also determined by comparing the wet weight of the expectorated bolus with the dry weight of the expectorated bolus after 24 h at 105 C in the air dry oven. The wet weight of the total bolus, determined from every session, was calculated by adding the wet weight of the expectorated bolus to the wet weight of the debris (by subtracting the weight of the 25 mL of rinsing water). 2.7. Statistical analysis To quantitate relationships for the respective gelatine gel and chocolate data sets of each dependent variable with the number of chews, simple linear regression was undertaken on the data from each matrix treatment. This was followed by multiple regressions with dummy variables to compare the gelatine gel and chocolate for each dependant variable. 2.7.1. Simple linear regression To undertake simple linear regression the Y series (dependent variable) was transformed using natural log if it gave a stronger linear t:

Y a bx1 cx2 dx2 x1

where Y is the dependent variable, x1 is the chew number, and x2 the variable that was 0 for gel and 1 for chocolate. Hence, the linear model was t to the gelatine gel curve (a was the intercept of the gelatine gel curve, and b was the slope of the gelatine curve), and the values of the constants necessary to modify this t to the chocolate curve were determined. When the constant c was signicant (P < 0.05), the intercept was signicantly different between the gelatine gel and chocolate curves, and when d was signicant (P < 0.05), the slope was signicantly different between gelatine gel and chocolate curves. The level of statistical difference between the gelatine gel and chocolate for slope and intercept is provided in the text describing the results for each parameter. SPSS (version 16.0 for Windows) (SPSS Inc., USA) was used for all statistical analysis. This analysis is equivalent to an ANCOVA. 3. Results and discussion 3.1. Changes in the size distribution of peanut particles during the chewing sequence The gelatine gel matrices required a greater number of chews prior to swallowing than the chocolate matrices (Fig. 2). The breakdown of peanuts also followed contrasting pathways between matrices (Fig. 2). At early (510 chews) and middle stages (15 20 chews) of the chewing sequence, a greater proportion of large particles (P4 mm) were present in the gelatine bolus than in the chocolate bolus. After 25 chews, which was nearing the point at which chocolate bolus was swallowed, the differences were much smaller. At the point of swallowing, the peanut particle size distributions were similar in both types of bolus, which is in agreement with Hutchings et al. (2011), where eight subjects were studied. The gelatine gel matrix was more soft, cohesive, springy, and chewy than chocolate matrix (Table 2). These textural differences appear to have caused different trends in peanut particle size parameters (d50, b, and specic surface area) between matrices during the chewing sequence (Sections 3.1.13.1.3). 3.1.1. Changes in the d50 The d50 of peanut particles decreased signicantly with number of chews according to the following relationships for both chocolate (Eq. (4)) and gelatine gel (Eq. (5)) (where x is the number of chews). The d50 relationship was an exponential decay (Fig. 3A) (weaker ts were observed without natural log transformation):

Y a bx

Chocolate : lnd50 0:564 0:044 SE 0:020x 0:002 SE r 0:84; Pintercept < 0:0005; Pslope < 0:0005 4

where Y is the dependant variable, x the number of chews, a the intercept, and b the slope. The intercept and slope were signicant when a and b were signicantly different from 0 (P < 0.05). The correlation coefcient (r) was also calculated to evaluate the strength of each linear t for the respective gelatine gel and chocolate data sets for each variable. The correlation coefcient (r), and the significance of slope and intercept of each linear regression is provided in italics below each regression. Given that 17 out of the 18 data sets were normally distributed according to the KolmogorovSmirnov test for normality (with Lillifors signicance correction), where P > 0.05, the regression was considered an appropriate tool (if data are presented in a natural log form, the test for normality was undertaken on the logged data). 2.7.2. Multiple regression To undertake multiple regression the Y series (dependent variable) was transformed using natural log if it gave a stronger linear t, and the following linear model was tted to each linear curve:

Gelatine gel : lnd50 1:122 0:113 SE 0:023x 0:005 SE r 0:72; Pintercept < 0:0005; Pslope < 0:0005 5
There was a signicant difference in the intercept (P < 0.005) but not the slope (P > 0.05) for the ln (d50) value between the chocolate and gelatine gel boluses (Fig. 3A). With the exception of the boluses at the point of swallowing, the d50 of the peanuts was smaller in the chocolate than in the gelatine gel bolus (Fig. 3A). Once the bolus was ready for swallowing, d50 was similar for the chocolate and the gelatine gel. The exponential decline in d50 is typical of data that have been obtained using homogenous foods in previous studies involving peanuts (Kawashima et al., 2009), Brazil nuts and carrots (Lucas and Luke, 1986), and an articial test food, Optosil (Olthoff et al., 1984). An exponential decline is likely to occur because large particles are more easily broken into small particles than small particles are broken into ne particles (Lucas, 2004). This can be

740

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

5 chews
100

100 80 60 40 20 0

10 chews

100 80 60 40 20 0

15 chews

Cumulative area (%)

80 60 40 20 0 0
100

6
100 80 60 40 20 0

6 100 80 60 40 20 0

20 chews

25 chews

Bolus ready for swallowing

Cumulative area (%)

80 60 40 20 0 0 1 2 3 4 5 6

Chocolate matrix Gelatine gel matrix 0 1 2 3 4 5 6

Particle diameter (mm)

Particle diameter (mm)

Particle diameter (mm)

Fig. 2. Cumulative size distribution of the area of peanut particles in the food bolus at different points in the chewing sequence. Chocolate matrix containing peanuts: d (swallow after 26 1 chews), gelatine gel matrix containing peanuts: s (swallow after 43 1 chews) (mean SE).

Table 2 Textural properties of the matrices (mean SE). Matrices Gelatine gel Chocolate Hardness (N) 252 8 400 8 Cohesiveness 0.89 0.01 0.17 0.01 Springiness (mm) 10.1 0.1 2.03 0.23 Chewiness (mJ) 2270 90 150 20

explained by the selection function a measure of the probability a given food particle will make contact with the teeth for fracture during a chewing stroke (Lucas et al., 2002, 2004). Large particles have been found to have a higher selection function than small particles (Lucas and Luke, 1983; van der Glas et al., 1987). Hence, once particles reach a certain size the chewing process becomes rather inefcient at reducing particle size any further. In this study, particle size reduction appears to slow substantially after 20 chews for the chocolate, and after 25 chews for the gelatine gel (Figs. 2 and 3A). After this number of chews the peanut particles appear have been reduced to a size where the probability that the teeth will encounter and fracture them is small. 3.1.2. Changes in broadness (b) of the peanut particle size distribution The broadness value (b) of the peanut particle size distribution increased signicantly with the number of chews according to the following relationship for chocolate (Eq. (6)) and gelatine gel (Eq. (7)) (Fig. 3B) (where x is the number of chews):

mately form a bolus with a narrow distribution of contained particles. This result is in agreement with work by Lucas and Luke (1983) where the spread in raw carrot particle size of the bolus decreased during the chewing sequence.

3.1.3. Changes in specic surface area The specic surface area of the peanut particles increased signicantly with the number of chews according to the following relationship for chocolate (Eq. (8)) and gelatine gel (Eq. (9)) (Fig. 3C) (where x is the number of chews):

Chocolate : SSA 971 292 SE 130x 16 SE r 0:87; Pintercept < 0:0005; Pslope < 0:0005 8

Gelatine gel : SSA 341 192 SE 81x 8 SE r 0:90; Pintercept > 0:05; Pslope < 0:0005 9
There was no signicant difference in the intercept (P > 0.05), but a signicant difference in the slope, between the specic surface area of the peanut particles in the gelatine gel and chocolate boluses (P < 0.05) (Fig. 3C). The specic surface area of peanut particles inside the chocolate was greater than the specic surface area of those inside the gelatine gel until the bolus was ready for swallowing (Fig. 3C). The increase in specic surface area of the peanut particles during the chewing sequence is likely to be caused by an increase in the number of particles as size is reduced. It has been shown in previous work that the total surface area of particles within the bolus increases as a result of mastication (Prinz and Lucas, 1997; Okiyama et al., 2003).

Chocolate : b 0:900 0:040 SE 0:099x 0:002 SE r 0:70; Pintercept < 0:0005; Pslope < 0:0005 Gelatine gel : b 0:847 0:023 SE 0:004x 0:001 SE r 0:64; Pintercept < 0:0005; Pslope < 0:0005 7
There was no signicant difference in the intercept (P > 0.05), but a signicant difference in the slope (P < 0.05) for the broadness value (b) between the gelatine gel and chocolate boluses (Fig. 3B). The broadness value (b) was higher in the chocolate bolus than in the gelatine gel throughout the chewing sequence, i.e. there was a narrower spread of peanut particle size in the chocolate. The trends in the broadness value (b) show that the spread in peanut particle size is reduced uniformly during chewing to ulti-

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

741

1.4 1.2 1.0 ln(d50) (mm) 0.8 0.6 0.4

(A)

1.3

(B)

1.2

1.1

b
1.0 0.9 0.2 0.0 0 10 20 30 40 50 Number of chews 5000 ln (% Peanut retention (drywt/drywt)) 5.0 0.8 0 10 20 30 40 50 Number of chews

Specific surface area (mm2/g)

(C)
4000 3000

(D)
4.5

4.0

2000

1000

3.5

0 0 10 20 30 40 50

3.0 0 10 20 30 40 50

Number of chews Number of chews

Fig. 3. ln (d50 (mm)) (A), broadness (b) (B), specic surface area (mm2/g) (C), and ln (peanut retention (% dry mass basis)) (D) of the peanut particles in the food bolus at different points in the chewing sequence (mean SE). Chocolate matrix containing peanuts: d, gelatine gel matrix containing peanuts: s.

3.2. Peanut retention The retention of the peanuts (on a dry mass basis) decreased signicantly with the number of chews according to the following relationships for chocolate (Eq. (10)) and gelatine gel (Eq. (11)) (Fig. 3D) (where x is the number of chews). Data followed an exponential decay shape (weaker ts were observed without the natural log transformation of the data):

Smaller particles are likely to be shifted more easily into the oropharynx, and will exhibit greater loss during washing across the 355 lm sieve (used in this study to isolate the peanut particles for image analysis). 3.3. Total dry weight of the bolus Total dry weight of the bolus (dry weight of expectorated bolus + dry weight of debris) decreased signicantly with the number of chews for the chocolate (Eq. (12)), but not for the gelatine gel (Eq. (13)) (Fig. 4):

Chocolate : ln%Peanut dry wt retention 4:212 0:066 SE 0:033x 0:004 SE r 0:89; Pintercept < 0:0005; Pslope < 0:0005 Gelatine gel : ln%Peanut dry wt retention 4:369 0:065 SE 0:023x 0:003 SE r 0:87; Pintercept < 0:0005; Pslope < 0:0005 11
For the peanut retention there was no signicant difference in the intercept (P > 0.05) between the gelatine gel and chocolate boluses, but there was a signicant difference between their slopes (P < 0.05). Similar levels of peanut particle retention have been reported in boluses at the end of the chewing sequence (Jalabert-Malbos et al., 2007; Flynn et al., 2011). Losses are suggested to occur due to intermediate swallowing (Jalabert-Malbos et al., 2007), fat loss into saliva, transportation to the oropharynx (where particles will not be collected using a rinse after expectoration), and during washing of the particles across sieves for particle size assessment (Flynn et al., 2011). Greater peanut loss in the chocolate than in the gelatine gel during the chewing sequence (but not at the swallow point) may be a function of a smaller particle size at each interval.

10

12.0

Total bolus dry weight (g)

11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 0 10 20 30 40 50 60

Number of chews
Fig. 4. Total dry weight of boluses recovered (expectorated bolus + debris) at different points in the chewing sequence. Chocolate matrix containing peanuts: d, gelatine gel matrix containing peanuts: s.

742

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

Chocolate : Total dry wt 11:51 0:14 SE 0:04x 0:01 SE r 0:86; Pintercept < 0:0005; Pslope < 0:0005 12

the relationships below (Eqs. (14) and (15)) (where x is the number of chews):

Gelatine gel : Total dry wt 9:03 0:13 SE 0:00x 0:01 SE r 0:10; Pintercept < 0:0005; Pslope > 0:05 13
For the total dry weight, a signicant difference in the intercept (P < 0.0005) and the slope (P < 0.0005) was found between the chocolate and gelatine gel. In comparison with peanut retention (Fig. 3D), where approximately only 30% of the original dry weight was retained by the end of the chewing sequence, the total dry weight loss of the bolus (matrix and peanuts) appears small (Fig. 4). In the case of gelatine gel, the total dry weight remains around 9 g throughout the chewing sequence. For the chocolate, the total dry weight dropped from approximately 11.5 g to 10.5 g. Peyron et al. (2004) measured particle retention after one quarter, one half, and a full chewing sequence with peanuts, almonds, pistachio nuts, carrots, radish, and cauliower. A signicant reduction in particle weight retention from one-quarter to one-half of a sequence, and from one-half to a full sequence was found for all six foods. Smaller loss in this study may be because continuous matrices such as gelatine or chocolate can be contained more easily by the mouth than discrete particles.

Chocolate : Total wet wt of total bolus 11:79 0:16 SE 0:13x 0:01 SE r 0:91; Pintercept < 0:0005; Pslope < 0:0005 Gelatine gel : Total wet wt of total bolus 12:88 0:12 SE 0:12x 0:01 SE r 0:96; Pintercept < 0:0005; Pslope < 0:0005 15
For the total bolus wet weight, a signicant difference in the intercept (P < 0.0005), but no signicant difference in the slope (P > 0.05), was found between the chocolate and the gelatine gel. The increase in total bolus weight can be attributed to saliva addition. 3.4.2. Wet weight of expectorated bolus The wet weight of the expectorated bolus increased signicantly for the gelatine gel but not for the chocolate (Fig. 5B) with the number of chews, as shown by Eqs. (16) and (17) (where x is the number of chews):

14

Chocolate : Wet wt of expectorated bolus 11:21 0:28 SE 0:03x 0:02 SE r 0:31; Pintercept < 0:0005; Pslope > 0:05 Gelatine gel : Wet wt of expectorated bolus 12:55 0:14 SE 0:09x 0:01 SE r 0:93; Pintercept < 0:0005; Pslope > 0:05
17 16 Expectorated bolus (g) 15 14 13 12 11 10

3.4. Wet weight of the bolus 3.4.1. Total wet weight of the bolus Total wet weight of the bolus (wet weight of expectorated bolus + wet weight of debris) increased signicantly for the chocolate and gelatine gel (Fig. 5A) with the number of chews according to
19 18 Total bolus weight (g) 17 16 15 14 13 12 11 0 10 20 30 40 50 Number of chews

16

17

(A)

(B)

10

20

30

40

50

Number of chews Moisture content (gH20/100gtotalweight)

(C)

100 80 60 40 20 0 0 10 20 30 40 50

(D)

3 Debris weight (g)

0 0 10 20 30 40 50 Number of chews

60

Number of chews

Fig. 5. Total wet weight of the bolus (expectorated bolus + debris) (A), wet weight of expectorated bolus (B), debris wet weight (C), and bolus moisture content (expectorated bolus) (D) at different points in the chewing sequence. Chocolate matrix containing peanuts: d, gelatine gel matrix containing peanuts: s (mean SE).

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744

743

For the weight of expectorate bolus, the intercept (P < 0.0005) and the slope (P < 0.0005) were signicantly different between chocolate and gelatine gel (Fig. 5B). The expectorated bolus was greater for the gelatine gel than the chocolate matrix during the chewing sequence, and increased at a greater rate. The increase in wet weight of the expectorated bolus can also be attributed to saliva addition. 3.4.3. Wet weight of debris The weight of debris increased signicantly for both matrices with the number of chews (Fig. 5C) according to the following relationships (Eqs. (18) and (19)) (where x is the number of chews):

Chocolate : Debris wt 0:58 0:24SE 0:10x 0:01 SE r 0:79; Pintercept < 0:05; Pslope < 0:0005 18

Gelatine gel : Debris wt 0:45 0:14 SE 0:03x 0:01 SE 19 r 0:70; Pintercept < 0:0005; Pslope < 0:0005
For debris weight, there was no signicant difference in the intercept between chocolate and gelatine gel (P > 0.05), but the slope was signicantly different (P < 0.0005). The weight of debris was greater for the chocolate than the gelatine gel throughout the chewing sequence, and increased at a greater rate (Fig. 5C). Signicant differences between the gelatine gel and chocolate boluses in the expectorated bolus and weight of debris is likely to be due to differences in the rheological properties of the matrices (Fig. 5B and C). It is probable that during mastication a substantial portion of chocolate adhered to the walls of the mouth. 3.5. Bolus moisture content The moisture content expectorated bolus increased signicantly with the number of chews for the chocolate and the gelatine gel (Fig. 5D) according to the relationships below (Eqs. (20) and (21)) (where x is the number of chews):

onstrated a greater spread in size than particles inside the chocolate bolus during much of the chewing process. Given the difference in the breakdown pathways between matrices, it is highly likely that the selection function differs between matrices. Differences in the selection function between matrices not only appear to affect differences in d50 and the specic surface area of the peanut particles, but also the broadness (b) of the peanut particle distribution. As peanut particles inside the chocolate are reduced in size more quickly than inside the gelatine gel as the chewing sequence progresses, the spread in size is also smaller (higher b values) as larger particles are more easily size selected inside the chocolate matrix. Two alternative hypothesizes are offered for the differences observed in breakdown rates. Firstly, given that cohesiveness, chewiness, and springiness were all greater in the gelatine gel than in the chocolate matrix (Table 2), it is likely to be more difcult for the molar teeth of the subject to effectively fracture or induce ow of the gelatine gel to access the peanuts. The chocolate is also likely to soften and melt at temperatures found inside the mouth (Do et al., 2007), and deform easily in response to stress. Secondly, breakdown pathways may differ because the gelatine gel requires greater masticatory work. Peanut particles may be isolated from both matrices early in the chewing sequence and then masticated separately from the matrix. Given the large differences in cohesiveness, chewiness, and springiness between matrices, more time may be spent breaking down the gelatine gel matrix than the chocolate matrix to produce a safe bolus to swallow. As a result, less time would be spent masticating the peanuts in the case of the gelatine gel, and consequently the rate at which particle size is reduced would be slower in the early and middle stages of the chewing sequence. 3.7. The selection of a single subject to function as a chewing device By using rigorous measures to select a subject with a high standard of dental and oral health, and relatively normal and consistent mastication behavior, this study used a single subject as a chewing device as an alternative to a chewing robot. This approach had a number of advantages, as it enabled multiple variables to be tested at regular intervals of the chewing sequence, as particle size (in terms of d50, broadness (b), and specic surface area of peanut particles), dry weight, wet weight, peanut weight retention, and moisture content, were all assessed. It also minimized variability in the data resulting from variation between individuals to simplify analysis, and acted as a biological technique for masticating the food samples in a reproducible way. Some recent chewing simulators have demonstrated similarity in particle reduction between in vitro and in vivo tests involving peanuts and carrots (Salles et al., 2007; Mishellany-Dutour et al., 2011), however their usefulness studies such as this remains questionable. Given that the human mouth dynamically reacts to varying food properties (Peyron et al., 2002; Foster et al., 2006), such robots are unlikely to be able to accurately compare the breakdown dynamics between two or more types of food. The disadvantage of this approach is that results are not representative for a population, and consequently, conclusions can only be drawn for this subject. However, given the selection technique employed, and the similarity in results at the point of swallowing between this subject and the eight subjects studied by Hutchings et al. (2011), this research provides a strong basis for further work into the dynamics of breakdown in heterogeneous foods. 4. Conclusion The properties of one food component altered the pathway of particle breakdown of another food component when heteroge-

Chocolate : %Moisture 1:6x 0:1 SE r 0:97; Pintercept > 0:05; Pslope < 0:0005 Gelatine gel : %Moisture 43:1 3:6 SE 0:9x 0:1 SE r 0:89; Pintercept < 0:0005; Pslope < 0:0005 21
For moisture content there was a signicant difference in the intercept (P < 0.0005) between the chocolate and gelatine gel, but no signicant difference in the slope (P < 0.05). The initial difference in moisture content continued throughout the sequence, and the nal moisture content of the gelatine bolus at the point of swallowing was approximately twice the moisture content of the chocolate bolus at the point of swallowing (Fig. 5D). The increase in moisture content for chocolate and gelatine gel can again be attributed to saliva addition. Saliva content has been shown to increase in several studies during the chewing sequence (Mioche et al., 2002; Loret et al., 2009). The similar rates of increase in total bolus weight and moisture content (Fig. 5A and D) between matrices suggest that saliva ow was stimulated similarly by both matrices. 3.6. The inuence of the matrices on the dynamics of peanut particle size reduction These results show that for this subject, the properties of the matrix altered the pathway of particle breakdown during the mastication of a heterogeneous food system. In this case, peanut particle size during the chewing sequence differed between the chocolate and the gelatine gel boluses (Figs. 2 and 3A and B), where peanut particles inside the gelatine gel bolus were larger and dem-

20

744

S.C. Hutchings et al. / Journal of Food Engineering 109 (2012) 736744 Lenfant, F., Loret, C., Pineau, N., Hartmann, C., Martin, N., 2009. Perception of oral food breakdown. The concept of sensory trajectory. Appetite 52 (3), 659667. Loret, C., Hartmann, C., Martin, C., 2009. Rheology and sensory characterization of a swallowed food bolus. Paper presented at the fth International Symposium on Food Rheology and Structure, Zurich, Switzerland. Lucas, P.W., 2004. Dental Functional Morphology: How Teeth Work. Cambridge University Press, Cambridge, UK. Lucas, P.W., Luke, D.A., 1983. Methods for analysing the breakdown of food in human mastication. Archives of Oral Biology 28 (9), 813819. Lucas, P.W., Luke, D.A., 1986. Is food particle size a criterion for the initiation of swallowing? Journal of Oral Rehabilitation 13 (2), 127136. Lucas, P.W., Prinz, J.F., Agrawal, K.R., Bruce, I.C., 2002. Food physics and oral physiology. Food Quality and Preference 13 (4), 203213. Mishellany, A., Woda, A., Labas, R., Peyron, M.A., 2006. The challenge of mastication: preparing a bolus suitable for deglutition. Dysphagia 21 (2), 8794. Mishellany-Dutour, A., Peyron, M.A., Croze, J., Francois, O., Hartmann, C., Alric, M., Woda, A., 2011. Comparison of food boluses prepared in vivo and by the AM2 mastication simulator. Food Quality and Preference 22 (4), 326331. Mioche, L., Hiiemae, K.M., Palmer, J.B., 2002. A posteroanterior video uorographic study of the intra-oral management of food in man. Archives of Oral Biology 47 (4), 267280. Okiyama, S., Ikebe, K., Nokubi, T., 2003. Association between masticatory performance and maximal occlusal force in young men. Journal of Oral Rehabilitation 30 (3), 278282. Olthoff, L.W., van der Bilt, A., Bosman, F., Kleizen, H.H., 1984. Distribution of particle sizes in food comminuted by human mastication. Archives of Oral Biology 29 (11), 899903. Parada, J., Aguilera, J.M., 2009. In vitro digestibility and glycemic response of potato starch is related to granule size and degree of gelatinization. Journal of Food Science 74 (1), E34E38. Peyron, M.A., Gierczynski, I., Woda, A., Loret, C., Dardevet, D., Hartmann, C., Martin, N., 2009. Enclosed cues in food bolus physical properties responsible for swallowing. Dysphagia 24 (4), 454455. Peyron, M.A., Lassauzay, C., Woda, A., 2002. Effects of increased hardness on jaw movement and muscle activity during chewing of visco-elastic model foods. Experimental Brain Research 142 (1), 4151. Peyron, M.A., Mishellany, A., Woda, A., 2004. Particle size distribution of food boluses after mastication of six natural foods. Journal of Dental Research 83 (7), 578582. Pons, M., Fiszman, S.M., 1996. Instrumental texture prole analysis with particular reference to gelled systems. Journal of Texture Studies 27 (6), 597624. Prinz, J., Lucas, P.W., 1997. An optimization model for mastication and swallowing in mammals. Proceedings of the Royal Society B: Biological Sciences 264 (1389), 17151721. Ranawana, V., Monro, J.A., Mishra, S., Henry, C.J.K., 2010. Degree of particle size breakdown during mastication may be a possible cause of inter individual glycemic variability. Nutrition Research 30 (4), 246254. Read, N.W., Welch, I.M., Austen, C.J., Barnish, C., Bartlett, C.E., Baxter, A.J., Brown, G., Compton, M.E., Hume, K.E., Storie, I., Worlding, J., 1986. Swallowing food without chewing a simple way to reduce postprandial glycemia. British Journal of Nutrition 55 (1), 4347. Szczesniak, A.S., 1963. Classication of textural characteristics. Journal of Food Science 28 (4), 385389. Salles, C., Tarrega, A., Mielle, P., Maratray, J., Gorria, P., Liaboeuf, J., Liodenot, J.J., 2007. Development of a chewing simulator for food breakdown and the analysis of in vitro avor compound release in a mouth environment. Journal of Food Engineering 82 (2), 189198. Suzuki, H., Fukushima, M., Okamoto, S., Takahashi, O., Shimbo, T., Kurose, T., Yamada, Y., Inagaki, N., Seino, Y., Fukui, T., 2005. Effects of thorough mastication on postprandial plasma glucose concentrations in nonobese Japanese subjects. Metabolism-Clinical and Experimental 54 (12), 15931599. Taylor, A.J., 1996. Volatile avor release from foods during eating. Critical Reviews in Food Science and Nutrition 36 (8), 765784. van der Glas, H.W., van der Bilt, A., Olthoff, L.W., Bosman, F., 1987. Measurement of selection chances and breakage functions during chewing in man. Journal of Dental Research 66 (10), 15471550. Woda, A., Mishellany-Dutour, A., Batier, L., Francois, O., Meunier, J.P., Reynaud, B., Alric, M., Peyron, M.A., 2010. Development and validation of a mastication simulator. Journal of Biomechanics 43 (9), 16671673. Xu, W.L., Pap, J.S., Bronlund, J., 2008. Design of a biologically inspired parallel robot for foods chewing. IEEE Transactions on Industrial Electronics 55 (2), 832841. Xu, W., Bronlund, J.E., 2010. Mastication Robots: Biological Inspiration to Implementation. Springer-Verlag, Berlin, Germany. Yven, C., Guessasma, S., Chaunier, L., Della Valle, G., Salles, C., 2010. The role of mechanical properties of brittle airy foods on the masticatory performance. Journal of Food Engineering 101 (1), 8591.

neous foods were masticated by a carefully selected single subject. Embedding peanuts inside two different matrices (chocolate and gelatine gel) resulted in peanut particles being broken down by different processes in the two matrices chewed by this subject. The matrices induced different dynamic trends in terms of the parameters of the peanut particles (d50, broadness (b), specic surface area), and in terms of the mass of bolus components. Results strongly suggest that the type of matrix inuences the ability of the subject to effectively select and breakdown the embedded peanuts. Acknowledgements This research was funded by the New Zealand Foundation for Research, Science and Technology under Contract C02X0401, Lifestyle Foods for Energy Balance. The authors greatly appreciate the time given by the subject who took part. References
Alfonso, M., Neyraud, E., Blanc, O., Peyron, M.A., Dranseld, E., 2002. Relationship between taste and chewing patterns of visco-elastic model foods. Journal of Sensory Studies 17 (2), 193206. Bourne, M., 2002. Food Texture and Viscosity: Concept and Measurement, second ed. Academic Press, New York, USA. Brown, W.E., Braxton, D., 2000. Dynamics of food breakdown during eating in relation to perceptions of texture and preference. A study on biscuits. Food Quality and Preference 11 (4), 259267. Brown, W.E., Langley, K.R., Martin, A., 1994. Characterisation of patterns of chewing behaviour in human subjects and their inuence on texture perception. Journal of Texture Studies 25 (4), 455468. Do, T.A.L., Hargreaves, J.M., Wolf, B., Hort, J., Mitchell, J.R., 2007. Impact of particle size distribution on rheological and textural properties of chocolate models with reduced fat content. Journal of Food Science 72 (9), E541E552. Flynn, C.S., Foster, K.D., Bronlund, J.E., Lentle, R.G., Jones, J.R., Morgenstern, M.P., 2011. Identication of multiple compartments present during the mastication of solid food. Archives of Oral Biology 56 (4), 345352. Foster, K.D., Woda, A., Peyron, M.A., 2006. Effect of texture of plastic and elastic model foods on the parameters of mastication. Journal of Neurophysiology 95 (6), 34693479. Gonzalez, R., Montoya, I., Carcel, J., 2001. Review: the use of electromyography on food texture assessment. Food Science and Technology International 7 (6), 461 471. Howel, A.H., Brudevold, F., 1950. Vertical forces used during chewing food. Journal of Dental Research 29, 133136. Hutchings, J.B., Lillford, P.J., 1988. The perception of food texture the philosophy of the breakdown path. Journal of Texture Studies 19 (2), 103115. Hutchings, S.C., Bronlund, J.E., Lentle, R.G., Foster, K.D., Jones, J.R., Morgenstern, M.P., 2009. Variation of bite size with different types of food bars and implications for serving methods in mastication studies. Food Quality and Preference 20 (6), 456460. Hutchings, S.C., Foster, K.D., Bronlund, J.E., Lentle, R.G., Jones, J.R., Morgenstern, M.P., 2011. Mastication of heterogeneous foods: peanuts inside two different food matrices. Food Quality and Preference 22 (4), 332339. Jalabert-Malbos, M.L., Mishellany-Dutour, A., Woda, A., Peyron, M.A., 2007. Particle size distribution in the food bolus after mastication of natural foods. Food Quality and Preference 18 (5), 803812. Kawashima, K., Miura, H., Kato, H., Yoshida, K., Yoshiro, T., 2009. The study of communition behaviour of food on buccal and lingual side during mastication. Journal of Medical Dental Science 56 (4), 131138. Kohyama, K., Mioche, L., 2004. Chewing behavior observed at different stages of mastication for six foods, studied by electromyography and jaw kinematics in young and elderly subjects. Journal of Texture Studies 35 (4), 395414. Kohyama, K., Mioche, L., Martin, J.F., 2002. Chewing patterns of various texture foods studied by electromyography in young and elderly populations. Journal of Texture Studies 33 (4), 269283. Lassauzay, C., Peyron, M.A., Albuisson, E., Dranseld, E., Woda, A., 2000. Variability of the masticatory process during chewing of elastic model foods. European Journal of Oral Sciences 108 (6), 484492.