- Email: [email protected]
Effects of added glutamate on liking for novel food flavors
Appetite 42 (2004) 143–150 www.elsevier.com/locate/appet
Research Report
Effects of added glutamate on liking for novel food flavors John Prescott* Sensory Science Research Centre, University of Otago, Dunedin, New Zealand Received 15 September 2002; revised 28 June 2003; accepted 28 August 2003
Abstract Adding glutamate to foods increases their umami quality, their acceptability and their consumption. The functional significance of this palatability is unclear. Other highly palatable substances, e.g. sugar and fats, also increase liking for novel flavors with which they are repeatedly paired, especially when ingested. This is thought to reflect the rewarding effects of sugar and fat energy, post-ingestion. To determine if a liking for novel flavors can also be conditioned using glutamate, 44 subjects rated 10 ml samples of three novel soups for liking and familiarity, both before and after seven daily exposures to each of two soup flavors—one with added monosodium L -glutamate (MSG) (0.5% w/w; MSG þ ) and one without (MSG 2 ). During exposure, subjects received either a 250 ml bowl of soup (Consume group) or a 10 ml sample (Taste group). There were no significant differences as a function of samples or groups, despite some trends for changes in liking to be higher in the consumed MSG þ condition. In a second experiment, 69 subjects were divided into three groups (Consume MSG þ ; Consume MSG 2 ; Taste MSG þ ) in which they received nine exposures to one novel soup flavor. The Consume MSG þ group showed a significantly greater increase in liking than either the Consume MSG 2 or the Taste MSG þ groups, which did not differ. Changes in familiarity ratings reflected amount consumed, not MSG content. Pairing glutamate with a novel flavor can condition liking for that flavor. While post-ingestive effects of glutamate may be rewarding, flavor conditioning cannot be ruled out. q 2003 Elsevier Ltd. All rights reserved. Keywords: Glutamate; Umami; Preference; Flavor; Post-ingestional effects; Taste
Introduction The mammalian gustatory system responds to a set of basic qualities—traditionally, sweet, sour, salty and bitter— that elicit relatively fixed hedonic responses. Many species, including humans, prefer sweetness and reject bitterness. These responses are evident from birth, or shortly thereafter, and are probably innate (Steiner, Glaser, Hawilo, & Berridge, 2001). Their origin appears to lie in the adaptive value in signalling the nutritional implications of taste qualities. The palatability of any taste compound, and the pattern of responsiveness of taste cells in the rat brainstem, is inversely related to its toxicity (Scott & Mark, 1987). Highly toxic compounds are rejected as unpalatable, primarily due to bitterness; highly nutritive substances with low toxicity are preferred, mainly because they are sweet. This neural and behavioral organisation has led to the hypothesis that preferences for tastes are a means by which * Address: School of Psychology, James Cook University, Cairns, Qld 4870, Australia. E-mail address: [email protected] (J. Prescott). 0195-6663/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2003.08.013
physiological wellbeing is maintained (Scott, 1992). Consistent with this, metabolic states can modulate taste palatability—preferred levels of salt in foods increase following salt depletion (Beauchamp, Bertino, Burke, & Engleman, 1990), while sweetness becomes less pleasant following glucose consumption (Cabanac, 1971). Glutamate, derived from the non-essential amino acid, glutamic acid, is an important contributor to food flavors. The addition of sauces (soy; Worcestershire), cheese (especially Parmesan), tomatoes, mushrooms, or meat, fish and vegetable stocks to foods all have the effect of increasing glutamate levels (Ninomiya, 1998; Yamaguchi, 1991). Since early 1900s, monosodium L -glutamate (MSG) has been commercially manufactured for use as a flavor enhancer, and there is ample evidence that adding MSG to suitable foods increases their palatability (Bellisle et al., 1991; Yamaguchi & Takahashi, 1984a,b) and consumption (Bellisle, 1998; Rogers & Blundell, 1990; Schiffman, 1998). Such hedonic changes are probably partly mediated by changes in the sensory properties of the foods, including increase in richness, savouriness and mouthfeel qualities (Fuke & Shimizu, 1993; Prescott, 2001; Yamaguchi, 1991).
144
J. Prescott / Appetite 42 (2004) 143–150
Although adding MSG also increases the saltiness of the foods (because of the Naþ), this cannot account for other sensory and hedonic changes. Soups containing added glutamate were found to be preferred to those without, even when the amount of Naþ was equivalent (Okiyama & Beauchamp, 1998). In addition, calcium glutamate is also able to enhance palatability (Bellisle, Dartois, & Broyer, 1992). There is considerable evidence that the taste quality associated with glutamate in foods, known as umami (a Japanese word, translated approximately as ‘savoury deliciousness’), is recognised and coded as a unique quality by the gustatory system (Bayliss & Rolls, 1991; Faurion, 1991; Ninomiya, Tanimukai, Yoshida, & Funakoshi, 1991; Rolls, 2000; Rolls, Critchley, Wakeman, & Mason, 1996; Yamaguchi & Komata, 1987). Recently a strong candidate for a specific umami taste receptor was identified. Chaudhari, Landin, and Roper (2000) showed that a variant of a particular glutamate receptor is present preferentially in rat taste buds, and responds specifically to compounds possessing umami quality. The ability of umami to elicit characteristic hedonic responses in human neonates, prior to experience with foods (Steiner et al., 2001) and across cultures (Prescott, 1998) is also suggestive of a basic taste. It has been proposed that the positive hedonic value of the umami taste, and its distinct coding in the taste system, reflect its function as a signal for protein in foods (Naim, Ohara, Kare, & Levinson, 1991). Consistent with this notion, persons with low nutritional/protein status preferred higher concentrations of MSG in solution than did those who were better nourished (Murphy, 1987). Other findings tend to be at odds with this interpretation, however. MSG preference in rats was highest during protein repletion (Torii, Mimura, & Yugari, 1987)—just the reverse of what would be expected if umami preference was a way of encouraging protein consumption. In addition to its role in proteins, however, glutamate is also involved in a variety of essential metabolic processes, including synthesis of other amino acids, as a precursor of glutamine (involved in amino acid homeostasis and gluconeogenesis) and glutathione (an important antioxidant), and in carbohydrate metabolism (Mori, Kawada, Ono, & Torii, 1991; Reeds, Burrin, Stoll, & Jahoor, 2000; Windmueller & Spaeth, 1975). Glutamate may also be a source of energy, since dietary glutamate was found to be the largest contributor to intestinal energy generation in studies of piglet nutrition and gut metabolism (Reeds et al., 2000). Consistent with these findings, in humans, umami taste, like sweetness, can trigger increase in pancreatic secretions (Naim et al., 1991), a cephalic phase response associated with preparation for nutrient utilisation (Mattes, 1997). One indication of the metabolic importance of a nutrient is its ability to produce a liking for foods or flavors with which it is repeatedly associated. Recovery from protein (Baker & Booth, 1989) and amino acid (Booth & Simson,
1971) depletion was found, in animal studies, to enhance preference for flavors/foods with which recovery from the depletion was associated. Conditioned liking has also been observed as a result of pairing of flavors with the postingestional consequences of nutrients that supply energy (Booth, 1991; Sclafani, 1995; Zellner, 1991). In animal studies, ingested or infused carbohydrates (Ackroff & Sclafani, 1994; Mehiel & Bolles, 1984; Perez, Lucas, & Sclafani, 1998), fats (Lucas & Sclafani, 1989) and proteins (Perez, Lucas, & Sclafani, 1996) were able to increase liking for flavors with which they were paired. Human research has also shown that novel flavors were preferred when paired with the consumption of carbohydrates (Birch, McPhee, Steinberg, & Sullivan, 1990), fat (Johnson, McPhee, & Birch, 1991; Kern, McPhee, Fisher, Johnson, & Birch, 1993) or protein (Gibson, Wainwright, & Booth, 1995). Kern et al. (1993) found increased liking for a flavor associated with fat intake, as compared to simple pairing of the flavor with the same fat level without ingestion. The usual interpretation of this effect is that it is mediated by the post-ingestive rewarding effects of the provision of energy. However, increased liking for flavors can also be demonstrated both through simple exposure to a novel flavor (Sullivan & Birch, 1990) and also through association with an already preferred stimulus, such as sucrose (flavor – flavor, or evaluative, conditioning), even when postingestional consequences can be ruled out (Zellner, Rozin, Aron, & Kulish, 1983). The present research aimed to determine if glutamate can condition increased liking for novel flavors to which it is added. One previous study has addressed the issue of whether addition of MSG to foods produces changes in acceptability with repeat exposure. Although Bellisle (1998) added MSG to relatively novel foods and found an increased intake over four weekly exposures, relative to these same foods with no added MSG, the study did not control for an effect purely due to the increased palatability of foods with MSG. By comparing liking for a flavor to which MSG is added in consumed and not consumed conditions, the present research also sought to determine if post-ingestive absorption of glutamate is responsible for any conditioned liking effects.
Methods Experiment 1 Subjects The participants were 11 males and 33 females (mean age: 24.5 years; range: 19 –48 years), students and staff at the University of Otago. Of these, two failed to attend all laboratory sessions, and their data were excluded from analysis. At the time of testing, no participant reported that they were severely restricting their diet, or the presence of
J. Prescott / Appetite 42 (2004) 143–150
current or recent illness that could interfere with perception of odors or tastes. Samples Pilot studies were carried out to identify three soup flavors that were both relatively novel and of low – moderate initial palatability, and to ensure that the impact of added MSG did not differ between the selected soups. Soups using two vegetables—chickpea and spinach—were prepared using two different soup bases. The soup bases, chosen because of their novelty for these subjects, were a dried fungus base and a dried flower base, both used in Chinese cooking. Added to water, these formed the soup stock in which the other ingredients were cooked. The soup/base combinations used were chickpea/dried flower base; chickpea/dried fungus base; and spinach/dried fungus base. All soups were prepared prior to the study, frozen, and then thawed overnight prior to serving at a constant temperature of 50 8C. Prior to freezing, MSG (0.5% w/w) was added to half the soup samples. This amount is within the range typically added to foods to enhance flavor (Yamaguchi, 1991). Procedure Subjects were divided into two groups. One group consumed 250 ml bowls of the soup during exposure sessions (Consume group), while the second group received 10 ml samples (Taste group). The second group was a control for simple exposure to the soup flavor plus MSG, similar to the group used by Kern et al. (1993). Subjects attended a pre-test session, 14 exposure sessions in total, and a post-test. As far as possible, exposure sessions were conducted on consecutive weekdays, from 12 to 2 p.m., and subjects were instructed not to consume any food 2 h prior to each session. In both the pre- and post-tests, subjects tasted 10 ml of the three soup flavors (without added MSG) and rated liking and familiarity for these on 100 mm line scales with descriptive anchors at each end (dislike extremely/like extremely; extremely unfamiliar/extremely familiar). Samples were presented monadically in balanced order, and subjects rinsed with water between samples. Since the ability to observe changes in familiarity and liking for these soups depended on initial levels being low – moderate, this pre-test rating also acted to identify unsuitable subjects with high preference and familiarity ratings (rating . 60 mm on a 100 mm scale) prior to the study commencing. No one was eliminated from the study at this stage. One of the three soup flavors acted as a control for changes that might occur over time, independent of exposure, and was received only in the pre- and post-tests. During the exposure sessions, subjects in both groups received two of the three soups — one with 0.5% added MSG (MSG þ ) and one without added MSG (MSG 2 )— on alternate days, that is, seven exposures per soup flavor. The association of the different soup flavors with the
145
MSG þ , MSG 2 , and Control conditions was balanced across subjects. Subjects in the Consume group received bowls of soup (250 ml) in each session, while those in the Taste group received 10 ml samples. To mask the purpose of the study and to ensure that attention was paid to the soups during each session, participants were asked to rate the flavor intensity of the soups on 100 mm line scales with descriptive anchors at each end (Extremely weak/Extremely strong). These data were not analysed. Analysis Familiarity and liking ratings were subjected to repeated measures ANOVA (SPSS V. 10), with Groups (Consume/ Taste) as a between-subjects factor, and Samples (MSG þ / MSG 2 /Control) and Exposure (Pre-exposure/post-exposure) as within-subjects factors. Post-hoc comparisons of means were undertaken using Tukey’s HSD Test. Results Familiarity. Initial ratings of familiarity were low – moderate (mean values: 33.9 –50 mm on a 100 mm scale, depending on group and sample type). Fig. 1 shows mean changes from pre- to post-test in familiarity ratings for each soup (MSG þ ; MSG 2 ; Control) in the two groups (Consume; Taste). Familiarity increased following exposure (Fð1; 42Þ ¼ 49:266; p , 0:0001), but there were no effects of either Samples or Groups. None of the confidence intervals (CIs; 95%) for the posttest –pre-test difference scores contained zero, reflecting familiarity increases for all samples, irrespective of MSG content, or whether the samples were consumed or tasted. Clearly, even a single exposure (Control samples) was sufficient to significantly increase familiarity. Liking. Subjects initially rated the samples as moderately disliked (mean values: 36.1– 39.6 on a 100 mm scale, depending on group and sample type). Fig. 2 shows the mean changes from pre- to post-test in liking ratings for each soup (MSG þ ; MSG 2 ; Control) in the two groups (Consume; Taste). All samples in both groups showed mean increases in liking from pre- to post-test, reflected in an ANOVA main effect of Exposure (Fð1; 42Þ ¼ 22:025; p , 0:0001). Despite the apparently greater increase in liking for the MSG þ sample in the Consume group (Fig. 2), there were no significant main effects or interactions involving Groups or Samples. Although this experiment failed to find significant differences between groups or samples in changes in liking for these soups, we conducted a second experiment based on the apparent trend towards an effect for the consumed MSG þ sample. In considering the change scores following exposure, only these scores had 95% CIs that did not contain zero, suggesting that liking only reliably increased for the MSG þ samples when consumed. In this experiment, the fact that subjects were exposed to two somewhat similar, unfamiliar soups (but differing in
146
J. Prescott / Appetite 42 (2004) 143–150
Fig. 1. Mean (þSEM) changes (post-test–pre-test) in ratings of familiarity of the flavor of the samples (MSG þ ; MSG 2 ; Control) in the Consume and Taste groups in Experiment 1.
MSG content) on alternate days may have diluted any differences, especially between samples. Hence, the second experiment used an experimental design in which separate groups of subjects each received exposure to only one soup
flavor. Controls for the post-ingestive effects of added glutamate consisted of one group that consumed an MSG 2 soup, and another that received limited exposure (10 ml samples) to an MSG þ soup.
Fig. 2. Mean (þ SEM) changes (post-test– pre-test) in ratings of liking for the flavor of the samples (MSG þ ; MSG 2 ; Control) in the Consume and Taste groups in Experiment 1.
J. Prescott / Appetite 42 (2004) 143–150
Experiment 2 Subjects Participants were 20 males and 49 females (mean age: 32.9 years; range: 19 – 63 years), students and staff at the University of Otago. None had participated in Experiment 1. Another eight participants failed to attend all sessions, and their data were excluded from analysis. At the time of testing, no participant reported that they were severely restricting their diet, or the presence of current or recent illness that could interfere with perception of odors or tastes. Samples Two soup/base combinations from Experiment 1 (Chickpea/dried flower base and Spinach/dried fungus base) were used. All soups were prepared prior to the study, frozen, and then thawed overnight prior to serving at a constant temperature of 50 8C. MSG (0.5% w/w) was added to half of the soup samples prior to freezing. Procedure Subjects were divided into three groups—two Consume groups (MSG þ ; MSG 2 ) and a Taste group (MSG þ ). There were 24 subjects in Group 1 (Consume MSG þ ), 24 in Group 2 (Consume MSG 2 ), and 21 in Group 3 (Taste MSG þ ). During the pre-test and post-test, subjects tasted 10 ml of the two soup flavors (without added MSG) and rated their liking and familiarity for these on 100 mm line scales. Sample order was balanced and subjects rinsed with
147
water between samples. Following the pre-test, subjects attended nine exposure sessions: three sessions per week, on successive weeks. Sessions were conducted from 12 to 2 p.m. each day. During these exposure sessions, each subject received only one soup flavor. Allocation of each flavor to control or exposure conditions was balanced across subjects. The two Consume groups received bowls of soup (250 ml) while the Taste group received 10 ml samples. Subjects were instructed not to consume any food 2 h prior to each session. As in Experiment 1, following consumption or tasting, subjects rated the flavor intensity of the soups on a 100 mm line scale with descriptive anchors at each end. Analysis Familiarity and liking ratings were subjected to repeated measures ANOVA (SPSS V. 10), with Exposure (Preexposure/post-exposure) and Samples (Exposed/Nonexposed) as within-subjects factors, and Groups (Consume MSG þ /Consume MSG 2 /Taste MSG þ ) as the between-subjects factor. Post-hoc comparisons of means were undertaken using Tukey’s HSD Test. Results Familiarity. As in Experiment 1, subjects were initially relatively unfamiliar with these flavors (mean values: 25.4– 37.1 mm on 100 mm scale, depending on group and sample type). Fig. 3 shows the mean changes from pre- to post-test in familiarity ratings for the exposed and non-exposed
Fig. 3. Mean (þ SEM) changes (post-test–pre-test) in ratings of familiarity of the flavor of the exposed and non-exposed samples in the Consume MSG þ , Consume MSG 2 and Taste MSG þ groups in Experiment 2.
148
J. Prescott / Appetite 42 (2004) 143–150
samples in the three groups (Consume MSG þ ; Consume MSG 2 ; Taste MSG þ ). There was a significant increase in familiarity from pretest to post-test (Exposure main effect: Fð1; 66Þ ¼ 155:48; p , 0:0001) in particular for those flavors that received repeated exposure (Exposure £ Samples interaction; Fð1; 66Þ ¼ 26:01; p , 0:0001). While there was no overall difference between Groups ðFð2; 66Þ ¼ 0:374Þ; the Group £ Exposure effect (Fð2; 66Þ ¼ 8:52; p , 0:005) was significant, with both Consume groups showing greater overall increase in familiarity from pre- to post-test than the Taste group. Liking. Initial ratings of liking were low –moderate (mean values: 32.9– 35.8 mm on a 100 mm scale, depending on group and sample type). Fig. 4 shows the mean changes from pre- to post-test in liking ratings for the exposed and non-exposed samples in the three groups (Consume MSG þ ; Consume MSG 2 ; Taste MSG þ ). ANOVA revealed significant increases in liking ratings from pre- to post-test (Fð1; 66Þ ¼ 13:14; p , 0:005), especially for the exposed flavors (Exposure £ Samples interaction: Fð1; 66Þ ¼ 23:73; p , 0:0001). Neither the Groups main effect ðFð2; 66Þ ¼ 0:18Þ; nor the Groups £ Exposure interaction ðFð2; 66Þ ¼ 0:40Þ were significant. However, there were significant effects for Samples (Fð1; 66Þ ¼ 16:46; p , 0:0001), Samples £ Groups (Fð2; 66Þ ¼ 8:37; p , 0:005), and Exposure £ Samples £ Groups (Fð2; 66Þ ¼ 6:38; p , 0:005). Post-hoc comparisons of means revealed that, for the exposed flavors, there was
a significant increase in liking ratings for the Consume MSG þ group, but not for either the Consume MSG 2 or the Taste MSG þ groups. While the three groups did not differ in pre-test ratings, post-test ratings for the Consume MSG þ group were substantially greater than those of the other groups, and suggestive of a relatively high degree of liking (mean: 63.17). By contrast, the mean post-test ratings of the Consume MSG 2 and the Taste MSG þ groups remained modest (44.29 and 47.76, respectively). For the non-exposed flavors, the only significant difference was higher post-test ratings in the Consume MSG 2 group than in the Consume MSG þ group.
Discussion As expected, familiarity ratings were increased by exposure to these novel flavors. These changes were even apparent in the case of the single exposure received by the ‘non-exposed’ flavors. While increased familiarity can facilitate liking through a reduction in neophobia (Pliner, 1982), familiarity changes alone cannot account for the present pattern of liking results. In Experiment 2, familiarity was a graded function of exposure to the sample, and not of MSG content, since it was significantly greater in the two Consume groups than in the Taste group. In Experiment 2, although all exposed samples showed an increase in liking for the associated flavor, this was significantly greater for the consumed MSG þ samples
Fig. 4. Mean (þSEM) changes (post-test–pre-test) in ratings of liking for the flavor of the exposed and non-exposed samples in the Consume MSG þ , Consume MSG 2 and Taste MSG þ groups in Experiment 2.
J. Prescott / Appetite 42 (2004) 143–150
than for either the tasted MSG þ samples or the consumed MSG 2 samples. The mean ratings indicated, moreover, that this sample became relatively well-liked following exposure. This experiment thus provides evidence that pairing a novel food flavor with glutamate can condition a liking for that flavor. Previous findings that ingested carbohydrates (Birch et al., 1990) and fat (Johnson et al., 1991; Kern et al., 1993) are able to condition liking for novel flavors have been interpreted in terms of the reward value of the energy that these nutrients provide. The fact that in Experiment 2 the consumed MSG þ soup was liked more than the tasted MSG þ soup implies that post-ingestional absorption of additional glutamate in foods is also primarily responsible for the enhanced liking and therefore similarly rewarding. In an attempt to control increased palatability associated with the effects of added MSG, the consumed MSG þ samples in both experiments were compared with the (presumably) equally palatable MSG þ samples that were only tasted. Since the energy due to carbohydrates contained within the basic soup formulation might also be expected to condition liking, comparisons were made between the consumed MSG þ samples and consumed samples that did not contain added MSG. It is unlikely that the natural glutamate content played a significant role here. Spinach, for example, contains approximately 0.05% glutamate (Ninomiya, 1998), around 10% of that added to the soup. One uncontrolled factor that limits the extent to which post-ingestive effects can be inferred is the absolute degree of exposure to the soup plus MSG in the different conditions/groups. It is possible that, in contrast to the nine discrete exposure sessions to a small sample in Experiment 2’s MSG þ Taste group, each of the multiple spoonfuls of the consumed MSG þ soup within each session acted as a separate, discrete exposure. Multiple exposures within a session can increase flavor liking through flavor –flavor conditioning (Zellner et al., 1983). Moreover, the degree of flavor liking has been shown to be a function of the number of discrete exposures (Pliner, 1982). Hence, the possibility exists that the consumed MSG þ samples were simply more effective at increasing liking through higher numbers of pairings of the novel flavor with the positive palatability of added MSG (that is, through flavor – flavor conditioning). The present study does not, of course, rule out the possibility that post-ingestive conditioning was responsible for the observed changes in liking. While it is not clear how the degree of exposure might be made equivalent in taste and consume conditions in this type of study, this issue might be resolved through manipulation of states of hunger. Post-ingestive conditioning in animal models is responsive to whether or not an animal is food-deprived (Ackroff, Vigorito, & Sclafani, 1990; Fedorchak & Bolles, 1987). Fedorchak and Bolles (1987) found not only that caloriemediated preferences to be stronger than flavor-mediated preferences, but that the former are also enhanced by
149
deprivation during preference testing, whereas flavormediated preferences are not. Kern et al. (1993) used subjects who were overnight food-deprived, both during exposure and at pre- and post-test, and found association with fat to be effective at increasing liking for novel flavors in this state, whereas they were ineffective in a state of satiety. Although the subjects in the present experiments were asked to refrain from eating for 2 h prior to each session, no measures of hunger were obtained. The present studies demonstrate that the addition of glutamate, like sugar and fat, can condition a liking for novel flavors. This may be the mechanism behind previous findings of increase in consumption of glutamate containing novel foods over time (Bellisle, 1998). The conditioned liking may reflect the role that dietary glutamate plays as a precursor in key metabolic processes or its utilisation as an energy source in gut metabolism. However, comparisons of subjects in deprived and non-deprived states during conditioning and testing may be the crucial test of whether post-ingestive processes are involved in such conditioned liking.
Acknowledgements Thanks to Bianca Turnbull who collected the data in these experiments. This research was supported by a grant from the International Glutamate Technical Committee.
References Ackroff, K., & Sclafani, A. (1994). Flavor preferences conditioned by intragastric infusions of dilute polycose solutions. Physiology and Behavior, 55, 957 –962. Ackroff, K., Vigorito, M., & Sclafani, A. (1990). Fat appetite in rats: The response of infant and adult rats to nutritive and non-nutritive oil emulsions. Appetite, 15, 171–188. Baker, B. J., & Booth, D. A. (1989). Genuinely olfactory preferences conditioned by protein repletion. Appetite, 13, 223–227. Bayliss, L. L., & Rolls, E. T. (1991). Responses of neurons in the primate taste cortex to glutamate. Physiology and Behavior, 49, 973–979. Beauchamp, G. K., Bertino, M., Burke, D., & Engleman, K. (1990). Experimental sodium depletion and salt taste in normal human volunteers. American Journal of Clinical Nutrition, 51, 881– 889. Bellisle, F. (1998). Nutritional effects of umami in the human diet. Food Reviews International, 14, 309– 320. Bellisle, F., Dartois, A. M., & Broyer, M. (1992). Two studies on the acceptability of calcium glutamate as a potassium-free sodium substitute in children. Journal of Renal Nutrition, 2, 42–46. Bellisle, F., Monneuse, M. O., Chabert, M., Laure-Achagiotis, C., Lanteaume, M. T., & Louis-Sylvestre, J. (1991). Monosodium glutamate as a palatability enhancer in the European diet. Physiology and Behavior, 49, 869–874. Birch, L. L., McPhee, L., Steinberg, L., & Sullivan, S. (1990). Conditioned flavor preferences in young children. Physiology and Behavior, 47, 501– 505. Booth, D. A. (1991). Learned ingestive motivation and the pleasures of the palate. In R. C. Bolles (Ed.), The hedonics of taste (pp. 29 –58). Hillsdale, NJ: Lawrence Erlbaum.
150
J. Prescott / Appetite 42 (2004) 143–150
Booth, D. A., & Simson, P. C. (1971). Food preferences acquired by association with variations in amino acid nutrition. Quarterly Journal of Experimental Psychology, 23, 135– 145. Cabanac, M. (1971). Physiological role of pleasure. Science, 173, 1103–1107. Chaudhari, N., Landin, A. M., & Roper, S. D. (2000). A metabotropic glutamate receptor variant functions as a taste receptor. Nature Neuroscience, 3, 113–119. Faurion, A. (1991). Are umami taste receptor sites structurally related to glutamate CNS receptor sites? Physiology and Behavior, 49, 905–912. Fedorchak, P. M., & Bolles, R. C. (1987). Hunger enhances the expression of calorie- but not taste-mediated conditioned flavor preferences. Journal of Experimental Psychology: Animal Behavior Processes, 13, 73–79. Fuke, S., & Shimizu, T. (1993). Sensory and preference aspects of umami. Trends in Food Science and Technology, 4, 246–251. Gibson, E. L., Wainwright, C. J., & Booth, D. A. (1995). Disguised protein in lunch after low-protein breakfast conditions food-flavor preferences dependent on recent lack of protein intake. Physiology and Behavior, 58, 363 –371. Johnson, S. L., McPhee, L., & Birch, L. L. (1991). Conditioned preferences: Young children prefer flavors associated with high dietary fat. Physiology and Behavior, 50, 1245–1251. Kern, D. L., McPhee, L., Fisher, J., Johnson, S., & Birch, L. L. (1993). The postingestive consequences of fat condition preferences for flavors associated with high dietary fat. Physiology and Behavior, 54, 71– 76. Lucas, F., & Sclafani, A. (1989). Favor preferences conditioned by intragastric fat infusions in rats. Physiology and Behavior, 46, 403– 412. Mattes, R. D. (1997). Physiologic responses to sensory stimulation by food: Nutritional implications. Journal of the American Dietetic Association, 97, 406 –413. Mehiel, R., & Bolles, R. C. (1984). Learned flavor preferences based on caloric outcomes. Animal Learning and Behavior, 12, 421– 427. Mori, M., Kawada, T., Ono, T., & Torii, K. (1991). Taste preference and protein nutrition and L -amino acid homeostasis in male Sprague– Dawley rats. Physiology and Behavior, 49, 987–996. Murphy, C. (1987). Flavor preference for monosodium glutamate and casein hydrolysate in young and elderly persons. In Y. Kawamura, & M. R. Kare (Eds.), Umami: A basic taste (pp. 139–151). New York: Marcel Dekker. Naim, M., Ohara, I., Kare, M. R., & Levinson, M. (1991). Interaction of MSG taste with nutrition: Perspectives in consummatory behavior and digestion. Physiology and Behavior, 49, 1019–1024. Ninomiya, K. (1998). Natural occurrence. Food Reviews International, 14, 177– 211. Ninomiya, Y., Tanimukai, T., Yoshida, S., & Funakoshi, M. (1991). Gustatory neural responses in preweanling mice. Physiology and Behavior, 49, 913– 918. Okiyama, A., & Beauchamp, G. K. (1998). Taste dimensions of monosodium glutamate (MSG) in a food system: Role of glutamate in young American subjects. Physiology and Behavior, 65, 177–181. Perez, C., Lucas, F., & Sclafani, A. (1996). Carbohydrate- and proteinconditioned flavor preferences: Effects of nutrient preloads. Physiology and Behavior, 59, 467–474. Perez, C., Lucas, F., & Sclafani, A. (1998). Increased flavor acceptance and preference conditioned by the postingestive actions of glucose. Physiology and Behavior, 64, 483–492. Pliner, P. (1982). The effects of mere exposure on liking for edible substances. Appetite, 3, 283 –290.
Prescott, J. (1998). Comparisons of taste perceptions and preferences of Japanese and Australian consumers: Overview and implications for cross-cultural sensory research. Food Quality and Preference, 9, 393 –402. Prescott, J. (2001). Taste hedonics and the role of umami. Food Australia, 53, 550–554. Reeds, P. J., Burrin, D. G., Stoll, B., & Jahoor, F. (2000). Intestinal glutamate metabolism. Journal of Nutrition, 130, 978S–982S. Rogers, P. J., & Blundell, J. (1990). Umami and appetite: Effects of monosodium glutamate on hunger and food intake in human subjects. Physiology and Behavior, 48, 801–804. Rolls, E. T. (2000). The representation of umami taste in the taste cortex. Journal of Nutrition, 130, 960S–965S. Rolls, E. T., Critchley, H., Wakeman, E. A., & Mason, R. (1996). Responses of neurons in the primate taste cortex to the glutamate ion and to inosine 50 -monophosphate. Physiology and Behavior, 59, 991 –1000. Schiffman, S. S. (1998). Sensory enhancement of foods for the elderly with monosodium glutamate. Food Reviews International, 14, 321–334. Sclafani, A. (1995). How food preferences are learned—animal laboratory models. Proceedings of the Nutrition Society, 54, 419–427. Scott, T. R. (1992). Taste, feeding, and pleasure. Progress in Psychobiology and Physiological Psychology, 15, 231–291. Scott, T. R., & Mark, G. P. (1987). The taste system encodes stimulus toxicity. Brain Research, 414, 197–203. Steiner, J. E., Glaser, D., Hawilo, M. E., & Berridge, K. C. (2001). Comparative expression of hedonic impact: Affective reactions to taste by human infants and other primates. Neuroscience and Biobehavioral Reviews, 25, 53 –74. Sullivan, S. A., & Birch, L. L. (1990). Pass the sugar, pass the salt: Experience dictates preference. Developmental Psychology, 26, 546 –551. Torii, K., Mimura, T., & Yugari, Y. (1987). Biochemical mechanism of umami taste perception and effect of dietary protein on the taste preference for amino acids and sodium chloride in rats. In Y. Kawamura, & M. R. Kare (Eds.), Umami: A basic taste (pp. 125 –138). New York: Marcel Dekker. Windmueller, H. G., & Spaeth, A. E. (1975). Intestinal metabolism of glutamine and glutamate from the lumen as compared to glutamine from blood. Archives of Biochemistry and Biophysics, 171, 662–672. Yamaguchi, S. (1991). Basic properties of umami and effects on humans. Physiology and Behavior, 49, 833–841. Yamaguchi, S., & Komata, Y. (1987). Independence and primacy of umami as compared with the four basic tastes. Annals of the New York Academy of Science, 510, 725–726. Yamaguchi, S., & Takahashi, C. (1984a). Interactions of monosodium glutamate and sodium chloride on saltiness and palatability. Journal of Food Science, 49, 82 –85. Yamaguchi, S., & Takihashi, C. (1984b). Hedonic functions of monosodium glutamate and four basic taste substances used at various concentration levels in single and complex systems. Agricultural and Biological Chemistry, 48, 1077– 1081. Zellner, D. A. (1991). How foods get to be liked: Some general mechanisms and some special cases. In R. C. Bolles (Ed.), The hedonics of taste (pp. 199 –217). Hillsdale, NJ: Lawrence Erlbaum. Zellner, D. A., Rozin, P., Aron, M., & Kulish, C. (1983). Conditioned enhancement of human’s liking for flavor by pairing with sweetness. Learning and Motivation, 14, 338 –350.
Research Report
Effects of added glutamate on liking for novel food flavors John Prescott* Sensory Science Research Centre, University of Otago, Dunedin, New Zealand Received 15 September 2002; revised 28 June 2003; accepted 28 August 2003
Abstract Adding glutamate to foods increases their umami quality, their acceptability and their consumption. The functional significance of this palatability is unclear. Other highly palatable substances, e.g. sugar and fats, also increase liking for novel flavors with which they are repeatedly paired, especially when ingested. This is thought to reflect the rewarding effects of sugar and fat energy, post-ingestion. To determine if a liking for novel flavors can also be conditioned using glutamate, 44 subjects rated 10 ml samples of three novel soups for liking and familiarity, both before and after seven daily exposures to each of two soup flavors—one with added monosodium L -glutamate (MSG) (0.5% w/w; MSG þ ) and one without (MSG 2 ). During exposure, subjects received either a 250 ml bowl of soup (Consume group) or a 10 ml sample (Taste group). There were no significant differences as a function of samples or groups, despite some trends for changes in liking to be higher in the consumed MSG þ condition. In a second experiment, 69 subjects were divided into three groups (Consume MSG þ ; Consume MSG 2 ; Taste MSG þ ) in which they received nine exposures to one novel soup flavor. The Consume MSG þ group showed a significantly greater increase in liking than either the Consume MSG 2 or the Taste MSG þ groups, which did not differ. Changes in familiarity ratings reflected amount consumed, not MSG content. Pairing glutamate with a novel flavor can condition liking for that flavor. While post-ingestive effects of glutamate may be rewarding, flavor conditioning cannot be ruled out. q 2003 Elsevier Ltd. All rights reserved. Keywords: Glutamate; Umami; Preference; Flavor; Post-ingestional effects; Taste
Introduction The mammalian gustatory system responds to a set of basic qualities—traditionally, sweet, sour, salty and bitter— that elicit relatively fixed hedonic responses. Many species, including humans, prefer sweetness and reject bitterness. These responses are evident from birth, or shortly thereafter, and are probably innate (Steiner, Glaser, Hawilo, & Berridge, 2001). Their origin appears to lie in the adaptive value in signalling the nutritional implications of taste qualities. The palatability of any taste compound, and the pattern of responsiveness of taste cells in the rat brainstem, is inversely related to its toxicity (Scott & Mark, 1987). Highly toxic compounds are rejected as unpalatable, primarily due to bitterness; highly nutritive substances with low toxicity are preferred, mainly because they are sweet. This neural and behavioral organisation has led to the hypothesis that preferences for tastes are a means by which * Address: School of Psychology, James Cook University, Cairns, Qld 4870, Australia. E-mail address: [email protected] (J. Prescott). 0195-6663/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2003.08.013
physiological wellbeing is maintained (Scott, 1992). Consistent with this, metabolic states can modulate taste palatability—preferred levels of salt in foods increase following salt depletion (Beauchamp, Bertino, Burke, & Engleman, 1990), while sweetness becomes less pleasant following glucose consumption (Cabanac, 1971). Glutamate, derived from the non-essential amino acid, glutamic acid, is an important contributor to food flavors. The addition of sauces (soy; Worcestershire), cheese (especially Parmesan), tomatoes, mushrooms, or meat, fish and vegetable stocks to foods all have the effect of increasing glutamate levels (Ninomiya, 1998; Yamaguchi, 1991). Since early 1900s, monosodium L -glutamate (MSG) has been commercially manufactured for use as a flavor enhancer, and there is ample evidence that adding MSG to suitable foods increases their palatability (Bellisle et al., 1991; Yamaguchi & Takahashi, 1984a,b) and consumption (Bellisle, 1998; Rogers & Blundell, 1990; Schiffman, 1998). Such hedonic changes are probably partly mediated by changes in the sensory properties of the foods, including increase in richness, savouriness and mouthfeel qualities (Fuke & Shimizu, 1993; Prescott, 2001; Yamaguchi, 1991).
144
J. Prescott / Appetite 42 (2004) 143–150
Although adding MSG also increases the saltiness of the foods (because of the Naþ), this cannot account for other sensory and hedonic changes. Soups containing added glutamate were found to be preferred to those without, even when the amount of Naþ was equivalent (Okiyama & Beauchamp, 1998). In addition, calcium glutamate is also able to enhance palatability (Bellisle, Dartois, & Broyer, 1992). There is considerable evidence that the taste quality associated with glutamate in foods, known as umami (a Japanese word, translated approximately as ‘savoury deliciousness’), is recognised and coded as a unique quality by the gustatory system (Bayliss & Rolls, 1991; Faurion, 1991; Ninomiya, Tanimukai, Yoshida, & Funakoshi, 1991; Rolls, 2000; Rolls, Critchley, Wakeman, & Mason, 1996; Yamaguchi & Komata, 1987). Recently a strong candidate for a specific umami taste receptor was identified. Chaudhari, Landin, and Roper (2000) showed that a variant of a particular glutamate receptor is present preferentially in rat taste buds, and responds specifically to compounds possessing umami quality. The ability of umami to elicit characteristic hedonic responses in human neonates, prior to experience with foods (Steiner et al., 2001) and across cultures (Prescott, 1998) is also suggestive of a basic taste. It has been proposed that the positive hedonic value of the umami taste, and its distinct coding in the taste system, reflect its function as a signal for protein in foods (Naim, Ohara, Kare, & Levinson, 1991). Consistent with this notion, persons with low nutritional/protein status preferred higher concentrations of MSG in solution than did those who were better nourished (Murphy, 1987). Other findings tend to be at odds with this interpretation, however. MSG preference in rats was highest during protein repletion (Torii, Mimura, & Yugari, 1987)—just the reverse of what would be expected if umami preference was a way of encouraging protein consumption. In addition to its role in proteins, however, glutamate is also involved in a variety of essential metabolic processes, including synthesis of other amino acids, as a precursor of glutamine (involved in amino acid homeostasis and gluconeogenesis) and glutathione (an important antioxidant), and in carbohydrate metabolism (Mori, Kawada, Ono, & Torii, 1991; Reeds, Burrin, Stoll, & Jahoor, 2000; Windmueller & Spaeth, 1975). Glutamate may also be a source of energy, since dietary glutamate was found to be the largest contributor to intestinal energy generation in studies of piglet nutrition and gut metabolism (Reeds et al., 2000). Consistent with these findings, in humans, umami taste, like sweetness, can trigger increase in pancreatic secretions (Naim et al., 1991), a cephalic phase response associated with preparation for nutrient utilisation (Mattes, 1997). One indication of the metabolic importance of a nutrient is its ability to produce a liking for foods or flavors with which it is repeatedly associated. Recovery from protein (Baker & Booth, 1989) and amino acid (Booth & Simson,
1971) depletion was found, in animal studies, to enhance preference for flavors/foods with which recovery from the depletion was associated. Conditioned liking has also been observed as a result of pairing of flavors with the postingestional consequences of nutrients that supply energy (Booth, 1991; Sclafani, 1995; Zellner, 1991). In animal studies, ingested or infused carbohydrates (Ackroff & Sclafani, 1994; Mehiel & Bolles, 1984; Perez, Lucas, & Sclafani, 1998), fats (Lucas & Sclafani, 1989) and proteins (Perez, Lucas, & Sclafani, 1996) were able to increase liking for flavors with which they were paired. Human research has also shown that novel flavors were preferred when paired with the consumption of carbohydrates (Birch, McPhee, Steinberg, & Sullivan, 1990), fat (Johnson, McPhee, & Birch, 1991; Kern, McPhee, Fisher, Johnson, & Birch, 1993) or protein (Gibson, Wainwright, & Booth, 1995). Kern et al. (1993) found increased liking for a flavor associated with fat intake, as compared to simple pairing of the flavor with the same fat level without ingestion. The usual interpretation of this effect is that it is mediated by the post-ingestive rewarding effects of the provision of energy. However, increased liking for flavors can also be demonstrated both through simple exposure to a novel flavor (Sullivan & Birch, 1990) and also through association with an already preferred stimulus, such as sucrose (flavor – flavor, or evaluative, conditioning), even when postingestional consequences can be ruled out (Zellner, Rozin, Aron, & Kulish, 1983). The present research aimed to determine if glutamate can condition increased liking for novel flavors to which it is added. One previous study has addressed the issue of whether addition of MSG to foods produces changes in acceptability with repeat exposure. Although Bellisle (1998) added MSG to relatively novel foods and found an increased intake over four weekly exposures, relative to these same foods with no added MSG, the study did not control for an effect purely due to the increased palatability of foods with MSG. By comparing liking for a flavor to which MSG is added in consumed and not consumed conditions, the present research also sought to determine if post-ingestive absorption of glutamate is responsible for any conditioned liking effects.
Methods Experiment 1 Subjects The participants were 11 males and 33 females (mean age: 24.5 years; range: 19 –48 years), students and staff at the University of Otago. Of these, two failed to attend all laboratory sessions, and their data were excluded from analysis. At the time of testing, no participant reported that they were severely restricting their diet, or the presence of
J. Prescott / Appetite 42 (2004) 143–150
current or recent illness that could interfere with perception of odors or tastes. Samples Pilot studies were carried out to identify three soup flavors that were both relatively novel and of low – moderate initial palatability, and to ensure that the impact of added MSG did not differ between the selected soups. Soups using two vegetables—chickpea and spinach—were prepared using two different soup bases. The soup bases, chosen because of their novelty for these subjects, were a dried fungus base and a dried flower base, both used in Chinese cooking. Added to water, these formed the soup stock in which the other ingredients were cooked. The soup/base combinations used were chickpea/dried flower base; chickpea/dried fungus base; and spinach/dried fungus base. All soups were prepared prior to the study, frozen, and then thawed overnight prior to serving at a constant temperature of 50 8C. Prior to freezing, MSG (0.5% w/w) was added to half the soup samples. This amount is within the range typically added to foods to enhance flavor (Yamaguchi, 1991). Procedure Subjects were divided into two groups. One group consumed 250 ml bowls of the soup during exposure sessions (Consume group), while the second group received 10 ml samples (Taste group). The second group was a control for simple exposure to the soup flavor plus MSG, similar to the group used by Kern et al. (1993). Subjects attended a pre-test session, 14 exposure sessions in total, and a post-test. As far as possible, exposure sessions were conducted on consecutive weekdays, from 12 to 2 p.m., and subjects were instructed not to consume any food 2 h prior to each session. In both the pre- and post-tests, subjects tasted 10 ml of the three soup flavors (without added MSG) and rated liking and familiarity for these on 100 mm line scales with descriptive anchors at each end (dislike extremely/like extremely; extremely unfamiliar/extremely familiar). Samples were presented monadically in balanced order, and subjects rinsed with water between samples. Since the ability to observe changes in familiarity and liking for these soups depended on initial levels being low – moderate, this pre-test rating also acted to identify unsuitable subjects with high preference and familiarity ratings (rating . 60 mm on a 100 mm scale) prior to the study commencing. No one was eliminated from the study at this stage. One of the three soup flavors acted as a control for changes that might occur over time, independent of exposure, and was received only in the pre- and post-tests. During the exposure sessions, subjects in both groups received two of the three soups — one with 0.5% added MSG (MSG þ ) and one without added MSG (MSG 2 )— on alternate days, that is, seven exposures per soup flavor. The association of the different soup flavors with the
145
MSG þ , MSG 2 , and Control conditions was balanced across subjects. Subjects in the Consume group received bowls of soup (250 ml) in each session, while those in the Taste group received 10 ml samples. To mask the purpose of the study and to ensure that attention was paid to the soups during each session, participants were asked to rate the flavor intensity of the soups on 100 mm line scales with descriptive anchors at each end (Extremely weak/Extremely strong). These data were not analysed. Analysis Familiarity and liking ratings were subjected to repeated measures ANOVA (SPSS V. 10), with Groups (Consume/ Taste) as a between-subjects factor, and Samples (MSG þ / MSG 2 /Control) and Exposure (Pre-exposure/post-exposure) as within-subjects factors. Post-hoc comparisons of means were undertaken using Tukey’s HSD Test. Results Familiarity. Initial ratings of familiarity were low – moderate (mean values: 33.9 –50 mm on a 100 mm scale, depending on group and sample type). Fig. 1 shows mean changes from pre- to post-test in familiarity ratings for each soup (MSG þ ; MSG 2 ; Control) in the two groups (Consume; Taste). Familiarity increased following exposure (Fð1; 42Þ ¼ 49:266; p , 0:0001), but there were no effects of either Samples or Groups. None of the confidence intervals (CIs; 95%) for the posttest –pre-test difference scores contained zero, reflecting familiarity increases for all samples, irrespective of MSG content, or whether the samples were consumed or tasted. Clearly, even a single exposure (Control samples) was sufficient to significantly increase familiarity. Liking. Subjects initially rated the samples as moderately disliked (mean values: 36.1– 39.6 on a 100 mm scale, depending on group and sample type). Fig. 2 shows the mean changes from pre- to post-test in liking ratings for each soup (MSG þ ; MSG 2 ; Control) in the two groups (Consume; Taste). All samples in both groups showed mean increases in liking from pre- to post-test, reflected in an ANOVA main effect of Exposure (Fð1; 42Þ ¼ 22:025; p , 0:0001). Despite the apparently greater increase in liking for the MSG þ sample in the Consume group (Fig. 2), there were no significant main effects or interactions involving Groups or Samples. Although this experiment failed to find significant differences between groups or samples in changes in liking for these soups, we conducted a second experiment based on the apparent trend towards an effect for the consumed MSG þ sample. In considering the change scores following exposure, only these scores had 95% CIs that did not contain zero, suggesting that liking only reliably increased for the MSG þ samples when consumed. In this experiment, the fact that subjects were exposed to two somewhat similar, unfamiliar soups (but differing in
146
J. Prescott / Appetite 42 (2004) 143–150
Fig. 1. Mean (þSEM) changes (post-test–pre-test) in ratings of familiarity of the flavor of the samples (MSG þ ; MSG 2 ; Control) in the Consume and Taste groups in Experiment 1.
MSG content) on alternate days may have diluted any differences, especially between samples. Hence, the second experiment used an experimental design in which separate groups of subjects each received exposure to only one soup
flavor. Controls for the post-ingestive effects of added glutamate consisted of one group that consumed an MSG 2 soup, and another that received limited exposure (10 ml samples) to an MSG þ soup.
Fig. 2. Mean (þ SEM) changes (post-test– pre-test) in ratings of liking for the flavor of the samples (MSG þ ; MSG 2 ; Control) in the Consume and Taste groups in Experiment 1.
J. Prescott / Appetite 42 (2004) 143–150
Experiment 2 Subjects Participants were 20 males and 49 females (mean age: 32.9 years; range: 19 – 63 years), students and staff at the University of Otago. None had participated in Experiment 1. Another eight participants failed to attend all sessions, and their data were excluded from analysis. At the time of testing, no participant reported that they were severely restricting their diet, or the presence of current or recent illness that could interfere with perception of odors or tastes. Samples Two soup/base combinations from Experiment 1 (Chickpea/dried flower base and Spinach/dried fungus base) were used. All soups were prepared prior to the study, frozen, and then thawed overnight prior to serving at a constant temperature of 50 8C. MSG (0.5% w/w) was added to half of the soup samples prior to freezing. Procedure Subjects were divided into three groups—two Consume groups (MSG þ ; MSG 2 ) and a Taste group (MSG þ ). There were 24 subjects in Group 1 (Consume MSG þ ), 24 in Group 2 (Consume MSG 2 ), and 21 in Group 3 (Taste MSG þ ). During the pre-test and post-test, subjects tasted 10 ml of the two soup flavors (without added MSG) and rated their liking and familiarity for these on 100 mm line scales. Sample order was balanced and subjects rinsed with
147
water between samples. Following the pre-test, subjects attended nine exposure sessions: three sessions per week, on successive weeks. Sessions were conducted from 12 to 2 p.m. each day. During these exposure sessions, each subject received only one soup flavor. Allocation of each flavor to control or exposure conditions was balanced across subjects. The two Consume groups received bowls of soup (250 ml) while the Taste group received 10 ml samples. Subjects were instructed not to consume any food 2 h prior to each session. As in Experiment 1, following consumption or tasting, subjects rated the flavor intensity of the soups on a 100 mm line scale with descriptive anchors at each end. Analysis Familiarity and liking ratings were subjected to repeated measures ANOVA (SPSS V. 10), with Exposure (Preexposure/post-exposure) and Samples (Exposed/Nonexposed) as within-subjects factors, and Groups (Consume MSG þ /Consume MSG 2 /Taste MSG þ ) as the between-subjects factor. Post-hoc comparisons of means were undertaken using Tukey’s HSD Test. Results Familiarity. As in Experiment 1, subjects were initially relatively unfamiliar with these flavors (mean values: 25.4– 37.1 mm on 100 mm scale, depending on group and sample type). Fig. 3 shows the mean changes from pre- to post-test in familiarity ratings for the exposed and non-exposed
Fig. 3. Mean (þ SEM) changes (post-test–pre-test) in ratings of familiarity of the flavor of the exposed and non-exposed samples in the Consume MSG þ , Consume MSG 2 and Taste MSG þ groups in Experiment 2.
148
J. Prescott / Appetite 42 (2004) 143–150
samples in the three groups (Consume MSG þ ; Consume MSG 2 ; Taste MSG þ ). There was a significant increase in familiarity from pretest to post-test (Exposure main effect: Fð1; 66Þ ¼ 155:48; p , 0:0001) in particular for those flavors that received repeated exposure (Exposure £ Samples interaction; Fð1; 66Þ ¼ 26:01; p , 0:0001). While there was no overall difference between Groups ðFð2; 66Þ ¼ 0:374Þ; the Group £ Exposure effect (Fð2; 66Þ ¼ 8:52; p , 0:005) was significant, with both Consume groups showing greater overall increase in familiarity from pre- to post-test than the Taste group. Liking. Initial ratings of liking were low –moderate (mean values: 32.9– 35.8 mm on a 100 mm scale, depending on group and sample type). Fig. 4 shows the mean changes from pre- to post-test in liking ratings for the exposed and non-exposed samples in the three groups (Consume MSG þ ; Consume MSG 2 ; Taste MSG þ ). ANOVA revealed significant increases in liking ratings from pre- to post-test (Fð1; 66Þ ¼ 13:14; p , 0:005), especially for the exposed flavors (Exposure £ Samples interaction: Fð1; 66Þ ¼ 23:73; p , 0:0001). Neither the Groups main effect ðFð2; 66Þ ¼ 0:18Þ; nor the Groups £ Exposure interaction ðFð2; 66Þ ¼ 0:40Þ were significant. However, there were significant effects for Samples (Fð1; 66Þ ¼ 16:46; p , 0:0001), Samples £ Groups (Fð2; 66Þ ¼ 8:37; p , 0:005), and Exposure £ Samples £ Groups (Fð2; 66Þ ¼ 6:38; p , 0:005). Post-hoc comparisons of means revealed that, for the exposed flavors, there was
a significant increase in liking ratings for the Consume MSG þ group, but not for either the Consume MSG 2 or the Taste MSG þ groups. While the three groups did not differ in pre-test ratings, post-test ratings for the Consume MSG þ group were substantially greater than those of the other groups, and suggestive of a relatively high degree of liking (mean: 63.17). By contrast, the mean post-test ratings of the Consume MSG 2 and the Taste MSG þ groups remained modest (44.29 and 47.76, respectively). For the non-exposed flavors, the only significant difference was higher post-test ratings in the Consume MSG 2 group than in the Consume MSG þ group.
Discussion As expected, familiarity ratings were increased by exposure to these novel flavors. These changes were even apparent in the case of the single exposure received by the ‘non-exposed’ flavors. While increased familiarity can facilitate liking through a reduction in neophobia (Pliner, 1982), familiarity changes alone cannot account for the present pattern of liking results. In Experiment 2, familiarity was a graded function of exposure to the sample, and not of MSG content, since it was significantly greater in the two Consume groups than in the Taste group. In Experiment 2, although all exposed samples showed an increase in liking for the associated flavor, this was significantly greater for the consumed MSG þ samples
Fig. 4. Mean (þSEM) changes (post-test–pre-test) in ratings of liking for the flavor of the exposed and non-exposed samples in the Consume MSG þ , Consume MSG 2 and Taste MSG þ groups in Experiment 2.
J. Prescott / Appetite 42 (2004) 143–150
than for either the tasted MSG þ samples or the consumed MSG 2 samples. The mean ratings indicated, moreover, that this sample became relatively well-liked following exposure. This experiment thus provides evidence that pairing a novel food flavor with glutamate can condition a liking for that flavor. Previous findings that ingested carbohydrates (Birch et al., 1990) and fat (Johnson et al., 1991; Kern et al., 1993) are able to condition liking for novel flavors have been interpreted in terms of the reward value of the energy that these nutrients provide. The fact that in Experiment 2 the consumed MSG þ soup was liked more than the tasted MSG þ soup implies that post-ingestional absorption of additional glutamate in foods is also primarily responsible for the enhanced liking and therefore similarly rewarding. In an attempt to control increased palatability associated with the effects of added MSG, the consumed MSG þ samples in both experiments were compared with the (presumably) equally palatable MSG þ samples that were only tasted. Since the energy due to carbohydrates contained within the basic soup formulation might also be expected to condition liking, comparisons were made between the consumed MSG þ samples and consumed samples that did not contain added MSG. It is unlikely that the natural glutamate content played a significant role here. Spinach, for example, contains approximately 0.05% glutamate (Ninomiya, 1998), around 10% of that added to the soup. One uncontrolled factor that limits the extent to which post-ingestive effects can be inferred is the absolute degree of exposure to the soup plus MSG in the different conditions/groups. It is possible that, in contrast to the nine discrete exposure sessions to a small sample in Experiment 2’s MSG þ Taste group, each of the multiple spoonfuls of the consumed MSG þ soup within each session acted as a separate, discrete exposure. Multiple exposures within a session can increase flavor liking through flavor –flavor conditioning (Zellner et al., 1983). Moreover, the degree of flavor liking has been shown to be a function of the number of discrete exposures (Pliner, 1982). Hence, the possibility exists that the consumed MSG þ samples were simply more effective at increasing liking through higher numbers of pairings of the novel flavor with the positive palatability of added MSG (that is, through flavor – flavor conditioning). The present study does not, of course, rule out the possibility that post-ingestive conditioning was responsible for the observed changes in liking. While it is not clear how the degree of exposure might be made equivalent in taste and consume conditions in this type of study, this issue might be resolved through manipulation of states of hunger. Post-ingestive conditioning in animal models is responsive to whether or not an animal is food-deprived (Ackroff, Vigorito, & Sclafani, 1990; Fedorchak & Bolles, 1987). Fedorchak and Bolles (1987) found not only that caloriemediated preferences to be stronger than flavor-mediated preferences, but that the former are also enhanced by
149
deprivation during preference testing, whereas flavormediated preferences are not. Kern et al. (1993) used subjects who were overnight food-deprived, both during exposure and at pre- and post-test, and found association with fat to be effective at increasing liking for novel flavors in this state, whereas they were ineffective in a state of satiety. Although the subjects in the present experiments were asked to refrain from eating for 2 h prior to each session, no measures of hunger were obtained. The present studies demonstrate that the addition of glutamate, like sugar and fat, can condition a liking for novel flavors. This may be the mechanism behind previous findings of increase in consumption of glutamate containing novel foods over time (Bellisle, 1998). The conditioned liking may reflect the role that dietary glutamate plays as a precursor in key metabolic processes or its utilisation as an energy source in gut metabolism. However, comparisons of subjects in deprived and non-deprived states during conditioning and testing may be the crucial test of whether post-ingestive processes are involved in such conditioned liking.
Acknowledgements Thanks to Bianca Turnbull who collected the data in these experiments. This research was supported by a grant from the International Glutamate Technical Committee.
References Ackroff, K., & Sclafani, A. (1994). Flavor preferences conditioned by intragastric infusions of dilute polycose solutions. Physiology and Behavior, 55, 957 –962. Ackroff, K., Vigorito, M., & Sclafani, A. (1990). Fat appetite in rats: The response of infant and adult rats to nutritive and non-nutritive oil emulsions. Appetite, 15, 171–188. Baker, B. J., & Booth, D. A. (1989). Genuinely olfactory preferences conditioned by protein repletion. Appetite, 13, 223–227. Bayliss, L. L., & Rolls, E. T. (1991). Responses of neurons in the primate taste cortex to glutamate. Physiology and Behavior, 49, 973–979. Beauchamp, G. K., Bertino, M., Burke, D., & Engleman, K. (1990). Experimental sodium depletion and salt taste in normal human volunteers. American Journal of Clinical Nutrition, 51, 881– 889. Bellisle, F. (1998). Nutritional effects of umami in the human diet. Food Reviews International, 14, 309– 320. Bellisle, F., Dartois, A. M., & Broyer, M. (1992). Two studies on the acceptability of calcium glutamate as a potassium-free sodium substitute in children. Journal of Renal Nutrition, 2, 42–46. Bellisle, F., Monneuse, M. O., Chabert, M., Laure-Achagiotis, C., Lanteaume, M. T., & Louis-Sylvestre, J. (1991). Monosodium glutamate as a palatability enhancer in the European diet. Physiology and Behavior, 49, 869–874. Birch, L. L., McPhee, L., Steinberg, L., & Sullivan, S. (1990). Conditioned flavor preferences in young children. Physiology and Behavior, 47, 501– 505. Booth, D. A. (1991). Learned ingestive motivation and the pleasures of the palate. In R. C. Bolles (Ed.), The hedonics of taste (pp. 29 –58). Hillsdale, NJ: Lawrence Erlbaum.
150
J. Prescott / Appetite 42 (2004) 143–150
Booth, D. A., & Simson, P. C. (1971). Food preferences acquired by association with variations in amino acid nutrition. Quarterly Journal of Experimental Psychology, 23, 135– 145. Cabanac, M. (1971). Physiological role of pleasure. Science, 173, 1103–1107. Chaudhari, N., Landin, A. M., & Roper, S. D. (2000). A metabotropic glutamate receptor variant functions as a taste receptor. Nature Neuroscience, 3, 113–119. Faurion, A. (1991). Are umami taste receptor sites structurally related to glutamate CNS receptor sites? Physiology and Behavior, 49, 905–912. Fedorchak, P. M., & Bolles, R. C. (1987). Hunger enhances the expression of calorie- but not taste-mediated conditioned flavor preferences. Journal of Experimental Psychology: Animal Behavior Processes, 13, 73–79. Fuke, S., & Shimizu, T. (1993). Sensory and preference aspects of umami. Trends in Food Science and Technology, 4, 246–251. Gibson, E. L., Wainwright, C. J., & Booth, D. A. (1995). Disguised protein in lunch after low-protein breakfast conditions food-flavor preferences dependent on recent lack of protein intake. Physiology and Behavior, 58, 363 –371. Johnson, S. L., McPhee, L., & Birch, L. L. (1991). Conditioned preferences: Young children prefer flavors associated with high dietary fat. Physiology and Behavior, 50, 1245–1251. Kern, D. L., McPhee, L., Fisher, J., Johnson, S., & Birch, L. L. (1993). The postingestive consequences of fat condition preferences for flavors associated with high dietary fat. Physiology and Behavior, 54, 71– 76. Lucas, F., & Sclafani, A. (1989). Favor preferences conditioned by intragastric fat infusions in rats. Physiology and Behavior, 46, 403– 412. Mattes, R. D. (1997). Physiologic responses to sensory stimulation by food: Nutritional implications. Journal of the American Dietetic Association, 97, 406 –413. Mehiel, R., & Bolles, R. C. (1984). Learned flavor preferences based on caloric outcomes. Animal Learning and Behavior, 12, 421– 427. Mori, M., Kawada, T., Ono, T., & Torii, K. (1991). Taste preference and protein nutrition and L -amino acid homeostasis in male Sprague– Dawley rats. Physiology and Behavior, 49, 987–996. Murphy, C. (1987). Flavor preference for monosodium glutamate and casein hydrolysate in young and elderly persons. In Y. Kawamura, & M. R. Kare (Eds.), Umami: A basic taste (pp. 139–151). New York: Marcel Dekker. Naim, M., Ohara, I., Kare, M. R., & Levinson, M. (1991). Interaction of MSG taste with nutrition: Perspectives in consummatory behavior and digestion. Physiology and Behavior, 49, 1019–1024. Ninomiya, K. (1998). Natural occurrence. Food Reviews International, 14, 177– 211. Ninomiya, Y., Tanimukai, T., Yoshida, S., & Funakoshi, M. (1991). Gustatory neural responses in preweanling mice. Physiology and Behavior, 49, 913– 918. Okiyama, A., & Beauchamp, G. K. (1998). Taste dimensions of monosodium glutamate (MSG) in a food system: Role of glutamate in young American subjects. Physiology and Behavior, 65, 177–181. Perez, C., Lucas, F., & Sclafani, A. (1996). Carbohydrate- and proteinconditioned flavor preferences: Effects of nutrient preloads. Physiology and Behavior, 59, 467–474. Perez, C., Lucas, F., & Sclafani, A. (1998). Increased flavor acceptance and preference conditioned by the postingestive actions of glucose. Physiology and Behavior, 64, 483–492. Pliner, P. (1982). The effects of mere exposure on liking for edible substances. Appetite, 3, 283 –290.
Prescott, J. (1998). Comparisons of taste perceptions and preferences of Japanese and Australian consumers: Overview and implications for cross-cultural sensory research. Food Quality and Preference, 9, 393 –402. Prescott, J. (2001). Taste hedonics and the role of umami. Food Australia, 53, 550–554. Reeds, P. J., Burrin, D. G., Stoll, B., & Jahoor, F. (2000). Intestinal glutamate metabolism. Journal of Nutrition, 130, 978S–982S. Rogers, P. J., & Blundell, J. (1990). Umami and appetite: Effects of monosodium glutamate on hunger and food intake in human subjects. Physiology and Behavior, 48, 801–804. Rolls, E. T. (2000). The representation of umami taste in the taste cortex. Journal of Nutrition, 130, 960S–965S. Rolls, E. T., Critchley, H., Wakeman, E. A., & Mason, R. (1996). Responses of neurons in the primate taste cortex to the glutamate ion and to inosine 50 -monophosphate. Physiology and Behavior, 59, 991 –1000. Schiffman, S. S. (1998). Sensory enhancement of foods for the elderly with monosodium glutamate. Food Reviews International, 14, 321–334. Sclafani, A. (1995). How food preferences are learned—animal laboratory models. Proceedings of the Nutrition Society, 54, 419–427. Scott, T. R. (1992). Taste, feeding, and pleasure. Progress in Psychobiology and Physiological Psychology, 15, 231–291. Scott, T. R., & Mark, G. P. (1987). The taste system encodes stimulus toxicity. Brain Research, 414, 197–203. Steiner, J. E., Glaser, D., Hawilo, M. E., & Berridge, K. C. (2001). Comparative expression of hedonic impact: Affective reactions to taste by human infants and other primates. Neuroscience and Biobehavioral Reviews, 25, 53 –74. Sullivan, S. A., & Birch, L. L. (1990). Pass the sugar, pass the salt: Experience dictates preference. Developmental Psychology, 26, 546 –551. Torii, K., Mimura, T., & Yugari, Y. (1987). Biochemical mechanism of umami taste perception and effect of dietary protein on the taste preference for amino acids and sodium chloride in rats. In Y. Kawamura, & M. R. Kare (Eds.), Umami: A basic taste (pp. 125 –138). New York: Marcel Dekker. Windmueller, H. G., & Spaeth, A. E. (1975). Intestinal metabolism of glutamine and glutamate from the lumen as compared to glutamine from blood. Archives of Biochemistry and Biophysics, 171, 662–672. Yamaguchi, S. (1991). Basic properties of umami and effects on humans. Physiology and Behavior, 49, 833–841. Yamaguchi, S., & Komata, Y. (1987). Independence and primacy of umami as compared with the four basic tastes. Annals of the New York Academy of Science, 510, 725–726. Yamaguchi, S., & Takahashi, C. (1984a). Interactions of monosodium glutamate and sodium chloride on saltiness and palatability. Journal of Food Science, 49, 82 –85. Yamaguchi, S., & Takihashi, C. (1984b). Hedonic functions of monosodium glutamate and four basic taste substances used at various concentration levels in single and complex systems. Agricultural and Biological Chemistry, 48, 1077– 1081. Zellner, D. A. (1991). How foods get to be liked: Some general mechanisms and some special cases. In R. C. Bolles (Ed.), The hedonics of taste (pp. 199 –217). Hillsdale, NJ: Lawrence Erlbaum. Zellner, D. A., Rozin, P., Aron, M., & Kulish, C. (1983). Conditioned enhancement of human’s liking for flavor by pairing with sweetness. Learning and Motivation, 14, 338 –350.