- Email: [email protected]
Inter-individual retronasal aroma release variability during cheese consumption: Role of food oral processing
Food Research International 64 (2014) 692–700
Contents lists available at ScienceDirect
Food Research International journal homepage: www.elsevier.com/locate/foodres
Inter-individual retronasal aroma release variability during cheese consumption: Role of food oral processing H. Labouré a,b,c,⁎, M. Repoux a,b,c, P. Courcoux d, G. Feron a,b,c, E. Guichard a,b,c a
INRA, UMR1324 Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France CNRS, UMR6265 Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France Université de Bourgogne, UMR Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France d LUNAM University, ONIRIS, USC “Sensometrics and Chemometrics Laboratory”, F-44322 Nantes, France b c
a r t i c l e
i n f o
Article history: Received 30 April 2014 Accepted 20 July 2014 Available online 14 August 2014 Keywords: In vivo aroma release Inter individual variability Food oral processing Model cheeses Chewing activity Mouth coating
a b s t r a c t The aim of our study was to explain inter-individual differences on in vivo aroma release during cheese consumption by oral physiological parameters. To reach this objective, 34 subjects were recruited. Their salivary flow, oral volume and velum opening were determined. Six cheddar-based melted cheeses with different fat levels and firmness were flavoured with nonan-2one (NO) and ethyl propanoate (EP). During their consumption (free protocol), in vivo retro nasal aroma release was followed by Atmospheric Pressure Chemical Ionisation–Mass Spectrometry (APCI–MS). Chewing activity was evaluated by electromyography recordings. Bolus saliva content, mouth-coating, and bolus rheology were also determined. Based on the quantity of aroma released before and after swallowing, subjects can be clustered into three groups: the first one (HRG) is characterized by a large quantity of aroma release whatever the aroma compound; the second one (MRG) showed a large release for EP and a lower one for NO; the third group (LRG) was characterized by a low quantity of aroma release whatever the compound. Whatever the group of subjects, fat and firmness effects differed according to the aroma compound. EP release increased with firmness and fat content, whereas NO release was not affected by firmness and decreased when fat content increased. Physiological parameters which better differentiated the three groups of subjects according to their release behaviour were chewing activity, mouth coating and frequency of velum opening. Subjects from HRG were differentiated from LRG subjects by a higher chewing activity, and more frequent velum opening. Subjects from MRG presented a lower mouth coating explaining their lower release of NO, the more hydrophobic compound. This study shows that the total amount of aroma released in the nasal cavity during food consumption depends not only on the characteristics of the product but also on the oral physiology of the subjects and on their food oral processing. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Retro-nasal perception occurs in humans while eating and is one of the key factors of food acceptability by consumers. During food consumption, aroma compounds responsible for aroma perception have to reach the olfactory receptors after being transferred from food to saliva then to the oro-nasal cavity. Both the composition and the structure of food have been shown to highly impact the release of aroma compounds (Linforth, Baek, & Taylor, 1999; Gierczynski, Laboure, & Guichard, 2008). According to their physico-chemical properties, aroma compounds are more or less released from the food matrix, ⁎ Corresponding author at: INRA, UMR1324 Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France. Tel.: +33 380 69 35 28; fax: +33 380 69 32 27. E-mail address: [email protected] (H. Labouré).
http://dx.doi.org/10.1016/j.foodres.2014.07.024 0963-9969/© 2014 Elsevier Ltd. All rights reserved.
due to molecular interactions with the macromolecules (proteins, polyosides) or to different solubility in water and oil (Guichard, 2002). Additionally great inter-individual variability on in vivo retro nasal aroma release during food consumption has been underlined (Gierczynski et al., 2008; Mestres, Kieffer, & Buettner, 2006; Pionnier, Chabanet, Mioche, Le Quere, & Salles, 2004). The different causes of this variability have not been extensively investigated yet due to the complexity of in-mouth mechanisms (Salles et al., 2011). Several authors have already observed relations between some physiological parameters and aroma release. On liquid samples, Buettner et al. (Buettner, Beer, Hannig, Settles, & Schieberle, 2002) found that velum opening, swallowing and adsorption to oral mucosa are able to modulate retro nasal aroma release. During mint tablet consumption, the inmouth air cavity (IMAC) volume changes after swallowing were also found to impact the total amount of aroma release (Mishellany-
H. Labouré et al. / Food Research International 64 (2014) 692–700
Dutour et al., 2012), but this parameter was not evaluated during solid food consumption. Using in vitro systems simulating eating conditions, the dilution effect of saliva has been also evidenced (van Ruth & Roozen, 2000). However the real impact of saliva on aroma release during real food consumption has not been elucidated yet due to a combined effect with that of the product (Mestres et al., 2006). A mathematical model for in vivo aroma release has been designed to take account of the product saliva mixing and the post-deglutive pharyngeal residue, showing the important impact on saliva dilution effect on aroma release (Doyennette, de Loubens, Deleris, Souchon, & Trelea, 2011). This model was validated with a hydrophilic compound, diacetyl, released from glucose syrup. An extension of this mechanistic model was further developed considering the role of the mastication process which showed that, among other parameters, the rate of saliva incorporated into the bolus greatly influenced aroma release (Doyennette et al., 2014). In the case of solid foods, a lot of literature is available evidencing that in-mouth process parameters related to chewing behaviours such as chewing work, chewing time, and burst numbers are highly positively correlated with in vivo aroma release (Haahr et al., 2004; Hansson, Giannouli, & Van Ruth, 2003; Mioche, Bourdiol, Monier, & Martin, 2002). Inter-individual variations of these parameters could explain inter-individual differences in aroma release (Blissett, Hort, & Taylor, 2006; Tarrega, Yven, Sémon, & Salles, 2008). Specific consumption patterns of subjects have been underlined as their consequences on aroma release profiles (Mestres, Moran, Jordan, & Buettner, 2005; Mestres et al., 2006). These authors showed a specific aroma release profile when subjects consume a soft gel by pressing with the tongue in the frontal part of the oral cavity against the hard palate and another one when the same panellists were instructed to chew the same product with jaw opening. This inter-individual specificity in aroma release profiles has been confirmed with subjects eating either liquid or solid foods (Ruijschop, Burgering, Jacobs, & Boelrijk, 2009). This study showed that subjects with relatively high retro-nasal aroma release intensity for liquid food products have also a relatively high release for a solid food product. However no mention was made on the nature of the aroma compound and its consequences on aroma release patterns, even if other studies mentioned different release patterns for aroma compounds differing in their hydrophobicity (Doyennette, de Loubens, Deleris, Souchon, & Trelea, 2011; Doyennette et al., 2011; Repoux et al., 2012). Moreover, most of the different studies were conducted on a limited number of subjects and only some specific oral physiological parameters were measured. In this context, the aim of our study was to investigate interindividual differences on in vivo aroma release from solid foods, namely model cheeses and to identify oral physiological variables which could explain this inter-individual variability. We mainly focused on chewing activity, bolus rheology, saliva flow and oral physiology. We choose to analyse the release of two aroma compounds differing in their hydrophobicity from model cheese varying in firmness and fat content. In a first step and in order to discriminate the subjects in function of the quantity of aroma realised, a hierarchical cluster analysis has been performed separately for each aroma compound because of their very different level of release; then the two classifications obtained were compared. In a second step, the effect of food matrix composition on the aroma release has been studied in each group of subjects and compared. Finally, the oral physiology and food oral processing parameters have been compared between the groups in order to explain the differences in release patterns.
693
2.1. Food products Six cheddar-based melted cheeses (model cheeses) were developed (Repoux, Labouré, et al., 2012). They differed by their firmness and their fat content (Fig. 1). Three firmness levels, measured by a compression test, were obtained: a very firm one (FF: higher than 10 000 N), a firm one (F between 6000 and 8000 N) and a soft one (S between 400 and 3000 N). Two levels of fat content have been studied, a low one (l: 25%) and a high one (h: 50%). These model cheeses were flavoured with two aroma compounds: nonan-2-one (NO: logP = 2.9; vp = 83.2 Pa at 25 °C) and ethyl propanoate (EP: logP = 1.4; vp = 4.79.103 Pa at 25 °C). They were added during cheese production at levels adjusted to achieve final concentrations in the cheeses around 6 ppm for nonan-2-one and 25 ppm for ethyl propanoate. 2.2. Subjects Thirty four naïve subjects (15 females and 19 males aged between 22 and 59 years; mean age: 39 years) participated in this study. Selection criteria were availability for the duration of the study, motivation, no medicine consumption, and a good dental status (no denture wearers and few artificial crowns, dental bridges and dental implants). The subjects were not allowed to smoke, eat or drink during one hour before the test session. All subjects were informed of the observational nature of this study, gave their signed consent and received compensatory indemnities for their participation in two sessions, each lasting about 2 h. 2.3. In vivo aroma release 2.3.1. Materials In vivo aroma release was measured using Atmospheric Pressure Chemical Ionisation–Mass Spectrometry APCI–MS (Gierczynski et al., 2008). Air from the nose was sampled from one nostril at an average flow rate of 37 mL·min−1 and introduced into the APCI source of an ion trap Esquire-LC mass spectrometer (Bruker Daltonique, Wissembourg, France) via a fused silica capillary tubing (i.d. = 0.53 mm) heated at 150 °C and to which a 5 kV positive ion corona pin discharge was applied (Le Quere, Gierczynski, Langlois, & Semon, 2006). For the oral movement measurements during mint consumption, compounds were monitored according to their protonated molecular ion (MH+): acetone (m/z = 59) and menthone (m/z = 155). For the experiments during cheese consumption, three compounds were monitored according to their protonated molecular ion (MH+): ethyl propanoate (m/z = 103) and nonan-2-one (m/z = 143), the two aroma compounds added to
2. Material and methods The study protocol was submitted to an Ethics Committee and was approved on 17 April 2008 by the Comité de Protection des Personnes Est-1 (N° 2008/15) and on 8 August 2008 by the Direction Générale de la Santé - France (N° DGS2008-0196).
Fig. 1. Rheological properties of the cheese products differing in firmness (S for soft, F for firm and FF for very firm) and fat level (l for low fat and h for high fat).
694
H. Labouré et al. / Food Research International 64 (2014) 692–700
the model cheeses, and acetone (m/z = 59), which was used as an indicator of the subject's breath (Hodgson, Linforth, & Taylor, 2003). 2.3.2. Methods Each subject was asked to position the plastic tube in one nostril and to breathe normally. After an initial swallow, the subject was instructed to place the piece of cheese ([2.25 ∗ 2.25 ∗ 1.1] cm; m = 6 g) in the mouth, and to consume it freely while keeping the lips closed. The times of swallowing were recorded using a sensor placed on the subject's throat. The products were presented in random order at 17 °C. Bread, apple and water were used as mouth cleansers between two tests. Three replicates per cheese and per subject were performed. 2.3.3. Data analysis After smoothing the curves to eliminate signal fluctuations due to the subject's breathing pattern, two release phases were identified. The mastication step (m) extended from placing the cheese in the mouth to the first swallowing, and the post-swallowing step (ps) extended from the first swallowing to the time at which the signal returned to its baseline level. The area under the curve was calculated for each aroma compound during each release phase. 2.4. Physiological parameters 2.4.1. Salivary flow (at rest and stimulated) Resting saliva flow (rSF) was determined by using the method used by Gaviao, Engelen, & van der Bilt (2004). Non-stimulated saliva flow was collected by instructing the subject to spit out the saliva every 30 s into a pre-weighed cup over a period of 5 min. Two measurements were performed by subjects, one per session. For stimulated saliva flow, the subjects were instructed to chew a piece of Parafilm™ (0.5 g ± 0.2 g) for a period of 1 min then spit out the saliva after 30 s, continue chewing for a further 30 s and then spit out both the saliva and the piece of Parafilm ™ (Repoux, Sémon, Feron, Guichard, & Labouré, 2012). Three stimulated saliva flow rate (sSF) replicates were performed during one session. 2.4.2. Oral volume and IMAC volume changes An Eccovision® acoustic pharyngometer (Hood Laboratories, USA) was used to measure oral volume and was composed of four components: a wave tube, an electronic platform, a mouthpiece and a disposable filter. Reflectance pharyngometry was performed with a twomicrophone imaging acoustic pharyngometer device, as described recently (Mishellany-Dutour et al., 2012; Poette et al. 2013). This device consists of two microphones and one horn driver mounted on a wave tube and connected to a PC-compatible computer with signal conversion capabilities. The signal was converted into the surface change (cm2) as a function of the length of the oral cavity (in cm). The subjects held the mouthpiece in their mouth with their teeth against the flange and their tongue in a low position. To prevent air leaks, which could cause measurement errors, the subjects placed their lips over the flange, sealing the mouthpiece. IMAC volume changes after swallowing were measured following the procedure described previously (MishellanyDutour et al., 2012). 2.4.3. Velum opening The influence of oral movements on velum opening was experimented on subjects while consuming a commercial mint tablet under an imposed protocol (Repoux, Sémon, Feron, Guichard, & Labouré, 2012). The subjects placed the tablet under the tongue. Then they were given a series of instructions in the following order: (1) to swallow (S1), (2) to make tongue movements (TM) without swallowing and mouth opening, (3) to swallow (S3), (4) to make jaw movements (JM) and to swallow (S3). Menthone release during mint tablet consumption was followed by APCI–MS. This measurement was analysed qualitatively and quantitatively. For the qualitative analysis, an event was coded ‘1’ if the instruction induced menthone release
and ‘0’ if the instruction did not induce any release t. For the quantitative analysis we calculated the area under the curve corresponding to menthone release, at each event, then the sum of the areas for all the events and the sum of the area for the swallowing events only. 2.5. Food oral process measurements 2.5.1. Chewing activity Chewing activity was monitored during cheese consumption, simultaneously to aroma release. The muscle activity of the superficial masseter and temporis muscles (left and right) during chewing was recorded by electromyography (EMG) using gold surface electrodes (Grass technologies, West Warwick, RI, U.S.A.), at 382 Hz, then the signal was amplified and digitalized (Mioche et al., 2002). From the EMG data the following parameters were collected: the number of chewing cycles, the duration of chewing cycle (expressed in s) and the maximal amplitude (expressed in mV) which correspond to the maximal value of amplitudes of the different chewing cycles registered in a whole chewing sequence and total muscle work (expressed in mV s) which is the sum of muscle work per chewing cycle, that is the surface area of the muscle activity signal. 2.5.2. Bolus saliva content The percentage of dry matter and water content were determined using an infrared dryer for all the cheeses and boluses obtained just before swallowing. For each subject and each cheese, the percentage of saliva incorporated (Sal) into the bolus was calculated as follows: Bolus water content ð%Þ Saliva incorporated ¼ Cheese dry matter ð%Þ Bolus dry matter ð%Þ ð% into the bolusÞ −Cheese water content ð%Þ:
2.5.3. Mouth coating Mouth coating (MCo) defined as the residual food that sticks to the oral surface after food ingestion, was quantified by the “mouth rinse” method (Pivk, Ulrih, Juillerat, & Raspor, 2008). The lipids of the residual food were quantified by the intensity of curcumin fluorescence in the rinse water as previously described (Repoux, Labouré, et al., 2012; Repoux et al., 2012). Briefly, each subject was asked to place a piece of cheese (6 g at 17 °C) containing curcumin (30 ppm in the final cheeses) in the mouth and to masticate normally until they needed to swallow. He/she swallowed without any cleaning movement and then rinsed his/her mouth twice with 4 ml of water at a temperature of 50 °C for 30 s and spit it into a vial. The fluorescence intensity of the curcumin in the two cumulated spittle was quantified using a Perkin Elmer 1420 Multilabel Counter Victor 3 V (Perlin Elmer, Courtaboeuf, France) at an excitation wavelength of 450 nm and an emission wavelength of 510 nm. The quantification of the coating was evaluated in reference to calibration curves performed for each cheese. 2.6. Rheological properties of food bolus Subjects were asked to chew the cheese samples until they were ready to swallow and to spit out the bolus into a truncated syringe. An aliquot of 3 mL of bolus at the syringe bottom was used for compression test to get a constant volume regardless of the cheese employed. The rheological properties of the bolus were measured by a compression test described elsewhere (Yven et al., 2012). From the compression curve, three phases were distinguished after the loading phase. During the first phase named “aggregated phase” the bolus behaves like a solid characterized by a failure threshold Sagg (Pa). During the second phase (“flow phase”), the suspension begins to flow. From a mechanical point of view, the viscous effects are no longer negligible in relation to the yield stress effects described by Sflow
H. Labouré et al. / Food Research International 64 (2014) 692–700
(Pa). During the third one (“particle phase”), the mechanical response is governed by the particles. At the end of compression, hend (mm) denotes the final height and Send (mm2) denotes the area generated under the maximal force (50 N). The hend/Send ratio (mm−1) stands for spreadability of the bolus; the lesser is the ratio, the greater is the spreadability. 2.7. Data analysis 2.7.1. Subjects classification In a first step, the 34 subjects were clustered based on the area under the curve obtained during the mastication phase and the postswallowing phase, for each of the 6 food products. A hierarchical clustering of subjects using the Euclidean distance and Ward's criterion was performed independently for each of the two aroma compounds because of their very different levels of release. 2.7.2. Characterization of subjects groups Three way ANOVAs with interactions (group, fat, firmness, group ∗ fat, group ∗ firmness, fat ∗ firmness) were performed to evaluate differences between release groups on the quantity of aroma released, the food oral processing parameters (mouth coating, saliva incorporated in the bolus, number of chewing cycle, maximal amplitude and total chewing work), the oral physiological parameters (oral volume, salivary flow at rest and stimulated) and the rheological characteristics of the bolus. Factors studied were considered significant when their p value was lower than 0.05. A Principal Component Analysis (PCA) based on the correlations was performed on the mean quantity of aroma release per subjects and per consumption phase in order to compare the different groups of subjects for each aroma compound. Statistical analyses were performed using STATISTICA® software, version 10 (StatSoft, France). 3. Results 3.1. Subjects classification 3.1.1. Ethyl propanoate Two groups of subjects (A and B) composed of 15 and 19 subjects respectively were identified by the hierarchical cluster analysis based on quantity of EP release during mastication and post-swallowing phase. Subjects from group A release significantly more EP than subjects from group B. Subjects from group A constitute a high release group, and subjects from group B a low release group for EP. The group effect is slightly higher in post-swallowing phase (Fgroup(1,5) = 33.8; p b 0.01) than in mastication phase (Fgroup(1,5) = 25.7; p b 0.01). 3.1.2. Nonan-2-one Two groups of subjects (C and D) composed of 7 and 27 subjects respectively were identified by the hierarchical cluster analysis based on quantity of NO release during mastication and post-swallowing phase. In each phase, subjects from group C release significantly more NO than subjects from group D. Subjects from group C constitute a high release group, and subjects from group D a low release group for NO. Compared to the groups based on EP release, the group effect is significantly higher in post-swallowing phase (Fgroup(1,5) = 71.4; p b 0.001) than in mastication phase (Fgroup(1,5) = 11.3; p b 0.05). This observation could be explained by the greater hydrophobicity and lower volatility of NO compared to EP and thus by the low quantity of NO released during the mastication phase. 3.1.3. Comparison of the HCA Looking at the subjects who belong to the different groups (A or B for EP and C or D for NO), we noticed that all the 7 subjects from group C (high release of NO) also belong to group A (high release of EP) and
695
that all the 19 subjects who belong to group B (low release of EP) also belong to group D (low release of NO). Only 8 subjects did not present the same level of release for the 2 aroma compounds, i.e. they release EP at a high level (they belong to group A) and NO at a low level (they belong to group D). This allowed us to identify three different groups of subjects: a high release group (HRG) composed of 7 subjects who release a high amount of the two aroma compounds, a mix release group (MRG) composed of 8 subjects who release a high amount of EP and a low amount of NO, and a low release group (LRG) composed of 19 subjects who release a low amount of the two aroma compounds. A principal component analysis, performed on the quantity of aroma compounds release during mastication and post-swallowing phase illustrates the three groups of subjects and their distribution as a function of the aroma compound (Fig. 2). For each aroma compound, the different release groups are mainly discriminated according to the first principal component, which represents most of the information (50.65% for EP and 56.36% for NO). For EP, HRG and MRG are on the negative part and LRG on the positive part of this first axis, whereas for NO, HRG is on the negative part and MRG and LRG on the positive part of this axis. This first axis is correlated to the quantity of aroma release during the post-swallowing phase whatever the cheese. The second axis explains only 18.5 (EP) and 16.3% (NO) of the variance. It is mainly defined by the quantity of aroma release during the mastication phase. The three groups of subjects were not well discriminated along this axis. 3.2. Influence of the effects related to the products on aroma release for each group of subjects In this part only the quantity of aroma released during the postswallowing phase will be discussed, firstly because the three groups of subjects are well discriminated during this post-swallowing phase of aroma release and not during the mastication phase; secondly because aroma compounds are mainly released in this phase, and thirdly because the quantity of aroma released in this post-swallowing phase is highly correlated with the total amount of aroma release (Repoux, Sémon, Feron, Guichard, & Labouré, 2012). 3.2.1. EP release Firmness effect is significant for the whole group and within each group (Fig. 3). For each group of subjects, the total amount of EP released after swallowing significantly increased with both the firmness (FFirmness(2,190) = 51.9; p b 0.0001) (Fig. 3a) and the fat content (F fat (1, 190) = 102.8; p b 0.00001) (Fig. 3b) of the cheeses. The firmness ∗ group interaction is also significant (FFirmness ∗ group(4, 190) = 7.2; p b 0.0001) because firmness effect is higher for HRG and MRG than for LRG. For HRG and MRG, the three levels of firmness are indeed significantly different at the 5% level whereas only the release from the softer cheese S is significantly lower than from the firmer cheeses (F and FF) for LRG. Similarly, fat effect is significant for the whole group and within each group. The fat ∗ group interaction is also significant (FFat ∗ group(2, 190) = 5.0; p b 0.001) because fat effect is greater in HRG than in LRG and MRG (Fig. 3b). 3.2.2. NO release NO release increases significantly with the firmness for the whole group (FFirmness(2,190) = 3.7; p b 0.05) but firmness effect is not significant within each group except for LRG (Fig. 3c). In LRG, as observed for EP, the release from the softer cheese S is significantly lower than from the firmer cheeses (F and FF). Contrary to EP, NO release decreased when fat content increased because of its hydrophobicity (Fig. 3c). The fat effect is higher than the firmness effect (Ffat(1, 190) = 54; p b 0.00001). Fat effect is significant and in the same direction for all groups but the interaction group ∗ fat is also significant (Ffat ∗ group(2,190) = 13.0; p b 0.0001) because the size of the effect is higher for HRG and MRG than for LRG (Fig. 3d).
696
H. Labouré et al. / Food Research International 64 (2014) 692–700
Fig. 2. PCA on the amount of aroma release represented by the area under the curve (AUC) in the mastication phase (m) and in the post-swallowing phase (ps) for each subject consuming each product (S = soft, F = firm, h = high fat, l = low fat). Subjects are represented by their release group. Upper figures for ethyl propanoate (EP), under figures for nonan-2-one (NO). Left: representation of the variables (quantity of aroma release for each cheese in the mastication phase and in the post-swallowing phase), right: representation of the means of all the subjects for each group (only the name of the group is reported: HRG: high release group, MRG: mix release group, LRG: low release group).
3.3. Influence of the subject physiological characteristics on aroma release for each group of subjects 3.3.1. Physiological parameters No significant difference was found for the saliva flows (at rest and stimulated) and IMAC volume changes between the groups of subjects. The only physiological parameter which presents significant differences between the groups of subjects is the velum opening. Subjects
were characterized for their velum opening by following the release of menthone during mint consumption. The release patterns are significantly different according to the groups. Even if all the subjects released menthone at the swallowing events, the area under the curve during these events are significantly different according to the groups (Fgroup(2,31) = 5.52; p b 0.001). Subjects from HRG released more menthone than subjects from the two others groups. The total amount of menthone released (during swallowing events but also at tongue
Fig. 3. Firmness effect and fat effect on EP (on the left) and NO (on the right) release per group. Vertical lines represent the confidence interval at 0.95.
H. Labouré et al. / Food Research International 64 (2014) 692–700
Fig. 4. Influence of group, fat and firmness on food oral processing parameters. Results of three way ANOVAs and 2-way interactions for food oral processing parameters.
697
698
H. Labouré et al. / Food Research International 64 (2014) 692–700
and jaw movements) has been calculated and is also significantly different between the groups (Fgroup(2,31) = 4.38; p b 0.05). Subjects from HRG release significantly more menthone than subjects from LRG but subjects from MRG present an intermediate behaviour not significantly different from that of the two others. Most of the subjects of HRG and MRG (86% and 88% respectively) release menthone at all the events whereas most of the subjects from LRG release only during the swallowing events.
3.3.2. Food oral processing parameters 3.3.2.1. Chewing activity. Globally, all the chewing parameters studied present significant differences between the release groups (Fig. 4). Subjects with an intermediate level of aroma release (MRG) present the lowest values for all of these parameters. As already reported (Repoux, Labouré et al., 2012), chewing activity varies in function of both fat content and firmness, with a greater chewing activity for low fat and very firm cheeses.
3.3.2.2. Bolus saliva content. The percentage of saliva incorporated into the bolus (bolus saliva content) does not vary significantly between the different release groups (Fig. 4). This parameter is only influenced by the cheese firmness, more saliva being incorporated into the bolus for the firmer cheeses. But even for these firmer cheeses, the group factor is not significantly different.
3.3.2.3. Mouth coating. The residual food that sticks to the oral surface (Mouth coating) varies significantly between the groups of subjects, with the lowest values for MRG, that is subjects with an intermediate level of aroma release. For the whole group of subjects and for each group, mouth coating increases significantly with fat content and firmness. 3.3.3. Rheological properties of food bolus The rheological parameters measured (Sagg, Sflow, hend/Send) do not present any significant difference between the release groups (Fig. 5). However for high fat cheeses, we observed a significant difference between the release groups on Sagg [F(2,43) = 3.7; p b 0.05] which is significantly smaller for MRG than for LRG, HRG being intermediate. This means that the failure threshold is lower for MRG than for LRG and thus boluses obtained for subjects belonging to MRG are less structured than those obtained for subjects belonging to LRG. Moreover, concerning cheeses, all parameters vary significantly according to the fat level. Their values are higher for low fat than for high fat cheeses. Only hend/Send varies according to the firmness level indicating that the spreadability is higher for soft cheeses (S) than for firm cheeses (F and FF). 4. Discussion The aim of this study was to explain inter individual variability in the quantity of aroma release by differences in the physiology of the subjects. In this objective, a classification of subjects was done, based on
Fig. 5. Influence of release group, fat and firmness on rheological bolus parameters. Results of three ways ANOVAs and 2-way interactions for bolus rheological parameters.
H. Labouré et al. / Food Research International 64 (2014) 692–700
the quantity of aroma release during each phase of consumption (mastication and post-swallowing), for each product and each aroma compound. Three groups of subjects were identified: one group with a high level of release whatever the phase of consumption and the aroma compounds (HRG), one group with a low level of aroma release whatever the phase of consumption and the aroma compounds (LRG), and one group with an intermediary level of release whatever the phase of consumption and the aroma compounds (MRG). We then compared the physiological and the food oral processing parameters between the different groups of subjects. In the present study, the three groups of subjects could not be differentiated by their saliva flow, neither at rest nor stimulated. This parameter does not seem to greatly impact on in vivo aroma release as was already reported (Pionnier et al., 2004), even if it has been demonstrated as an important parameter during simulated in-mouth process (Doyennette et al., 2014; van Ruth & Roozen, 2000). Mishellany-Dutour et al. (2012) found a relation between the inmouth air cavity (IMAC) volume changes after swallowing and the amount of menthone release during mint consumption however we were not able in the present study to find such effect on subjects' discrimination. This result suggests that other physiological parameters such as the velum opening induced by chewing movements may impact on the amount of aroma release during the consumption of solid food. This confirms other observations from Ruijschop et al. (2009) showing that subjects could be segmented on the morphology of their retronasal aroma release curves whatever the type of product (liquid or solid). However in our case the amount of menthone release could not explain the aroma release patterns of MRG group which differed according to the aroma compound. The parameters which better discriminate the groups of subjects are the food oral processing parameters and velum opening. The group with the highest amount of aroma release (HRG) is characterized by a great chewing activity (number of chewing cycles, chewing duration, maximal amplitude and total muscle work), high mouth coating and velum opening at each chewing event. The group with the lowest amount of aroma release (LRG) presents a lower chewing activity than HRG and a velum opening limited to the swallowing events. The intermediate group (MRG) is characterized by the lowest chewing activity, the lowest amount of product remaining in the mouth (mouth coating) but by a velum opening at each chewing event. The particularity of this group of subjects is to present a different release behaviour for the two aroma compounds. The high differences in release observed between HRG and LRG can be logically explained by the differences in velum opening. Subjects from HRG open their velum at each chewing event whereas the majority of the subjects of LRG open their velum only after swallowing. The more frequent opening of the velum would allow more frequent renewal of the gaseous phase in the mouth and therefore a larger amount of aroma would be extracted from the matrix and then transferred to the nasal cavity. This result confirms the specific role of the velum in aroma release (Buettner et al., 2008). HRG is also significantly differentiated from LRG by the chewing activity. A lot of data in the literature reported a positive correlation between the concentration of volatile in the nose during food consumption and the number of chews, total chewing work (Hansson et al., 2003; Pionnier et al., 2004) and chewing duration (Blissett et al., 2006; Tarrega et al., 2008). In addition, the velum opening could be linked to the amplitude of muscle contraction during chewing. A greater muscle contraction could favour the velum opening, which is the case for HRG group which differs from LRG group by greater amplitude of muscle contraction and a more frequent velum opening leading to a larger quantity of aroma release. However for MRG group, the frequent opening of the velum cannot be explained by the low amplitude of chewing. Because of the greater chewing activity of HRG compared to LRG, we can suppose that subjects from HRG will present boluses significantly
699
different to those produced by LRG, which is not the case. In fact the bolus properties seem to be more a result of chewing strategy as previously reported (Yven et al., 2012) and less related to aroma release parameters (Feron et al., 2014). Our subjects are the same than those reported by Yven et al. (2012). Looking at their chewing strategies, we found in each release group both subjects who adapted their chewing process to cheese characteristics and subjects which did not. Subjects who adapted their chewing process to cheese characteristics produced boluses with relatively similar rheological properties whatever the texture of the cheese. On the contrary, rheological properties of the bolus produced by subjects who did not adapt their chewing process to cheese texture differed significantly (Yven et al., 2012). As these two types of subjects were present in each release group, a great variability in the rheological properties of the bolus was observed within each group of subjects, which could explain that we were not able to find significant differences between rheological properties of the bolus, even if significant differences in chewing behaviour are observed. Due to the greater chewing activity of HRG (longer chewing duration, higher number of chewing) compared to LRG and MRG, we expected a greater quantity of saliva incorporated into the bolus for HRG than for the other group (van der Bilt, Engelen, Abbink, & Pereira, 2007; C. Yven, Culioli, & Mioche, 2005) which was not the case. However, this result is in accordance with Gaviao et al. (2004) that did not see any significant correlations between the salivary flow rate obtained from chewing on a food and the number of chewing cycles needed to prepare that food for swallowing. Authors suggested that subjects being used to their respective amounts of saliva, their respective swallowing threshold is not influenced by their amount of saliva. So a relative large salivary flow does not necessarily lead to swallow the food after a relative small number of chewing cycles. In agreement with Mioche et al. (2002) and Tarrega, Yven, Semon, & Salles (2011), more saliva was incorporated into the bolus for the hardest cheese for which chewing duration and number of chewing cycle are greater. Moreover we did not find any significant effect of fat content on the amount of saliva incorporated into the bolus even if the chewing parameters were higher for low fat cheeses than for high fat ones. Mioche et al. (2002) showed yet that saliva incorporated to the bolus is largely determined by chewing duration and to a small extent by muscular activity. According to our classification, HRG and MRG were differentiated on their amount of release of the more hydrophobic compound, nonan-2one (NO). The main parameters which significantly differentiated these two groups are the chewing parameters and mouth coating. The amount of product remaining in the mouth after food ingestion is significantly lower for MRG than for HRG. Nonan-2-one has a great affinity for fat and is more released during the post-swallowing phase (Repoux, Labouré et al., 2012). Thus the amount of nonan-2-one released could be logically related to the amount of product remaining in the mouth. Considering the less hydrophobic compound, ethyl propanoate (EP), subjects from MRG are closer to those belonging to HRG. This compound is more volatile than NO, has less affinity for fat and thus is more released during mastication (Repoux, Labouré et al., 2012). During the mastication step, we observed that velum opening was similar for these groups, which explains the similarity in the release patterns for EP between HRG and MRG. 5. Conclusion Our study allowed us to point out the main physiological parameters which explain the different amounts of aroma release during food consumption. Subjects who present the highest level of release whatever the aroma compound have the highest chewing activity, highest mouth coating and a velum opening at each mouth movement. Subjects with the lowest level of release whatever the aroma compounds present the lowest chewing activity and a velum opening at swallowing events only. An intermediate group of subjects was identified with a level of
700
H. Labouré et al. / Food Research International 64 (2014) 692–700
aroma release close to the high release group for the more hydrophilic compound due to a velum opening at each mouth movement during food consumption and close to low release group for the more hydrophobic compound due to a lower mouth coating after swallowing. A better understanding of food oral processing depending on both products and subjects allowed us to explain inter-individual differences in aroma release. These differences could then impact aroma perception and food preferences. Aknowledgements The authors gratefully acknowledge the French National Research Agency (ANR-07-PNRA-014), Fromageries Bel SA, Soredab (groupe Soparind Bongrain), the Association Nationale de la Recherche et de la Technologie (ANRT), the Regional Council of Burgundy France and the FEDER (European Funding for Regional Economical Development contract no. RB 29000350) for the financial support, ChemoSens Platform (Centre des Sciences du Goût et de l'Alimentation, Dijon, France) for the technical assistance in APCI–MS analyses, Laboratoire de Rhéologie (Université Joseph Fourier Grenoble, France) for the bolus rheology, Claude Yven for the electromyography, and the panellists who participated in this study. References Blissett, A., Hort, J., & Taylor, A. J. (2006). Influence of chewing and swallowing behavior on volatile release in two confectionary systems. Journal of Texture Studies, 37(5), 476–496. Buettner, A., Beer, A., Hannig, C., Settles, M., & Schieberle, P. (2002). Physiological and analytical studies on flavor perception dynamics as induced by the eating and swallowing process. Food Quality and Preference, 13(7–8), 497–504. Buettner, A., Otto, S., Beer, A., Mestres, M., Schieberle, P., & Hummel, T. (2008). Dynamics of retronasal aroma perception during consumption: Cross-linking on-line breath analysis with medico-analytical tools to elucidate a complex process. Food Chemistry, 108(4), 1234–1246. Doyennette, M., de Loubens, C., Deleris, I., Souchon, I., & Trelea, I. C. (2011). Mechanisms explaining the role of viscosity and post-deglutitive pharyngeal residue on in vivo aroma release: A combined experimental and modeling study. [Article]. Food Chemistry, 128(2), 380–390, http://dx.doi.org/10.1016/j.foodchem.2011.03.039. Doyennette, M., Déléris, I., Féron, G., Guichard, E., Souchon, I., & Trelea, I. C. (2014). Main individual and product characteristics influencing in-mouth flavour release during eating masticated food products with different textures: Mechanistic modelling and experimental validation. Journal of Theoretical Biology, 340, 209–221. Doyennette, M., Deleris, I., Saint-Eve, A., Gasiglia, A., Souchon, I., & Trelea, I. C. (2011). The dynamics of aroma compound transfer properties in cheeses during simulated eating conditions. Food Research International, 44(10), 3174–3181, http://dx.doi.org/10. 1016/j.foodres.2011.07.034. Feron, G., Ayed, C., Qannari, E. M., Courcoux, P., Labouré, H., & Guichard, E. (2014). Understanding aroma release from model cheeses by a statistical multiblock approach on oral processing. Plos One, 9(4), e93113, http://dx.doi.org/10.1371/journal.pone.0093113. Gaviao, M. B.D., Engelen, L., & van der Bilt, A. (2004). Chewing behavior and salivary secretion. European Journal of Oral Sciences, 112(1), 19–24. Gierczynski, I., Laboure, H., & Guichard, E. (2008). In vivo aroma release of milk gels of different hardnesses: Inter-individual differences and their consequences on aroma perception. Journal of Agricultural and Food Chemistry, 56(5), 1697–1703. Guichard, E. (2002). Interactions between flavor compounds and food ingredients and their influence on flavor perception. Food Reviews International, 18(1), 49–70. Haahr, A. -M., Bardow, A., Thomsen, C. E., Jensen, S. B., Nauntofte, B., & Bakke, M. (2004). Release of peppermint flavour compounds from chewing gum: Effect of oral functions. Physiology & Behavior, 82(2–3), 531–540. Hansson, A., Giannouli, P., & Van Ruth, S. (2003). The influence of gel strength on aroma release from pectin gels in a model mouth and in vivo, monitored with proton-
transfer-reaction mass spectrometry. Journal of Agricultural and Food Chemistry, 51(16), 4732–4740. Hodgson, M., Linforth, R. S. T., & Taylor, A. J. (2003). Simultaneous real-time measurements of mastication, swallowing, nasal airflow, and aroma release. Journal of Agricultural and Food Chemistry, 51(17), 5052–5057. Le Quere, J. -L., Gierczynski, I., Langlois, D., & Semon, E. (2006). Nosespace with an ion trap mass spectrometer: Quantitative aspects (book section). In W. L. P. Bredie, & M.A. Petersen (Eds.), (Amsterdam (NLD) ed.). Flavour science: Recent advances and trends, Vol. 43. (pp. 589–592). . Oxford (GBR): Elsevier. Linforth, R. S. T., Baek, I., & Taylor, A. J. (1999). Simultaneous instrumental and sensory analysis of volatile release from gelatine and pectine/gelatine gels. Food Chemistry, 65, 77–83. Mestres, M., Kieffer, R., & Buettner, A. (2006). Release and perception of ethyl butanoate during and after consumption of whey protein gels: Relation between textural and physiological parameters. Journal of Agricultural and Food Chemistry, 54(5), 1814–1821. Mestres, M., Moran, N., Jordan, A., & Buettner, A. (2005). Aroma release and retronasal perception during and after consumption of flavored whey protein gels with different textures. 1. In vivo release analysis. Journal of Agricultural and Food Chemistry, 53(2), 403–409. Mioche, L., Bourdiol, P., Monier, S., & Martin, J. F. (2002). The relationship between chewing activity and food bolus properties obtained from different meat textures. Food Quality and Preference, 13(7–8), 583–588. Mishellany-Dutour, A., Woda, A., Labouré, H., Bourdiol, P., Lachaze, P., & Guichard, E. (2012). Retro-nasal aroma release is correlated with variations in the in-mouth air cavity volume after empty deglutition. PloS One, 7(7), e41276, http://dx.doi.org/10. 1371/journal.pone.0041276. Pionnier, E., Chabanet, C., Mioche, L., Le Quere, J. L., & Salles, C. (2004). 1. In vivo aroma release during eating of a model cheese: Relationships with oral parameters. Journal of Agricultural and Food Chemistry, 52(3), 557–564. Pivk, U., Ulrih, N.P., Juillerat, M.A., & Raspor, P. (2008). Assessing lipid coating of the human oral cavity after ingestion of fatty foods. Journal of Agricultural and Food Chemistry, 56(2), 507–511. Poette, J., Mekoue, J., Neyraud, E., Berdeaux, O., Renault, A., & Guichard, E. (2013). Fat sensitivity in humans: oleic acid detection threshold is linked to saliva composition and oral volume. Flavour and Fragrance Journal, 29(1), 39–49. Repoux, M., Labouré, H., Courcoux, P., Andriot, I., Sémon, E., & Yven, C. (2012). Combined effect of cheese characteristics and food oral processing on in vivo aroma release. Flavour and fragrance journal, 27(6 (special issue paper)), 414–423, http://dx.doi. org/10.1002/ffj.3110. Repoux, M., Sémon, E., Feron, G., Guichard, E., & Labouré, H. (2012). Inter-individual variability in aroma release during sweet mint consumption. Flavour and Fragrance Journal, 27(1), 40–46, http://dx.doi.org/10.1002/ffj.2077. Repoux, M., Septier, C., Palicki, O., Guichard, E., Feron, G., & Labouré, H. (2012). Solid cheese consumption: Quantification of oral coating. Archives of Oral Biology, 57(1), 81–86, http://dx.doi.org/10.1016/j.archoralbio.2011.07.011. Ruijschop, R., Burgering, M. J. M., Jacobs, M.A., & Boelrijk, A. E. M. (2009). Retro-nasal aroma release depends on both subject and product differences: A link to food intake regulation? Chemical Senses, 34(5), 395–403, http://dx.doi.org/10.1093/chemse/ bjp011. Salles, C., Chagnon, M. -C., Feron, G., Guichard, E., Labouré, H., & Morzel, M. (2011). Inmouth mechanisms leading to flavor release and perception. Critical Reviews in Food Science and Nutrition, 51(1), 67–90, http://dx.doi.org/10.1080/ 10408390903044693. Tarrega, A., Yven, C., Sémon, E., & Salles, C. (2008). Aroma release and chewing activity during eating different model cheeses. International Dairy Journal, 18(8), 849–857. Tarrega, A., Yven, C., Semon, E., & Salles, C. (2011). In-mouth aroma compound release during cheese consumption: Relationship with food bolus formation. International Dairy Journal, 21(5), 358–364, http://dx.doi.org/10.1016/j.idairyj.2010.12.010. van der Bilt, A., Engelen, L., Abbink, J., & Pereira, L. J. (2007). Effects of adding fluids to solid foods on muscle activity and number of chewing cycles. European Journal of Oral Sciences, 115, 198–205. van Ruth, S. M., & Roozen, J. P. (2000). Influence of mastication and saliva on aroma release in a model mouth system. Food Chemistry, 71(3), 339–345. Yven, C., Culioli, J., & Mioche, L. (2005). Meat bolus properties in relation with meat texture and chewing context. Meat Science, 70, 365–371. Yven, C., Patarin, J., Magnin, A., Labouré, H., Repoux, M., & Guichard, E. (2012). Consequences of individual chewing strategies on bolus rheological properties at the swallowing threshold. Journal of Texture Studies, 43(4), 309–318, http://dx.doi.org/ 10.1111/j.1745-4603.2011.00340.x.
Contents lists available at ScienceDirect
Food Research International journal homepage: www.elsevier.com/locate/foodres
Inter-individual retronasal aroma release variability during cheese consumption: Role of food oral processing H. Labouré a,b,c,⁎, M. Repoux a,b,c, P. Courcoux d, G. Feron a,b,c, E. Guichard a,b,c a
INRA, UMR1324 Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France CNRS, UMR6265 Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France Université de Bourgogne, UMR Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France d LUNAM University, ONIRIS, USC “Sensometrics and Chemometrics Laboratory”, F-44322 Nantes, France b c
a r t i c l e
i n f o
Article history: Received 30 April 2014 Accepted 20 July 2014 Available online 14 August 2014 Keywords: In vivo aroma release Inter individual variability Food oral processing Model cheeses Chewing activity Mouth coating
a b s t r a c t The aim of our study was to explain inter-individual differences on in vivo aroma release during cheese consumption by oral physiological parameters. To reach this objective, 34 subjects were recruited. Their salivary flow, oral volume and velum opening were determined. Six cheddar-based melted cheeses with different fat levels and firmness were flavoured with nonan-2one (NO) and ethyl propanoate (EP). During their consumption (free protocol), in vivo retro nasal aroma release was followed by Atmospheric Pressure Chemical Ionisation–Mass Spectrometry (APCI–MS). Chewing activity was evaluated by electromyography recordings. Bolus saliva content, mouth-coating, and bolus rheology were also determined. Based on the quantity of aroma released before and after swallowing, subjects can be clustered into three groups: the first one (HRG) is characterized by a large quantity of aroma release whatever the aroma compound; the second one (MRG) showed a large release for EP and a lower one for NO; the third group (LRG) was characterized by a low quantity of aroma release whatever the compound. Whatever the group of subjects, fat and firmness effects differed according to the aroma compound. EP release increased with firmness and fat content, whereas NO release was not affected by firmness and decreased when fat content increased. Physiological parameters which better differentiated the three groups of subjects according to their release behaviour were chewing activity, mouth coating and frequency of velum opening. Subjects from HRG were differentiated from LRG subjects by a higher chewing activity, and more frequent velum opening. Subjects from MRG presented a lower mouth coating explaining their lower release of NO, the more hydrophobic compound. This study shows that the total amount of aroma released in the nasal cavity during food consumption depends not only on the characteristics of the product but also on the oral physiology of the subjects and on their food oral processing. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Retro-nasal perception occurs in humans while eating and is one of the key factors of food acceptability by consumers. During food consumption, aroma compounds responsible for aroma perception have to reach the olfactory receptors after being transferred from food to saliva then to the oro-nasal cavity. Both the composition and the structure of food have been shown to highly impact the release of aroma compounds (Linforth, Baek, & Taylor, 1999; Gierczynski, Laboure, & Guichard, 2008). According to their physico-chemical properties, aroma compounds are more or less released from the food matrix, ⁎ Corresponding author at: INRA, UMR1324 Centre des Sciences du Goût et de l'Alimentation, F-21000 Dijon, France. Tel.: +33 380 69 35 28; fax: +33 380 69 32 27. E-mail address: [email protected] (H. Labouré).
http://dx.doi.org/10.1016/j.foodres.2014.07.024 0963-9969/© 2014 Elsevier Ltd. All rights reserved.
due to molecular interactions with the macromolecules (proteins, polyosides) or to different solubility in water and oil (Guichard, 2002). Additionally great inter-individual variability on in vivo retro nasal aroma release during food consumption has been underlined (Gierczynski et al., 2008; Mestres, Kieffer, & Buettner, 2006; Pionnier, Chabanet, Mioche, Le Quere, & Salles, 2004). The different causes of this variability have not been extensively investigated yet due to the complexity of in-mouth mechanisms (Salles et al., 2011). Several authors have already observed relations between some physiological parameters and aroma release. On liquid samples, Buettner et al. (Buettner, Beer, Hannig, Settles, & Schieberle, 2002) found that velum opening, swallowing and adsorption to oral mucosa are able to modulate retro nasal aroma release. During mint tablet consumption, the inmouth air cavity (IMAC) volume changes after swallowing were also found to impact the total amount of aroma release (Mishellany-
H. Labouré et al. / Food Research International 64 (2014) 692–700
Dutour et al., 2012), but this parameter was not evaluated during solid food consumption. Using in vitro systems simulating eating conditions, the dilution effect of saliva has been also evidenced (van Ruth & Roozen, 2000). However the real impact of saliva on aroma release during real food consumption has not been elucidated yet due to a combined effect with that of the product (Mestres et al., 2006). A mathematical model for in vivo aroma release has been designed to take account of the product saliva mixing and the post-deglutive pharyngeal residue, showing the important impact on saliva dilution effect on aroma release (Doyennette, de Loubens, Deleris, Souchon, & Trelea, 2011). This model was validated with a hydrophilic compound, diacetyl, released from glucose syrup. An extension of this mechanistic model was further developed considering the role of the mastication process which showed that, among other parameters, the rate of saliva incorporated into the bolus greatly influenced aroma release (Doyennette et al., 2014). In the case of solid foods, a lot of literature is available evidencing that in-mouth process parameters related to chewing behaviours such as chewing work, chewing time, and burst numbers are highly positively correlated with in vivo aroma release (Haahr et al., 2004; Hansson, Giannouli, & Van Ruth, 2003; Mioche, Bourdiol, Monier, & Martin, 2002). Inter-individual variations of these parameters could explain inter-individual differences in aroma release (Blissett, Hort, & Taylor, 2006; Tarrega, Yven, Sémon, & Salles, 2008). Specific consumption patterns of subjects have been underlined as their consequences on aroma release profiles (Mestres, Moran, Jordan, & Buettner, 2005; Mestres et al., 2006). These authors showed a specific aroma release profile when subjects consume a soft gel by pressing with the tongue in the frontal part of the oral cavity against the hard palate and another one when the same panellists were instructed to chew the same product with jaw opening. This inter-individual specificity in aroma release profiles has been confirmed with subjects eating either liquid or solid foods (Ruijschop, Burgering, Jacobs, & Boelrijk, 2009). This study showed that subjects with relatively high retro-nasal aroma release intensity for liquid food products have also a relatively high release for a solid food product. However no mention was made on the nature of the aroma compound and its consequences on aroma release patterns, even if other studies mentioned different release patterns for aroma compounds differing in their hydrophobicity (Doyennette, de Loubens, Deleris, Souchon, & Trelea, 2011; Doyennette et al., 2011; Repoux et al., 2012). Moreover, most of the different studies were conducted on a limited number of subjects and only some specific oral physiological parameters were measured. In this context, the aim of our study was to investigate interindividual differences on in vivo aroma release from solid foods, namely model cheeses and to identify oral physiological variables which could explain this inter-individual variability. We mainly focused on chewing activity, bolus rheology, saliva flow and oral physiology. We choose to analyse the release of two aroma compounds differing in their hydrophobicity from model cheese varying in firmness and fat content. In a first step and in order to discriminate the subjects in function of the quantity of aroma realised, a hierarchical cluster analysis has been performed separately for each aroma compound because of their very different level of release; then the two classifications obtained were compared. In a second step, the effect of food matrix composition on the aroma release has been studied in each group of subjects and compared. Finally, the oral physiology and food oral processing parameters have been compared between the groups in order to explain the differences in release patterns.
693
2.1. Food products Six cheddar-based melted cheeses (model cheeses) were developed (Repoux, Labouré, et al., 2012). They differed by their firmness and their fat content (Fig. 1). Three firmness levels, measured by a compression test, were obtained: a very firm one (FF: higher than 10 000 N), a firm one (F between 6000 and 8000 N) and a soft one (S between 400 and 3000 N). Two levels of fat content have been studied, a low one (l: 25%) and a high one (h: 50%). These model cheeses were flavoured with two aroma compounds: nonan-2-one (NO: logP = 2.9; vp = 83.2 Pa at 25 °C) and ethyl propanoate (EP: logP = 1.4; vp = 4.79.103 Pa at 25 °C). They were added during cheese production at levels adjusted to achieve final concentrations in the cheeses around 6 ppm for nonan-2-one and 25 ppm for ethyl propanoate. 2.2. Subjects Thirty four naïve subjects (15 females and 19 males aged between 22 and 59 years; mean age: 39 years) participated in this study. Selection criteria were availability for the duration of the study, motivation, no medicine consumption, and a good dental status (no denture wearers and few artificial crowns, dental bridges and dental implants). The subjects were not allowed to smoke, eat or drink during one hour before the test session. All subjects were informed of the observational nature of this study, gave their signed consent and received compensatory indemnities for their participation in two sessions, each lasting about 2 h. 2.3. In vivo aroma release 2.3.1. Materials In vivo aroma release was measured using Atmospheric Pressure Chemical Ionisation–Mass Spectrometry APCI–MS (Gierczynski et al., 2008). Air from the nose was sampled from one nostril at an average flow rate of 37 mL·min−1 and introduced into the APCI source of an ion trap Esquire-LC mass spectrometer (Bruker Daltonique, Wissembourg, France) via a fused silica capillary tubing (i.d. = 0.53 mm) heated at 150 °C and to which a 5 kV positive ion corona pin discharge was applied (Le Quere, Gierczynski, Langlois, & Semon, 2006). For the oral movement measurements during mint consumption, compounds were monitored according to their protonated molecular ion (MH+): acetone (m/z = 59) and menthone (m/z = 155). For the experiments during cheese consumption, three compounds were monitored according to their protonated molecular ion (MH+): ethyl propanoate (m/z = 103) and nonan-2-one (m/z = 143), the two aroma compounds added to
2. Material and methods The study protocol was submitted to an Ethics Committee and was approved on 17 April 2008 by the Comité de Protection des Personnes Est-1 (N° 2008/15) and on 8 August 2008 by the Direction Générale de la Santé - France (N° DGS2008-0196).
Fig. 1. Rheological properties of the cheese products differing in firmness (S for soft, F for firm and FF for very firm) and fat level (l for low fat and h for high fat).
694
H. Labouré et al. / Food Research International 64 (2014) 692–700
the model cheeses, and acetone (m/z = 59), which was used as an indicator of the subject's breath (Hodgson, Linforth, & Taylor, 2003). 2.3.2. Methods Each subject was asked to position the plastic tube in one nostril and to breathe normally. After an initial swallow, the subject was instructed to place the piece of cheese ([2.25 ∗ 2.25 ∗ 1.1] cm; m = 6 g) in the mouth, and to consume it freely while keeping the lips closed. The times of swallowing were recorded using a sensor placed on the subject's throat. The products were presented in random order at 17 °C. Bread, apple and water were used as mouth cleansers between two tests. Three replicates per cheese and per subject were performed. 2.3.3. Data analysis After smoothing the curves to eliminate signal fluctuations due to the subject's breathing pattern, two release phases were identified. The mastication step (m) extended from placing the cheese in the mouth to the first swallowing, and the post-swallowing step (ps) extended from the first swallowing to the time at which the signal returned to its baseline level. The area under the curve was calculated for each aroma compound during each release phase. 2.4. Physiological parameters 2.4.1. Salivary flow (at rest and stimulated) Resting saliva flow (rSF) was determined by using the method used by Gaviao, Engelen, & van der Bilt (2004). Non-stimulated saliva flow was collected by instructing the subject to spit out the saliva every 30 s into a pre-weighed cup over a period of 5 min. Two measurements were performed by subjects, one per session. For stimulated saliva flow, the subjects were instructed to chew a piece of Parafilm™ (0.5 g ± 0.2 g) for a period of 1 min then spit out the saliva after 30 s, continue chewing for a further 30 s and then spit out both the saliva and the piece of Parafilm ™ (Repoux, Sémon, Feron, Guichard, & Labouré, 2012). Three stimulated saliva flow rate (sSF) replicates were performed during one session. 2.4.2. Oral volume and IMAC volume changes An Eccovision® acoustic pharyngometer (Hood Laboratories, USA) was used to measure oral volume and was composed of four components: a wave tube, an electronic platform, a mouthpiece and a disposable filter. Reflectance pharyngometry was performed with a twomicrophone imaging acoustic pharyngometer device, as described recently (Mishellany-Dutour et al., 2012; Poette et al. 2013). This device consists of two microphones and one horn driver mounted on a wave tube and connected to a PC-compatible computer with signal conversion capabilities. The signal was converted into the surface change (cm2) as a function of the length of the oral cavity (in cm). The subjects held the mouthpiece in their mouth with their teeth against the flange and their tongue in a low position. To prevent air leaks, which could cause measurement errors, the subjects placed their lips over the flange, sealing the mouthpiece. IMAC volume changes after swallowing were measured following the procedure described previously (MishellanyDutour et al., 2012). 2.4.3. Velum opening The influence of oral movements on velum opening was experimented on subjects while consuming a commercial mint tablet under an imposed protocol (Repoux, Sémon, Feron, Guichard, & Labouré, 2012). The subjects placed the tablet under the tongue. Then they were given a series of instructions in the following order: (1) to swallow (S1), (2) to make tongue movements (TM) without swallowing and mouth opening, (3) to swallow (S3), (4) to make jaw movements (JM) and to swallow (S3). Menthone release during mint tablet consumption was followed by APCI–MS. This measurement was analysed qualitatively and quantitatively. For the qualitative analysis, an event was coded ‘1’ if the instruction induced menthone release
and ‘0’ if the instruction did not induce any release t. For the quantitative analysis we calculated the area under the curve corresponding to menthone release, at each event, then the sum of the areas for all the events and the sum of the area for the swallowing events only. 2.5. Food oral process measurements 2.5.1. Chewing activity Chewing activity was monitored during cheese consumption, simultaneously to aroma release. The muscle activity of the superficial masseter and temporis muscles (left and right) during chewing was recorded by electromyography (EMG) using gold surface electrodes (Grass technologies, West Warwick, RI, U.S.A.), at 382 Hz, then the signal was amplified and digitalized (Mioche et al., 2002). From the EMG data the following parameters were collected: the number of chewing cycles, the duration of chewing cycle (expressed in s) and the maximal amplitude (expressed in mV) which correspond to the maximal value of amplitudes of the different chewing cycles registered in a whole chewing sequence and total muscle work (expressed in mV s) which is the sum of muscle work per chewing cycle, that is the surface area of the muscle activity signal. 2.5.2. Bolus saliva content The percentage of dry matter and water content were determined using an infrared dryer for all the cheeses and boluses obtained just before swallowing. For each subject and each cheese, the percentage of saliva incorporated (Sal) into the bolus was calculated as follows: Bolus water content ð%Þ Saliva incorporated ¼ Cheese dry matter ð%Þ Bolus dry matter ð%Þ ð% into the bolusÞ −Cheese water content ð%Þ:
2.5.3. Mouth coating Mouth coating (MCo) defined as the residual food that sticks to the oral surface after food ingestion, was quantified by the “mouth rinse” method (Pivk, Ulrih, Juillerat, & Raspor, 2008). The lipids of the residual food were quantified by the intensity of curcumin fluorescence in the rinse water as previously described (Repoux, Labouré, et al., 2012; Repoux et al., 2012). Briefly, each subject was asked to place a piece of cheese (6 g at 17 °C) containing curcumin (30 ppm in the final cheeses) in the mouth and to masticate normally until they needed to swallow. He/she swallowed without any cleaning movement and then rinsed his/her mouth twice with 4 ml of water at a temperature of 50 °C for 30 s and spit it into a vial. The fluorescence intensity of the curcumin in the two cumulated spittle was quantified using a Perkin Elmer 1420 Multilabel Counter Victor 3 V (Perlin Elmer, Courtaboeuf, France) at an excitation wavelength of 450 nm and an emission wavelength of 510 nm. The quantification of the coating was evaluated in reference to calibration curves performed for each cheese. 2.6. Rheological properties of food bolus Subjects were asked to chew the cheese samples until they were ready to swallow and to spit out the bolus into a truncated syringe. An aliquot of 3 mL of bolus at the syringe bottom was used for compression test to get a constant volume regardless of the cheese employed. The rheological properties of the bolus were measured by a compression test described elsewhere (Yven et al., 2012). From the compression curve, three phases were distinguished after the loading phase. During the first phase named “aggregated phase” the bolus behaves like a solid characterized by a failure threshold Sagg (Pa). During the second phase (“flow phase”), the suspension begins to flow. From a mechanical point of view, the viscous effects are no longer negligible in relation to the yield stress effects described by Sflow
H. Labouré et al. / Food Research International 64 (2014) 692–700
(Pa). During the third one (“particle phase”), the mechanical response is governed by the particles. At the end of compression, hend (mm) denotes the final height and Send (mm2) denotes the area generated under the maximal force (50 N). The hend/Send ratio (mm−1) stands for spreadability of the bolus; the lesser is the ratio, the greater is the spreadability. 2.7. Data analysis 2.7.1. Subjects classification In a first step, the 34 subjects were clustered based on the area under the curve obtained during the mastication phase and the postswallowing phase, for each of the 6 food products. A hierarchical clustering of subjects using the Euclidean distance and Ward's criterion was performed independently for each of the two aroma compounds because of their very different levels of release. 2.7.2. Characterization of subjects groups Three way ANOVAs with interactions (group, fat, firmness, group ∗ fat, group ∗ firmness, fat ∗ firmness) were performed to evaluate differences between release groups on the quantity of aroma released, the food oral processing parameters (mouth coating, saliva incorporated in the bolus, number of chewing cycle, maximal amplitude and total chewing work), the oral physiological parameters (oral volume, salivary flow at rest and stimulated) and the rheological characteristics of the bolus. Factors studied were considered significant when their p value was lower than 0.05. A Principal Component Analysis (PCA) based on the correlations was performed on the mean quantity of aroma release per subjects and per consumption phase in order to compare the different groups of subjects for each aroma compound. Statistical analyses were performed using STATISTICA® software, version 10 (StatSoft, France). 3. Results 3.1. Subjects classification 3.1.1. Ethyl propanoate Two groups of subjects (A and B) composed of 15 and 19 subjects respectively were identified by the hierarchical cluster analysis based on quantity of EP release during mastication and post-swallowing phase. Subjects from group A release significantly more EP than subjects from group B. Subjects from group A constitute a high release group, and subjects from group B a low release group for EP. The group effect is slightly higher in post-swallowing phase (Fgroup(1,5) = 33.8; p b 0.01) than in mastication phase (Fgroup(1,5) = 25.7; p b 0.01). 3.1.2. Nonan-2-one Two groups of subjects (C and D) composed of 7 and 27 subjects respectively were identified by the hierarchical cluster analysis based on quantity of NO release during mastication and post-swallowing phase. In each phase, subjects from group C release significantly more NO than subjects from group D. Subjects from group C constitute a high release group, and subjects from group D a low release group for NO. Compared to the groups based on EP release, the group effect is significantly higher in post-swallowing phase (Fgroup(1,5) = 71.4; p b 0.001) than in mastication phase (Fgroup(1,5) = 11.3; p b 0.05). This observation could be explained by the greater hydrophobicity and lower volatility of NO compared to EP and thus by the low quantity of NO released during the mastication phase. 3.1.3. Comparison of the HCA Looking at the subjects who belong to the different groups (A or B for EP and C or D for NO), we noticed that all the 7 subjects from group C (high release of NO) also belong to group A (high release of EP) and
695
that all the 19 subjects who belong to group B (low release of EP) also belong to group D (low release of NO). Only 8 subjects did not present the same level of release for the 2 aroma compounds, i.e. they release EP at a high level (they belong to group A) and NO at a low level (they belong to group D). This allowed us to identify three different groups of subjects: a high release group (HRG) composed of 7 subjects who release a high amount of the two aroma compounds, a mix release group (MRG) composed of 8 subjects who release a high amount of EP and a low amount of NO, and a low release group (LRG) composed of 19 subjects who release a low amount of the two aroma compounds. A principal component analysis, performed on the quantity of aroma compounds release during mastication and post-swallowing phase illustrates the three groups of subjects and their distribution as a function of the aroma compound (Fig. 2). For each aroma compound, the different release groups are mainly discriminated according to the first principal component, which represents most of the information (50.65% for EP and 56.36% for NO). For EP, HRG and MRG are on the negative part and LRG on the positive part of this first axis, whereas for NO, HRG is on the negative part and MRG and LRG on the positive part of this axis. This first axis is correlated to the quantity of aroma release during the post-swallowing phase whatever the cheese. The second axis explains only 18.5 (EP) and 16.3% (NO) of the variance. It is mainly defined by the quantity of aroma release during the mastication phase. The three groups of subjects were not well discriminated along this axis. 3.2. Influence of the effects related to the products on aroma release for each group of subjects In this part only the quantity of aroma released during the postswallowing phase will be discussed, firstly because the three groups of subjects are well discriminated during this post-swallowing phase of aroma release and not during the mastication phase; secondly because aroma compounds are mainly released in this phase, and thirdly because the quantity of aroma released in this post-swallowing phase is highly correlated with the total amount of aroma release (Repoux, Sémon, Feron, Guichard, & Labouré, 2012). 3.2.1. EP release Firmness effect is significant for the whole group and within each group (Fig. 3). For each group of subjects, the total amount of EP released after swallowing significantly increased with both the firmness (FFirmness(2,190) = 51.9; p b 0.0001) (Fig. 3a) and the fat content (F fat (1, 190) = 102.8; p b 0.00001) (Fig. 3b) of the cheeses. The firmness ∗ group interaction is also significant (FFirmness ∗ group(4, 190) = 7.2; p b 0.0001) because firmness effect is higher for HRG and MRG than for LRG. For HRG and MRG, the three levels of firmness are indeed significantly different at the 5% level whereas only the release from the softer cheese S is significantly lower than from the firmer cheeses (F and FF) for LRG. Similarly, fat effect is significant for the whole group and within each group. The fat ∗ group interaction is also significant (FFat ∗ group(2, 190) = 5.0; p b 0.001) because fat effect is greater in HRG than in LRG and MRG (Fig. 3b). 3.2.2. NO release NO release increases significantly with the firmness for the whole group (FFirmness(2,190) = 3.7; p b 0.05) but firmness effect is not significant within each group except for LRG (Fig. 3c). In LRG, as observed for EP, the release from the softer cheese S is significantly lower than from the firmer cheeses (F and FF). Contrary to EP, NO release decreased when fat content increased because of its hydrophobicity (Fig. 3c). The fat effect is higher than the firmness effect (Ffat(1, 190) = 54; p b 0.00001). Fat effect is significant and in the same direction for all groups but the interaction group ∗ fat is also significant (Ffat ∗ group(2,190) = 13.0; p b 0.0001) because the size of the effect is higher for HRG and MRG than for LRG (Fig. 3d).
696
H. Labouré et al. / Food Research International 64 (2014) 692–700
Fig. 2. PCA on the amount of aroma release represented by the area under the curve (AUC) in the mastication phase (m) and in the post-swallowing phase (ps) for each subject consuming each product (S = soft, F = firm, h = high fat, l = low fat). Subjects are represented by their release group. Upper figures for ethyl propanoate (EP), under figures for nonan-2-one (NO). Left: representation of the variables (quantity of aroma release for each cheese in the mastication phase and in the post-swallowing phase), right: representation of the means of all the subjects for each group (only the name of the group is reported: HRG: high release group, MRG: mix release group, LRG: low release group).
3.3. Influence of the subject physiological characteristics on aroma release for each group of subjects 3.3.1. Physiological parameters No significant difference was found for the saliva flows (at rest and stimulated) and IMAC volume changes between the groups of subjects. The only physiological parameter which presents significant differences between the groups of subjects is the velum opening. Subjects
were characterized for their velum opening by following the release of menthone during mint consumption. The release patterns are significantly different according to the groups. Even if all the subjects released menthone at the swallowing events, the area under the curve during these events are significantly different according to the groups (Fgroup(2,31) = 5.52; p b 0.001). Subjects from HRG released more menthone than subjects from the two others groups. The total amount of menthone released (during swallowing events but also at tongue
Fig. 3. Firmness effect and fat effect on EP (on the left) and NO (on the right) release per group. Vertical lines represent the confidence interval at 0.95.
H. Labouré et al. / Food Research International 64 (2014) 692–700
Fig. 4. Influence of group, fat and firmness on food oral processing parameters. Results of three way ANOVAs and 2-way interactions for food oral processing parameters.
697
698
H. Labouré et al. / Food Research International 64 (2014) 692–700
and jaw movements) has been calculated and is also significantly different between the groups (Fgroup(2,31) = 4.38; p b 0.05). Subjects from HRG release significantly more menthone than subjects from LRG but subjects from MRG present an intermediate behaviour not significantly different from that of the two others. Most of the subjects of HRG and MRG (86% and 88% respectively) release menthone at all the events whereas most of the subjects from LRG release only during the swallowing events.
3.3.2. Food oral processing parameters 3.3.2.1. Chewing activity. Globally, all the chewing parameters studied present significant differences between the release groups (Fig. 4). Subjects with an intermediate level of aroma release (MRG) present the lowest values for all of these parameters. As already reported (Repoux, Labouré et al., 2012), chewing activity varies in function of both fat content and firmness, with a greater chewing activity for low fat and very firm cheeses.
3.3.2.2. Bolus saliva content. The percentage of saliva incorporated into the bolus (bolus saliva content) does not vary significantly between the different release groups (Fig. 4). This parameter is only influenced by the cheese firmness, more saliva being incorporated into the bolus for the firmer cheeses. But even for these firmer cheeses, the group factor is not significantly different.
3.3.2.3. Mouth coating. The residual food that sticks to the oral surface (Mouth coating) varies significantly between the groups of subjects, with the lowest values for MRG, that is subjects with an intermediate level of aroma release. For the whole group of subjects and for each group, mouth coating increases significantly with fat content and firmness. 3.3.3. Rheological properties of food bolus The rheological parameters measured (Sagg, Sflow, hend/Send) do not present any significant difference between the release groups (Fig. 5). However for high fat cheeses, we observed a significant difference between the release groups on Sagg [F(2,43) = 3.7; p b 0.05] which is significantly smaller for MRG than for LRG, HRG being intermediate. This means that the failure threshold is lower for MRG than for LRG and thus boluses obtained for subjects belonging to MRG are less structured than those obtained for subjects belonging to LRG. Moreover, concerning cheeses, all parameters vary significantly according to the fat level. Their values are higher for low fat than for high fat cheeses. Only hend/Send varies according to the firmness level indicating that the spreadability is higher for soft cheeses (S) than for firm cheeses (F and FF). 4. Discussion The aim of this study was to explain inter individual variability in the quantity of aroma release by differences in the physiology of the subjects. In this objective, a classification of subjects was done, based on
Fig. 5. Influence of release group, fat and firmness on rheological bolus parameters. Results of three ways ANOVAs and 2-way interactions for bolus rheological parameters.
H. Labouré et al. / Food Research International 64 (2014) 692–700
the quantity of aroma release during each phase of consumption (mastication and post-swallowing), for each product and each aroma compound. Three groups of subjects were identified: one group with a high level of release whatever the phase of consumption and the aroma compounds (HRG), one group with a low level of aroma release whatever the phase of consumption and the aroma compounds (LRG), and one group with an intermediary level of release whatever the phase of consumption and the aroma compounds (MRG). We then compared the physiological and the food oral processing parameters between the different groups of subjects. In the present study, the three groups of subjects could not be differentiated by their saliva flow, neither at rest nor stimulated. This parameter does not seem to greatly impact on in vivo aroma release as was already reported (Pionnier et al., 2004), even if it has been demonstrated as an important parameter during simulated in-mouth process (Doyennette et al., 2014; van Ruth & Roozen, 2000). Mishellany-Dutour et al. (2012) found a relation between the inmouth air cavity (IMAC) volume changes after swallowing and the amount of menthone release during mint consumption however we were not able in the present study to find such effect on subjects' discrimination. This result suggests that other physiological parameters such as the velum opening induced by chewing movements may impact on the amount of aroma release during the consumption of solid food. This confirms other observations from Ruijschop et al. (2009) showing that subjects could be segmented on the morphology of their retronasal aroma release curves whatever the type of product (liquid or solid). However in our case the amount of menthone release could not explain the aroma release patterns of MRG group which differed according to the aroma compound. The parameters which better discriminate the groups of subjects are the food oral processing parameters and velum opening. The group with the highest amount of aroma release (HRG) is characterized by a great chewing activity (number of chewing cycles, chewing duration, maximal amplitude and total muscle work), high mouth coating and velum opening at each chewing event. The group with the lowest amount of aroma release (LRG) presents a lower chewing activity than HRG and a velum opening limited to the swallowing events. The intermediate group (MRG) is characterized by the lowest chewing activity, the lowest amount of product remaining in the mouth (mouth coating) but by a velum opening at each chewing event. The particularity of this group of subjects is to present a different release behaviour for the two aroma compounds. The high differences in release observed between HRG and LRG can be logically explained by the differences in velum opening. Subjects from HRG open their velum at each chewing event whereas the majority of the subjects of LRG open their velum only after swallowing. The more frequent opening of the velum would allow more frequent renewal of the gaseous phase in the mouth and therefore a larger amount of aroma would be extracted from the matrix and then transferred to the nasal cavity. This result confirms the specific role of the velum in aroma release (Buettner et al., 2008). HRG is also significantly differentiated from LRG by the chewing activity. A lot of data in the literature reported a positive correlation between the concentration of volatile in the nose during food consumption and the number of chews, total chewing work (Hansson et al., 2003; Pionnier et al., 2004) and chewing duration (Blissett et al., 2006; Tarrega et al., 2008). In addition, the velum opening could be linked to the amplitude of muscle contraction during chewing. A greater muscle contraction could favour the velum opening, which is the case for HRG group which differs from LRG group by greater amplitude of muscle contraction and a more frequent velum opening leading to a larger quantity of aroma release. However for MRG group, the frequent opening of the velum cannot be explained by the low amplitude of chewing. Because of the greater chewing activity of HRG compared to LRG, we can suppose that subjects from HRG will present boluses significantly
699
different to those produced by LRG, which is not the case. In fact the bolus properties seem to be more a result of chewing strategy as previously reported (Yven et al., 2012) and less related to aroma release parameters (Feron et al., 2014). Our subjects are the same than those reported by Yven et al. (2012). Looking at their chewing strategies, we found in each release group both subjects who adapted their chewing process to cheese characteristics and subjects which did not. Subjects who adapted their chewing process to cheese characteristics produced boluses with relatively similar rheological properties whatever the texture of the cheese. On the contrary, rheological properties of the bolus produced by subjects who did not adapt their chewing process to cheese texture differed significantly (Yven et al., 2012). As these two types of subjects were present in each release group, a great variability in the rheological properties of the bolus was observed within each group of subjects, which could explain that we were not able to find significant differences between rheological properties of the bolus, even if significant differences in chewing behaviour are observed. Due to the greater chewing activity of HRG (longer chewing duration, higher number of chewing) compared to LRG and MRG, we expected a greater quantity of saliva incorporated into the bolus for HRG than for the other group (van der Bilt, Engelen, Abbink, & Pereira, 2007; C. Yven, Culioli, & Mioche, 2005) which was not the case. However, this result is in accordance with Gaviao et al. (2004) that did not see any significant correlations between the salivary flow rate obtained from chewing on a food and the number of chewing cycles needed to prepare that food for swallowing. Authors suggested that subjects being used to their respective amounts of saliva, their respective swallowing threshold is not influenced by their amount of saliva. So a relative large salivary flow does not necessarily lead to swallow the food after a relative small number of chewing cycles. In agreement with Mioche et al. (2002) and Tarrega, Yven, Semon, & Salles (2011), more saliva was incorporated into the bolus for the hardest cheese for which chewing duration and number of chewing cycle are greater. Moreover we did not find any significant effect of fat content on the amount of saliva incorporated into the bolus even if the chewing parameters were higher for low fat cheeses than for high fat ones. Mioche et al. (2002) showed yet that saliva incorporated to the bolus is largely determined by chewing duration and to a small extent by muscular activity. According to our classification, HRG and MRG were differentiated on their amount of release of the more hydrophobic compound, nonan-2one (NO). The main parameters which significantly differentiated these two groups are the chewing parameters and mouth coating. The amount of product remaining in the mouth after food ingestion is significantly lower for MRG than for HRG. Nonan-2-one has a great affinity for fat and is more released during the post-swallowing phase (Repoux, Labouré et al., 2012). Thus the amount of nonan-2-one released could be logically related to the amount of product remaining in the mouth. Considering the less hydrophobic compound, ethyl propanoate (EP), subjects from MRG are closer to those belonging to HRG. This compound is more volatile than NO, has less affinity for fat and thus is more released during mastication (Repoux, Labouré et al., 2012). During the mastication step, we observed that velum opening was similar for these groups, which explains the similarity in the release patterns for EP between HRG and MRG. 5. Conclusion Our study allowed us to point out the main physiological parameters which explain the different amounts of aroma release during food consumption. Subjects who present the highest level of release whatever the aroma compound have the highest chewing activity, highest mouth coating and a velum opening at each mouth movement. Subjects with the lowest level of release whatever the aroma compounds present the lowest chewing activity and a velum opening at swallowing events only. An intermediate group of subjects was identified with a level of
700
H. Labouré et al. / Food Research International 64 (2014) 692–700
aroma release close to the high release group for the more hydrophilic compound due to a velum opening at each mouth movement during food consumption and close to low release group for the more hydrophobic compound due to a lower mouth coating after swallowing. A better understanding of food oral processing depending on both products and subjects allowed us to explain inter-individual differences in aroma release. These differences could then impact aroma perception and food preferences. Aknowledgements The authors gratefully acknowledge the French National Research Agency (ANR-07-PNRA-014), Fromageries Bel SA, Soredab (groupe Soparind Bongrain), the Association Nationale de la Recherche et de la Technologie (ANRT), the Regional Council of Burgundy France and the FEDER (European Funding for Regional Economical Development contract no. RB 29000350) for the financial support, ChemoSens Platform (Centre des Sciences du Goût et de l'Alimentation, Dijon, France) for the technical assistance in APCI–MS analyses, Laboratoire de Rhéologie (Université Joseph Fourier Grenoble, France) for the bolus rheology, Claude Yven for the electromyography, and the panellists who participated in this study. References Blissett, A., Hort, J., & Taylor, A. J. (2006). Influence of chewing and swallowing behavior on volatile release in two confectionary systems. Journal of Texture Studies, 37(5), 476–496. Buettner, A., Beer, A., Hannig, C., Settles, M., & Schieberle, P. (2002). Physiological and analytical studies on flavor perception dynamics as induced by the eating and swallowing process. Food Quality and Preference, 13(7–8), 497–504. Buettner, A., Otto, S., Beer, A., Mestres, M., Schieberle, P., & Hummel, T. (2008). Dynamics of retronasal aroma perception during consumption: Cross-linking on-line breath analysis with medico-analytical tools to elucidate a complex process. Food Chemistry, 108(4), 1234–1246. Doyennette, M., de Loubens, C., Deleris, I., Souchon, I., & Trelea, I. C. (2011). Mechanisms explaining the role of viscosity and post-deglutitive pharyngeal residue on in vivo aroma release: A combined experimental and modeling study. [Article]. Food Chemistry, 128(2), 380–390, http://dx.doi.org/10.1016/j.foodchem.2011.03.039. Doyennette, M., Déléris, I., Féron, G., Guichard, E., Souchon, I., & Trelea, I. C. (2014). Main individual and product characteristics influencing in-mouth flavour release during eating masticated food products with different textures: Mechanistic modelling and experimental validation. Journal of Theoretical Biology, 340, 209–221. Doyennette, M., Deleris, I., Saint-Eve, A., Gasiglia, A., Souchon, I., & Trelea, I. C. (2011). The dynamics of aroma compound transfer properties in cheeses during simulated eating conditions. Food Research International, 44(10), 3174–3181, http://dx.doi.org/10. 1016/j.foodres.2011.07.034. Feron, G., Ayed, C., Qannari, E. M., Courcoux, P., Labouré, H., & Guichard, E. (2014). Understanding aroma release from model cheeses by a statistical multiblock approach on oral processing. Plos One, 9(4), e93113, http://dx.doi.org/10.1371/journal.pone.0093113. Gaviao, M. B.D., Engelen, L., & van der Bilt, A. (2004). Chewing behavior and salivary secretion. European Journal of Oral Sciences, 112(1), 19–24. Gierczynski, I., Laboure, H., & Guichard, E. (2008). In vivo aroma release of milk gels of different hardnesses: Inter-individual differences and their consequences on aroma perception. Journal of Agricultural and Food Chemistry, 56(5), 1697–1703. Guichard, E. (2002). Interactions between flavor compounds and food ingredients and their influence on flavor perception. Food Reviews International, 18(1), 49–70. Haahr, A. -M., Bardow, A., Thomsen, C. E., Jensen, S. B., Nauntofte, B., & Bakke, M. (2004). Release of peppermint flavour compounds from chewing gum: Effect of oral functions. Physiology & Behavior, 82(2–3), 531–540. Hansson, A., Giannouli, P., & Van Ruth, S. (2003). The influence of gel strength on aroma release from pectin gels in a model mouth and in vivo, monitored with proton-
transfer-reaction mass spectrometry. Journal of Agricultural and Food Chemistry, 51(16), 4732–4740. Hodgson, M., Linforth, R. S. T., & Taylor, A. J. (2003). Simultaneous real-time measurements of mastication, swallowing, nasal airflow, and aroma release. Journal of Agricultural and Food Chemistry, 51(17), 5052–5057. Le Quere, J. -L., Gierczynski, I., Langlois, D., & Semon, E. (2006). Nosespace with an ion trap mass spectrometer: Quantitative aspects (book section). In W. L. P. Bredie, & M.A. Petersen (Eds.), (Amsterdam (NLD) ed.). Flavour science: Recent advances and trends, Vol. 43. (pp. 589–592). . Oxford (GBR): Elsevier. Linforth, R. S. T., Baek, I., & Taylor, A. J. (1999). Simultaneous instrumental and sensory analysis of volatile release from gelatine and pectine/gelatine gels. Food Chemistry, 65, 77–83. Mestres, M., Kieffer, R., & Buettner, A. (2006). Release and perception of ethyl butanoate during and after consumption of whey protein gels: Relation between textural and physiological parameters. Journal of Agricultural and Food Chemistry, 54(5), 1814–1821. Mestres, M., Moran, N., Jordan, A., & Buettner, A. (2005). Aroma release and retronasal perception during and after consumption of flavored whey protein gels with different textures. 1. In vivo release analysis. Journal of Agricultural and Food Chemistry, 53(2), 403–409. Mioche, L., Bourdiol, P., Monier, S., & Martin, J. F. (2002). The relationship between chewing activity and food bolus properties obtained from different meat textures. Food Quality and Preference, 13(7–8), 583–588. Mishellany-Dutour, A., Woda, A., Labouré, H., Bourdiol, P., Lachaze, P., & Guichard, E. (2012). Retro-nasal aroma release is correlated with variations in the in-mouth air cavity volume after empty deglutition. PloS One, 7(7), e41276, http://dx.doi.org/10. 1371/journal.pone.0041276. Pionnier, E., Chabanet, C., Mioche, L., Le Quere, J. L., & Salles, C. (2004). 1. In vivo aroma release during eating of a model cheese: Relationships with oral parameters. Journal of Agricultural and Food Chemistry, 52(3), 557–564. Pivk, U., Ulrih, N.P., Juillerat, M.A., & Raspor, P. (2008). Assessing lipid coating of the human oral cavity after ingestion of fatty foods. Journal of Agricultural and Food Chemistry, 56(2), 507–511. Poette, J., Mekoue, J., Neyraud, E., Berdeaux, O., Renault, A., & Guichard, E. (2013). Fat sensitivity in humans: oleic acid detection threshold is linked to saliva composition and oral volume. Flavour and Fragrance Journal, 29(1), 39–49. Repoux, M., Labouré, H., Courcoux, P., Andriot, I., Sémon, E., & Yven, C. (2012). Combined effect of cheese characteristics and food oral processing on in vivo aroma release. Flavour and fragrance journal, 27(6 (special issue paper)), 414–423, http://dx.doi. org/10.1002/ffj.3110. Repoux, M., Sémon, E., Feron, G., Guichard, E., & Labouré, H. (2012). Inter-individual variability in aroma release during sweet mint consumption. Flavour and Fragrance Journal, 27(1), 40–46, http://dx.doi.org/10.1002/ffj.2077. Repoux, M., Septier, C., Palicki, O., Guichard, E., Feron, G., & Labouré, H. (2012). Solid cheese consumption: Quantification of oral coating. Archives of Oral Biology, 57(1), 81–86, http://dx.doi.org/10.1016/j.archoralbio.2011.07.011. Ruijschop, R., Burgering, M. J. M., Jacobs, M.A., & Boelrijk, A. E. M. (2009). Retro-nasal aroma release depends on both subject and product differences: A link to food intake regulation? Chemical Senses, 34(5), 395–403, http://dx.doi.org/10.1093/chemse/ bjp011. Salles, C., Chagnon, M. -C., Feron, G., Guichard, E., Labouré, H., & Morzel, M. (2011). Inmouth mechanisms leading to flavor release and perception. Critical Reviews in Food Science and Nutrition, 51(1), 67–90, http://dx.doi.org/10.1080/ 10408390903044693. Tarrega, A., Yven, C., Sémon, E., & Salles, C. (2008). Aroma release and chewing activity during eating different model cheeses. International Dairy Journal, 18(8), 849–857. Tarrega, A., Yven, C., Semon, E., & Salles, C. (2011). In-mouth aroma compound release during cheese consumption: Relationship with food bolus formation. International Dairy Journal, 21(5), 358–364, http://dx.doi.org/10.1016/j.idairyj.2010.12.010. van der Bilt, A., Engelen, L., Abbink, J., & Pereira, L. J. (2007). Effects of adding fluids to solid foods on muscle activity and number of chewing cycles. European Journal of Oral Sciences, 115, 198–205. van Ruth, S. M., & Roozen, J. P. (2000). Influence of mastication and saliva on aroma release in a model mouth system. Food Chemistry, 71(3), 339–345. Yven, C., Culioli, J., & Mioche, L. (2005). Meat bolus properties in relation with meat texture and chewing context. Meat Science, 70, 365–371. Yven, C., Patarin, J., Magnin, A., Labouré, H., Repoux, M., & Guichard, E. (2012). Consequences of individual chewing strategies on bolus rheological properties at the swallowing threshold. Journal of Texture Studies, 43(4), 309–318, http://dx.doi.org/ 10.1111/j.1745-4603.2011.00340.x.