12
This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA
Overview
Download & View 12 as PDF for free.
More details
- Words: 2,603
- Pages: 8
E. T. Contis et al. (Editors) Food Flavors: Formation, Analysis and Packaging Influences © 1998 Elsevier Science B.V. All rights reserved
Comparisons of volatile compounds released during consumption Cheddar cheeses by different consumers
117
of
CM. Delahunty, P.J. O'Riordan, E.M. Sheehan and P.A. Morrissey Department of Nutrition, University College, Cork, Ireland
Abstract Methods exist for measuring volatile compounds released in the mouth during food consumption, however little work has compared the volatile compounds released during consumption by different consumers or related individual differences to consumers' chewing patterns and saliva production rates. In this work, eight consumers were chosen and each consumed six Cheddar cheeses during Buccal Headspace Analysis (BHA). Released volatile compounds were measured for each cheese and for each consumer. Electromyography was used to record each consumers chewing style, and their saliva production rate was also measured. It was found that although there were differences in consumers' chewing styles and saliva production rates, the volatile profiles obtained by BHA, for each individual, were similar for each cheese when compared with the other cheeses examined.
1. INTRODUCTION It is the volatile compounds of a food, released in sufficient concentration during consumption, which stimulate the olfactory epithelium and induce perceived odor. Recent flavor research has emphasised the importance of volatile release from a food matrix and shown how volatile release is related to consumer flavor perception. This work is often driven by a food industry which must reduce costs or meet the demands of diet conscious, but discerning, consumers who wish to reduce fat and salt intake. Substitution for these ingredients is necessary to restore removed flavor and regulate flavor release. However, understanding of flavor release is made difficult by the complexity of the interactions between foods and consumers. Each volatile compound has different physiochemical properties and its release is influenced by interactions with other food matrix variables such as moisture, fat, protein, carbohydrate, and other soluble (salt, sugars) and non-soluble materials. In addition, food breakdown and mixing with saliva during consumption, respiratory air flow over and around the food and temperature and pH changes occurring in the consumers mouth will influence volatile release and subsequent flavor perception. It is also known that individual consumers expression of flavor differs as a result of physiological, psychological and social differences [1,2]. Therefore an underlying question which remains unanswered is; what part of flavor differences between foods result from volatile release dynamics from the food matrix, and what part result from differences between consumers? Conclusions reached in response to this question have been mixed.
118 There are model systems which measure volatile compounds released while mimicking conditions in the mouth [3-5]. Other methods measure volatile release directly during consumption using mass spectrometry of breath [6,7] and indirectly by trapping volatiles on adsorbents, such as Tenax, before analysis [8-10]. Soeting and Heidema [6] showed thirtyfold differences in the relative quantities of 2-pentanone which was measured directly from the breath of different consumers. Van Ruth et al [10] also found subject specific volatile profiles were released during consumption of vegetables. Taylor et al [11] trapped volatiles released from mint sweets during consumption and also found differences between subjects in terms of the quantities of volatiles released. However, they concluded that there were similarities between the relative concentrations of volatiles released for each subject. Delahunty et al [12], who analyzed Buccal Headspace Analysis [BHA;13] data using Principal Components Analysis [PCA;14] to examine the volatile profiles released during consumption of cheeses by three different consumers, found product specific volatile release was most important. However, three consumers were too few to draw any firm conclusions. Workers studying food texture have developed methods such as electromyography [15] which measure muscle activity during mastication of a food matrix. From these measurements they have shown mastication patterns and can calculate the amount of work done by a consumer during consumption. These methods, which show considerable differences between consumers' mastication characteristics, have recently been related to differences between consumers' temporal perception of flavor intensity measured by timeintensity sensory analysis [16]. Other physiological parameters, such as the influence of saliva [4] and air flow through the mouth [17] have also been investigated. The present study was carried out to investigate discrepancies in the literature relating to the differences between consumers' interactions with foods and the relationships found between physiological measures during chewing and individuals' differences in flavor perception. In order to achieve this, similar varieties of a complex food were chosen and a multivariate technique, PCA, was used to examine the volatile profiles released.
2. EXPERIMENTAL 2.1. Samples and consumers Six Cheddar cheeses, in 5kg blocks, of equal age (6-8 months) were obtained from 4 different producers. Eight consumers, 3 female and 5 male, aged between 22 and 28 were used for all studies. 2.2. Buccal Headspace Analysis Buccal Headspace Analysis of each cheese was carried out for each consumer in triplicate. For this method a 50 g cheese sample was consumed in 10 x 5g pieces in a normal way, allowing 30 s for the consumption of each piece. During the entire consumption time (5 min) volatile compounds released were displaced through the nose by vacuum and trapped on a Tenax-TA trap. The order of sample analysis was balanced for consumers, cheeses and day of consumption [18]. A blank buccal headspace sample was taken each day for each consumer. Traps were thermally desorbed using a Teckmar Purge and Trap 3000 concentrator (Teckmar, Cincinnati, OH, USA). Desorbed volatiles were identified and quantified using gas chromatography-mass spectrometry (GC-MS) with a Varian Saturn GC-3400CX
119 incorporating a Varian Saturn 3 GC/MS detector (Varian chromatography systems, Mitchell drive, Walnut Creek, CA, USA).The column was a DB-5ms, 30m x 0.257mm fused silica capillary column, with a film thickness of 0.25 |Lim (J & W scientific, Folsom, CA, USA) 2.3 Mastication behaviour The activity of the consumers left and right masseter muscles during chewing was recorded by Electromyography [15]. The electromyograph record was measured for 1 cheese over a period of 5 min (10 x 30 s for each 5g piece ), in triplicate, for each of the eight consumers. Each individuals electromyogram was integrated using a poly VIEW data acquisition and analysis system (Grass instrument division, Astro-Med Inc., East Greenwich Avenue, West Warwick, UK) 2.4 Saliva production Consumers unstimulated saliva production was measured by allowing their saliva to drip into a beaker for a 5 min period. The consumers swallowed immediately before collection and forcefully spat out at the end [19]. The stimulated saliva production was measured by dividing the volume of saliva produced by each consumer in response to 50g of cheese (10 x 5g) by the chewing time required by the consumer for that cheese [19]. Each measurement was repeated four times. 2.5 Data Analysis Buccal headspace data was analyzed by PC A, using the Unscrambler v 6.0 (CAMO AS, N-7041 Trondheim, Norway), of the log transformed peak areas of volatile compounds. Electromyography data was analyzed by Analysis of Variance (ANOVA) using SPSS v 6.1 (SPSS Inc. Chicago, IL 60611, USA) of the totals for chew number, chew time, chew rate and chew work. Saliva production data was analyzed by ANOVA of the unstimulated and stimulated saliva flow rates. Differences between cheeses and between subjects were investigated using ANOVA. Relationships between data sets were investigated by linear and Partial Least Squares regression [PLS;20], using the Unscrambler v 6.0.
3. RESULTS AND DISCUSSION In the present study the quantities and balance of volatile compounds released during consumption of a food, by different consumers, was compared. For this purpose Cheddar cheese was chosen as this represents a complex protein matrix containing fat and moisture. To minimize product related compositional differences, and therefore to maximize the influence of consumer related differences to volatile release from one food type, cheeses of equal age were chosen. Eighteen volatile compounds were selected from chromatograms of BHA of all cheeses and the amounts of each present were quantified. Both Figures 1 and 2 depict two PCA's. The first (in italics) was calculated using individual consumers' headspace data (triplicates averaged) and the second using the average of the 8 consumers.
120 Figure 1. PC A scores on PC's 1 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H) (see text for explanation).The pooled SD for the analysis is represented by an ellipse on cheese 1. 4 T
3B 4C
13H ^^
c o a.
4G
-4
5C
"^ fr ^^ Ic ,2iP2f 4H 2F 3G
5F
6E IG
-im
4A
IB
2C
2 6D
OH
5 6G
-4 -^
Principal component 1 {?>9Vo)(24%) Principal Components (PC) 1 and 3, which accounted for 39% and 15% of the explained variance, respectively, of the PC A of compound peak areas, showed significant differences {p = 0.017 and p = 0.021) between cheeses (Table 1 and Figure 1). Differences between consumers (p = 0.050), which accounted for \3% of the explained experimental variance, were found on PC3 (Table 1 and Figure 2). Table 1 ANOVA between cheeses (1 - 6 ) and between consumers ( subjects A • H)on Principal Components 1 - 4 of the PC A Principal component Cheese Subject PCI PC2 PC3 PC4
0.017 0.657 0.021 0.055
0.873 0.072 0.050 0.351
The volatile compounds which distinguished the cheeses from one another on these components are shown in the PC loadings plots (Figures 3 and 4).The differences found between cheeses on PCI, which was the most important as it contained the highest proportion of the experimental variance (39%), were caused mostly by the quantities of compounds released during consumption by each consumer rather than by their balance.
121 Figure 3. PC A loadings for 18 volatile compounds on PC's 1 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H). 0.5
T 2-heptanone
cyclohexanel cyclohexane2 cpdl3 cpdl4
o OH
-0.5
cpd3 cpd5
toluene
cpd2
dmds
cpdl7
0.5
heptane
(Eodecane
ethyl butyrate cpdl2
cpdl6 -0.5 -^
Principal component 1 (39%) However, differences between cheeses found on PC's 2 and 3 were caused mostly by differences in the balance of the compounds released. This can be determined from the relative positions of the volatile compounds in the loadings plots (Figure 4). In a previous study physiological differences between consumers have been related to differences in flavor perception [16]. In this study very significant differences were found between consumers mastication characteristics and also between their saliva production rates (p = 0.000 for all parameters apart from chew rate (p = 0.021) and chew work (p = 0.045)) (Table 2 and Figure 5). Using linear regression chew number and chew work were found to relate to saliva production rate during cheese consumption for 5 of the 8 subjects (r = 0.98 and r = 0.84, respectively). However by using PLS and linear regression, no significant relationships were found between the measured physiological characteristics and total volatile release. Sensory evaluation of the cheeses is not reported in this study and therefore no conclusions can be made with regard to consumers' expressions of flavor perception. Further work is also required to investigate the dynamics of volatile release during time of consumption.
4. CONCLUSION Some differences were found between the quantities of volatile compounds released during cheese consumption by different consumers. Very significant differences were found between consumers mastication characteristics and between saliva production rates during cheese consumption. Despite these differences, the distribution of experimental variance explained
122 Table 2 Mastication behaviour and saliva production rates of 8 consumers ( subjects A-H) Subject TCN
A B C D E F G H pooled SD P
224.33 138.67 348.00 324.33 228.67 215.67 184.33 282.33 22.84 0.000
Electromyography^ TCT CR
141.26 169.07 212.27 198.53 144.00 212.07 145.71 180.99 28.35 0.000
1.59 1.08 1.67 1.64 1.68 1.02 1.27 1.55 0.12 0.021
CW
2435.86 1011.45 3823.63 4497.6 2066.55 1194.73 866.71 3010.14 1403.50 0.045
Saliva flow rate'' Unstim. Stimulated
0.77 0.43 0.59 1.06 0.53 0.70 0.76 0.35 0.12 0.000
5.30 4.06 1.69 7.31 4.52 4.89 3.25 3.48 0.65 0.000
^ TCN = total chew number; TCT = total chew time (sec); CR = chew rate (chew / sec); CW = chew work b Unstim. = unstimulated saliva production rate (mL / min); Stimulated saliva production rate (mL / min)
Figure 5. Mastication behaviour of 3 consumers during consumption of one 5 g piece of cheese. CN = chew number; CT = chew time (sec); CR = chew rate (chew / sec); CW = chew work. Pat (A) CN = 23 CT= 14.66 CR=1.57 C W = 133.03
a
Chew Time (sec)
123 Figure 3. PCA scores on PC's 2 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H). The pooled SD for the analysis is represented by an ellipse on cheese 6. 4 T
4C 2A IG 6C
C
ID 3D
2B 2F
o
r^
4A 2C
OH
5E 6F 5D 6Bry
ex
ifP
5G
3F6D
,^
-4 -^
Principal component 2 (\3%)(]2%) Figure 4. PCA loadings of 18 volatile compounds on PC's 2 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H). 0.5 T 2-heptanone
cyclohexanel cycldhexane2 pctane cdd7 cpdl2 cpdl4
o a.
|cpd3
cpf
-0.5 dmds
dodecane
heptane
cpd2
0.5 toluene
cpdl7 ethylbutyrate cpdl2 -0.5 -^
cpdl6
Principal component 2 (13%)
124 by the PCA showed that Cheddar cheese of equal age could be identified by their product specific volatile release. Therefore, the volatile profile for a particular cheese at the end of consumption was found to be similar in all consumers.
5. ACKNOWLEDGEMENT This work was part funded by the Department of Agriculture, Food and Forestry, Ireland, under the Food Industry Sub-Programme of EU Structural Funds.
6. REFERENCES 1 D. Lancet, In: Sensory Transduction (D.P. Corey and S.D. Roper, eds.). Pp. 73, Rockefeller University, New York, 1992. 2 J.R. Piggott, Fd. Qual. Pref, 5 (1994) 167. 3 W.E. Lee III, J. Fd. Sci., 51 (1986) 249. 4 D.D. Roberts and T.E. Acree, J. Agric. Fd. Chem., 43 (1995) 2179. 5 K. Napi, F. Kropf and H. Klostermeyer, Z Lebensm Unters Forsch, 201 (1995) 62. 6 W.J. Soeting and J. Heidema, Chem. Senses, 13:4 (1988) 607. 7 R.S.T. Linforth, K.E. Ingham and A.J.Taylor, In: Flavour Science: Recent Developments (A.J. Taylor and D.S. Mottram, eds.). Pp. 361, Royal Society of Chemistry, Oxford, 1997. 8 R.S.T. Linforth and A.J.Taylor, Fd. Chem., 48 (1993) 115. 9 CM. Delahunty, J.R. Piggott, J.M. Conner and A. Paterson, In: Trends in Flavour Research (H. Maarse and D.G. van der Heij, eds.). Pp. 47, Elsevier Applied Science, Amsterdam, 1994. 10 S.M. Van Ruth, J.P. Roozen and J.L. Cozijnsen, Fd. Chem., 53 (1995) 15. 11 A.J. Taylor, R.S.T. Linforth, K.E. Ingham and A.R. Clawson, In: Bioflavour '95 (P. Etievant and P. Schreier, eds.). Pp. 45, INRA, Paris, 1995. 12 CM. Delahunty, F. Crowe and P.A. Morrissey, In: Flavour Science: Recent Developments (A.J. Taylor and D.S. Mottram, eds.). Pp. 339, Royal Society of Chemistry, Oxford, 1997. 13 CM. Delahunty, J.R. Piggott, J.M. Conner and A. Paterson, J. Sci. Fd. Agric, 71 (1996) 273. 14 J.R. Piggott and K. Sharman. In: Statistical Procedures in Food Research (J.R. Piggott, ed.) Pp. 181, Elsevier Applied Science, London, 1986. 15 M.M. Boyar and D. Kilcast, J. Fd. Sci., 51 (1986) 859. 16 W.E. Brown, C Dauchel and I. Wakeling, J. Texture Stud., 27 (1996) 433. 17 M. Harrison and B.P. Hills, Int. J. Fd. Sci. Tech., 32 (1997) 1. 18 H.J.H. MacFie, N. Bratchell, K. Greenhoff and I.V. ValHs, 1989. J. Sens. Stud., 4 (1989) 129. 19 S. Watanabe and C Dawes, Arch. Oral Biol., 33:1 (1988), 1. 20 M. Martens and H. Martens, In: Statistical Procedures in Food Research (J.R. Piggott, ed.) Pp. 293, Elsevier Applied Science, London, 1986.
Comparisons of volatile compounds released during consumption Cheddar cheeses by different consumers
117
of
CM. Delahunty, P.J. O'Riordan, E.M. Sheehan and P.A. Morrissey Department of Nutrition, University College, Cork, Ireland
Abstract Methods exist for measuring volatile compounds released in the mouth during food consumption, however little work has compared the volatile compounds released during consumption by different consumers or related individual differences to consumers' chewing patterns and saliva production rates. In this work, eight consumers were chosen and each consumed six Cheddar cheeses during Buccal Headspace Analysis (BHA). Released volatile compounds were measured for each cheese and for each consumer. Electromyography was used to record each consumers chewing style, and their saliva production rate was also measured. It was found that although there were differences in consumers' chewing styles and saliva production rates, the volatile profiles obtained by BHA, for each individual, were similar for each cheese when compared with the other cheeses examined.
1. INTRODUCTION It is the volatile compounds of a food, released in sufficient concentration during consumption, which stimulate the olfactory epithelium and induce perceived odor. Recent flavor research has emphasised the importance of volatile release from a food matrix and shown how volatile release is related to consumer flavor perception. This work is often driven by a food industry which must reduce costs or meet the demands of diet conscious, but discerning, consumers who wish to reduce fat and salt intake. Substitution for these ingredients is necessary to restore removed flavor and regulate flavor release. However, understanding of flavor release is made difficult by the complexity of the interactions between foods and consumers. Each volatile compound has different physiochemical properties and its release is influenced by interactions with other food matrix variables such as moisture, fat, protein, carbohydrate, and other soluble (salt, sugars) and non-soluble materials. In addition, food breakdown and mixing with saliva during consumption, respiratory air flow over and around the food and temperature and pH changes occurring in the consumers mouth will influence volatile release and subsequent flavor perception. It is also known that individual consumers expression of flavor differs as a result of physiological, psychological and social differences [1,2]. Therefore an underlying question which remains unanswered is; what part of flavor differences between foods result from volatile release dynamics from the food matrix, and what part result from differences between consumers? Conclusions reached in response to this question have been mixed.
118 There are model systems which measure volatile compounds released while mimicking conditions in the mouth [3-5]. Other methods measure volatile release directly during consumption using mass spectrometry of breath [6,7] and indirectly by trapping volatiles on adsorbents, such as Tenax, before analysis [8-10]. Soeting and Heidema [6] showed thirtyfold differences in the relative quantities of 2-pentanone which was measured directly from the breath of different consumers. Van Ruth et al [10] also found subject specific volatile profiles were released during consumption of vegetables. Taylor et al [11] trapped volatiles released from mint sweets during consumption and also found differences between subjects in terms of the quantities of volatiles released. However, they concluded that there were similarities between the relative concentrations of volatiles released for each subject. Delahunty et al [12], who analyzed Buccal Headspace Analysis [BHA;13] data using Principal Components Analysis [PCA;14] to examine the volatile profiles released during consumption of cheeses by three different consumers, found product specific volatile release was most important. However, three consumers were too few to draw any firm conclusions. Workers studying food texture have developed methods such as electromyography [15] which measure muscle activity during mastication of a food matrix. From these measurements they have shown mastication patterns and can calculate the amount of work done by a consumer during consumption. These methods, which show considerable differences between consumers' mastication characteristics, have recently been related to differences between consumers' temporal perception of flavor intensity measured by timeintensity sensory analysis [16]. Other physiological parameters, such as the influence of saliva [4] and air flow through the mouth [17] have also been investigated. The present study was carried out to investigate discrepancies in the literature relating to the differences between consumers' interactions with foods and the relationships found between physiological measures during chewing and individuals' differences in flavor perception. In order to achieve this, similar varieties of a complex food were chosen and a multivariate technique, PCA, was used to examine the volatile profiles released.
2. EXPERIMENTAL 2.1. Samples and consumers Six Cheddar cheeses, in 5kg blocks, of equal age (6-8 months) were obtained from 4 different producers. Eight consumers, 3 female and 5 male, aged between 22 and 28 were used for all studies. 2.2. Buccal Headspace Analysis Buccal Headspace Analysis of each cheese was carried out for each consumer in triplicate. For this method a 50 g cheese sample was consumed in 10 x 5g pieces in a normal way, allowing 30 s for the consumption of each piece. During the entire consumption time (5 min) volatile compounds released were displaced through the nose by vacuum and trapped on a Tenax-TA trap. The order of sample analysis was balanced for consumers, cheeses and day of consumption [18]. A blank buccal headspace sample was taken each day for each consumer. Traps were thermally desorbed using a Teckmar Purge and Trap 3000 concentrator (Teckmar, Cincinnati, OH, USA). Desorbed volatiles were identified and quantified using gas chromatography-mass spectrometry (GC-MS) with a Varian Saturn GC-3400CX
119 incorporating a Varian Saturn 3 GC/MS detector (Varian chromatography systems, Mitchell drive, Walnut Creek, CA, USA).The column was a DB-5ms, 30m x 0.257mm fused silica capillary column, with a film thickness of 0.25 |Lim (J & W scientific, Folsom, CA, USA) 2.3 Mastication behaviour The activity of the consumers left and right masseter muscles during chewing was recorded by Electromyography [15]. The electromyograph record was measured for 1 cheese over a period of 5 min (10 x 30 s for each 5g piece ), in triplicate, for each of the eight consumers. Each individuals electromyogram was integrated using a poly VIEW data acquisition and analysis system (Grass instrument division, Astro-Med Inc., East Greenwich Avenue, West Warwick, UK) 2.4 Saliva production Consumers unstimulated saliva production was measured by allowing their saliva to drip into a beaker for a 5 min period. The consumers swallowed immediately before collection and forcefully spat out at the end [19]. The stimulated saliva production was measured by dividing the volume of saliva produced by each consumer in response to 50g of cheese (10 x 5g) by the chewing time required by the consumer for that cheese [19]. Each measurement was repeated four times. 2.5 Data Analysis Buccal headspace data was analyzed by PC A, using the Unscrambler v 6.0 (CAMO AS, N-7041 Trondheim, Norway), of the log transformed peak areas of volatile compounds. Electromyography data was analyzed by Analysis of Variance (ANOVA) using SPSS v 6.1 (SPSS Inc. Chicago, IL 60611, USA) of the totals for chew number, chew time, chew rate and chew work. Saliva production data was analyzed by ANOVA of the unstimulated and stimulated saliva flow rates. Differences between cheeses and between subjects were investigated using ANOVA. Relationships between data sets were investigated by linear and Partial Least Squares regression [PLS;20], using the Unscrambler v 6.0.
3. RESULTS AND DISCUSSION In the present study the quantities and balance of volatile compounds released during consumption of a food, by different consumers, was compared. For this purpose Cheddar cheese was chosen as this represents a complex protein matrix containing fat and moisture. To minimize product related compositional differences, and therefore to maximize the influence of consumer related differences to volatile release from one food type, cheeses of equal age were chosen. Eighteen volatile compounds were selected from chromatograms of BHA of all cheeses and the amounts of each present were quantified. Both Figures 1 and 2 depict two PCA's. The first (in italics) was calculated using individual consumers' headspace data (triplicates averaged) and the second using the average of the 8 consumers.
120 Figure 1. PC A scores on PC's 1 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H) (see text for explanation).The pooled SD for the analysis is represented by an ellipse on cheese 1. 4 T
3B 4C
13H ^^
c o a.
4G
-4
5C
"^ fr ^^ Ic ,2iP2f 4H 2F 3G
5F
6E IG
-im
4A
IB
2C
2 6D
OH
5 6G
-4 -^
Principal component 1 {?>9Vo)(24%) Principal Components (PC) 1 and 3, which accounted for 39% and 15% of the explained variance, respectively, of the PC A of compound peak areas, showed significant differences {p = 0.017 and p = 0.021) between cheeses (Table 1 and Figure 1). Differences between consumers (p = 0.050), which accounted for \3% of the explained experimental variance, were found on PC3 (Table 1 and Figure 2). Table 1 ANOVA between cheeses (1 - 6 ) and between consumers ( subjects A • H)on Principal Components 1 - 4 of the PC A Principal component Cheese Subject PCI PC2 PC3 PC4
0.017 0.657 0.021 0.055
0.873 0.072 0.050 0.351
The volatile compounds which distinguished the cheeses from one another on these components are shown in the PC loadings plots (Figures 3 and 4).The differences found between cheeses on PCI, which was the most important as it contained the highest proportion of the experimental variance (39%), were caused mostly by the quantities of compounds released during consumption by each consumer rather than by their balance.
121 Figure 3. PC A loadings for 18 volatile compounds on PC's 1 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H). 0.5
T 2-heptanone
cyclohexanel cyclohexane2 cpdl3 cpdl4
o OH
-0.5
cpd3 cpd5
toluene
cpd2
dmds
cpdl7
0.5
heptane
(Eodecane
ethyl butyrate cpdl2
cpdl6 -0.5 -^
Principal component 1 (39%) However, differences between cheeses found on PC's 2 and 3 were caused mostly by differences in the balance of the compounds released. This can be determined from the relative positions of the volatile compounds in the loadings plots (Figure 4). In a previous study physiological differences between consumers have been related to differences in flavor perception [16]. In this study very significant differences were found between consumers mastication characteristics and also between their saliva production rates (p = 0.000 for all parameters apart from chew rate (p = 0.021) and chew work (p = 0.045)) (Table 2 and Figure 5). Using linear regression chew number and chew work were found to relate to saliva production rate during cheese consumption for 5 of the 8 subjects (r = 0.98 and r = 0.84, respectively). However by using PLS and linear regression, no significant relationships were found between the measured physiological characteristics and total volatile release. Sensory evaluation of the cheeses is not reported in this study and therefore no conclusions can be made with regard to consumers' expressions of flavor perception. Further work is also required to investigate the dynamics of volatile release during time of consumption.
4. CONCLUSION Some differences were found between the quantities of volatile compounds released during cheese consumption by different consumers. Very significant differences were found between consumers mastication characteristics and between saliva production rates during cheese consumption. Despite these differences, the distribution of experimental variance explained
122 Table 2 Mastication behaviour and saliva production rates of 8 consumers ( subjects A-H) Subject TCN
A B C D E F G H pooled SD P
224.33 138.67 348.00 324.33 228.67 215.67 184.33 282.33 22.84 0.000
Electromyography^ TCT CR
141.26 169.07 212.27 198.53 144.00 212.07 145.71 180.99 28.35 0.000
1.59 1.08 1.67 1.64 1.68 1.02 1.27 1.55 0.12 0.021
CW
2435.86 1011.45 3823.63 4497.6 2066.55 1194.73 866.71 3010.14 1403.50 0.045
Saliva flow rate'' Unstim. Stimulated
0.77 0.43 0.59 1.06 0.53 0.70 0.76 0.35 0.12 0.000
5.30 4.06 1.69 7.31 4.52 4.89 3.25 3.48 0.65 0.000
^ TCN = total chew number; TCT = total chew time (sec); CR = chew rate (chew / sec); CW = chew work b Unstim. = unstimulated saliva production rate (mL / min); Stimulated saliva production rate (mL / min)
Figure 5. Mastication behaviour of 3 consumers during consumption of one 5 g piece of cheese. CN = chew number; CT = chew time (sec); CR = chew rate (chew / sec); CW = chew work. Pat (A) CN = 23 CT= 14.66 CR=1.57 C W = 133.03
a
Chew Time (sec)
123 Figure 3. PCA scores on PC's 2 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H). The pooled SD for the analysis is represented by an ellipse on cheese 6. 4 T
4C 2A IG 6C
C
ID 3D
2B 2F
o
r^
4A 2C
OH
5E 6F 5D 6Bry
ex
ifP
5G
3F6D
,^
-4 -^
Principal component 2 (\3%)(]2%) Figure 4. PCA loadings of 18 volatile compounds on PC's 2 and 3 for 6 cheeses assessed by BHA using 8 subjects (A-H). 0.5 T 2-heptanone
cyclohexanel cycldhexane2 pctane cdd7 cpdl2 cpdl4
o a.
|cpd3
cpf
-0.5 dmds
dodecane
heptane
cpd2
0.5 toluene
cpdl7 ethylbutyrate cpdl2 -0.5 -^
cpdl6
Principal component 2 (13%)
124 by the PCA showed that Cheddar cheese of equal age could be identified by their product specific volatile release. Therefore, the volatile profile for a particular cheese at the end of consumption was found to be similar in all consumers.
5. ACKNOWLEDGEMENT This work was part funded by the Department of Agriculture, Food and Forestry, Ireland, under the Food Industry Sub-Programme of EU Structural Funds.
6. REFERENCES 1 D. Lancet, In: Sensory Transduction (D.P. Corey and S.D. Roper, eds.). Pp. 73, Rockefeller University, New York, 1992. 2 J.R. Piggott, Fd. Qual. Pref, 5 (1994) 167. 3 W.E. Lee III, J. Fd. Sci., 51 (1986) 249. 4 D.D. Roberts and T.E. Acree, J. Agric. Fd. Chem., 43 (1995) 2179. 5 K. Napi, F. Kropf and H. Klostermeyer, Z Lebensm Unters Forsch, 201 (1995) 62. 6 W.J. Soeting and J. Heidema, Chem. Senses, 13:4 (1988) 607. 7 R.S.T. Linforth, K.E. Ingham and A.J.Taylor, In: Flavour Science: Recent Developments (A.J. Taylor and D.S. Mottram, eds.). Pp. 361, Royal Society of Chemistry, Oxford, 1997. 8 R.S.T. Linforth and A.J.Taylor, Fd. Chem., 48 (1993) 115. 9 CM. Delahunty, J.R. Piggott, J.M. Conner and A. Paterson, In: Trends in Flavour Research (H. Maarse and D.G. van der Heij, eds.). Pp. 47, Elsevier Applied Science, Amsterdam, 1994. 10 S.M. Van Ruth, J.P. Roozen and J.L. Cozijnsen, Fd. Chem., 53 (1995) 15. 11 A.J. Taylor, R.S.T. Linforth, K.E. Ingham and A.R. Clawson, In: Bioflavour '95 (P. Etievant and P. Schreier, eds.). Pp. 45, INRA, Paris, 1995. 12 CM. Delahunty, F. Crowe and P.A. Morrissey, In: Flavour Science: Recent Developments (A.J. Taylor and D.S. Mottram, eds.). Pp. 339, Royal Society of Chemistry, Oxford, 1997. 13 CM. Delahunty, J.R. Piggott, J.M. Conner and A. Paterson, J. Sci. Fd. Agric, 71 (1996) 273. 14 J.R. Piggott and K. Sharman. In: Statistical Procedures in Food Research (J.R. Piggott, ed.) Pp. 181, Elsevier Applied Science, London, 1986. 15 M.M. Boyar and D. Kilcast, J. Fd. Sci., 51 (1986) 859. 16 W.E. Brown, C Dauchel and I. Wakeling, J. Texture Stud., 27 (1996) 433. 17 M. Harrison and B.P. Hills, Int. J. Fd. Sci. Tech., 32 (1997) 1. 18 H.J.H. MacFie, N. Bratchell, K. Greenhoff and I.V. ValHs, 1989. J. Sens. Stud., 4 (1989) 129. 19 S. Watanabe and C Dawes, Arch. Oral Biol., 33:1 (1988), 1. 20 M. Martens and H. Martens, In: Statistical Procedures in Food Research (J.R. Piggott, ed.) Pp. 293, Elsevier Applied Science, London, 1986.
